Wind Resource Mapping in the Maldives SITE INSTALLATION REPORTS MARCH 2018 This report was prepared by DNV GL, under contract to The World Bank. It is one of several outputs from the wind Resource Mapping and Geospatial Planning Maldives [Project ID: P146018]. This activity is funded and supported by the Energy Sector Management Assistance Program (ESMAP), a multi-donor trust fund administered by The World Bank, under a global initiative on Renewable Energy Resource Mapping. Further details on the initiative can be obtained from the ESMAP website. This document is an interim output from the above-mentioned project, and the content is the sole responsibility of the consultant authors. Users are strongly advised to exercise caution when utilizing the information and data contained, as this may include preliminary data and/or findings, and the document has not been subject to full peer review. Final outputs from this project will be marked as such, and any improved or validated wind resource data will be incorporated into the Global Wind Atlas. Copyright © 2018 THE WORLD BANK Washington DC 20433 Telephone: +1-202-473-1000 Internet: www.worldbank.org The World Bank, comprising the International Bank for Reconstruction and Development (IBRD) and the International Development Association (IDA), is the commissioning agent and copyright holder for this publication. However, this work is a product of the consultants listed, and not of World Bank staff. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work and accept no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for non-commercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: +1-202-522-2625; e-mail: pubrights@worldbank.org. Furthermore, the ESMAP Program Manager would appreciate receiving a copy of the publication that uses this publication for its source sent in care of the address above, or to esmap@worldbank.org. Contents: 1. Site Installation Report: Thulusdho 2. Site Installation Report: Hoarafushi RENEWABLE ENERGY WIND MAPPING FOR THE MALDIVES LIDAR Site Installation Report - Thulusdhoo The World Bank Document No.: 702909-AUME-R05 Issue : C, Status : FINAL Date: 6 March 2018 IMPORTANT NOTICE AND DISCLAIMER 1. This document is intended for the sole use of the Client as detailed on the front page of this document to whom the document is addressed and who has entered into a written agreement with the DNV GL entity issuing this document (“DNV GL”). To the extent permitted by law, neither DNV GL nor any group company (the "Group") assumes any responsibility whether in contract, tort including without limitation negligence, or otherwise howsoever, to third parties (being persons other than the Client), and no company in the Group other than DNV GL shall be liable for any loss or damage whatsoever suffered by virtue of any act, omission or default (whether arising by negligence or otherwise) by DNV GL, the Group or any of its or their servants, subcontractors or agents. This document must be read in its entirety and is subject to any assumptions and qualifications expressed therein as well as in any other relevant communications in connection with it. This document may contain detailed technical data which is intended for use only by persons possessing requisite expertise in its subject matter. 2. This document is protected by copyright and may only be reproduced and circulated in accordance with the Document Classification and associated conditions stipulated or referred to in this document and/or in DNV GL’s written agreement with the Client. No part of this document may be disclosed in any public offering memorandum, prospectus or stock exchange listing, circular or announcement without the express and prior written consent of DNV GL. A Document Classification permitting the Client to redistribute this document shall not thereby imply that DNV GL has any liability to any recipient other than the Client. 3. This document has been produced from information relating to dates and periods referred to in this document. This document does not imply that any information is not subject to change. Except and to the extent that checking or verification of information or data is expressly agreed within the written scope of its services, DNV GL shall not be responsible in any way in connection with erroneous information or data provided to it by the Client or any third party, or for the effects of any such erroneous information or data whether or not contained or referred to in this document. 4. Any wind or energy forecasts estimates or predictions are subject to factors not all of which are within the scope of the probability and uncertainties contained or referred to in this document and nothing in this document guarantees any particular wind speed or energy output. KEY TO DOCUMENT CLASSIFICATION For disclosure only to named individuals within the Client’s Strictly Confidential : organisation. For disclosure only to individuals directly concerned with the Private and Confidential : subject matter of the document within the Client’s organisation. Commercial in Confidence : Not to be disclosed outside the Client’s organisation. DNV GL only : Not to be disclosed to non-DNV GL staff Distribution for information only at the discretion of the Client (subject to the above Important Notice and Disclaimer and the Client’s Discretion : terms of DNV GL’s written agreement with the Client). Available for information only to the general public (subject to Published : the above Important Notice and Disclaimer). Project name: Renewable Energy Wind Mapping for the Maldives DNV GL - Energy Report title: LIDAR Site Installation Report - Thulusdhoo Renewables Advisory Customer: The World Bank, 9665 Chesapeake Drive, Suite 435 1818 H Street, N.W. San Diego, CA 92123 Washington, DC 20433 Tel: 703-795-8103 Contact person: Sandeep Kohli Enterprise No.: 94-340223694- Date of issue: 6 March 2018 340223694-3402236 Project No.: 702909 Document No.: 702909-AUME-R05 Issue/Status: C / FINAL Task and objective: provide a permanent record of the site characteristics and measurement equipment for the LIDAR at Thulusdhoo. Prepared by: Verified by: Approved by: Fowzi Dahhan Kevin Bleibler Trenton Gilbert Engineer, Renewables Advisory Head of Section, Measurements Head of Section, Developer Support Renewables Advisory Services (Pacific), Renewables Advisory ☐ Strictly Confidential Keywords: ☐ Private and Confidential World Bank, ESMAP, Maldives, wind, measurement, ☐ Commercial in Confidence LIDAR, Site Installation Report ☐ DNV GL only ☒ Client’s Discretion ☐ Published Reference to part of this report which may lead to misinterpretation is not permissible. Issue Date Reason for Issue Prepared by Verified by Approved by A 9 Jun 2017 PRELIMINARY DRAFT FD KB TG B 30 Aug 2017 Revised version - DRAFT FD KB TG C 6 Mar 2018 FINAL FD KB TG Table of contents 1 INTRODUCTION .............................................................................................................. 2 2 SITE INFORMATION ........................................................................................................ 2 2.1 Site Location 2 2.2 Site Description 4 3 SITE EQUIPMENT ............................................................................................................ 4 3.1 LIDAR Unit 4 3.2 Auxiliary Meteorological Station 8 3.3 Power Supply 12 3.4 Communications 14 4 OPERATIONS AND MAINTENANCE ................................................................................... 15 4.1 Data Acquisition and Management 15 4.2 Direction Measurements 15 4.3 Scheduled maintenance 16 4.4 Unscheduled Maintenance 16 4.5 Servicing and Re-verification 16 5 REFERENCES ................................................................................................................ 17 APPENDIX A – MAPPING ............................................................................................................. 18 APPENDIX B – PANORAMIC PHOTOGRAPHS OF SITE ...................................................................... 22 APPENDIX C – LIDAR DATA SHEET ............................................................................................... 24 APPENDIX D – LIDAR PERFORMANCE VERIFICATION REPORT .......................................................... 27 DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 4 1 INTRODUCTION The World Bank (the “Client”) has retained Garrad Hassan America, Inc. (“DNV GL”) to provide a validated mesoscale wind atlas for the Maldives, including associated deliverables and wind energy development training courses. During Phase 1 of the project, which has been completed, preliminary mesoscale mapping was carried out covering the entire country [1]. Phase 2 of the project, which is currently underway, involves the installation of two Light Detection and Ranging (LIDAR) based wind measurement sites in the country. Meteorological data collected at these sites over a two-year period will provide the basis for validating the mesoscale modeling outputs from Phase 1. LIDAR units and associated equipment have been commissioned at the two measurement sites, Hoarafushi and Thulusdhoo, in April 2017. This report documents the installation of the LIDAR site at Thulusdhoo, and presents the characteristics of the site, as well as details of the measurement equipment, power supply and data acquisition system. 2 SITE INFORMATION 2.1 Site Location The Thulusdhoo site is located in the central east of the island of Thulusdhoo in the Kaafu Atoll, in the northern central Maldives, approximately 26 km north-north-east of Male. A map of the site location is shown in Figure 2-1 below. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 2 Area shown Male MALDIVES Figure 2-1 Location of Thulusdhoo LIDAR site DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 3 It should be noted that there is an area of recently reclaimed land on the western side of Thulusdhoo Island. The satellite imagery shown in Figure 2-1 predates this and therefore the reclaimed land does not appear on the map. Figure A-1 in Appendix A presents a map of the site vicinity, showing 0 and 3 m elevation contours based on SRTM-1 data [2]. It should be noted that SRTM-1 data can be influenced by trees and buildings, and may not accurately reflect the ground elevation. It should also be noted that since the available SRTM-1 data predates the recent land reclamation, the reclaimed land does not appear on the contour map. Figure A-2 in Appendix A presents imagery from Google Earth showing the reclaimed land for reference. Figure A-3 in Appendix A presents a map of estimated surface roughness zones in the vicinity of the site. It should be noted that this map does take into consideration the reclaimed land. Table 2-1 presents details of the site location coordinates and elevation. Table 2-1 Site location summary Coordinates Coordinates (Geographic)2 (UTM, Zone 43 N)2 Commissioning Elevation Site name Date [m ASL]1 Latitude Longitude Eastings Northings [degrees] [degrees] [m] [m] Thulusdhoo 6 April 2017 0 to 3 m 4.37453 73.65218 350442 483661 1. Approximate elevation only, due to inaccuracy of SRTM-1 data 2. Datum: WGS 84 2.2 Site Description The site is situated within a diesel electricity generation compound belonging to STELCO. The LIDAR location consists of a clear area bounded by the powerhouse building to the north, and compound boundary fences to the east, south, and west, with vegetation of varying heights along the fences. There are trees with heights of approximately 6 m on the western fence line. Outside the compound there is a tall tree to the southwest with a height of approximately 13 m, and a school building to the south with a height of approximately 10 m. The surrounding terrain is flat with trees and low-lying buildings, with the area of recently reclaimed land, which is sandy and has minimal ground cover, located further to the west. The site is located within approximately 300 m of the shoreline to the north, east and south. Panoramic photographs of the site are shown in 0Appendix B. 3 SITE EQUIPMENT 3.1 LIDAR Unit The LIDAR unit installed at the Thulusdhoo site is a ZephIR 300 LIDAR remote sensing device. The device employs a continuous wave laser to measure horizontal and vertical wind speed and wind direction at a specified range of heights. Table 3-1 provides a summary of key information about the LIDAR unit. Additional details on the LIDAR specifications are provided in Appendix C. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 4 Table 3-1 LIDAR unit summary Manufacturer Model Serial Number Zephir Ltd ZephIR 300 597 The unit is secured to a concrete pad footing and is enclosed by fencing. Figure 3-1 to Figure 3-3 show photographs of the LIDAR unit as installed on site. Figure 3-1 LIDAR unit at Thulusdhoo – general view of enclosure DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 5 Figure 3-2 View within enclosure Figure 3-3 Close up of LIDAR unit The LIDAR unit is equipped with a washing system for cleaning the window, as shown in Figure 3-4. The washing system is connected to an external cleaning fluid supply bottle, as shown in Figure 3-5. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 6 Figure 3-4 Top of LIDAR unit showing sensor window and washing system Figure 3-5 Base of LIDAR unit showing cleaning fluid supply bottle The LIDAR simultaneously records 10-minute average, maximum, minimum, and standard deviation statistics for the horizontal and vertical wind speeds, and average and standard deviation statistics for wind direction, at eleven specified measurement heights between 10 m and 200 m above the top of the unit. The LIDAR measurement configuration is summarised in Table 3-2 below. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 7 Table 3-2 LIDAR configuration summary Measurement Measurement Measurement heights Unit Channels parameter interval [m AGL] Mean / minimum / Horizontal wind m/s maximum / 11, 20, 30, 39, speed standard deviation 10 minutes 50, 60, 80, (continuous 100, 120, 150, Vertical wind speed m/s Mean scan) 200 Wind direction1 Degrees Mean 1. The unit is aligned at a bearing of approximately 1° from true north. The LIDAR unit has been subjected to independent testing and performance verification in accordance with the second edition of the reviewed IEC 61400-12-1 standard, Annex L [2][3]. A copy of the performance verification report is presented in Appendix D. 3.2 Auxiliary Meteorological Station The LIDAR unit also encompasses a small auxiliary meteorological station to verify the wind direction measured by the LIDAR and record other atmospheric parameters including wind speed, temperature, air pressure, relative humidity and precipitation. The meteorological station is mounted on an existing pole on the wall on the southern boundary of the compound. The elevated mounting arrangement aims to minimise the effect of any flow disruption that may be caused by surrounding buildings and other obstructions. Figure 3-6 below shows a photograph of the pole at the wall of the compound. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 8 Figure 3-6 View of pole carrying meteorological station Figure 3-7 shows the location of the meteorological station in relation to the LIDAR enclosure. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 9 Figure 3-7 LIDAR unit enclosure with pole carrying meteorological station highlighted in the background Figure 3-8 below shows a close up photograph of the meteorological station taken during the installation process. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 10 Figure 3-8 Close up of meteorological station The meteorological station measurement configuration is summarised in Table 3-3. Table 3-3 Meteorological station configuration summary Measurement height Measurement Measurement Unit Channels [m AGL] parameter interval Air temperature °C Mean Pressure Millibar Mean Relative humidity % Mean 6 1 10 minutes Precipitation % Mean Wind speed m/s Mean Wind direction2 Degrees Mean 1. Approximate height. 2. Meteorological station aligned approximately to magnetic north. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 11 3.3 Power Supply The LIDAR unit is powered by the local mains electricity supply, with back-up power provided by a battery system consisting of two 12 V, 100 Ah deep cycle rechargeable batteries, which are float charged by the mains supply via a charge regulator. The batteries are intended to provide a backup power supply for between 12 and 24 hours in the event of a mains power supply interruption, and are housed within a weatherproof enclosure mounted on the concrete pad adjacent to the LIDAR device, as shown in Figure 3-7 above. A more detailed view of the battery enclosure is shown in Figure 3-9. Figure 3-9 Battery enclosure adjacent to LIDAR unit The batteries located inside the enclosure are shown in Figure 3-10. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 12 Figure 3-10 Batteries inside enclosure adjacent to LIDAR unit A schematic of the battery enclosure configuration is shown in Figure 3-11. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 13 Figure 3-11 Schematic of battery enclosure configuration Table 3-4 summarizes the key specifications of the power supply system. Table 3-4 Power supply system summary Battery Battery Charger 2 x 12V 100 Ah, AGM deep cycle Victron Energy Blue Power 12/25 - 12V, 25A 3.4 Communications The system uses an external 3G modem and directional antenna for data transmission. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 14 4 OPERATIONS AND MAINTENANCE The LIDAR unit will operate for a period of two years. 4.1 Data Acquisition and Management All recorded data is sent on a daily basis to DNV GL for evaluation and archiving. The data is transmitted automatically to an email address which is the data receipt portal for the DNV GL “Resource Panorama” service. The data is then subjected to regular quality control by a data analyst. DNV GL will provide both raw and quality controlled data from the LIDAR device. 4.2 Direction Measurements The Zephir 300 LIDAR device uses measurements from both its scanning laser and met station to measure the wind direction. A laser scan determines the axis of the wind direction. The direction of the wind along that axis is then determined in conjunction with the met station reading. The following table illustrates the process, for a northeast wind: Laser Measurement Met Station Wind Direction Reported Wind Direction If the met direction in the above diagram pointed in the opposite direction, the reported direction would be the other axial possibility in the first diagram. For example: Laser Measurement Met Station Wind Direction Reported Wind Direction The Met Station direction reading only needs to be very coarse for this purposes as is only used to determine the direction ‘phase’. However, if the met station is shadowed or if it is placed in areas of wind recirculation then there is the potential for a 180 degrees direction ambiguity to result. Given the proximity of the LIDAR device and met station to buildings, trees and other obstacles, the potential for a 180 degree ambiguity in the wind direction measurements exists. Care should be taken when using the raw data from the LIDAR device that instances where the 180 degree direction ambiguity exists are corrected. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 15 DNV GL will correct periods where there is evidence of a 180 degree ambiguity in the quality controlled data set provided. 4.3 Scheduled maintenance Regular inspections and maintenance of the LIDAR device will be undertaken by trained local staff. The purpose of the inspections will be to: a) Check the operating status of the LIDAR device b) Check the LIDAR device and ancillary equipment have not been damaged or soiled c) Clean the LIDAR viewing window d) Refill the water reservoir e) Report findings to DNV GL 4.4 Unscheduled Maintenance If a LIDAR unit shows signs of fault or failure, or if any issues with the communication system are detected, as flagged through remote analysis of data, a corrective maintenance program will be initiated. Once initiated, the corrective maintenance program will involve a multi-level approach as follows. As a first step, local staff on the islands will perform a visual inspection of the unit and correct any issues if possible. If further inspection, troubleshooting, or maintenance of the unit is required, a corrective maintenance team consisting of in-country trained personnel from DNV GL’s Local Partner will visit the site. In cases where the issue cannot be resolved by local staff, and where appropriate, expert personnel from DNV GL will travel from Australia to the Maldives to perform the necessary corrective maintenance. DNV GL will remotely assist with the interventions at all stages of the process. 4.5 Servicing and Re-verification The manufacturer recommends that the LIDAR device is returned to the manufacturer every 3 years for servicing and re-verification. