GUIDELINES FOR COMPARISON

OF ZEPHIR AGAINST A
METEOROLOGICAL MAST
 1. POSITIONING OF LIDAR ADJACENT TO MAST

Many customers perform ZephIR/mast comparisons in order to check lidar calibration and performance.
Care is required when carrying out such comparisons; these guidelines should help to achieve good
agreement and hence provide confidence in the technology.

For a standard lattice or pull-up mast, the lidar is best positioned close to the mast. A separation of
between 3m and 10m is recommended. The laser beam emerges from the lidar window at an angle of 30
degrees to the vertical, and scans in a cone (figure 1). It is important that the beam does not illuminate any
moving objects such as cups or vanes throughout the scan. Guy wires will no doubt intersect the scan and
they should be under full tension to limit movement. The ZephIR should be positioned to minimise
intersection with any guy wires. Ideally the