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 16 5 REFERENCES [1] “Mesoscale Wind Modeling Report 1- Interim wind atlas for Maldives”, DNV GL, 702909-AUME-R- 01-D, 2 July 2015. [2] National Aeronautics and Space Administration (NASA), “Shuttle Radar Topography Mission (SRTM) -1 arc second resolution”, data accessed using Global Mapper 18 software, 25 May 2017. [3] International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power performance measurements of electricity producing wind turbines. Ed. 2. CD. International Electronic Commission.3, June 2013. DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 17 APPENDIX A – MAPPING DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 18 Figure A-1 Map of Thulusdhoo LIDAR site showing terrain elevation contours DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 19 Reclaimed land Source: Google Earth Figure A-2 Satellite imagery of Thulusdhoo LIDAR site showing recently reclaimed land on western side of island DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 20 Figure A-3 Map of Thulusdhoo LIDAR site showing surface roughness zones DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 21 APPENDIX B – PANORAMIC PHOTOGRAPHS OF SITE DNV GL – Report No. 702909-AUME-R05, Rev. A, Status: FINAL – www.dnvgl.com Page 22 Panoramic view from location of LIDAR at Thulusdhoo N W (350442, 483661) E S DNV GL – Report No. 702909-AUME-R05, Rev. A, Status: FINAL – www.dnvgl.com Page 23 APPENDIX C – LIDAR DATA SHEET DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 24 DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 25 DNV GL – Report No. 702909-AUME-R05, Rev. C, Status: FINAL – www.dnvgl.com Page 26 APPENDIX D – LIDAR PERFORMANCE VERIFICATION REPORT DNV GL – Report No. 702909-AUME-R05, Rev. A, Status: FINAL – www.dnvgl.com Page 27 ZP 597 Independent analysis and reporting of ZephIR Lidar performance verification at Pershore test site, including IEC compliant validation analysis The World Bank Report No.: 702909-AUME-R-07, Rev. B Date: 6 March 2018 IMPORTANT NOTICE AND DISCLAIMER 1. This document is intended for the sole use of the Client as detailed on the front page of this document to whom the document is addressed and who has entered into a written agreement with the DNV GL entity issuing this document (“DNV GL”). To the extent permitted by law, neither DNV GL nor any group company (the "Group") assumes any responsibility whether in contract, tort including without limitation negligence, or otherwise howsoever, to third parties (being persons other than the Client), and no company in the Group other than DNV GL shall be liable for any loss or damage whatsoever suffered by virtue of any act, omission or default (whether arising by negligence or otherwise) by DNV GL, the Group or any of its or their servants, subcontractors or agents. This document must be read in its entirety and is subject to any assumptions and qualifications expressed therein as well as in any other relevant communications in connection with it. This document may contain detailed technical data which is intended for use only by persons possessing requisite expertise in its subject matter. 2. This document is protected by copyright and may only be reproduced and circulated in accordance with the Document Classification and associated conditions stipulated or referred to in this document and/or in DNV GL’s written agreement with the Client. No part of this document may be disclosed in any public offering memorandum, prospectus or stock exchange listing, circular or announcement without the express and prior written consent of DNV GL. A Document Classification permitting the Client to redistribute this document shall not thereby imply that DNV GL has any liability to any recipient other than the Client. 3. This document has been produced from information relating to dates and periods referred to in this document. This document does not imply that any information is not subject to change. Except and to the extent that checking or verification of information or data is expressly agreed within the written scope of its services, DNV GL shall not be responsible in any way in connection with erroneous information or data provided to it by the Client or any third party, or for the effects of any such erroneous information or data whether or not contained or referred to in this document. 4. Any wind or energy forecasts estimates or predictions are subject to factors not all of which are within the scope of the probability and uncertainties contained or referred to in this document and nothing in this document guarantees any particular wind speed or energy output. KEY TO DOCUMENT CLASSIFICATION For disclosure only to named individuals within the Client’s Strictly Confidential : organisation. For disclosure only to individuals directly concerned with the Private and Confidential : subject matter of the document within the Client’s organisation. Commercial in Confidence : Not to be disclosed outside the Client’s organisation. DNV GL only : Not to be disclosed to non-DNV GL staff Distribution for information only at the discretion of the Client (subject to the above Important Notice and Disclaimer and the Client’s Discretion : terms of DNV GL’s written agreement with the Client). Available for information only to the general public (subject to Published : the above Important Notice and Disclaimer). DNV GL Project name: ZP 597 DNV GL - Energy Report title: Independent analysis and reporting of ZephIR Renewables Advisory Lidar performance verification at Pershore test Suite 25, Level 8, site, including IEC compliant validation analysis 401 Docklands Drive, Docklands, Customer: The World Bank, Victoria 3008, Australia 1818 H Street, N.W. Tel: +61 3 9600 1993 Washington, DC 20433 Contact person: Sandeep Kohli Date of issue: 6 March 2018 Project No.: 702909 Report No.: 702909-AUME-R-07, Rev. B Task and objective: Independent analysis and reporting of ZephIR Lidar performance verification at Pershore test site, including IEC compliant validation analysis Prepared by: Verified by: Approved by: M Quan F Dahhan T Gilbert Engineer Engineer Principal Engineer, Head of Section ☐ Strictly Confidential Keywords: ☐ Private and Confidential ZephIR, Lidar, performance verification ☐ Commercial in Confidence ☐ DNV GL only ☒ Client’s Discretion ☐ Published Reference to part of this report which may lead to misinterpretation is not permissible. Rev. No. Date Reason for Issue Prepared by Verified by Approved by A 30 Aug 2017 Draft (electronic version, only) M Quan F Dahhan, B Schmidt T Gilbert B 6 Mar 2018 Final M Quan F Dahhan, B Schmidt T Gilbert DNV GL Table of contents 1 INTRODUCTION .............................................................................................................. 2 2 DESCRIPTION OF THE TEST SITE...................................................................................... 3 2.1 The test site 3 2.2 Measuring equipment 4 2.2.1 Meteorological mast: layout, sensors and data acquisition 4 2.2.2 The ZephIR Lidar 7 3 LIDAR PERFORMANCE VERIFICATION (LPV) APPROACH ....................................................... 8 3.1 Common test conditions and data filtering 8 3.2 Sector filtering 8 3.3 Lidar specific filtering 9 3.4 Data coverage requirements for accuracy assessment 9 3.5 LPV evaluation 9 4 RESULTS ..................................................................................................................... 11 4.1 System availability 11 4.2 Data availability 12 4.3 Data filtering 12 4.4 Wind speed comparison 12 4.5 Wind direction comparison 16 4.6 Performance verification according to revised IEC standard, Annex L 17 4.6.1 Performance verification uncertainty 19 5 IMPORTANT REMARKS AND LIMITATIONS ........................................................................ 25 6 CONCLUSION ............................................................................................................... 26 7 REFERENCES ................................................................................................................ 28 8 GLOSSARY ................................................................................................................... 29 Appendices KEY PERFORMANCE INDICATORS AND ACCEPTANCE CRITERIA, IN LINE WITH [2] 30 PERSHORE/THROCKMORTON MET MAST DETAILS .................................................... 32 TIME SERIES OF WIND SPEED .............................................................................. 34 WIND DIRECTION ................................................................................................ 35 CUP CALIBRATION CERTIFICATES, TAKEN FROM[6] ................................................. 38 IEC ANNEX L UNCERTAINTY ANALYSES................................................................... 51 ENVIRONMENTAL PARAMETERS ACCORDING TO IEC ANNEX L, DURING THE CAMPAIGN, REF. 70 M LEVEL ......................................................................................... 52 DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page i LIST OF TABLES Table 1: List of meteorological sensors and individual anemometers installed at the mast during verification campaign, as of Appendix B, and list of calibration factors for cup anemometers. The valid calibration certificates are attached to this report in Appendix E. 6 Table 2: Height settings of ZP300 Lidar and reference mast. Levels for wind speed and wind direction comparisons are highlighted in bold letters. 7 Table 3: Number of 10 minute data points after filtering used for WS comparison at each of the four (4) levels. 11 Table 4: Summary of system and data availabilities for ZP 597 at respective heights 11 Table 5: Regression results comparison for ZPH 597; acceptance relevant results are colour shaded. Note the regression lines are forced through the origin. 14 Table 6: Summary of absolute wind speed differences between cups and Lidar 14 Table 7: Summary of WD comparison results for both comparison levels 17 Table 8: Statistical parameters of wind speed deviation 18 Table 9: Uncertainty calculation for ZP 597 at 20 m level 21 Table 10: Uncertainty calculation for ZP 597 at 45 m level 22 Table 11: Uncertainty calculation for ZP 597 at 70 m level 23 Table 12: Uncertainty calculation for ZP 597 at 91 m level 24 Table 13: List of KPIs and ACs relevant for System and Data Availability assessment 30 Table 14: List of KPIs and ACs relevant for Wind Data Accuracy assessment 31 LIST OF FIGURES Figure 1: Map of the Pershore test site near Throckmorton, UK. The position of the reference mast is marked by a red dot. 3 Figure 2: Schematic of the sensor level and boom distribution at the 90.5 m mast, as taken from [1]. See Table 1 for sensor distribution according to the alphanumeric label per boom (A to N) and the actually valid serial numbers. 5 Figure 3: Typical setup of ZephIR Lidars next to the reference mast at Pershore. 7 Figure 4: Wind direction sectors used to select undisturbed wind speed data from oppositely arranged cup carrying booms for comparison. 9 Figure 5: Plots of linear wind speed regression results for 20, 45, 70 and 91 m (note that regression slopes are forced through the origin) 15 Figure 6: Regression plot of wind direction comparisons at 44 m (left) and 88 m (right) 16 Figure 7: Comparison of the horizontal wind speed component for ZP 597 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) 18 Figure 8: Bin-wise comparison of the horizontal wind speed component for ZP 597 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) 19 DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 1 1 INTRODUCTION DNV GL has prepared an independent analysis and report of a ZephIR Lidar performance verification. In this analysis and report, the ZephIR Lidar with the serial number ZP 597 will be discussed. The verification measurements for this device were performed by ZephIR Ltd. at their test site in Pershore, UK between 2016-12-23 to 2017-01-31. The met tower was equipped with classical anemometry components (cup anemometers, wind vanes etc.) serving as the verification reference for the Lidar wind speed and wind direction comparisons. Those comparisons were performed in line with a Remote Sensing (RS) best practice verification approach as developed within the EU-FP7-Projekt NORSEWInD [2] against corresponding Key Performance Indicators (KPIs) and Acceptance Criteria (ACs; compare Appendix A). In addition, a performance verification and uncertainty calculation is carried out in accordance with the second edition of the reviewed IEC 61400-12-1 standard, Annex L [4]. DNV GL is accredited according to ISO 17025 for measurements on wind turbines and for wind resource measurements and energy assessments. DNV GL is also a full member of the network of measurement institutes in Europe ‘MEASNET’. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 2 2 DESCRIPTION OF THE TEST SITE 2.1 The test site The following description and figures of the Pershore test site, which is a disused air field, are taken from a technical report by ZephIR Ltd. [1]: The terrain in the vicinity of the mast is flat and covered with sparse low growing vegetation. A number of hangars and outbuildings exist in sectors between 260° and 317° at distances between 300m and 700m from the mast. These buildings are estimated not to exceed 14m in height. Approximately 500 m to the North-East lies the small village of Throckmorton which consists of a few scattered farms and houses. 700 m to the South-West of the mast between 190° and 240° lies an area of spoil heaps and filtration pools associated with a mining operation. On a wider scale the site is surrounded by flat arable land that is devoid of any dense closed canopy forest. The larger conurbations of Pershore and Evesham lie at distances of 5km and 9km to the South West and South East respectively. Figure 1: Map of the Pershore test site near Throckmorton, UK. The position of the reference mast is marked by a red dot. The site specifications given in the above description have been verified during a site visit by a DNV GL expert on 2015-09-01, see [8]. Further details on the site are given in [1], a 360° photo round is shown in Appendix B. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 3 2.2 Measuring equipment In the following sections technical details and specifications of the measuring equipment are described. This description covers the meteorological reference mast (met mast) including its sensors and data acquisition system as well as the tested Lidar. The following items regarding the meteorological measurement systems have been verified during the above mentioned site visit: • Site suitability and exact positions of mast and Lidar test stand • Mast height, measurement levels and boom orientations • Distribution and mounting of sensors at the mast • Validity of MEASNET [6] calibrations of cups and correct application of calibration factors and offsets • Wind vane offsets • Data acquisition components, logger configuration • Data storage and data provision 2.2.1 Meteorological mast: layout, sensors and data acquisition The following description is taken from [1]: The mast has been constructed to be fully compliant with the 2005 edition of IEC 61400-12-1 [3] and the terrain of the test site falls within requirements for testing without a site calibration. All cup anemometers installed on the reference mast are class 1A instruments as defined by [3] and have undergone individual rotor specific MEASNET [6] calibration at a MEASNET certified wind tunnel. All boom and upright dimensions have been determined using the lattice porosity and mast dimensions provided by the manufacturer and in compliance with [3] to operate within a maximum flow distortion of 0.5% at the wind measurement locations. The directional vanes are installed with their North markings aligned along the booms towards the mast. The boom orientation is compensated for in the data logger. The main mast installation documents (as presented in [1]) are included for reference in Appendix B and the instrument calibration certificates are included in Appendix E. Those calibrations belong to the most recently changed anemometers (see Table 1), hence being valid for the wind speed sensors of the met tower during this verification campaign. The met mast is a guyed 90.5 m triangular lattice tower with a face width of 0.7 m. The MEASNET calibrated [6] cup anemometers (cups) of type Vector Instruments A 100 LM and Thies Frist Class Advanced (TFCA) are mounted on booms aside the mast at heights of 20.5 m, 45.5 m and 70.5 m and in a top mounting position at 90.5 m A.G.L., see Figure 2. Those mounting arrangements are consistent with the IEA [Error! Reference source not found.] and IEC [3] recommendations for the use of cup anemometry at masts. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 4 Figure 2: Schematic of the sensor level and boom distribution at the 90.5 m mast, as taken from [1]. See Table 1 for sensor distribution according to the alphanumeric label per boom (A to N) and the actually valid serial numbers. The legend in Figure 2 describes the sensor at each positionTable 1 lists the sensors operating during the campaign period. Respective calibration certificates for each sensor are given in Appendix E. The photo in Figure 2 shows mast anemometry levels between 20 and 90.5 m AGL. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 5 The position of the test stand (Lidar / met mast) in terms of the WGS84 standard is: • Lat N 52° 08' 35" • Lon W 002° 02' 14" Label A E F G H K L Channel WS_2R WS_4R WS_3V WS_6R WS_5V WS_8R WS_7V Thies First Thies First Thies First Thies First Vector Vector Vector Model Class Class Class Class A100LM A100LM A100LM Advanced Advanced Advanced Advanced S/N 0916477 8920 08157941 08157939 11202 11203 08157940 Installation Date 19/11/2016 30/10/2014 29/10/2015 29/10/2015 30/10/2014 30/10/2014 29/10/2015 Height 91.5 70.5 70.5 45.5 45.5 20.5 20.5 Orientation (⁰ ) Mast 300 300 120 300 120 300 120 to Instrument Calibration 14/10/2016 22/08/2014 17/08/2015 17/08/2015 22/08/2014 22/08/2014 17/08/2015 Date DWG Slope 0.04596 0.09714 0.046 0.04602 0.09767 0.09745 0.04603 Offset 0.2519 0.2096 0.2568 0.2448 0.1875 0.1839 0.2482 Calibration - 30/08/2014 05/10/2015 05/10/2015 30/08/2014 30/08/2014 05/10/2015 Date SOH Slope - 0.09825 0.04688 0.04677 0.09893 0.09915 0.04687 Offset - 0.12616 0.17228 0.1863 0.1049 0.12139 0.18083 Slope 0.04596 0.097695 0.04644 0.04648 0.0983 0.0983 0.04645 Applied Offset 0.2519 0.16788 0.2145 0.2014 0.1462 0.152645 0.2145 Table 1: List of meteorological sensors and individual anemometers installed at the mast during verification campaign, as of Appendix B, and list of calibration factors for cup anemometers. The valid calibration certificates are attached to this report in Appendix E. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 6 2.2.2 The ZephIR Lidar The Lidar under test is a ZephIR of type Z 300 Doppler Wind Lidar, employing a CW laser (continuous wave laser) that has specifically been designed to measure wind speeds at heights in the boundary layer of the atmosphere. The serial number of the lidar device is ZP 597. During the verification campaign the Lidar system was configured to record wind speed measurements at 11 different levels between 10 and 300 m. The actual Lidar measurement heights were 10, 20, 38, 45, 70, 91, 120, 149, 200, 250 and 300 m above ground. The four heights at 20, 45, 70 and 91 m were used for the comparison to the cup/mast reference measurements. Figure 3 shows an array of ZephIR Lidars under test being typically located to the East of the base of the met mast, and Table 2 lists wind speed and wind direction measurement and comparison levels as given and selected for the performance verification. Figure 3: Typical setup of ZephIR Lidars next to the reference mast at Pershore. Height Settings (relative to ground level) ZP300 Meas. 10 20 38 45 70 91 120 149 200 250 300 Levels [m] Mast/WS-Cup 20 45 70 91 Levels [m] Mast/WD-Vane 44 88 Levels [m] 1 Standard height in ZephIR ZP300 devises (automatically recorded) Table 2: Height settings of ZP300 Lidar and reference mast. Levels for wind speed and wind direction comparisons are highlighted in bold letters. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 7 3 LIDAR PERFORMANCE VERIFICATION (LPV) APPROACH 3.1 Common test conditions and data filtering In the process of the LPV trial the following test conditions and filters are applied • All comparisons are based on 10-minute average wind values returned from wind vanes and MEASNET calibrated cup anemometers installed on the reference mast (primary reference) and concurrent wind direction and wind speed data from the Lidar under test. • All data collected during periods of possible icing at cup anemometers, i.e. temperatures below 2 °C and humidity of above 80% are excluded. • All data collected during periods of precipitation (i.e. when precipitation is detected by the watch sensor with a ten minute averaged period) are excluded. • All other reported data (particularly wind speed) within undisturbed free-stream wind direction sector relative to the reference mast as well to the Lidar are used in the comparison analysis. • For the validation of Lidar wind speeds against the mast the wind speeds from the cup anemometers at 20 m, 45 m, 70 m and 91 m are used. The Lidar data are selected according to the sector screening of the cup data prior to comparison, see following section. • 3.2 Sector filtering The orientation of cup carrying booms at the mast is to the North West at one side and to the South East on the other side. Hence, wind speed data needs to be screened at wind directions between 85° and 155° for cups on the Northwest side of the mast and between 265° and 335° for cups on the Southeast side of the mast. This sector screening of 70° per boom directions accounts for downwind mast wake effects on the boom mounted instruments, see sector sketch in Figure 4. Top goal-post mounted Sonic and cup anemometers at 91.5 m are treated in a similar way such that the wind data is screened from the other anemometers’ wake effects. However, data from only the 91.5 m top mounted cup anemometer is used in this case. If cup data from both boom directions is available (i.e. for wind directions out of the remaining two sectors), the wind speed average of the two oppositely mounted instruments is used as reference for the comparison with the Lidar wind speeds. Wind data is further screened within the two disturbance sectors whereby wind speed data from a single cup, i.e. from the one mounted on the upwind directed boom is considered valid, only. Wind speed data measured at 91.5 m and 70.5 m are screened against wind direction data at the 88 m vane. Instruments at the 45.5 m and 20.5 m are screened against directional data from the 43.5 m vane. For the validation of ZephIR wind speeds against the mast, only wind speeds from the cup anemometers are used as reference. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 8 Figure 4: Wind direction sectors used to select undisturbed wind speed data from oppositely arranged cup carrying booms for comparison. 3.3 Lidar specific filtering An automatic, first-pass filter was applied to the lidar data whereby: • The vertical windspeed was capped between 2 and -2 m/s; • The tilt angle was limited to a maximum of 3°; • The minimum count of points was 18. All recorded data at any given timestamp where these criteria are not satisfied is excluded. A maximum horizontal speed limit was not applied given manual horizontal wind speed filtering will be applied in the analysis. 3.4 Data coverage requirements for accuracy assessment The following data coverage definitions are prescribed for the LPV: • The overall minimum number of 10 minute data points after filtering (according to sections 3.1 and 3.2) for the WS ranges [all > 3 m/s] and [4 to 16 m/s] should not be lower than 600. • At least 200 10-minute data points should to be in the WS range between 4 and 8 m/s and 200 data points between 8 and 12 m/s. These data coverage requirements are regarded as achievable for a typical test period of 4 weeks. 3.5 LPV evaluation The performance of the LIDAR under test is evaluated for its system and data availability as well as for its wind data accuracy, based on a number of Key Performance Indicators (KPI) and according Acceptance Criteria (AC). DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 9 The evaluation approach in terms of the applicable KPIs and according ACs is outlined in Appendix A, where KPIs and ACs for system and data availability are listed in Table 13 those for wind data quality in Table 14. The performance assessment of the given KPIs and respective Acceptance Criteria regarding Availability and Accuracy is executed at each reference level present, in this case at each of the four (4) met tower’s 1st Class reference anemometry levels which are 20 m, 45 m, 70 m and 91 m above ground level. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 10 4 RESULTS For the treated Lidar Performance Verification (LPV) campaign data were provided for the period from 2016-12-09, 11:30 until 2017-01-31, 23:50. So the campaign was completed after 39.6 days. The wind speed ranges covered and used for comparison are 0.7 to 18.7 m/s at the upper level (91 m) and 0.7 to 14.4 m/s at the lower level (20 m). # of Data Points per Height Level WS range ZPH597 20 45 70 91 All >= 3m/s 2509 2880 3107 3179 4 - 8 m/s 1612 1833 1932 1831 8 - 12 m/s 250 418 638 826 4 - 16 m/s 1880 2311 2677 2811 Table 3: Number of 10 minute data points after filtering used for WS comparison at each of the four (4) levels. The completeness requirements as of section 3.4 are fulfilled for all WS ranges. 4.1 System availability The system availability as applied to the Lidar device is defined by a percentage of the maximum possible number of ten-minute periods within the above mentioned total campaign duration. The total number of 10 min intervals for the campaign duration of ZP 597 is 5704 corresponding to 39.6 days. As Lidar ten-minute data entries were present (regardless of the data validity) the Lidar device achieved a system availability of almost 100 % see table below. ✓ The Acceptance Criterion for Data Availability (KPI SACA) to be ≥95 % is successfully met. LIDAR ZPH 597 Period 23/12/2016 09:20 to 31/01/2017 23:50 Test height / m 20 45 70 91 Max. # of 10-min points in period 5704 5704 5704 5704 Data present 5701 5701 5701 5701 System availability 99.9% 99.9% 99.9% 99.9% Total # of 10-min valid data 5564 5534 5439 5357 Data availability 97.5% 97.0% 95.4% 93.9% # after ext filtering for WD, WS 1917 2028 2163 2195 Data availability for comparison 33.6% 35.6% 37.9% 38.5% Table 4: Summary of system and data availabilities for ZP 597 at respective heights DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 11 4.2 Data availability Table 4 shows the data availability for the treated comparison measurement levels between 20 and 91 m A.G.L. A data availability of 93.9 % to 97.5 % relative to the maximum possible number of ten-minute periods. ✓ The Acceptance Criterion for Data Availability (KPI DACA) to be ≥90 % is successfully met for all measurement levels. Data for individual heights were treated as available when they show a numeric value in contrast to a value being flagged as NaN (not a number), 9999, or not fully operational. The difference in number of available data between the rows “system” and “data availability” as shown in Table 4 reflect the reduction of valid data according to internal system filtering. 4.3 Data filtering The data from both the Lidar and the mast were filtered for external parameters: • wind direction to avoid non-valid wind speed sectors being influenced by e.g. mast wake effects, (refer to section 3.2); • wind speed, clipping wind speeds below 3 m/s; and • air temperature and humidity measurements recorded at 43.5 m and 88 m respectively at the mast were used to filter out potential icing of the mast cup anemometers. WS and wind direction (WD) measurements were excluded whenever the temperature dropped below 2° and relative humidity exceeded 80 %. After the application of those filters the number of ten-minute data points remaining to be processed was reduced to a percentage between 44.0 % at 20 m and 55.7 % at 91 m (refer to Table 4). 4.4 Wind speed comparison Cup anemometers are regarded as the current industry standard for wind speed measurements at wind farm sites. Measurements with cup anemometers must therefore be considered the standard reference against which any new measurement device needs to be judged. Wind speed as treated in this LPV process are assessed by means of Linear Regressions through the origin of the form y = m x + b and b=:0 between Lidar (y-axis) wind speeds and cup (x-axis) wind speeds for the four mentioned heights, and were derived from the comparison of data from the following wind speed ranges a) 4 to 16 m/s 1 b) all above 3 m/s according to the following acceptance criteria 1) slope (m) (KPI Xmws) between 0.98 and 1.02 for all WS ranges a) and b) 2) R2 > 0.97 (KPI R2mws) for all WS ranges a) and b) as prescribed in and Appendix A. 1 In consistency with the IEC bin selection criteria the actual range spans from 3.75 to 16.25 since 4 m/s and 16 m/s are the central points of the corresponding 0.5 m/s wide bins. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 12 This campaign represents a series performance test of a technology proven Remote Sensing device. As the test campaign was limited in WS coverage for natural reasons, the core verification concentrates on a subset of statistically meaningful performance criteria (in terms of amount of available representative data) being treated relevant for acceptance. Results of wind speed comparisons The time series of wind speeds as recorded by the Lidar (for all 5 pre-set heights) covering 39.6 days, is overlapped by that of the met mast system. Two comparison heights (70m and 91m) are shown in Appendix C. Table 5 summarizes the wind speed regression results at all four (4) comparison heights showing that the ZephIR Lidar achieves a high level of accuracy compared to the cups at respective heights in terms of regression slopes (m) which are close to unity and good regression coefficient R2 (KPI R2mws). Figure 5 shows the corresponding regression plots for the wind speed range >= 3 m/s. The mean Lidar wind speeds averaged over all used values (KPI Cmwsd) resemble those of the cups very closely (see columns 5 and 6 from Table 5), yielding very low relative Campaign Mean WS Differences (KPI Cmwsd) at all measurement heights and WS ranges.Table 6 reflects the results according to the absolute wind speed error criterion. It shows that for the wind speed range of 3 to 16 m/s at all heights between 20 to 91 m, a small fraction of data ranging between 0.5% and 5.1 % of concurrent ten-minute data points exceed the prescribed wind speed difference threshold of 0.5 m/s which is below the allowed upper limit of 10 %. With respect to the linear WS regressions the following KPI’s Acceptance Criteria are passed ✓ Regression slope (KPI Xmws) between 0.98 and 1.02 at all treated levels and for all WS ranges; meeting the Acceptance Criteria. ✓ R2 (KPI R2mws) > 0.97 at all treated levels for both the WS ranges a) [all > 3 m/s] and b) [4 to 16 m/s]; meeting the Acceptance Criteria. ✓ The Acceptance Criterion for the relative Campaign Mean Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is successfully passed at the three highest assessment levels (45, 70 and 91 m) in both WS ranges. Furthermore, the following wind speed related Acceptance Criteria were met: ✓ Absolute Wind Speed Difference (KPI Awsd) at all comparison levels and for all analysed wind speed data between 3 and 16 m/s, see Table 6. The following deviation from applicable test conditions and performance criteria are reported. o The Acceptance Criterion for the relative Campaign Mean Wind Speed (KPI Cmwsd) is missed at the lowest measurement level (20 m) for both WS ranges. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 13 WS-avg WS-avg Mean rel. mean # values slope R2 20 m Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 1917 0.989 0.995 5.66 5.60 0.06 1.06% 4 - 16 m/s 1514 0.988 0.995 6.22 6.16 0.07 1.11% WS-avg WS-avg Mean rel. mean # values Slope R2 45 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 2028 0.991 0.995 6.29 6.24 0.05 0.82% 4 - 16 m/s 1669 0.991 0.994 6.87 6.81 0.06 0.89% WS-avg WS-avg Mean rel. mean # values slope R2 70 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R2mws KPI Cmwsd All >= 3m/s 2163 0.993 0.995 6.87 6.81 0.06 0.82% 4 - 16 m/s 1880 0.993 0.995 7.31 7.25 0.06 0.78% WS-avg WS-avg Mean rel. mean # values slope R2 91 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 2195 0.993 0.992 7.63 7.56 0.07 0.91% 4 - 16 m/s 1969 0.993 0.991 7.96 7.90 0.06 0.78% Table 5: Regression results comparison for ZPH 597; acceptance relevant results are colour shaded. Note the regression lines are forced through the origin. ZPH 594 Criterion for > 0.5 m/s for 3 to 16 m/s > 2% above 16 m/s abs WS error KPI Awsd Height Level total # identified # fraction % total # identified # fraction % 20 1917 10 0.5 0 0 N/A 45 2026 25 1.2 0 0 N/A ZPH597 70 2156 57 2.6 6 1 16.7 91 2178 99 4.5 13 4 30.8 Table 6: Summary of absolute wind speed differences between cups and Lidar DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 14 Figure 5: Plots of linear wind speed regression results for 20, 45, 70 and 91 m (note that regression slopes are forced through the origin) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 15 4.5 Wind direction comparison By comparing the wind direction as measured by the Lidar device at 91 m with the mast mounted wind vane at 88 m A.G.L., it is possible to see how well correlated the measures are, providing confidence that the Lidar is ‘seeing’ the same wind direction as the vane. In order to validate this comparison quantitatively a two variant regression solving for the slope m and the interception of the best-fit line with the y-axis b (according to y = m x + b) was performed, compare Appendix A. The results of such regression are shown in the x-y-plots in Figure 6 with the vane wind direction at 44 and 88 m on the x-axis and the Lidar direction at 45 and 91 m on the y-axis respectively. For this analysis the data was again filtered for Lidar and the cup wind speeds at 91 m, i.e. for WS >=3 m/s (to avoid false readings from the vane at low wind speeds), but not for possibly disturbed wind directions sectors. Note that a few 180° wind direction ambiguities were observed, when ZephIR Lidar data were correlated to the wind vane readings at 88 and 44 m (see Appendix D). These ambiguities were removed whenever there was a distinct 180° offset when compared with the mast vane wind direction measurements as the reference. This mast based correction is justified by the assumption, that a few 180° offset occurrences are related to lower wind speeds in combination with near ground site induced turbulence effects. Wind direction time series present during the course of the campaign period together with raw data correlations and WD distribution statistics can be found in Appendix D. Figure 6: Regression plot of wind direction comparisons at 44 m (left) and 88 m (right) The regression plots in Figure 6 reveal a close resemblance in measurement between the Lidar against the wind vanes for heights at 44 and 88 with an offset (in terms of a mean difference) of 3.1° at 45 m and 3.1° at 91 m. The directional offsets are within typical directional setup uncertainties for wind vanes and remotes sensing devices. Table 7 summarizes the WD comparison results for the acceptance relevant WD comparison levels at 88 and 44 m vanes, showing an equally good resemblance slope. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 16 WS filtering for WS > 3 m/s Height # values slope offset [⁰ ] R2 level - (KPI Xmwd) (KPI OFFmwd) (KPI R2mwd) 44 2161 0.997 3.089 0.982 ZPH 597 88 2289 0.992 3.110 0.982 Table 7: Summary of WD comparison results for both comparison levels ✓ The Acceptance Criteria for the respective KPIs for wind direction assessment (KPIs for Xmwd, OFFmwd, and R2mwd) are successfully passed. 4.6 Performance verification according to revised IEC standard, Annex L This subsection represents as a supplement to the standard Lidar DNV GL / NORSEWInD performance verification test with respect to a Remote Sensing Devices (RSD) validation approach as described in a draft version of the current edition of the IEC standard for power performance tests [4]. This approach is based on a wind speed bin averaged procedure in order to compare the horizontal wind speed measurements acquired by the RSD and the reference sensors at the mast. The objective of the IEC approach is to calculate the bin-wise deviation of the two sources and report the associated uncertainty. The bin averaging procedure was performed using 0.5 m/s wide wind speed bins centred on integers of from 4 to 16 m/s. In order to achieve statistical relevance this IEC approach requires • a minimum of three (3) 10-minute values available within each wind speed bin; and • a total amount of 180 hours of valid data (corresponding to a number of 1080 10-min values) Figure 7 shows the scatter plots of the wind speed comparison based on 10 min averages between the data pairs of the ZP 597 Lidar and the cups at 20, 45, 70 and 91 m. Additionally, the 10 minute averaged deviation for each data point of the two data sets is plotted (orange dots). Furthermore, the correlation coefficient, mean deviation and standard deviation of the deviations are shown in Table 8. The relative deviation of the data pairs was calculated in relation to the cup wind speeds as reference. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 17 Figure 7: Comparison of the horizontal wind speed component for ZP 597 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) Height Coeffcient of STD of Data Mean Deviation Level Determination Deviations Points [m] (R2) [m/s] [%] [%] # 20 0.9948 0.11 1.8 1.67 1608 45 0.9945 0.11 1.76 1.98 1755 ZPH597 70 0.9950 0.12 1.88 2.09 1829 91 0.9915 0.15 2.05 2.86 2026 Table 8: Statistical parameters of wind speed deviation DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 18 4.6.1 Performance verification uncertainty Bin-averaged wind speeds of ZP 594 RSD and the reference measurements is shown in Figure 8. The bin-averaged mean deviation (solid red line in the graphs) can be compared to the standard uncertainty of the cup anemometers combined with the statistical uncertainty of the comparison for each of the WS bins. Figure 8: Bin-wise comparison of the horizontal wind speed component for ZP 597 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 19 According to the IEC standard the verification uncertainty consists of five independent uncertainty components, which are summarized below: 1. Reference / anemometer uncertainty 2. Mean deviation of the remote sensor measurements and the reference measurements 3. Standard uncertainty of the measurement of the remote sensing device 4. Mounting uncertainty of the remote sensor at the verification test 5. Uncertainty due to non-homogenous flow The different uncertainty components are added in quadrature for each wind speed bin. The uncertainty due to non-homogenous flow between the measurement volume of the Lidar and at the met mast is assumed to be negligible due to the proximity of the Lidar to the mast and the benign terrain conditions at the Pershore test site. Details on the calculation of the separate uncertainty components are described in Appendix F. The results of the uncertainty calculation for the IEC compliant verification of the Lidar device at every comparison level are plotted in Figure 8. The finally combined uncertainties of the remote sensing RSD (VRSD) for the different WS bins and comparison levels show results values well below 2 % within most of the bins except for at the 20 m measurement level. For the current Lidar verification campaign the completeness requirement to yield 180 hours of valid and useable concurrent data (which translates into 7.5 days of data) in the WS range 4 and 16 m/s between the RSD and the reference cup is met for each comparison level. The additional data completeness requirement of yielding a minimum of 3 data pairs in each 0.5 m/s wide wind speed bin is fulfilled for most of the WS bins and comparison levels. Note that uncertainties are not calculated for wind speed bins with less than 3 data points. In Appendix G the environmental parameters - present during the performance verification test - are documented. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 20 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.96 184 3.99 4.61 3.50 0.193 0.014 -0.83 0.5 2.0 2.2 4.25 4.75 4.46 242 4.50 5.00 3.88 0.207 0.013 -0.95 0.5 1.8 2.1 4.75 5.25 4.97 243 5.02 5.41 4.32 0.174 0.011 -0.83 0.5 1.7 2.0 5.25 5.75 5.42 240 5.49 6.09 4.84 0.179 0.012 -1.32 0.5 1.6 2.2 5.75 6.25 5.89 152 5.97 6.24 5.29 0.192 0.016 -1.36 0.5 1.6 2.1 6.25 6.75 6.40 123 6.48 6.86 6.00 0.191 0.017 -1.15 0.5 1.5 2.0 6.75 7.25 6.92 101 6.99 7.45 6.36 0.187 0.019 -1.03 0.5 1.4 1.9 7.25 7.75 7.44 92 7.50 8.00 6.80 0.220 0.023 -0.83 0.5 1.4 1.7 7.75 8.25 7.89 63 7.98 8.36 7.54 0.199 0.025 -1.07 0.5 1.4 1.8 8.25 8.75 8.30 35 8.49 8.78 7.84 0.223 0.038 -2.22 0.5 1.3 2.7 8.75 9.25 8.85 35 8.95 9.38 8.38 0.243 0.041 -1.16 0.5 1.3 1.9 9.25 9.75 9.36 34 9.52 9.77 8.81 0.228 0.039 -1.71 0.5 1.3 2.2 9.75 10.25 9.84 29 9.96 10.33 9.36 0.216 0.040 -1.22 0.5 1.3 1.9 10.25 10.75 10.32 19 10.46 10.82 9.96 0.236 0.054 -1.32 0.5 1.3 2.0 10.75 11.25 10.82 11 10.98 11.26 10.38 0.254 0.076 -1.45 0.5 1.2 2.1 11.25 11.75 11.36 9 11.56 11.62 11.10 0.172 0.057 -1.75 0.5 1.2 2.2 11.75 12.25 11.85 7 11.95 12.05 11.67 0.130 0.049 -0.90 0.5 1.2 1.6 12.25 12.75 12.39 4 12.47 12.75 11.92 0.375 0.187 -0.64 0.5 1.2 2.1 12.75 13.25 * 2 * * * * * * * * * 13.25 13.75 13.42 4 13.49 13.76 13.12 0.320 0.160 -0.52 0.5 1.2 1.8 13.75 14.25 * 1 * * * * * * * * * 14.25 14.75 14.14 4 14.37 14.42 13.74 0.317 0.158 -1.58 0.5 1.2 2.3 14.75 15.25 * 0 * * * * * * * * * 15.25 15.75 * 0 * * * * * * * * * 15.75 16.25 * 0 * * * * * * * * * * Insufficient number of data points for uncertainty calculations Table 9: Uncertainty calculation for ZP 597 at 20 m level DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 21 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.99 162 4.00 4.58 3.42 0.210 0.016 -0.25 0.5 2.0 2.1 4.25 4.75 4.47 165 4.50 5.27 3.89 0.240 0.019 -0.65 0.5 1.8 2.0 4.75 5.25 4.97 164 5.03 5.37 4.23 0.209 0.016 -1.18 0.5 1.7 2.2 5.25 5.75 5.45 163 5.49 5.95 4.88 0.207 0.016 -0.73 0.5 1.6 1.9 5.75 6.25 5.99 182 6.03 6.37 5.47 0.180 0.013 -0.59 0.5 1.5 1.7 6.25 6.75 6.41 215 6.49 6.97 5.75 0.202 0.014 -1.19 0.5 1.5 2.0 6.75 7.25 6.92 166 6.98 7.38 6.36 0.186 0.014 -0.86 0.5 1.4 1.8 7.25 7.75 7.42 117 7.49 8.01 6.79 0.202 0.019 -0.88 0.5 1.4 1.7 7.75 8.25 7.92 87 8.00 8.33 7.48 0.189 0.020 -0.95 0.5 1.4 1.8 8.25 8.75 8.40 73 8.49 8.81 6.08 0.333 0.039 -1.09 0.5 1.3 1.9 8.75 9.25 8.90 70 8.98 9.31 8.32 0.186 0.022 -0.98 0.5 1.3 1.7 9.25 9.75 9.39 48 9.49 9.80 9.00 0.190 0.027 -0.99 0.5 1.3 1.7 9.75 10.25 9.88 39 10.00 10.65 9.38 0.237 0.038 -1.24 0.5 1.3 1.9 10.25 10.75 10.36 34 10.48 10.89 9.87 0.267 0.046 -1.19 0.5 1.2 1.9 10.75 11.25 10.86 27 11.00 11.24 10.11 0.258 0.050 -1.34 0.5 1.2 1.9 11.25 11.75 11.36 19 11.52 11.87 10.86 0.240 0.055 -1.33 0.5 1.2 1.9 11.75 12.25 11.85 7 11.89 12.22 11.53 0.241 0.091 -0.39 0.5 1.2 1.6 12.25 12.75 12.43 7 12.50 12.60 12.03 0.214 0.081 -0.58 0.5 1.2 1.6 12.75 13.25 12.84 13 13.01 13.20 12.52 0.215 0.060 -1.29 0.5 1.2 1.9 13.25 13.75 13.36 6 13.40 13.80 13.03 0.309 0.126 -0.28 0.5 1.2 1.6 13.75 14.25 14.09 4 14.07 14.53 13.69 0.356 0.178 0.15 0.5 1.2 1.8 14.25 14.75 * 2 * * * * * * * * * 14.75 15.25 14.86 3 15.06 14.88 14.83 0.026 0.015 -1.32 0.5 1.2 1.8 15.25 15.75 * 1 * * * * * * * * * 15.75 16.25 15.64 3 16.01 15.95 15.37 0.294 0.170 -2.34 0.5 1.1 2.9 * Insufficient number of data points for uncertainty calculations Table 10: Uncertainty calculation for ZP 597 at 45 m level DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 22 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.94 149 3.99 4.58 3.03 0.281 0.023 -1.23 0.5 2.0 2.5 4.25 4.75 4.43 158 4.48 5.17 3.51 0.272 0.022 -1.31 0.5 1.8 2.4 4.75 5.25 4.94 159 5.01 5.70 4.25 0.216 0.017 -1.25 0.5 1.7 2.2 5.25 5.75 5.46 153 5.52 5.96 4.73 0.219 0.018 -1.10 0.5 1.6 2.1 5.75 6.25 5.93 168 5.99 6.32 5.32 0.204 0.016 -1.07 0.5 1.6 2.0 6.25 6.75 6.45 149 6.51 6.86 5.92 0.194 0.016 -0.96 0.5 1.5 1.9 6.75 7.25 6.92 186 6.98 7.64 5.95 0.218 0.016 -0.90 0.5 1.4 1.8 7.25 7.75 7.42 188 7.49 7.95 6.93 0.183 0.013 -0.86 0.5 1.4 1.7 7.75 8.25 7.92 149 7.98 8.34 7.18 0.200 0.016 -0.71 0.5 1.4 1.6 8.25 8.75 8.39 126 8.47 9.13 7.97 0.211 0.019 -0.96 0.5 1.3 1.7 8.75 9.25 8.91 92 8.99 9.41 7.79 0.247 0.026 -0.90 0.5 1.3 1.7 9.25 9.75 9.49 52 9.50 10.09 8.89 0.238 0.033 -0.10 0.5 1.3 1.4 9.75 10.25 9.91 58 10.00 10.57 9.48 0.220 0.029 -0.92 0.5 1.3 1.7 10.25 10.75 10.36 44 10.46 10.82 10.08 0.187 0.028 -0.93 0.5 1.3 1.7 10.75 11.25 10.93 49 10.99 11.42 10.50 0.214 0.031 -0.49 0.5 1.2 1.4 11.25 11.75 11.51 26 11.53 11.84 11.00 0.216 0.042 -0.13 0.5 1.2 1.4 11.75 12.25 11.90 25 11.99 12.25 11.24 0.232 0.046 -0.78 0.5 1.2 1.6 12.25 12.75 12.47 16 12.47 12.85 12.11 0.265 0.066 -0.01 0.5 1.2 1.4 12.75 13.25 13.00 11 13.00 13.50 12.62 0.254 0.077 -0.01 0.5 1.2 1.4 13.25 13.75 13.50 10 13.62 13.81 13.26 0.186 0.059 -0.84 0.5 1.2 1.6 13.75 14.25 13.93 11 14.00 14.26 13.53 0.193 0.058 -0.50 0.5 1.2 1.4 14.25 14.75 14.51 6 14.47 15.11 14.28 0.307 0.125 0.34 0.5 1.2 1.6 14.75 15.25 15.26 3 15.02 15.47 15.09 0.193 0.112 1.64 0.5 1.2 2.2 15.25 15.75 * 2 * * * * * * * * * 15.75 16.25 * 2 * * * * * * * * * * Insufficient number of data points for uncertainty calculations Table 11: Uncertainty calculation for ZP 597 at 70 m level DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 23 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.91 118 4.02 4.71 3.08 0.361 0.033 -2.75 0.5 2.0 3.5 4.25 4.75 4.38 119 4.51 5.09 3.15 0.364 0.033 -2.82 0.5 1.8 3.5 4.75 5.25 4.94 102 5.02 5.38 3.93 0.207 0.021 -1.52 0.5 1.7 2.4 5.25 5.75 5.42 142 5.52 6.05 4.37 0.261 0.022 -1.83 0.5 1.6 2.5 5.75 6.25 5.94 137 6.01 6.45 4.83 0.257 0.022 -1.09 0.5 1.5 2.0 6.25 6.75 6.43 163 6.52 7.01 5.10 0.272 0.021 -1.27 0.5 1.5 2.0 6.75 7.25 6.91 151 7.00 7.59 5.13 0.376 0.031 -1.38 0.5 1.4 2.1 7.25 7.75 7.42 174 7.49 7.93 5.70 0.297 0.023 -0.97 0.5 1.4 1.8 7.75 8.25 7.95 178 8.00 9.28 7.40 0.195 0.015 -0.63 0.5 1.4 1.6 8.25 8.75 8.42 152 8.50 10.02 7.50 0.264 0.021 -0.93 0.5 1.3 1.7 8.75 9.25 8.88 130 9.00 9.34 5.91 0.332 0.029 -1.29 0.5 1.3 1.9 9.25 9.75 9.40 105 9.47 10.38 8.98 0.227 0.022 -0.74 0.5 1.3 1.6 9.75 10.25 9.93 56 9.99 10.44 9.36 0.222 0.030 -0.56 0.5 1.3 1.5 10.25 10.75 10.49 55 10.50 13.13 9.93 0.411 0.055 -0.09 0.5 1.2 1.4 10.75 11.25 10.98 56 11.03 11.43 10.25 0.243 0.033 -0.42 0.5 1.2 1.4 11.25 11.75 11.47 49 11.50 12.02 11.04 0.238 0.034 -0.23 0.5 1.2 1.4 11.75 12.25 11.92 37 11.97 12.51 11.33 0.293 0.048 -0.37 0.5 1.2 1.4 12.25 12.75 12.51 38 12.50 13.28 12.00 0.267 0.043 0.08 0.5 1.2 1.3 12.75 13.25 12.94 27 13.00 13.58 11.93 0.367 0.071 -0.45 0.5 1.2 1.5 13.25 13.75 13.47 25 13.47 13.94 12.91 0.233 0.047 -0.03 0.5 1.2 1.3 13.75 14.25 14.07 14 14.05 14.46 13.58 0.258 0.069 0.16 0.5 1.2 1.4 14.25 14.75 14.52 18 14.51 14.99 13.91 0.313 0.074 0.04 0.5 1.2 1.4 14.75 15.25 15.13 7 15.01 15.47 14.73 0.240 0.091 0.81 0.5 1.2 1.6 15.25 15.75 15.55 8 15.41 15.84 15.17 0.215 0.076 0.88 0.5 1.2 1.6 15.75 16.25 16.13 7 15.97 16.58 15.75 0.305 0.115 1.00 0.5 1.1 1.8 * Insufficient number of data points for uncertainty calculations Table 12: Uncertainty calculation for ZP 597 at 91 m level DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 24 5 IMPORTANT REMARKS AND LIMITATIONS Independently performed Lidar Performance Verifications (LPV) of individual Lidar devices as reported in this document present a reasonable means to assure overall system integrity of the Lidar unit after manufacturing, and are meant to give an indication of the quality of wind data produced by the Lidar. Furthermore, the IEC compliant bin-wise uncertainty implementation may serve as a traceable means to judge the uncertainty of the RSD as determined from a well-defined verification process. Any statement given in the context of system integrity and data quality related results within this report are limited to the given test site conditions, to the prevailing atmospheric (in particular wind) conditions and to the specific Lidar configuration as selected for this LPV campaign. For sites with non-benign terrain and atmospheric conditions, an LPV is not thought to replace the requirement for an on-site verification of a Lidar in real field campaigns. In this situation it may be necessary to conduct measurements in close proximity to an on-site mast over a reasonable period. . DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 25 6 CONCLUSION A ZephIR 300 Lidar and cup anemometer wind measurements were carried out at the Pershore UK Remote Sensing Test Site to validate Lidar wind data quality against a well-known high quality standard cup anemometer. Measurement heights of 20 m, 45 m, 70 m and 91 m were available for wind speed correlations (88/91 m for wind direction correlation) between a proximate met mast and a ZephIR 300 Lidar with the serial number ZP 597. The duration of the validation campaign was 39.6 days. The test period and wind data coverage is considered sufficient for the purposes of characterising the wind data performance of the ZephIR Lidar in the context of a Lidar Performance Verification. The overall system availability for the mentioned total campaign duration of 36.9 days for ZP 597 is 99.9 %. The data availabilities at the selected Lidar measurement levels 20 m, 45 m, 70 m and 91 m was in the range of 93.9 % to 97.5 %. These data coverage figures are relative to the number of maximum possible ten-minute data points for the total duration of the campaign. Wind speed (and direction) correlations were carried out for each of the four WS measurement heights (one for WD) mentioned above. The wind speeds of both techniques at all treated heights correlated very well, showing a very low level of scatter and an excellent resemblance of Lidar wind speeds to those of cups, in terms of linear regression slopes. In summary the following KPI related Acceptance Criteria are met. ✓ The Acceptance Criterion for System Availability (KPI SACA) to be ≥95 % is successfully passed (Table 4). ✓ The Acceptance Criterion for Data Availability (KPI DACA) to be ≥90 % is successfully met at all assessment levels (Table 4). ✓ Regression slope (KPI Xmws) between 0.98 and 1.02 at all treated levels and for all WS ranges, meeting the Acceptance Criteria (Table 5, column 2). ✓ R2 (KPI R2mws) > 0.97 at all treated levels for the WS ranges a) [all WS > 3 m/s]and b) [4 to 16 m/s], meeting the Acceptance Criteria (Table 5, column 3). ✓ The Acceptance Criterion for the relative Campaign Mean Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is successfully passed at the 45, 70 and 91 m assessment levels for both WS ranges. ✓ Absolute Wind Speed Difference (KPI Awsd) at all comparison levels and for all analysed wind speed data between 3 and 16 m/s where wind speed difference of greater than 0.5 m/s makes up < 5 % of the full dataset (Table 6). ✓ The Acceptance Criteria for the respective KPIs for wind direction assessment (KPIs for Xmwd, OFFmwd, and R2mwd) are successfully passed at all comparison levels. The following KPI related Acceptance Criteria are not met. o The Criterion for the relative Campaign Means Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is missed at the 20 m measurement level. The performance verification and uncertainty calculation has also been carried out in accordance with the IEC standard yielding a traceable uncertainty measure. The following deviation from the applicable IEC test conditions are reported: DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 26 o During the verification campaign the bins for the following measurement levels cannot be calculated given insufficient data points: ▪ 13 m/s, 14 m/s and 15 m/s onwards for the 20 m measurement level; ▪ 14.5 m/s and 15.5 m/s bins for the 45 m measurement level; ▪ Bins greater than 15 m/s for the 70 m measurement level. In summary, this Pershore validation campaign indicates that the ZephIR 300 Lidar with the serial number ZP 597 is able to reproduce cup anemometer wind speeds and wind vane directions at an accurate and acceptable level. DNV GL considers that the ZephIR 300 Lidar device under test (with the serial number ZP 597) can be used for formal wind potential and long-term wind resource assessments. Specifically, DNV GL concludes that this Lidar may be employed as a standalone measurement system – replacing a conventional met mast – given the following criteria are met: (1) The Lidar is deployed in a location with benign terrain and atmospheric conditions. (2) The long term data accuracy stability is assured by either: ▪ recording data for a period sufficient to obtain an adequate in-situ correlation to an onsite reference (e.g. a short met. mast); ▪ or – in case of lack of a suitable in-situ reference – by performing a post deployment performance verification campaign, provided a continuous system operation during the preceding deployment period. Finally, DNV GL recommends, that care needs to be taken with respect to the formal use of Lidar turbulence and extreme wind speed measures, not treated in this report but known to be different from classical anemometry measures. DNV GL likes to point out that good measurement and data collection practices need to be maintained for all wind speed measurements, be they Lidar or more conventional anemometry. Therefore, special care needs to be exercised in the transportation, installation and on-going maintenance of the Lidar as it may be exposed to a wide range of environmental conditions at different sites over time. A key element of any formal wind study is the traceability of the wind speed data uncertainty. Hence, a strict uncertainty assessment (which is not part of this report) should be employed. Furthermore it is recommended that thorough practices of documenting the salient features of Lidar installation and maintenance are instigated from the outset. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 27 7 REFERENCES 1. Barker, W., E. Burin des Roziers and S. Wylie, "Pershore: 91m Anemometer Mast Specification ", by ZephIR Ltd., issued: 05/11/2014. 2. Kindler, D., " Best Practice Test and Verification Procedure for Wind LiDARs on the Høvsøre Test Site”, GL GH-D Report WT 6960/09 for EU-Project NORSEWInD, Deliv. 1.1, June 2009. 3. International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power Performance Measurements of Grid Connected Wind Turbines. Ed. 1. International Electronic Commission. 3, rue de Varembé Geneva. Switzerland, Dec. 2005. 4. International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power performance measurements of electricity producing wind turbines. Ed. 2. CD. International Electronic Commission.3, June. 2013. 5. IEA Expert Group Study on Recommended Practices for Wind Turbine Testing and Evaluation 11. Wind speed measurement and use of cup anemometry, 1. Edition 1999. 6. MEASNET: “Cup Anemometer Calibration Procedure”. Version 1, September 1997. 7. W. Barker, E. Burin des Roziers, "Pershore: 91m Anemometer Mast Specification and Data Validation, Revision 2", by ZephIR Ltd., issued: 10/10/2012. 8. Stein, D. “Technical Note of Inspection of ZephIR’s Reference Met Mast and Lidar Test Site (exec. 2015 - 09-01) at Pershore/Throckmorton, UK” DNV GL Report, No. GLGH-4257 15 13307 267 T-0001, Rev. B, 2016-05-02. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 28 8 GLOSSARY The following table lists abbreviations and acronyms used in this report. Abbreviation Meaning Acronym AC Acceptance Criterion a.g.l. Above ground level DNV GL New company name, successor of legacy GL GH IEC International Electro-technical Commission IEA International Energy Agency GH-D GL Garrad Hassan Deutschland GmbH KPI Key Performance Indicator MM Meteorological Mast PAR Performance Assessment Requirement LPV Lidar Performance Verification TI Turbulence Intensity WD Wind direction WS Wind speed DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 29 KEY PERFORMANCE INDICATORS AND ACCEPTANCE CRITERIA, IN LINE WITH [2] Table 13: List of KPIs and ACs relevant for System and Data Availability assessment Acceptance Criteria KPI Definition / Rationale across total campaign duration SACA System Availability ≥95% The Lidar system is ready to function according to specifications and to deliver data, taking into account all time stamped data entries in the output data files including flagged data (e.g. by NaNs or 9999s) for the pre-defined total campaign length. The System Availability is the number of these time stamped data entries relative to the maximum possible number of data entries (for 10 minute intervals) within the pre-defined total campaign period. (Any conditions affecting the test’s data availability outside of the LIDAR system’s control is not to be included in this calculation. Such as: power outages, acts of nature causing system damage, communication outages, maintenance, etc.) DACA Data Availability ≥90% The Data Availability is defined as the number of valid data points returned by the Lidar unit as compared to maximum number of possible points that can be acquired during the test (Any conditions affecting the test’s data availability outside of the LIDAR system’s control is not to be included in this calculation. Such as: power outages, acts of nature causing system damage, communication outages, maintenance, etc.) MV Number of Maintenance Visits N/A Number of Visits to the Lidar system by either the manufacturer or an authorized third party to maintain and service the system. This is to be documented and reported. UO Number of Unscheduled Outages N/A Number Unscheduled Outages of the Lidar system in addition to scheduled service outages. Each outage needs to be documented regarding possible cause of outage, exact time / duration and action performed to overcome the Unscheduled outage. This is to be reported. CU Uptime of Communication System N/A To be documented and reported by the manufacturer. In the above table, during periods of maintenance; the system is deemed unavailable. * Undisturbed sectors: this means sectors with no significant flow distortion e.g. by wake effects of nearby wind turbines DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 30 Table 14: List of KPIs and ACs relevant for Wind Data Accuracy assessment Acceptance Criteria KPI Definition / Rationale Cmwsd Campaign Mean Wind Speed – Difference <1% Absolute difference of mean wind speeds between Lidar and reference as measured over the whole verification campaign duration, expressed as percentage relative to the Campaign Mean Wind Speed A threshold is imposed on the Difference. Analysis shall be applied to wind speed ranges a) 4 to 16 m/s b) all above 3 m/s given achieved data coverage requirements. Awsd Absolute Wind Speed Differences a) < 0.5 m/s Absolute 10 minute mean wind speed differences between Lidar and reference for all data points treated after filtering. b) within 5% A threshold is imposed on the Difference. Not more than 10% of Analysis shall be applied to wind speed ranges data to exceed the a) 3 to 16 m/s criteria above. b) all above 16 m/s given achieved data coverage requirements. Xmws Mean Wind Speed – Slope 0.98 – 1.02 Slope returned from single variant regression with the regression analysis constrained to pass through the origin. A tolerance is imposed on the Slope value. Analysis shall be applied to wind speed ranges a) all > 3 m/s b) 4 to 16 m/s given achieved data coverage requirements. R2mws Mean Wind Speed – Coefficient of Determination >0.97 Correlation Co-efficient returned from single variant regression A threshold is imposed on the Correlation Co-efficient value. Analysis shall be applied to wind speed ranges a) all > 3m/s b) 4 to 16 m/s given achieved data coverage requirements. Xmwd Mean Wind Direction – Slope 0.97 – 1.03 Slope returned from a two-variant regression. A tolerance is imposed on the Slope value. Analysis shall be applied to a) all wind speeds above 3 m/s regardless of coverage requirements. OFFmwd Mean Wind Direction – Offset (absolute value) < 5° In terms of the campaign mean difference in wind direction (same as for Mmwd) R2mwd Mean Wind Direction – Coefficient of Determination > 0.97 (same as for Mmwd) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 31 PERSHORE/THROCKMORTON MET MAST DETAILS 360° Panorama Photos, taken on 2015-09-01, see inspection report [8]: Met Mast Photo: DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 32 Met. Mast Sketch: Met. Mast Sensor Distribution Table: DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 33 TIME SERIES OF WIND SPEED Wind Speed time series for 91 m at the mast and ZP 597 including temperature recorded at the 88 m temperature and pressure sensor. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 34 WIND DIRECTION WD mast time series at 88 m and 44 m wind vane levels: DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 35 X-Y-plot of raw wind direction data for WS > 3 m/s (blue dots) and 180° ambiguity cleaned data (red dots) between wind vane and Lidar measures at 88/91 and 44/45 m DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 36 Mast ZP 597 10% 20% 30% 40% 10% 20% 30% 40% 91 m 0-3 3-6 6-9 >9m/s 0-3 3-6 6-9 >9m/s 45 m 10% 20% 30% 40% 10% 20% 30% 40% 0-3 3-6 6-9 >9m/s 0-3 3-6 6-9 >9m/s DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 37 CUP CALIBRATION CERTIFICATES, TAKEN FROM[6] Thies First Class Advanced Cup S/N 09164177 at 91.5 m, 300° orientation (Deutsche WindGuard Calibration) Vector A100 LM Cup S/N 8920 at 70.5 m, 300° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 38 Vector A100 LM Cup S/N 8920 at 70.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 39 Thies First Class Advanced Cup S/N 08157941 at 70.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 40 Thies First Class Advanced Cup S/N 08157941 at 70.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 41 Vector A100LM Cup S/N 11202 at 45.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 42 Vector A100LM Cup S/N 11202 at 45.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 43 DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 44 Thies First Class Advanced Cup S/N 08157939 at 45.5 m, 300° orientation (Deutsche WindGuard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 45 Thies First Class Advanced Cup S/N 08157939 at 45.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 46 Thies First Class Advanced Cup S/N 08157940 at 20.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 47 Thies First Class Advanced Cup S/N 08157940 at 20.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 48 Vector A100LM Cup S/N 11203 at 20.5 m, 300° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 49 Vector A100LM Cup S/N 11203 at 20.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 50 IEC ANNEX L UNCERTAINTY ANALYSES 1. Reference / anemometer uncertainty The anemometer uncertainty of the specific reference heights is calculated based on the wind tunnel calibration of the individual anemometer, the anemometer classification and the mounting effect at the met tower. 2. Mean deviation of the remote sensor measurements and the reference measurements This is the relative deviation between the bin averages of the RSD and the mast reference measurement divided by with the reference measurement. 3. Standard uncertainty of the measurement of the remote sensing device The standard deviation of the measurements was divided by the square root of the number of data records per bin. The relative uncertainty was calculated by dividing the value by the bin average wind speed of the mast (reference) measurement. 4. Mounting uncertainty of the remote sensor at the verification test The uncertainty of the remote sensing device due to non-ideal levelling was estimated to be 0.5 %. 5. Uncertainty due to non-homogenous flow The Lidar device is located in close proximity of the met tower just a few m to the East of the tower base. As a result the uncertainty due to non-homogenous flow within the measurement volume is considered to be negligible. DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 51 ENVIRONMENTAL PARAMETERS ACCORDING TO IEC ANNEX L, DURING THE CAMPAIGN, REF. 70 M LEVEL # PiA Quality BIN BIN Shear exponenet TI Wind direction Temperature Air Density Wind Veer Vmm Factor Lower Upper 70 m -20 m at 70 m at 88 m at 44 m at 44 m 88 m - 44 m at 70 m Average Average Std Average Std Average Std Average Std Average Std Average Std Average Std [m/s] [m/s] [m/s] [-] [-] [-] [-] [#] [#] [°] [°] [°C] [°C] [kg/m3] [kg/m3] [°] [°] 3.75 4.25 3.99 0.48 0.28 0.098 0.047 34.4 1.4 171.9 82.7 3.4 2.7 1.286 0.020 7.2 12.7 4.25 4.75 4.49 0.48 0.28 0.093 0.044 34.4 1.4 197.0 85.1 3.6 2.7 1.282 0.019 7.5 10.3 4.75 5.25 5.01 0.42 0.25 0.099 0.044 34.2 1.4 222.0 87.2 4.1 2.9 1.278 0.020 6.9 7.9 5.25 5.75 5.51 0.37 0.22 0.104 0.045 34.2 1.5 215.8 81.8 4.5 2.9 1.274 0.018 5.6 6.0 5.75 6.25 6.00 0.35 0.19 0.098 0.041 34.3 1.5 221.1 72.4 4.5 2.9 1.271 0.018 6.2 5.8 6.25 6.75 6.51 0.31 0.17 0.097 0.037 34.2 1.4 229.8 68.6 5.0 2.9 1.269 0.018 5.7 5.5 6.75 7.25 6.99 0.28 0.14 0.100 0.034 34.1 1.5 222.2 60.6 5.4 2.6 1.264 0.018 5.7 4.8 7.25 7.75 7.49 0.28 0.10 0.096 0.035 34.3 1.4 231.0 61.0 5.4 2.8 1.263 0.018 6.4 5.1 7.75 8.25 7.98 0.28 0.10 0.094 0.033 34.5 1.3 239.0 62.9 5.2 2.8 1.263 0.017 6.1 4.9 8.25 8.75 8.48 0.27 0.10 0.096 0.034 34.6 1.1 241.2 56.1 5.6 3.0 1.263 0.018 5.7 4.6 8.75 9.25 8.99 0.24 0.08 0.104 0.034 34.6 1.1 241.6 48.0 6.6 3.0 1.257 0.014 4.0 2.2 9.25 9.75 9.50 0.22 0.07 0.111 0.032 34.3 1.4 246.9 59.2 6.8 3.2 1.255 0.014 3.3 2.1 9.75 10.25 9.98 0.23 0.06 0.105 0.027 34.6 1.1 245.8 43.2 7.2 2.7 1.255 0.012 3.6 2.1 10.25 10.75 10.46 0.22 0.05 0.119 0.031 34.5 1.0 251.7 38.7 7.4 3.0 1.253 0.013 3.3 1.6 10.75 11.25 10.99 0.19 0.05 0.121 0.040 34.3 1.1 247.0 32.6 8.4 2.7 1.252 0.013 2.8 2.0 11.25 11.75 11.54 0.19 0.05 0.130 0.033 34.4 0.9 249.0 31.7 9.4 2.8 1.246 0.013 2.4 2.3 11.75 12.25 11.99 0.19 0.05 0.130 0.032 34.5 1.0 253.3 37.4 9.1 3.3 1.245 0.012 2.1 1.5 12.25 12.75 12.48 0.19 0.03 0.135 0.030 34.5 0.9 252.7 34.4 9.4 2.7 1.244 0.009 2.0 1.0 12.75 13.25 12.97 0.18 0.04 0.145 0.037 34.5 0.8 258.2 35.6 9.5 1.8 1.244 0.007 2.1 1.0 13.25 13.75 13.60 0.19 0.05 0.137 0.037 34.8 1.1 259.7 43.1 8.8 2.0 1.245 0.008 1.6 0.9 13.75 14.25 14.03 0.16 0.04 0.146 0.028 34.7 1.0 241.3 42.5 9.4 2.2 1.244 0.008 2.1 0.9 14.25 14.75 14.50 0.17 0.03 0.137 0.022 33.3 2.0 222.4 34.8 10.1 0.9 1.240 0.005 1.8 0.7 14.75 15.25 15.01 0.18 0.04 0.147 0.022 34.1 2.0 261.0 51.4 9.0 2.8 1.240 0.010 1.3 0.9 15.25 15.75 15.43 0.17 0.03 0.160 0.024 34.3 1.7 255.5 55.2 8.8 2.7 1.243 0.009 1.4 1.2 15.75 16.25 16.07 0.16 0.06 0.131 0.020 33.6 2.1 220.1 41.9 10.4 0.9 1.237 0.001 2.0 0.6 DNV GL – Report No. 702909-AUME-R-07, Rev. B – www.dnvgl.com Page 52 ABOUT DNV GL Driven by our purpose of safeguarding life, property and the environment, DNV GL enables organizations to advance the safety and sustainability of their business. We provide classification and technical assurance along with software and independent expert advisory services to the maritime, oil and gas, and energy industries. We also provide certification services to customers across a wide range of industries. Operating in more than 100 countries, our 16,000 professionals are dedicated to helping our customers make the world safer, smarter and greener. RENEWABLE ENERGY WIND MAPPING FOR THE MALDIVES LIDAR Site Installation Report - Hoarafushi The World Bank Document No.: 702909-AUME-R04 Issue : C, Status : FINAL Date: 6 March 2018 IMPORTANT NOTICE AND DISCLAIMER 1. This document is intended for the sole use of the Client as detailed on the front page of this document to whom the document is addressed and who has entered into a written agreement with the DNV GL entity issuing this document (“DNV GL”). To the extent permitted by law, neither DNV GL nor any group company (the "Group") assumes any responsibility whether in contract, tort including without limitation negligence, or otherwise howsoever, to third parties (being persons other than the Client), and no company in the Group other than DNV GL shall be liable for any loss or damage whatsoever suffered by virtue of any act, omission or default (whether arising by negligence or otherwise) by DNV GL, the Group or any of its or their servants, subcontractors or agents. This document must be read in its entirety and is subject to any assumptions and qualifications expressed therein as well as in any other relevant communications in connection with it. This document may contain detailed technical data which is intended for use only by persons possessing requisite expertise in its subject matter. 2. This document is protected by copyright and may only be reproduced and circulated in accordance with the Document Classification and associated conditions stipulated or referred to in this document and/or in DNV GL’s written agreement with the Client. No part of this document may be disclosed in any public offering memorandum, prospectus or stock exchange listing, circular or announcement without the express and prior written consent of DNV GL. A Document Classification permitting the Client to redistribute this document shall not thereby imply that DNV GL has any liability to any recipient other than the Client. 3. This document has been produced from information relating to dates and periods referred to in this document. This document does not imply that any information is not subject to change. Except and to the extent that checking or verification of information or data is expressly agreed within the written scope of its services, DNV GL shall not be responsible in any way in connection with erroneous information or data provided to it by the Client or any third party, or for the effects of any such erroneous information or data whether or not contained or referred to in this document. 4. Any wind or energy forecasts estimates or predictions are subject to factors not all of which are within the scope of the probability and uncertainties contained or referred to in this document and nothing in this document guarantees any particular wind speed or energy output. KEY TO DOCUMENT CLASSIFICATION For disclosure only to named individuals within the Client’s Strictly Confidential : organisation. For disclosure only to individuals directly concerned with the Private and Confidential : subject matter of the document within the Client’s organisation. Commercial in Confidence : Not to be disclosed outside the Client’s organisation. DNV GL only : Not to be disclosed to non-DNV GL staff Distribution for information only at the discretion of the Client (subject to the above Important Notice and Disclaimer and the Client’s Discretion : terms of DNV GL’s written agreement with the Client). Available for information only to the general public (subject to Published : the above Important Notice and Disclaimer). Project name: Renewable Energy Wind Mapping for the Maldives DNV GL - Energy Report title: LIDAR Site Installation Report - Hoarafushi Renewables Advisory Customer: The World Bank, 9665 Chesapeake Drive, Suite 435 1818 H Street, N.W. San Diego, CA 92123 Washington, DC 20433 Tel: 703-795-8103 Contact person: Sandeep Kohli Enterprise No.: 94-340223694- Date of issue: 6 March 2018 340223694-3402236 Project No.: 702909 Document No.: 702909-AUME-R04 Issue/Status: C / FINAL Task and objective: provide a permanent record of the site characteristics and measurement equipment for the LIDAR at Hoarafushi. Prepared by: Verified by: Approved by: Fowzi Dahhan Kevin Bleibler Trenton Gilbert Engineer, Renewables Advisory Head of Section, Measurements Head of Section, Developer Support Renewables Advisory Services (Pacific), Renewables Advisory ☐ Strictly Confidential Keywords: ☐ Private and Confidential World Bank, ESMAP, Maldives, wind, measurement, ☐ Commercial in Confidence LIDAR, Site Installation Report ☐ DNV GL only ☒ Client’s Discretion ☐ Published Reference to part of this report which may lead to misinterpretation is not permissible. Issue Date Reason for Issue Prepared by Verified by Approved by A 9 Jun 2017 PRELIMINARY DRAFT FD KB TG B 30 Aug 2017 Revised version - DRAFT FD KB TG C 6 Mar 2018 FINAL FD KB TG Table of contents 1 INTRODUCTION .............................................................................................................. 2 2 SITE INFORMATION ........................................................................................................ 2 2.1 Site Location 2 2.2 Site Description 4 3 SITE EQUIPMENT ............................................................................................................ 4 3.1 LIDAR Unit 4 3.2 Auxiliary Meteorological Station 8 3.3 Power Supply 11 3.4 Communications 13 4 OPERATIONS AND MAINTENANCE ................................................................................... 14 4.1 Data Acquisition and Management 14 4.2 Direction Measurements 14 4.3 Scheduled maintenance 15 4.4 Unscheduled Maintenance 15 4.5 Servicing and Re-verification 15 5 REFERENCES ................................................................................................................ 16 APPENDIX A – MAPPING ............................................................................................................. 17 APPENDIX B – PANORAMIC PHOTOGRAPHS OF SITE ...................................................................... 20 APPENDIX C – LIDAR DATA SHEET ............................................................................................... 22 APPENDIX D – LIDAR PERFORMANCE VERIFICATION REPORT .......................................................... 25 DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 4 1 INTRODUCTION The World Bank (the “Client”) has retained Garrad Hassan America, Inc. (“DNV GL”) to provide a validated mesoscale wind atlas for the Maldives, including associated deliverables and wind energy development training courses. During Phase 1 of the project, which has been completed, preliminary mesoscale mapping was carried out covering the entire country [1]. Phase 2 of the project, which is currently underway, involves the installation of two Light Detection and Ranging (LIDAR) based wind measurement sites in the country. Meteorological data collected at these sites over a two-year period will provide the basis for validating the mesoscale modeling outputs from Phase 1. LIDAR units and associated equipment have been commissioned at the two measurement sites, Hoarafushi and Thulusdhoo, in April 2017. This report documents the installation of the LIDAR site at Hoarafushi, and presents the characteristics of the site, as well as details of the measurement equipment, power supply and data acquisition system. 2 SITE INFORMATION 2.1 Site Location The Hoarafushi site is located on the east coast of the island of Hoarafushi in the Haa Alif Atoll, in the northern Maldives, approximately 320 km north of Male. A map of the site location is shown in Figure 2-1 below. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 2 Area shown Figure 2-1 Location of Hoarafushi LIDAR site DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 3 Figure A-1 in Appendix A presents a map of the site vicinity, showing 0 and 3 m elevation contours based on SRTM-1 data [2]. It should be noted that SRTM-1 data can be influenced by trees and buildings, and may not accurately reflect the ground elevation. Figure A-2 in Appendix A presents a map of estimated surface roughness zones in the vicinity of the site. Table 2-1 presents details of the site location coordinates and elevation. Table 2-1 Site location summary Coordinates Coordinates (Geographic)2 (UTM, Zone 43 N)2 Commissioning Elevation Site name Date [m ASL]1 Latitude Longitude Eastings Northings [degrees] [degrees] [m] [m] Hoarafushi 9 April 2017 0 to 3 m 6.98341 72.89785 267756 772433 1. Approximate elevation only, due to inaccuracy of SRTM-1 data 2. Datum: WGS 84 2.2 Site Description The site is situated within a diesel electricity generation compound belonging to FENAKA Corporation. The LIDAR location consists of a clear area bounded by garden beds to the north, south, and west, and some trees to the east with a height of approximately 6 m. There are two buildings within the compound located to the south and northeast. Outside the compound to the west there is a line of tall trees running from north to south with a height of approximately 15 m. The surrounding terrain to the west is flat with trees and low-lying buildings. The shoreline lies to the east of the site. Panoramic photographs of the site are shown in Appendix B. 3 SITE EQUIPMENT 3.1 LIDAR Unit The LIDAR unit installed at the Hoarafushi site is a ZephIR 300 LIDAR remote sensing device. The device employs a continuous wave laser to measure horizontal and vertical wind speed and wind direction at a specified range of heights. Table 3-1 provides a summary of key information about the LIDAR unit. Additional details on the LIDAR specifications are provided in Appendix C. Table 3-1 LIDAR unit summary Manufacturer Model Serial Number Zephir Ltd ZephIR 300 594 The unit is secured to a concrete pad footing and is enclosed by fencing. Figure 3-1 to Figure 3-3 show photographs of the LIDAR unit as installed on site. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 4 Figure 3-1 LIDAR unit at Hoarafushi – general view of enclosure Figure 3-2 View within enclosure DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 5 Figure 3-3 Close up of LIDAR unit The LIDAR unit is equipped with a washing system for cleaning the window, as shown in Figure 3-4 below. The washing system is connected to an external cleaning fluid supply bottle, which can be seen in Figure 3-2. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 6 Figure 3-4 Top of LIDAR unit showing sensor window and washing system The LIDAR simultaneously records 10-minute average, maximum, minimum, and standard deviation statistics for the horizontal and vertical wind speeds, and average and standard deviation statistics for wind direction, at eleven specified measurement heights between 10 m and 200 m above the top of the unit. The LIDAR measurement configuration is summarised in Table 3-2 below. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 7 Table 3-2 LIDAR configuration summary Measurement Measurement Measurement heights Unit Channels parameter interval [m AGL] Mean / minimum / Horizontal wind m/s maximum / 11, 20, 30, 39, speed standard deviation 10 minutes 50, 60, 80, (continuous 100, 120, 150, Vertical wind speed m/s Mean scan) 200 Wind direction1 Degrees Mean 1. The unit is aligned at a bearing of approximately 3° from true north. The LIDAR unit has been subjected to independent testing and performance verification in accordance with the second edition of the reviewed IEC 61400-12-1 standard, Annex L [2][3]. A copy of the performance verification report is presented in Appendix D. 3.2 Auxiliary Meteorological Station The LIDAR unit also encompasses a small auxiliary meteorological station to verify the wind direction measured by the LIDAR and record other atmospheric parameters including wind speed, temperature, air pressure, relative humidity and precipitation. The meteorological station is mounted on a separate pole installed adjacent to the LIDAR enclosure so as to minimise the effect of any flow disruption that may be caused by surrounding buildings and other obstructions. Figure 3-5 and Figure 3-6 below show photographs of the pole in relation to the LIDAR enclosure, and Figure 3-7 shows a close up photograph of the meteorological station taken during the installation process. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 8 Figure 3-5 LIDAR unit enclosure with adjacent pole carrying meteorological station DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 9 Figure 3-6 LIDAR unit enclosure with adjacent pole carrying meteorological station – side view Figure 3-7 Close up of meteorological station DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 10 The meteorological station measurement configuration is summarised in Table 3-3. Table 3-3 Meteorological station configuration summary Measurement height Measurement Measurement Unit Channels [m AGL] parameter interval Air temperature °C Mean Pressure Millibar Mean Relative humidity % Mean 71 10 minutes Precipitation % Mean Wind speed m/s Mean Wind direction2 Degrees Mean 1. Approximate height. 2. Meteorological station aligned approximately to magnetic north. 3.3 Power Supply The LIDAR unit is powered by the local mains electricity supply, with back-up power provided by a battery system consisting of two 12 V, 100 Ah deep cycle rechargeable batteries, which are float charged by the mains supply via a charge regulator. The batteries are intended to provide a backup power supply for between 12 and 24 hours in the event of a mains power supply interruption, and are housed within a weatherproof enclosure mounted on the concrete pad adjacent to the LIDAR device, as shown in Figure 3-2 above. A more detailed view of the battery enclosure is shown in Figure 3-8. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 11 Figure 3-8 Battery enclosure adjacent to LIDAR unit The batteries located inside the enclosure are shown in Figure 3-9. Figure 3-9 Batteries inside enclosure adjacent to LIDAR unit A schematic of the battery enclosure configuration is shown in Figure 3-10. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 12 Figure 3-10 Schematic of battery enclosure configuration Table 3-4 summarizes the key specifications of the power supply system. Table 3-4 Power supply system summary Battery Battery Charger 2 x 12V 100 Ah, AGM deep cycle Victron Energy Blue Power 12/25 - 12V, 25A 3.4 Communications The system uses an external 3G modem and directional antenna for data transmission. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 13 4 OPERATIONS AND MAINTENANCE The LIDAR unit will operate for a period of two years. 4.1 Data Acquisition and Management All recorded data is sent on a daily basis to DNV GL for evaluation and archiving. The data is transmitted automatically to an email address which is the data receipt portal for the DNV GL “Resource Panorama” service. The data is then subjected to regular quality control by a data analyst. DNV GL will provide both raw and quality controlled data from the LIDAR device. 4.2 Direction Measurements The Zephir 300 LIDAR device uses measurements from both its scanning laser and met station to measure the wind direction. A laser scan determines the axis of the wind direction. The direction of the wind along that axis is then determined in conjunction with the met station reading. The following table illustrates the process, for a northeast wind: Laser Measurement Met Station Wind Direction Reported Wind Direction If the met direction in the above diagram pointed in the opposite direction, the reported direction would be the other axial possibility in the first diagram. For example: Laser Measurement Met Station Wind Direction Reported Wind Direction The Met Station direction reading only needs to be very coarse for this purposes as is only used to determine the direction ‘phase’. However, if the met station is shadowed or if it is placed in areas of wind recirculation then there is the potential for a 180 degrees direction ambiguity to result. Given the proximity of the LIDAR device and met station to buildings, trees and other obstacles, the potential for a 180 degree ambiguity in the wind direction measurements exists. Care should be taken when using the raw data from the LIDAR device that instances where the 180 degree direction ambiguity exists are corrected. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 14 DNV GL will correct periods where there is evidence of a 180 degree ambiguity in the quality controlled data set provided. 4.3 Scheduled maintenance Regular inspections and maintenance of the LIDAR device will be undertaken by trained local staff. The purpose of the inspections will be to: a) Check the operating status of the LIDAR device b) Check the LIDAR device and ancillary equipment have not been damaged or soiled c) Clean the LIDAR viewing window d) Refill the water reservoir e) Report findings to DNV GL 4.4 Unscheduled Maintenance If a LIDAR unit shows signs of fault or failure, or if any issues with the communication system are detected, as flagged through remote analysis of data, a corrective maintenance program will be initiated. Once initiated, the corrective maintenance program will involve a multi-level approach as follows. As a first step, local staff on the islands will perform a visual inspection of the unit and correct any issues if possible. If further inspection, troubleshooting, or maintenance of the unit is required, a corrective maintenance team consisting of in-country trained personnel from DNV GL’s Local Partner will visit the site. In cases where the issue cannot be resolved by local staff, and where appropriate, expert personnel from DNV GL will travel from Australia to the Maldives to perform the necessary corrective maintenance. DNV GL will remotely assist with the interventions at all stages of the process. 4.5 Servicing and Re-verification The manufacturer recommends that the LIDAR device is returned to the manufacturer every 3 years for servicing and re-verification. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 15 5 REFERENCES [1] “Mesoscale Wind Modeling Report 1- Interim wind atlas for Maldives”, DNV GL, 702909-AUME-R- 01-D, 2 July 2015. [2] National Aeronautics and Space Administration (NASA), “Shuttle Radar Topography Mission (SRTM) -1 arc second resolution”, data accessed using Global Mapper 18 software, 25 May 2017. [3] International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power performance measurements of electricity producing wind turbines. Ed. 2. CD. International Electronic Commission.3, June 2013. DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 16 APPENDIX A – MAPPING DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 17 Figure A-1 Map of Hoarafushi LIDAR site showing terrain elevation contours DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 18 Figure A-2 Map of Hoarafushi LIDAR site showing surface roughness zones DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 19 APPENDIX B – PANORAMIC PHOTOGRAPHS OF SITE DNV GL – Report No. 702909-AUME-R04, Rev. A, Status: FINAL – www.dnvgl.com Page 20 Panoramic view from location of LIDAR at Hoarafushi N W (267756, 772433) E S DNV GL – Report No. 702909-AUME-R04, Rev. A, Status: FINAL – www.dnvgl.com Page 21 APPENDIX C – LIDAR DATA SHEET DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 22 DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 23 DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 24 APPENDIX D – LIDAR PERFORMANCE VERIFICATION REPORT DNV GL – Report No. 702909-AUME-R04, Rev. C, Status: FINAL – www.dnvgl.com Page 25 ZP 594 Independent analysis and reporting of ZephIR Lidar performance verification at Pershore test site, including IEC compliant validation analysis The World Bank Report No.: 702909-AUME-R-06, Rev. B Date: 6 March 2018 IMPORTANT NOTICE AND DISCLAIMER 1. This document is intended for the sole use of the Client as detailed on the front page of this document to whom the document is addressed and who has entered into a written agreement with the DNV GL entity issuing this document (“DNV GL”). To the extent permitted by law, neither DNV GL nor any group company (the "Group") assumes any responsibility whether in contract, tort including without limitation negligence, or otherwise howsoever, to third parties (being persons other than the Client), and no company in the Group other than DNV GL shall be liable for any loss or damage whatsoever suffered by virtue of any act, omission or default (whether arising by negligence or otherwise) by DNV GL, the Group or any of its or their servants, subcontractors or agents. This document must be read in its entirety and is subject to any assumptions and qualifications expressed therein as well as in any other relevant communications in connection with it. This document may contain detailed technical data which is intended for use only by persons possessing requisite expertise in its subject matter. 2. This document is protected by copyright and may only be reproduced and circulated in accordance with the Document Classification and associated conditions stipulated or referred to in this document and/or in DNV GL’s written agreement with the Client. No part of this document may be disclosed in any public offering memorandum, prospectus or stock exchange listing, circular or announcement without the express and prior written consent of DNV GL. A Document Classification permitting the Client to redistribute this document shall not thereby imply that DNV GL has any liability to any recipient other than the Client. 3. This document has been produced from information relating to dates and periods referred to in this document. This document does not imply that any information is not subject to change. Except and to the extent that checking or verification of information or data is expressly agreed within the written scope of its services, DNV GL shall not be responsible in any way in connection with erroneous information or data provided to it by the Client or any third party, or for the effects of any such erroneous information or data whether or not contained or referred to in this document. 4. Any wind or energy forecasts estimates or predictions are subject to factors not all of which are within the scope of the probability and uncertainties contained or referred to in this document and nothing in this document guarantees any particular wind speed or energy output. KEY TO DOCUMENT CLASSIFICATION For disclosure only to named individuals within the Client’s Strictly Confidential : organisation. For disclosure only to individuals directly concerned with the Private and Confidential : subject matter of the document within the Client’s organisation. Commercial in Confidence : Not to be disclosed outside the Client’s organisation. DNV GL only : Not to be disclosed to non-DNV GL staff Distribution for information only at the discretion of the Client (subject to the above Important Notice and Disclaimer and the Client’s Discretion : terms of DNV GL’s written agreement with the Client). Available for information only to the general public (subject to Published : the above Important Notice and Disclaimer). DNV GL Project name: ZP 594 DNV GL - Energy Report title: Independent analysis and reporting of ZephIR Renewables Advisory Lidar performance verification at Pershore test Suite 25, Level 8, site, including IEC compliant validation analysis 401 Docklands Drive, Docklands, Customer: The World Bank, Victoria 3008, Australia 1818 H Street, N.W. Tel: +61 3 9600 1993 Washington, DC 20433 Contact person: Sandeep Kohli Date of issue: 6 March 2018 Project No.: 702909 Report No.: 702909-AUME-R-06, Rev. B Task and objective: Independent analysis and reporting of ZephIR Lidar performance verification at Pershore test site, including IEC compliant validation analysis Prepared by: Verified by: Approved by: M Quan F Dahhan T Gilbert Engineer Engineer Principal Engineer, Head of Section ☐ Strictly Confidential Keywords: ☐ Private and Confidential ZephIR, Lidar, performance verification ☐ Commercial in Confidence ☐ DNV GL only ☒ Client’s Discretion ☐ Published Reference to part of this report which may lead to misinterpretation is not permissible. Rev. No. Date Reason for Issue Prepared by Verified by Approved by A 30 Aug 2017 Draft (electronic version, only) M Quan F Dahhan, B Schmidt T Gilbert B 6 Mar 2018 Final M Quan F Dahhan, B Schmidt T Gilbert DNV GL Table of contents 1 INTRODUCTION .............................................................................................................. 2 2 DESCRIPTION OF THE TEST SITE...................................................................................... 3 2.1 The test site 3 2.2 Measuring equipment 4 2.2.1 Meteorological mast: layout, sensors and data acquisition 4 2.2.2 The ZephIR Lidar 7 3 LIDAR PERFORMANCE VERIFICATION (LPV) APPROACH ....................................................... 8 3.1 Common test conditions and data filtering 8 3.2 Sector filtering 8 3.3 Lidar specific filtering 9 3.4 Data coverage requirements for accuracy assessment 9 3.5 LPV evaluation 9 4 RESULTS ..................................................................................................................... 11 4.1 System availability 11 4.2 Data availability 11 4.3 Data filtering 12 4.4 Wind speed comparison 12 4.5 Wind direction comparison 16 4.6 Performance verification according to revised IEC standard, Annex L 17 4.6.1 Performance verification uncertainty 19 5 IMPORTANT REMARKS AND LIMITATIONS ........................................................................ 25 6 CONCLUSION ............................................................................................................... 26 7 REFERENCES ................................................................................................................ 28 8 GLOSSARY ................................................................................................................... 29 Appendices KEY PERFORMANCE INDICATORS AND ACCEPTANCE CRITERIA, IN LINE WITH [2] 30 PERSHORE/THROCKMORTON MET MAST DETAILS .................................................... 32 TIME SERIES OF WIND SPEED .............................................................................. 34 WIND DIRECTION ................................................................................................ 35 CUP CALIBRATION CERTIFICATES, TAKEN FROM[6] ................................................. 38 IEC ANNEX L UNCERTAINTY ANALYSES................................................................... 51 ENVIRONMENTAL PARAMETERS ACCORDING TO IEC ANNEX L, DURING THE CAMPAIGN, REF. 70 M LEVEL ......................................................................................... 52 DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page i LIST OF TABLES Table 1: List of meteorological sensors and individual anemometers installed at the mast during verification campaign, as of Appendix B, and list of calibration factors for cup anemometers. The valid calibration certificates are attached to this report in Appendix E. 6 Table 2: Level settings of ZP300 Lidar and reference mast. Levels for wind speed and wind direction comparisons are highlighted in bold letters. 7 Table 3: Number of 10 minute data points after filtering used for WS comparison at each of the four (4) levels. 11 Table 4: Summary of system and data availabilities for ZP 594 at respective levels 11 Table 5: Regression results comparison for ZPH 594; acceptance relevant results are colour shaded. Note the regression lines are forced through the origin. 14 Table 6: Summary of absolute wind speed differences between cups and Lidar 14 Table 7: Summary of WD comparison results for both comparison levels 17 Table 8: Statistical parameters of wind speed deviation 18 Table 9: Uncertainty calculation for ZP 594 at 20 m level 21 Table 10: Uncertainty calculation for ZP 594 at 45 m level 22 Table 11: Uncertainty calculation for ZP 594 at 70 m level 23 Table 12: Uncertainty calculation for ZP 594 at 91 m level 24 Table 13: List of KPIs and ACs relevant for System and Data Availability assessment 30 Table 14: List of KPIs and ACs relevant for Wind Data Accuracy assessment 31 LIST OF FIGURES Figure 1: Map of the Pershore test site near Throckmorton, UK. The position of the reference mast is marked by a red dot. 3 Figure 2: Schematic of the sensor level and boom distribution at the 90.5 m mast, as taken from [1]. See Table 1 for sensor distribution according to the alphanumeric label per boom (A to N) and the actually valid serial numbers. 5 Figure 3: Typical setup of ZephIR Lidars next to the reference mast at Pershore. 7 Figure 4: Wind direction sectors used to select undisturbed wind speed data from oppositely arranged cup carrying booms for comparison. 9 Figure 5: Plots of linear wind speed regression results for 20, 45, 70 and 91 m (note the regression results are forced through the origin) 15 Figure 6: Regression plot of wind direction comparisons at 45 m (left) and 88 m (right) 16 Figure 7: Comparison of the horizontal wind speed component for ZP 594 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) 18 Figure 8: Bin-wise comparison of the horizontal wind speed component for ZP 594 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) 19 DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 1 1 INTRODUCTION DNV GL has prepared an independent analysis and report of a ZephIR Lidar performance verification. In this analysis and report, the ZephIR Lidar with the serial number ZP 594 will be discussed. The verification measurements for this device were performed by ZephIR Ltd. at their test site in Pershore, UK between 2016-12-09 to 2017-01-31. The met tower was equipped with classical anemometry components (cup anemometers, wind vanes etc.) serving as the verification reference for the Lidar wind speed and wind direction comparisons. Those comparisons were performed in line with a Remote Sensing (RS) best practice verification approach as developed within the EU-FP7-Projekt NORSEWInD [2] against corresponding Key Performance Indicators (KPIs) and Acceptance Criteria (ACs; compare Appendix A). In addition, a performance verification and uncertainty calculation is carried out in accordance with the second edition of the reviewed IEC 61400-12-1 standard, Annex L [4]. DNV GL is accredited according to ISO 17025 for measurements on wind turbines and for wind resource measurements and energy assessments. DNV GL is also a full member of the network of measurement institutes in Europe ‘MEASNET’. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 2 2 DESCRIPTION OF THE TEST SITE 2.1 The test site The following description and figures of the Pershore test site, which is a disused air field, are taken from a technical report by ZephIR Ltd. [1]: The terrain in the vicinity of the mast is flat and covered with sparse low growing vegetation. A number of hangars and outbuildings exist in sectors between 260° and 317° at distances between 300m and 700m from the mast. These buildings are estimated not to exceed 14m in level. Approximately 500 m to the North-East lies the small village of Throckmorton which consists of a few scattered farms and houses. 700 m to the South-West of the mast between 190° and 240° lies an area of spoil heaps and filtration pools associated with a mining operation. On a wider scale the site is surrounded by flat arable land that is devoid of any dense closed canopy forest. The larger conurbations of Pershore and Evesham lie at distances of 5km and 9km to the South West and South East respectively. Figure 1: Map of the Pershore test site near Throckmorton, UK. The position of the reference mast is marked by a red dot. The site specifications given in the above description have been verified during a site visit by a DNV GL expert on 2015-09-01, see [8]. Further details on the site are given in [1], a 360° photo round is shown in Appendix B. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 3 2.2 Measuring equipment In the following sections technical details and specifications of the measuring equipment are described. This description covers the meteorological reference mast (met mast) including its sensors and data acquisition system as well as the tested Lidar. The following items regarding the meteorological measurement systems have been verified during the above mentioned site visit: • Site suitability and exact positions of mast and Lidar test stand • Mast level, measurement levels and boom orientations • Distribution and mounting of sensors at the mast • Validity of MEASNET [6] calibrations of cups and correct application of calibration factors and offsets • Wind vane offsets • Data acquisition components, logger configuration • Data storage and data provision 2.2.1 Meteorological mast: layout, sensors and data acquisition The following description is taken from [1]: The mast has been constructed to be fully compliant with the 2005 edition of IEC 61400-12-1 [3] and the terrain of the test site falls within requirements for testing without a site calibration. All cup anemometers installed on the reference mast are class 1A instruments as defined by [3] and have undergone individual rotor specific MEASNET [6] calibration at a MEASNET certified wind tunnel. All boom and upright dimensions have been determined using the lattice porosity and mast dimensions provided by the manufacturer and in compliance with [3] to operate within a maximum flow distortion of 0.5% at the wind measurement locations. The directional vanes are installed with their North markings aligned along the booms towards the mast. The boom orientation is compensated for in the data logger. The main mast installation documents (as presented in [1]) are included for reference in Appendix B and the instrument calibration certificates are included in Appendix E. Those calibrations belong to the most recently changed anemometers (see Table 1), hence being valid for the wind speed sensors of the met tower during this verification campaign. The met mast is a guyed 90.5 m triangular lattice tower with a face width of 0.7 m. The MEASNET calibrated [6] cup anemometers (cups) of type Vector Instruments A 100 LM and Thies Frist Class Advanced (TFCA) are mounted on booms aside the mast at levels of 20.5 m, 45.5 m and 70.5 m and in a top mounting position at 90.5 m A.G.L., see Figure 2. Those mounting arrangements are consistent with the IEA [5] and IEC [3] recommendations for the use of cup anemometry at masts. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 4 Figure 2: Schematic of the sensor level and boom distribution at the 90.5 m mast, as taken from [1]. See Table 1 for sensor distribution according to the alphanumeric label per boom (A to N) and the actually valid serial numbers. The legend in Figure 2 describes the sensor at each position. Table 1 lists the sensors operating during the campaign period. Respective calibration certificates for each sensor are given in Appendix E. The photo in Figure 2 shows mast anemometry levels between 20 and 90.5 m AGL. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 5 The position of the test stand (Lidar / met mast) in terms of the WGS84 standard is: • Lat N 52° 08' 35" • Lon W 002° 02' 14" Label A E F G H K L Channel WS_2R WS_4R WS_3V WS_6R WS_5V WS_8R WS_7V Thies First Thies First Thies First Thies First Vector Vector Vector Model Class Class Class Class A100LM A100LM A100LM Advanced Advanced Advanced Advanced S/N 0916477 8920 08157941 08157939 11202 11203 08157940 Installation Date 19/11/2016 30/10/2014 29/10/2015 29/10/2015 30/10/2014 30/10/2014 29/10/2015 Level 91.5 70.5 70.5 45.5 45.5 20.5 20.5 Orientation (⁰ ) Mast 300 300 120 300 120 300 120 to Instrument Calibration 14/10/2016 22/08/2014 17/08/2015 17/08/2015 22/08/2014 22/08/2014 17/08/2015 Date DWG Slope 0.04596 0.09714 0.046 0.04602 0.09767 0.09745 0.04603 Offset 0.2519 0.2096 0.2568 0.2448 0.1875 0.1839 0.2482 Calibration - 30/08/2014 05/10/2015 05/10/2015 30/08/2014 30/08/2014 05/10/2015 Date SOH Slope - 0.09825 0.04688 0.04677 0.09893 0.09915 0.04687 Offset - 0.12616 0.17228 0.1863 0.1049 0.12139 0.18083 Slope 0.04596 0.097695 0.04644 0.04648 0.0983 0.0983 0.04645 Applied Offset 0.2519 0.16788 0.2145 0.2014 0.1462 0.152645 0.2145 Table 1: List of meteorological sensors and individual anemometers installed at the mast during verification campaign, as of Appendix B, and list of calibration factors for cup anemometers. The valid calibration certificates are attached to this report in Appendix E. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 6 2.2.2 The ZephIR Lidar The Lidar under test is a ZephIR of type Z 300 Doppler Wind Lidar, employing a CW laser (continuous wave laser) that has specifically been designed to measure wind speeds at levels in the boundary layer of the atmosphere. The serial number of the lidar device is ZP 594. During the verification campaign the Lidar system was configured to record wind speed measurements at 11 different levels between 10 and 300 m. The actual Lidar measurement levels were 10, 20, 38, 45, 70, 91, 120, 149, 200, 250 and 300 m above ground. The four levels at 20, 45, 70 and 91 m were used for the comparison to the cup/mast reference measurements. Figure 3 shows an array of ZephIR Lidars under test being typically located to the East of the base of the met mast, and Table 2 lists wind speed and wind direction measurement and comparison levels as given and selected for the performance verification. Figure 3: Typical setup of ZephIR Lidars next to the reference mast at Pershore. Level Settings (relative to ground level) ZP300 Meas. 10 20 38 45 70 91 120 149 200 250 300 Levels [m] Mast/WS-Cup 20 45 70 91 Levels [m] Mast/WD-Vane 44 88 Levels [m] 1 Standard level in ZephIR ZP300 devises (automatically recorded) Table 2: Level settings of ZP300 Lidar and reference mast. Levels for wind speed and wind direction comparisons are highlighted in bold letters. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 7 3 LIDAR PERFORMANCE VERIFICATION (LPV) APPROACH 3.1 Common test conditions and data filtering In the process of the LPV trial the following test conditions and filters are applied: • All comparisons are based on 10-minute average wind values returned from wind vanes and MEASNET calibrated cup anemometers installed on the reference mast (primary reference) and concurrent wind direction and wind speed data from the Lidar under test. • All data collected during periods of possible icing at cup anemometers, i.e. temperatures below 2 °C and humidity of above 80% are excluded. • All data collected during periods of precipitation (i.e. when precipitation is detected by the watch sensor with a ten minute averaged period) are excluded. • All other reported data (particularly wind speed) within undisturbed free-stream wind direction sector relative to the reference mast as well to the Lidar are used in the comparison analysis. • For the validation of Lidar wind speeds against the mast the wind speeds from the cup anemometers at 20 m, 45 m, 70 m and 91 m are used. The Lidar data are selected according to the sector screening of the cup data prior to comparison, see following section. 3.2 Sector filtering The orientation of cup carrying booms at the mast is to the North West at one side and to the South East on the other side. Hence, wind speed data needs to be screened at wind directions between 85° and 155° for cups on the Northwest side of the mast and between 265° and 335° for cups on the Southeast side of the mast. This sector screening of 70° per boom directions accounts for downwind mast wake effects on the boom mounted instruments, see sector sketch in Figure 4. Top goal-post mounted Sonic and cup anemometers at 91.5 m are treated in a similar way such that the wind data is screened from the other anemometers’ wake effects. However, data from only the 91.5 m top mounted cup anemometer is used in this case. If cup data from both boom directions is available (i.e. for wind directions out of the remaining two sectors), the wind speed average of the two oppositely mounted instruments is used as reference for the comparison with the Lidar wind speeds. Wind data is further screened within the two disturbance sectors whereby wind speed data from a single cup, i.e. from the one mounted on the upwind directed boom is considered valid, only. Wind speed data measured at 91.5 m and 70.5 m are screened against wind direction data at the 88 m vane. Instruments at the 45.5 m and 20.5 m are screened against directional data from the 43.5 m vane. For the validation of ZephIR wind speeds against the mast, only wind speeds from the cup anemometers are used as reference. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 8 Figure 4: Wind direction sectors used to select undisturbed wind speed data from oppositely arranged cup carrying booms for comparison. 3.3 Lidar specific filtering An automatic, first-pass filter was applied to the lidar data whereby: • The vertical windspeed was capped between 2 and -2 m/s; • The tilt angle was limited to a maximum of 3°; • The minimum count of points was 18. All recorded data at any given timestamp where these criteria are not satisfied is excluded. A maximum horizontal speed limit was not applied given manual horizontal wind speed filtering will be applied in the analysis. 3.4 Data coverage requirements for accuracy assessment The following data coverage definitions are prescribed for the LPV: • The overall minimum number of 10 minute data points after filtering (according to sections 3.1 and 3.2) for the wind speed (WS) ranges [all > 3 m/s] and [4 to 16 m/s] should not be lower than 600. • At least 200 10-minute data points should to be in the WS range between 4 and 8 m/s and 200 data points between 8 and 12 m/s. These data coverage requirements are regarded as achievable for a typical test period of 4 weeks. 3.5 LPV evaluation The performance of the LIDAR under test is evaluated for its system and data availability as well as for its wind data accuracy, based on a number of Key Performance Indicators (KPI) and according Acceptance Criteria (AC). The evaluation approach in terms of the applicable KPIs and according ACs is outlined in Appendix A, where KPIs and ACs for system and data availability are listed in Table 13 those for wind data quality in Table 14. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 9 The performance assessment of the given KPIs and respective Acceptance Criteria regarding Availability and Accuracy is executed at each reference level present, in this case at each of the four (4) met tower’s 1st Class reference anemometry levels which are 20 m, 45 m, 70 m and 91 m above ground level. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 10 4 RESULTS For the treated Lidar Performance Verification (LPV) campaign data were provided for the period from 2016-12-09, 09:50 until 2017-01-31, 23:50. So the campaign was completed after 53.5 days. The wind speed ranges covered and used for comparison are 0.7 to 18.3 m/s at the upper level (91 m) and 0.7 to 14.5 m/s at the lower level (20 m). # of Data Points per Level Level WS range ZPH594 20 45 70 91 All >= 3m/s 3119 3854 4136 4230 4 - 8 m/s 2024 2559 2700 2579 8 - 12 m/s 248 469 771 1023 4 - 16 m/s 2293 3091 3572 3754 Table 3: Number of 10 minute data points after filtering used for WS comparison at each of the four (4) levels. The completeness requirements as of section 3.4 are fulfilled for all WS ranges. 4.1 System availability The system availability as applied to the Lidar device is defined by a percentage of the maximum possible number of ten-minute periods within the above mentioned total campaign duration. The total number of 10 min intervals for the campaign duration of ZP 594 is 7707 corresponding to 53.5 days. ✓ The Acceptance Criterion for System Availability (KPI SACA) to be ≥95 % is successfully met. LIDAR ZPH 594 Period 09/12/2016 11:30 to 31/01/2017 23:50 Test level / m 20 45 70 91 Max. # of 10-min points in period 7707 7707 7707 7707 Data present 7457 7457 7457 7457 System availability 96.8% 96.8% 96.8% 96.8% Total # of 10-min valid data 7289 7225 7070 6911 Data availability 94.6% 93.7% 91.7% 89.7% # after ext filtering for WD, WS 2533 2982 3159 2971 Data availability for comparison 32.9% 38.7% 41.0% 38.5% Table 4: Summary of system and data availabilities for ZP 594 at respective levels 4.2 Data availability Table 4 shows the data availability for the treated comparison measurement levels between 20 and 91 m A.G.L. A data availability of 89.7 % to 94.6 % is achieved relative to the maximum possible number of ten-minute periods. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 11 ✓ The Acceptance Criterion for Data Availability (KPI DACA) to be ≥90 % is successfully met for the 20, 45 and 70 m measurement levels. The following deviation from the applicable test condition is reported. o The Acceptance criterion for Data Availability (KPI DACA) to be ≥90 % is missed by a small margin at the 91 m measurement level. Data for individual levels were treated as available when they show a numeric value in contrast to a value being flagged as NaN (not a number), 9999, or not fully operational. The difference in number of available data between the rows “system” and “data availability” as shown in Table 4 reflect the reduction of valid data according to internal system filtering. 4.3 Data filtering The data from both the Lidar and the mast were filtered for external parameters: • wind direction to avoid non-valid wind speed sectors being influenced by e.g. mast wake effects, (refer to section 3.2); • wind speed, clipping wind speeds below 3 m/s; and • air temperature and humidity measurements recorded at 43.5 m and 88 m respectively at the mast were used to filter out potential icing of the mast cup anemometers. WS and wind direction (WD) measurements were excluded whenever the temperature dropped below 2° and relative humidity exceeded 80 %. After the application of those filters the number of ten-minute data points remaining to be processed was reduced to a percentage between 40.5 % at 20 m and 54.9 % at 91 m, (refer to Table 4). 4.4 Wind speed comparison Cup anemometers are regarded as the current industry standard for wind speed measurements at wind farm sites. Measurements with cup anemometers must therefore be considered the standard reference against which any new measurement device needs to be judged. Wind speed as treated in this LPV process are assessed by means of Linear Regressions through the origin of the form y = m x + b and b=:0 between Lidar (y-axis) wind speeds and cup (x-axis) wind speeds for the four mentioned heights, and were derived from the comparison of data from the following wind speed ranges a) 4 to 16 m/s 1 b) all above 3 m/s according to the following acceptance criteria 1) slope (m) (KPI Xmws) between 0.98 and 1.02 for all WS ranges a) and b) 2) R2 > 0.97 (KPI R2mws) for all WS ranges a) and b) as prescribed in and Appendix A. 1 In consistency with the IEC bin selection criteria the actual range spans from 3.75 to 16.25 since 4 m/s and 16 m/s are the central points of the corresponding 0.5 m/s wide bins. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 12 This campaign represents a series performance test of a technology proven Remote Sensing device. As the test campaign was limited in WS coverage for natural reasons, the core verification concentrates on a subset of statistically meaningful performance criteria (in terms of amount of available representative data) being treated relevant for acceptance. Results of wind speed comparisons The time series of wind speeds as recorded by the Lidar (for all 5 pre-set levels) covering 53.5 days, is overlapped by that of the met mast system. Two comparison levels (70m and 91m) are shown in Appendix C. Table 5 summarizes the wind speed regression results at all four (4) comparison levels showing that the ZephIR Lidar achieves a high level of accuracy compared to the cups at respective levels in terms of regression slopes (m) which are close to unity and good regression coefficient R2 (KPI R2mws). Figure 5 shows the corresponding regression plots for the wind speed range >= 3 m/s. The mean Lidar wind speeds averaged over all used values (KPI Cmwsd) resemble those of the cups very closely (see columns 5 and 6 from Table 5), yielding very low relative Campaign Mean WS Differences (KPI Cmwsd) at all measurement levels and WS ranges. Table 6 reflects the results according to the absolute wind speed error criterion. It shows that for the wind speed range of 3 to 16 m/s at all levels between 20 to 91 m, a small fraction of data ranging between 0.6 % and 3.8 % of concurrent ten- minute data points exceed the prescribed wind speed difference threshold of 0.5 m/s which is below the allowed upper limit of 10 %. With respect to the linear WS regressions the following KPI’s Acceptance Criteria are passed ✓ Regression slope (KPI Xmws) between 0.98 and 1.02 at all treated levels and for all WS ranges; meeting the Acceptance Criteria. ✓ R2 (KPI R2mws) > 0.97 at all treated levels for both the WS ranges a) [all > 3 m/s] and b) [4 to 16 m/s]; meeting the Acceptance Criteria. ✓ The Acceptance Criterion for the relative Campaign Mean Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is successfully passed at all assessment levels and for 20, 45 and 70 m measurement levels for both WS rangeslevel. Furthermore, the following wind speed related Acceptance Criteria were met: ✓ Absolute Wind Speed Difference (KPI Awsd) at all comparison levels and for all analysed wind speed data between 3 and 16 m/s, see Table 6. The following deviation from applicable test conditions and performance criteria are reported. o The Acceptance Criterion for the relative Campaign Mean Wind speed Difference (KPI Cmwsd) is missed at the 91 m level measurement level for both wind speed ranges. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 13 WS-avg WS-avg Mean rel. mean # values slope R2 20 m Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 2533 0.992 0.995 5.42 5.38 0.04 0.81% 4 - 16 m/s 1928 0.992 0.994 6.01 5.96 0.05 0.83% WS-avg WS-avg Mean rel. mean # values slope R2 45 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 2982 0.990 0.995 6.01 5.96 0.05 0.88% 4 - 16 m/s 2428 0.990 0.995 6.55 6.49 0.06 0.98% WS-avg WS-avg Mean rel. mean # values slope R2 70 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R2mws KPI Cmwsd All >= 3m/s 3159 0.993 0.995 6.60 6.55 0.05 0.80% 4 - 16 m/s 2749 0.993 0.994 7.01 6.95 0.06 0.79% WS-avg WS-avg Mean rel. mean # values slope R2 91 Cup LiDAR diff. diff. - - - [m/s] [m/s] [m/s] % WS-range KPI Xmws KPI R 2 mws KPI Cmwsd All >= 3m/s 2971 0.989 0.994 7.38 7.29 0.09 1.27% 4 - 16 m/s 2674 0.990 0.993 7.71 7.62 0.09 1.16% Table 5: Regression results comparison for ZPH 594; acceptance relevant results are colour shaded. Note the regression lines are forced through the origin. ZPH 594 Criterion for > 0.5 m/s for 3 to 16 m/s > 2% above 16 m/s abs WS error KPI Awsd Level Level total # identified # fraction % total # identified # fraction % 20 2533 11 0.4 0 0 N/A 45 2980 27 0.9 1 0 0.0 ZPH594 70 3152 50 1.6 7 1 14.3 91 2955 101 3.4 14 3 21.4 Table 6: Summary of absolute wind speed differences between cups and Lidar DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 14 Figure 5: Plots of linear wind speed regression results for 20, 45, 70 and 91 m (note the regression results are forced through the origin) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 15 4.5 Wind direction comparison By comparing the wind direction as measured by the Lidar device at 91 m with the mast mounted wind vane at 88 m A.G.L., it is possible to see how well correlated the measures are, providing confidence that the Lidar is ‘seeing’ the same wind direction as the vane. In order to validate this comparison quantitatively a two variant regression solving for the slope m and the interception of the best-fit line with the y-axis b (according to y = m x + b) was performed, compare Appendix A. The results of such regression are shown in the x-y-plots in Figure 6 with the vane wind direction at 44 and 88 m on the x-axis and the Lidar direction at 45 and 91 m on the y-axis respectively. For this analysis the data was again filtered for Lidar and the cup wind speeds at 91 m, i.e. for WS >=3 m/s (to avoid false readings from the vane at low wind speeds), but not for possibly disturbed wind directions sectors. Note that a few 180° wind direction ambiguities were observed, when ZephIR Lidar data were correlated to the wind vane readings at 88 and 44 m (see Appendix D). These ambiguities were removed whenever there was a distinct 180° offset when compared with the mast vane wind direction measurements as the reference. This mast based correction is justified by the assumption, that a few 180° offset occurrences are related to lower wind speeds in combination with near ground site induced turbulence effects. Wind direction time series present during the course of the campaign period together with raw data correlations and WD distribution statistics can be found in Appendix D. Figure 6: Regression plot of wind direction comparisons at 45 m (left) and 88 m (right) The regression plots in Figure 6 reveal a close resemblance in measurement between the Lidar against the wind vanes for levels at 44 and 88 with an offset (in terms of a mean difference) of 5.1° at 45 m and 2.1° at 91 m. It should be noted that the calculated directional offset for ZPH 594 at 45 m is 5.1°, which does not lie within the applicable OFFmwd bounds. The remaining directional offsets are however within typical directional setup uncertainties for wind vanes and remotes sensing devices. Table 7 summarizes the WD comparison results for the acceptance relevant WD comparison levels at 88 and 44 m vanes, showing an equally good resemblance slope. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 16 WS filtering for WS > 3 m/s Level # values slope offset [⁰ ] R2 level - (KPI Xmwd) (KPI OFFmwd) (KPI R2mwd) 45 3131 0.994 5.143 0.975 ZPH 594 88 3451 0.999 2.149 0.988 Table 7: Summary of WD comparison results for both comparison levels ✓ The Acceptance Criteria for the respective KPIs for wind direction assessment (KPIs for Xmwd and R2mwd) are successfully passed for both comparison levels. The Criteria for the KPI OFFmwd for the wind direction assessment is passed at higher level. o The criterion for the KPI OFFmwd for the wind direction assessment is missed by a small margin at the 45 m measurement level. 4.6 Performance verification according to revised IEC standard, Annex L This subsection represents as a supplement to the standard Lidar DNV GL / NORSEWInD performance verification test with respect to a Remote Sensing Devices (RSD) validation approach as described in a draft version of the current edition of the IEC standard for power performance tests [4]. This approach is based on a wind speed bin averaged procedure in order to compare the horizontal wind speed measurements acquired by the RSD and the reference sensors at the mast. The objective of the IEC approach is to calculate the bin-wise deviation of the two sources and report the associated uncertainty. The bin averaging procedure was performed using 0.5 m/s wide wind speed bins centred on integers of from 4 to 16 m/s. In order to achieve statistical relevance this IEC approach requires • a minimum of three (3) 10-minute values available within each wind speed bin; and • a total amount of 180 hours of valid data (corresponding to a number of 1080 10-min values) Figure 7 shows the scatter plots of the wind speed comparison based on 10 min averages between the data pairs of the ZP 594 Lidar and the cups at 20, 45, 70 and 91 m. Additionally, the 10 minute averaged deviation for each data point of the two data sets is plotted (orange dots). Furthermore, the correlation coefficient, mean deviation and standard deviation of the deviations are shown in Table 8. The relative deviation of the data pairs was calculated in relation to the cup wind speeds as reference. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 17 Figure 7: Comparison of the horizontal wind speed component for ZP 594 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) Level Coeffcient of STD of Data Mean Deviation Level Determination Deviations Points [m] (R2) [m/s] [%] [%] # 20 0.9945 0.10 1.73 1.55 2079 45 0.9949 0.11 1.74 1.79 2569 ZPH594 70 0.9952 0.12 1.76 1.81 2690 91 0.9931 0.15 2.07 2.45 2750 Table 8: Statistical parameters of wind speed deviation DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 18 4.6.1 Performance verification uncertainty Bin-averaged wind speeds of ZP 594 RSD and the reference measurements is shown in Figure 8. The bin-averaged mean deviation (solid red line in the graphs) can be compared to the standard uncertainty of the cup anemometers combined with the statistical uncertainty of the comparison for each of the WS bins. Figure 8: Bin-wise comparison of the horizontal wind speed component for ZP 594 – 20 m (top left), 45 m (top right), 70 m (bottom left), 91 m (bottom right) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 19 According to the IEC standard the verification uncertainty consists of five independent uncertainty components, which are summarized below: 1. Reference / anemometer uncertainty 2. Mean deviation of the remote sensor measurements and the reference measurements 3. Standard uncertainty of the measurement of the remote sensing device 4. Mounting uncertainty of the remote sensor at the verification test 5. Uncertainty due to non-homogenous flow The different uncertainty components are added in quadrature for each wind speed bin. The uncertainty due to non-homogenous flow between the measurement volume of the Lidar and at the met mast is assumed to be negligible due to the proximity of the Lidar to the mast and the benign terrain conditions at the Pershore test site. Details on the calculation of the separate uncertainty components are described in Appendix F. The results of the uncertainty calculation for the IEC compliant verification of the Lidar device at every comparison level are plotted in Figure 8. The finally combined uncertainties of the remote sensing RSD (VRSD) for the different WS bins and comparison levels show results values well below 2 % within most of the bins. For the current Lidar verification campaign the completeness requirement to yield 180 hours of valid and useable concurrent data (which translates into 7.5 days of data) in the WS range 4 and 16 m/s between the RSD and the reference cup is met for each comparison level. The additional data completeness requirement of yielding a minimum of 3 data pairs in each 0.5 m/s wide wind speed bin is fulfilled for most of the WS bins and comparison levels. Note that uncertainties are not calculated for wind speed bins with less than three data points. In Appendix G the environmental parameters - present during the performance verification test - are documented. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 20 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.96 293 3.99 4.44 3.46 0.190 0.011 -0.88 0.5 2.0 2.2 4.25 4.75 4.47 359 4.50 4.84 3.89 0.186 0.010 -0.70 0.5 1.8 2.0 4.75 5.25 4.97 330 5.01 5.47 4.33 0.177 0.010 -0.75 0.5 1.7 1.9 5.25 5.75 5.44 308 5.50 5.89 4.89 0.166 0.009 -0.96 0.5 1.6 2.0 5.75 6.25 5.90 179 5.97 6.35 5.48 0.181 0.014 -1.20 0.5 1.6 2.0 6.25 6.75 6.41 144 6.48 6.95 6.05 0.185 0.015 -1.03 0.5 1.5 1.9 6.75 7.25 6.91 125 6.98 7.41 6.37 0.215 0.019 -0.99 0.5 1.4 1.8 7.25 7.75 7.44 111 7.50 7.89 6.84 0.202 0.019 -0.89 0.5 1.4 1.8 7.75 8.25 7.91 75 7.97 8.33 7.30 0.206 0.024 -0.85 0.5 1.4 1.7 8.25 8.75 8.32 37 8.47 8.71 7.88 0.221 0.036 -1.70 0.5 1.3 2.3 8.75 9.25 8.89 35 8.95 9.29 8.46 0.218 0.037 -0.73 0.5 1.3 1.6 9.25 9.75 9.36 33 9.52 9.79 8.95 0.229 0.040 -1.68 0.5 1.3 2.2 9.75 10.25 9.92 27 9.97 10.64 9.55 0.284 0.055 -0.49 0.5 1.3 1.6 10.25 10.75 10.46 19 10.46 11.19 9.97 0.338 0.077 0.07 0.5 1.3 1.5 10.75 11.25 10.94 11 10.98 11.26 10.62 0.204 0.062 -0.33 0.5 1.2 1.5 11.25 11.75 11.58 9 11.56 12.16 11.12 0.278 0.093 0.15 0.5 1.2 1.5 11.75 12.25 11.95 7 11.95 12.22 11.72 0.207 0.078 0.00 0.5 1.2 1.5 12.25 12.75 12.39 4 12.47 12.61 12.16 0.186 0.093 -0.61 0.5 1.2 1.6 12.75 13.25 * 2 * * * * * * * * * 13.25 13.75 13.45 3 13.43 13.54 13.38 0.086 0.050 0.10 0.5 1.2 1.3 13.75 14.25 * 1 * * * * * * * * * 14.25 14.75 14.25 3 14.38 14.47 14.12 0.188 0.109 -0.90 0.5 1.2 1.7 14.75 15.25 * 0 * * * * * * * * * 15.25 15.75 * 0 * * * * * * * * * 15.75 16.25 * 0 * * * * * * * * * * Insufficient number of data points for uncertainty calculations Table 9: Uncertainty calculation for ZP 594 at 20 m level DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 21 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.97 275 3.99 4.67 3.34 0.202 0.012 -0.64 0.5 2.0 2.2 4.25 4.75 4.46 277 4.50 5.28 3.89 0.206 0.012 -0.93 0.5 1.8 2.1 4.75 5.25 4.96 269 5.01 5.37 4.31 0.192 0.012 -0.94 0.5 1.7 2.0 5.25 5.75 5.43 295 5.49 5.94 4.82 0.203 0.012 -1.00 0.5 1.6 2.0 5.75 6.25 5.96 293 6.01 6.33 5.38 0.168 0.010 -0.83 0.5 1.5 1.8 6.25 6.75 6.40 304 6.48 6.95 5.81 0.192 0.011 -1.23 0.5 1.5 2.0 6.75 7.25 6.91 221 6.98 7.33 6.20 0.195 0.013 -1.02 0.5 1.4 1.8 7.25 7.75 7.42 156 7.49 7.90 6.86 0.189 0.015 -0.91 0.5 1.4 1.8 7.75 8.25 7.92 114 8.00 8.45 7.49 0.205 0.019 -0.99 0.5 1.4 1.8 8.25 8.75 8.42 98 8.50 8.99 8.09 0.180 0.018 -0.94 0.5 1.3 1.7 8.75 9.25 8.87 83 8.99 9.27 8.26 0.176 0.019 -1.29 0.5 1.3 1.9 9.25 9.75 9.36 52 9.48 9.74 8.96 0.204 0.028 -1.27 0.5 1.3 1.9 9.75 10.25 9.89 41 10.00 10.58 9.51 0.189 0.029 -1.17 0.5 1.3 1.8 10.25 10.75 10.36 34 10.48 10.81 9.91 0.233 0.040 -1.19 0.5 1.2 1.8 10.75 11.25 10.86 26 11.01 11.25 10.00 0.265 0.052 -1.38 0.5 1.2 2.0 11.25 11.75 11.38 19 11.52 11.75 11.07 0.192 0.044 -1.16 0.5 1.2 1.8 11.75 12.25 11.92 7 11.89 12.34 11.37 0.329 0.124 0.23 0.5 1.2 1.7 12.25 12.75 12.43 7 12.50 12.72 12.18 0.214 0.081 -0.59 0.5 1.2 1.6 12.75 13.25 12.91 13 13.01 13.30 12.57 0.251 0.069 -0.72 0.5 1.2 1.6 13.25 13.75 13.38 6 13.40 13.64 13.22 0.148 0.060 -0.17 0.5 1.2 1.4 13.75 14.25 13.89 4 14.07 14.17 13.60 0.244 0.122 -1.26 0.5 1.2 2.0 14.25 14.75 * 2 * * * * * * * * * 14.75 15.25 15.00 3 15.06 15.21 14.79 0.209 0.120 -0.39 0.5 1.2 1.5 15.25 15.75 * 1 * * * * * * * * * 15.75 16.25 15.67 3 16.01 16.16 15.39 0.426 0.246 -2.14 0.5 1.1 2.9 * Insufficient number of data points for uncertainty calculations Table 10: Uncertainty calculation for ZP 594 at 45 m level DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 22 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.96 218 3.99 4.53 3.11 0.237 0.016 -0.89 0.5 2.0 2.3 4.25 4.75 4.44 252 4.49 5.13 3.51 0.229 0.014 -1.12 0.5 1.8 2.2 4.75 5.25 4.95 255 5.00 5.62 4.43 0.188 0.012 -1.07 0.5 1.7 2.1 5.25 5.75 5.46 250 5.51 6.16 4.83 0.201 0.013 -0.90 0.5 1.6 1.9 5.75 6.25 5.94 301 5.99 6.80 5.28 0.200 0.012 -0.95 0.5 1.6 1.9 6.25 6.75 6.45 248 6.50 7.12 5.80 0.193 0.012 -0.75 0.5 1.5 1.8 6.75 7.25 6.91 270 6.99 7.81 6.01 0.207 0.013 -1.03 0.5 1.4 1.9 7.25 7.75 7.41 270 7.49 8.02 6.91 0.184 0.011 -1.00 0.5 1.4 1.8 7.75 8.25 7.93 207 7.98 8.75 7.39 0.208 0.014 -0.64 0.5 1.4 1.6 8.25 8.75 8.40 158 8.48 8.89 8.01 0.201 0.016 -0.95 0.5 1.3 1.7 8.75 9.25 8.96 124 9.00 10.56 8.43 0.292 0.026 -0.35 0.5 1.3 1.5 9.25 9.75 9.45 68 9.49 10.92 8.99 0.300 0.036 -0.44 0.5 1.3 1.5 9.75 10.25 9.87 68 9.99 10.30 9.53 0.196 0.024 -1.17 0.5 1.3 1.8 10.25 10.75 10.35 47 10.45 10.71 9.97 0.186 0.027 -0.94 0.5 1.3 1.7 10.75 11.25 10.92 49 10.98 11.44 10.45 0.219 0.031 -0.55 0.5 1.2 1.5 11.25 11.75 11.45 27 11.53 11.77 10.70 0.225 0.043 -0.70 0.5 1.2 1.5 11.75 12.25 11.89 25 11.99 12.31 11.41 0.235 0.047 -0.80 0.5 1.2 1.6 12.25 12.75 12.49 16 12.47 12.84 12.12 0.203 0.051 0.15 0.5 1.2 1.4 12.75 13.25 12.99 11 13.00 13.44 12.65 0.270 0.081 -0.03 0.5 1.2 1.4 13.25 13.75 13.68 10 13.62 13.96 13.31 0.197 0.062 0.43 0.5 1.2 1.4 13.75 14.25 13.89 11 14.00 14.17 13.51 0.197 0.059 -0.77 0.5 1.2 1.5 14.25 14.75 14.52 6 14.47 14.73 14.36 0.156 0.064 0.37 0.5 1.2 1.4 14.75 15.25 15.10 3 15.02 15.30 14.89 0.204 0.118 0.55 0.5 1.2 1.6 15.25 15.75 * 2 * * * * * * * * * 15.75 16.25 * 2 * * * * * * * * * * Insufficient number of data points for uncertainty calculations Table 11: Uncertainty calculation for ZP 594 at 70 m level DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 23 Mounting Number of Mean Vcup VRSD BIN lower BIN upper Vrsd Vmm Vmaxrsd Vminrsd StdVrsd StdVrsd/√n uncertainty data sets deviation Uncertainty Uncertainty [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] [m/s] RSD [-] [%] [%] [%] [%] 3.75 4.25 3.91 167 4.02 4.71 3.12 0.311 0.024 -2.67 0.5 2.0 3.4 4.25 4.75 4.37 158 4.51 5.09 3.19 0.334 0.027 -3.15 0.5 1.8 3.7 4.75 5.25 4.91 154 5.01 5.45 4.02 0.212 0.017 -2.12 0.5 1.7 2.8 5.25 5.75 5.43 215 5.52 5.93 4.62 0.229 0.016 -1.78 0.5 1.6 2.5 5.75 6.25 5.92 211 6.01 6.50 5.03 0.231 0.016 -1.46 0.5 1.5 2.2 6.25 6.75 6.41 237 6.52 7.05 5.06 0.266 0.017 -1.58 0.5 1.5 2.2 6.75 7.25 6.90 233 7.00 7.89 5.37 0.304 0.020 -1.48 0.5 1.4 2.1 7.25 7.75 7.39 230 7.49 7.88 6.11 0.260 0.017 -1.36 0.5 1.4 2.0 7.75 8.25 7.92 246 8.00 8.40 7.44 0.173 0.011 -1.08 0.5 1.4 1.8 8.25 8.75 8.41 213 8.50 9.85 7.83 0.227 0.016 -1.09 0.5 1.3 1.8 8.75 9.25 8.87 161 9.00 9.41 8.38 0.206 0.016 -1.38 0.5 1.3 2.0 9.25 9.75 9.35 139 9.47 10.37 8.82 0.234 0.020 -1.21 0.5 1.3 1.8 9.75 10.25 9.86 76 9.99 10.49 9.40 0.238 0.027 -1.23 0.5 1.3 1.9 10.25 10.75 10.38 67 10.49 12.86 9.77 0.383 0.047 -1.08 0.5 1.2 1.8 10.75 11.25 10.91 62 11.02 11.40 10.31 0.229 0.029 -0.92 0.5 1.2 1.6 11.25 11.75 11.43 51 11.50 11.98 10.93 0.249 0.035 -0.65 0.5 1.2 1.5 11.75 12.25 11.89 37 11.97 12.29 10.98 0.250 0.041 -0.66 0.5 1.2 1.5 12.25 12.75 12.48 38 12.50 12.99 11.91 0.258 0.042 -0.11 0.5 1.2 1.3 12.75 13.25 12.98 27 13.00 13.79 12.26 0.314 0.060 -0.12 0.5 1.2 1.4 13.25 13.75 13.48 25 13.47 13.76 13.15 0.175 0.035 0.07 0.5 1.2 1.3 13.75 14.25 14.12 14 14.05 14.57 13.75 0.250 0.067 0.55 0.5 1.2 1.5 14.25 14.75 14.51 18 14.51 15.08 14.05 0.282 0.066 -0.02 0.5 1.2 1.3 14.75 15.25 15.09 7 15.01 15.47 14.67 0.330 0.125 0.50 0.5 1.2 1.6 15.25 15.75 15.58 8 15.41 15.92 15.30 0.206 0.073 1.08 0.5 1.2 1.7 15.75 16.25 16.09 7 15.97 16.51 15.63 0.290 0.110 0.77 0.5 1.1 1.6 * Insufficient number of data points for uncertainty calculations Table 12: Uncertainty calculation for ZP 594 at 91 m level DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 24 5 IMPORTANT REMARKS AND LIMITATIONS Independently performed Lidar Performance Verifications (LPV) of individual Lidar devices as reported in this document present a reasonable means to assure overall system integrity of the Lidar unit after manufacturing, and are meant to give an indication of the quality of wind data produced by the Lidar. Furthermore, the IEC compliant bin-wise uncertainty implementation may serve as a traceable means to judge the uncertainty of the RSD as determined from a well-defined verification process. Any statement given in the context of system integrity and data quality related results within this report are limited to the given test site conditions, to the prevailing atmospheric (in particular wind) conditions and to the specific Lidar configuration as selected for this LPV campaign. For sites with non-benign terrain and atmospheric conditions, an LPV is not thought to replace the requirement for an on-site verification. In this situation it may be necessary to conduct measurements in close proximity to an on-site mast over a reasonable period. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 25 6 CONCLUSION A ZephIR 300 Lidar and cup anemometer wind measurements were carried out at the Pershore UK Remote Sensing Test Site to validate Lidar wind data quality against a well-known high quality standard cup anemometer. Measurement levels of 20 m, 45 m, 70 m and 91 m were available for wind speed correlations (88/91 m for wind direction correlation) between a proximate met mast and a ZephIR 300 Lidar with the serial number ZP 594. The duration of the validation campaign period for was 53.5 days. The test period and wind data coverage is considered sufficient for the purposes of characterising the wind data performance of the ZephIR Lidar in the context of a Lidar Performance Verification. The overall system availability for the mentioned total campaign duration of 53.5 days for ZP 594 is 96.8 %. The data availabilities at the selected Lidar measurement levels 20 m, 45 m, 70 m and 91 m was in the range of 89.7 % to 94.6 %. These data coverage figures are relative to the number of maximum possible ten-minute data points for the total duration of the campaign. Wind speed (and direction) correlations were carried out for each of the four WS measurement levels (one for WD) mentioned above. The wind speeds of both techniques at all treated levels correlated very well, showing a very low level of scatter and an excellent resemblance of Lidar wind speeds to those of cups, in terms of linear regression slopes. In summary the following KPI related Acceptance Criteria are met. ✓ The Acceptance Criterion for System Availability (KPI SACA) to be ≥95 % is successfully passed (Table 4). ✓ The Acceptance Criterion for Data Availability (KPI DACA) to be ≥90 % is successfully met at the 20, 45 and 70 m measurement levels. ✓ Regression slope (KPI Xmws) between 0.98 and 1.02 at all treated levels and for all WS ranges, meeting the Acceptance Criteria (Table 5, column 2). ✓ R2 (KPI R2mws) > 0.97 at all treated levels for the WS ranges a) [all WS > 3 m/s]and b) [4 to 16 m/s], meeting the Acceptance Criteria (Table 5, column 3). ✓ The Acceptance Criterion for the relative Campaign Mean Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is successfully passed at the 20, 45 and 70 m measurement levels for both WS ranges. ✓ Absolute Wind Speed Difference (KPI Awsd) at all comparison levels and for all analysed wind speed data between 3 and 16 m/s where wind speed differences of greater than 0.5 m/s makes up < 5 % of the full dataset (Table 6). ✓ The Acceptance Criteria for the respective KPIs for wind direction assessment (KPIs for Xmwd and R2mwd) is successfully passed at only the for both measurement levels. The Criterion for KPI OFFmwd for the wind direction assessment is passed at the 88 m level. The following KPI related Acceptance Criteria are not met. o Acceptance Criterion for Data Availability (KPI DACA) to be ≥90 % is missed for the 91 m measurement level (Table 4). o The Acceptance Criterion for the relative Campaign Mean Wind Speed Difference (KPI Cmwsd) (see Table 5, column 7) is missed at the 91 m measurement level. o The Criterion for the KPI OFFmwd for the wind direction assessment is missed at the 45 m level. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 26 The performance verification and uncertainty calculation has also been carried out in accordance with the IEC standard yielding a traceable uncertainty measure. The following deviations from applicable IEC test conditions are reported: o During the verification campaign the bins for the following measurement levels cannot be calculated given insufficient data points: ▪ 13 m/s, 14 m/s and 15 m/s onwards for the 20 m measurement level; ▪ 14.5 m/s and 15.5 m/s bins for the 45 m measurement level; ▪ Bins greater than 15 m/s for the 70 m measurement level. In summary, this Pershore validation campaign indicates that the ZephIR 300 Lidar with the serial number ZP 594 is able to reproduce cup anemometer wind speeds and wind vane directions at an accurate and acceptable level. DNV GL considers that the ZephIR 300 Lidar device under test (with the serial number ZP 594) can be used for formal wind potential and long-term wind resource assessments. Specifically, DNVGL concludes that this Lidar may be employed as a standalone measurement system – replacing a conventional met mast – given the following criteria are met: (1) The Lidar is deployed in a location with benign terrain and atmospheric conditions. (2) The long term data accuracy stability is assured by either: ▪ recording data for a period sufficient to obtain an adequate in-situ correlation to an onsite reference (e.g. a short met. mast); ▪ or – in case of lack of a suitable in-situ reference – by performing a post deployment performance verification campaign, provided a continuous system operation during the preceding deployment period. Finally, DNV GL recommends, that care needs to be taken with respect to the formal use of Lidar turbulence and extreme wind speed measures, not treated in this report but known to be different from classical anemometry measures. DNV GL likes to point out that good measurement and data collection practices need to be maintained for all wind speed measurements, be they Lidar or more conventional anemometry. Therefore, special care needs to be exercised in the transportation, installation and on-going maintenance of the Lidar as it may be exposed to a wide range of environmental conditions at different sites over time. A key element of any formal wind study is the traceability of the wind speed data uncertainty. Hence, a strict uncertainty assessment (which is not part of this report) should be employed. Furthermore it is recommended that thorough practices of documenting the salient features of Lidar installation and maintenance are instigated from the outset. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 27 7 REFERENCES 1. Barker, W., E. Burin des Roziers and S. Wylie, "Pershore: 91m Anemometer Mast Specification ", by ZephIR Ltd., issued: 05/11/2014. 2. Kindler, D., " Best Practice Test and Verification Procedure for Wind LiDARs on the Høvsøre Test Site”, GL GH-D Report WT 6960/09 for EU-Project NORSEWInD, Deliv. 1.1, June 2009. 3. International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power Performance Measurements of Grid Connected Wind Turbines. Ed. 1. International Electronic Commission. 3, rue de Varembé Geneva. Switzerland, Dec. 2005. 4. International Standard: IEC 61400-12-1: Wind turbines – Part 12-1: Power performance measurements of electricity producing wind turbines. Ed. 2. CD. International Electronic Commission.3, June. 2013. 5. IEA Expert Group Study on Recommended Practices for Wind Turbine Testing and Evaluation 11. Wind speed measurement and use of cup anemometry, 1. Edition 1999. 6. MEASNET: “Cup Anemometer Calibration Procedure”. Version 1, September 1997. 7. W. Barker, E. Burin des Roziers, "Pershore: 91m Anemometer Mast Specification and Data Validation, Revision 2", by ZephIR Ltd., issued: 10/10/2012. 8. Stein, D. “Technical Note of Inspection of ZephIR’s Reference Met Mast and Lidar Test Site (exec. 2015 - 09-01) at Pershore/Throckmorton, UK” DNV GL Report, No. GLGH-4257 15 13307 267 T-0001, Rev. B, 2016-05-02. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 28 8 GLOSSARY The following table lists abbreviations and acronyms used in this report. Abbreviation Meaning Acronym AC Acceptance Criterion a.g.l. Above ground level DNV GL New company name, successor of legacy GL GH IEC International Electro-technical Commission IEA International Energy Agency GH-D GL Garrad Hassan Deutschland GmbH KPI Key Performance Indicator MM Meteorological Mast PAR Performance Assessment Requirement LPV Lidar Performance Verification TI Turbulence Intensity WD Wind direction WS Wind speed DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 29 KEY PERFORMANCE INDICATORS AND ACCEPTANCE CRITERIA, IN LINE WITH [2] Table 13: List of KPIs and ACs relevant for System and Data Availability assessment Acceptance Criteria KPI Definition / Rationale across total campaign duration SACA System Availability ≥95% The Lidar system is ready to function according to specifications and to deliver data, taking into account all time stamped data entries in the output data files including flagged data (e.g. by NaNs or 9999s) for the pre-defined total campaign length. The System Availability is the number of these time stamped data entries relative to the maximum possible number of data entries (for 10 minute intervals) within the pre-defined total campaign period. (Any conditions affecting the test’s data availability outside of the LIDAR system’s control is not to be included in this calculation. Such as: power outages, acts of nature causing system damage, communication outages, maintenance, etc.) DACA Data Availability ≥90% The Data Availability is defined as the number of valid data points returned by the Lidar unit as compared to maximum number of possible points that can be acquired during the test (Any conditions affecting the test’s data availability outside of the LIDAR system’s control is not to be included in this calculation. Such as: power outages, acts of nature causing system damage, communication outages, maintenance, etc.) MV Number of Maintenance Visits N/A Number of Visits to the Lidar system by either the manufacturer or an authorized third party to maintain and service the system. This is to be documented and reported. UO Number of Unscheduled Outages N/A Number Unscheduled Outages of the Lidar system in addition to scheduled service outages. Each outage needs to be documented regarding possible cause of outage, exact time / duration and action performed to overcome the Unscheduled outage. This is to be reported. CU Uptime of Communication System N/A To be documented and reported by the manufacturer. In the above table, during periods of maintenance; the system is deemed unavailable. * Undisturbed sectors: this means sectors with no significant flow distortion e.g. by wake effects of nearby wind turbines DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 30 Table 14: List of KPIs and ACs relevant for Wind Data Accuracy assessment Acceptance Criteria KPI Definition / Rationale Cmwsd Campaign Mean Wind Speed – Difference <1% Absolute difference of mean wind speeds between Lidar and reference as measured over the whole verification campaign duration, expressed as percentage relative to the Campaign Mean Wind Speed A threshold is imposed on the Difference. Analysis shall be applied to wind speed ranges a) 4 to 16 m/s b) all above 3 m/s given achieved data coverage requirements. Awsd Absolute Wind Speed Differences a) < 0.5 m/s Absolute 10 minute mean wind speed differences between Lidar and reference for all data points treated after filtering. b) within 5% A threshold is imposed on the Difference. Not more than 10% of Analysis shall be applied to wind speed ranges data to exceed the a) 3 to 16 m/s criteria above. b) all above 16 m/s given achieved data coverage requirements. Xmws Mean Wind Speed – Slope 0.98 – 1.02 Slope returned from single variant regression with the regression analysis constrained to pass through the origin. A tolerance is imposed on the Slope value. Analysis shall be applied to wind speed ranges a) all > 3 m/s b) 4 to 16 m/s given achieved data coverage requirements. R2mws Mean Wind Speed – Coefficient of Determination >0.97 Correlation Co-efficient returned from single variant regression A threshold is imposed on the Correlation Co-efficient value. Analysis shall be applied to wind speed ranges a) all > 3m/s b) 4 to 16 m/s given achieved data coverage requirements. Xmwd Mean Wind Direction – Slope 0.97 – 1.03 Slope returned from a two-variant regression. A tolerance is imposed on the Slope value. Analysis shall be applied to a) all wind speeds above 3 m/s regardless of coverage requirements. OFFmwd Mean Wind Direction – Offset (absolute value) < 5° In terms of the campaign mean difference in wind direction (same as for Mmwd) R2mwd Mean Wind Direction – Coefficient of Determination > 0.97 (same as for Mmwd) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 31 PERSHORE/THROCKMORTON MET MAST DETAILS 360° Panorama Photos, taken on 2015-09-01, see inspection report [8]: Met Mast Photo: DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 32 Met. Mast Sketch: Met. Mast Sensor Distribution Table: DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 33 TIME SERIES OF WIND SPEED Wind Speed time series for 91 m at the mast and ZP 594 including temperature recorded at the 88 m temperature and pressure sensor. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 34 WIND DIRECTION WD time series at 88 m and 44 m wind vane levels: DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 35 X-Y-plot of raw wind direction data for WS > 3 m/s (blue dots) and 180° ambiguity cleaned data (red dots) between wind vane and Lidar measures at 88/91 and 44/45 m DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 36 Mast ZP 594 91 m 10% 20% 30% 40% 10% 20% 30% 40% 0-3 3-6 6-9 >9m/s 0-3 3-6 6-9 >9m/s 45 m 10% 20% 30% 40% 10% 20% 30% 40% 0-3 3-6 6-9 >9m/s 0-3 3-6 6-9 >9m/s DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 37 CUP CALIBRATION CERTIFICATES, TAKEN FROM[6] Thies First Class Advanced Cup S/N 09164177 at 91.5 m, 300° orientation (Deutsche WindGuard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 38 Vector A100 LM Cup S/N 8920 at 70.5 m, 300° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 39 Vector A100 LM Cup S/N 8920 at 70.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 40 Thies First Class Advanced Cup S/N 08157941 at 70.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 41 Thies First Class Advanced Cup S/N 08157941 at 70.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 42 Vector A100LM Cup S/N 11202 at 45.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 43 Vector A100LM Cup S/N 11202 at 45.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 44 Thies First Class Advanced Cup S/N 08157939 at 45.5 m, 300° orientation (Deutsche WindGuard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 45 Thies First Class Advanced Cup S/N 08157939 at 45.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 46 Thies First Class Advanced Cup S/N 08157940 at 20.5 m, 120° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 47 Thies First Class Advanced Cup S/N 08157940 at 20.5 m, 120° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 48 Vector A100LM Cup S/N 11203 at 20.5 m, 300° orientation (Deutsche Wind Guard Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 49 Vector A100LM Cup S/N 11203 at 20.5 m, 300° orientation (Svend-Ole-Hansen Calibration) DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 50 IEC ANNEX L UNCERTAINTY ANALYSES 1. Reference / anemometer uncertainty The anemometer uncertainty of the specific reference levels is calculated based on the wind tunnel calibration of the individual anemometer, the anemometer classification and the mounting effect at the met tower. 2. Mean deviation of the remote sensor measurements and the reference measurements This is the relative deviation between the bin averages of the RSD and the mast reference measurement divided by with the reference measurement. 3. Standard uncertainty of the measurement of the remote sensing device The standard deviation of the measurements was divided by the square root of the number of data records per bin. The relative uncertainty was calculated by dividing the value by the bin average wind speed of the mast (reference) measurement. 4. Mounting uncertainty of the remote sensor at the verification test The uncertainty of the remote sensing device due to non-ideal levelling was estimated to be 0.5 %. 5. Uncertainty due to non-homogenous flow The Lidar device is located in close proximity of the met tower just a few m to the East of the tower base. As a result the uncertainty due to non-homogenous flow within the measurement volume is considered to be negligible. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 51 ENVIRONMENTAL PARAMETERS ACCORDING TO IEC ANNEX L, DURING THE CAMPAIGN, REF. 70 M LEVEL # PiA Quality BIN BIN Shear exponenet TI Wind direction Temperature Air Density Wind Veer Vmm Factor Lower Upper 70 m -20 m at 70 m at 88 m at 44 m at 44 m 88 m - 44 m at 70 m1 Average Average Std Average Std Average Std Average Std Average Std Average Std Average Std [m/s] [m/s] [m/s] [-] [-] [-] [-] [#] [#] [°] [°] [°C] [°C] [kg/m3] [kg/m3] [°] [°] 3.75 4.25 3.99 0.43 0.27 0.099 0.047 34.4 1.5 175.5 76.4 4.3 3.0 1.280 0.021 7.3 12.0 4.25 4.75 4.50 0.43 0.26 0.096 0.043 34.3 1.3 185.0 77.0 4.8 3.2 1.274 0.022 7.9 9.7 4.75 5.25 5.01 0.40 0.23 0.098 0.043 34.2 1.3 200.3 81.2 5.1 3.2 1.273 0.021 7.1 7.7 5.25 5.75 5.51 0.36 0.20 0.102 0.045 34.2 1.3 205.5 74.4 5.5 3.2 1.269 0.020 6.2 6.1 5.75 6.25 6.00 0.34 0.18 0.098 0.040 34.2 1.7 207.4 65.6 5.7 3.4 1.265 0.020 6.6 6.1 6.25 6.75 6.50 0.30 0.15 0.097 0.036 34.2 1.3 215.4 61.9 6.1 3.3 1.263 0.019 6.1 5.5 6.75 7.25 6.99 0.29 0.13 0.098 0.034 34.2 1.4 215.1 55.2 6.1 2.9 1.261 0.018 6.0 4.9 7.25 7.75 7.49 0.28 0.10 0.094 0.034 34.3 1.4 221.5 56.1 6.0 2.9 1.262 0.017 6.3 5.0 7.75 8.25 7.99 0.29 0.10 0.092 0.034 34.5 1.2 228.8 57.0 5.8 2.8 1.262 0.017 6.3 5.1 8.25 8.75 8.49 0.27 0.09 0.097 0.034 34.5 1.1 232.8 52.0 6.3 3.3 1.260 0.018 5.6 4.7 8.75 9.25 9.00 0.25 0.09 0.102 0.035 34.6 1.1 232.1 43.6 7.0 3.2 1.256 0.015 4.5 3.8 9.25 9.75 9.50 0.23 0.09 0.107 0.035 34.4 1.3 235.1 53.1 7.0 3.2 1.255 0.014 3.8 3.5 9.75 10.25 9.98 0.23 0.07 0.106 0.031 34.5 1.1 240.6 41.8 7.5 3.0 1.255 0.013 3.9 2.8 10.25 10.75 10.46 0.22 0.06 0.118 0.033 34.6 0.8 250.0 39.6 7.4 2.8 1.254 0.012 3.5 2.3 10.75 11.25 10.99 0.19 0.06 0.120 0.041 34.5 1.1 247.6 34.2 8.4 2.7 1.252 0.013 2.8 2.3 11.25 11.75 11.54 0.19 0.05 0.129 0.033 34.6 1.2 250.4 32.9 9.3 2.7 1.246 0.012 2.4 2.3 11.75 12.25 11.99 0.19 0.05 0.130 0.032 34.8 1.9 253.3 37.4 9.1 3.3 1.245 0.012 2.1 1.5 12.25 12.75 12.48 0.19 0.03 0.135 0.030 34.5 0.9 252.7 34.4 9.4 2.7 1.244 0.009 2.0 1.0 12.75 13.25 12.97 0.18 0.04 0.145 0.037 34.7 0.7 258.2 35.6 9.5 1.8 1.244 0.007 2.1 1.0 13.25 13.75 13.60 0.19 0.05 0.137 0.037 35.0 0.9 259.7 43.1 8.8 2.0 1.245 0.008 1.6 0.9 13.75 14.25 14.03 0.16 0.04 0.146 0.028 34.2 3.0 241.3 42.5 9.4 2.2 1.244 0.008 2.1 0.9 14.25 14.75 14.50 0.17 0.03 0.137 0.022 33.6 2.1 222.4 34.8 10.1 0.9 1.240 0.005 1.8 0.7 14.75 15.25 15.01 0.18 0.04 0.147 0.022 34.1 1.8 261.0 51.4 9.0 2.8 1.240 0.010 1.3 0.9 15.25 15.75 15.43 0.17 0.03 0.160 0.024 34.0 1.4 255.5 55.2 8.8 2.7 1.243 0.009 1.4 1.2 15.75 16.25 16.07 0.16 0.06 0.131 0.020 34.0 1.2 220.1 41.9 10.4 0.9 1.237 0.001 2.0 0.6 1. Packets in Average (PiA) quality factor records for the Lidar were missing for a total of approximately two days during the measurement campaign. DNV GL – Report No. 702909-AUME-R-06, Rev. B – www.dnvgl.com Page 52 ABOUT DNV GL Driven by our purpose of safeguarding life, property and the environment, DNV GL enables organizations to advance the safety and sustainability of their business. We provide classification and technical assurance along with software and independent expert advisory services to the maritime, oil and gas, and energy industries. We also provide certification services to customers across a wide range of industries. Operating in more than 100 countries, our 16,000 professionals are dedicated to helping our customers make the world safer, smarter and greener.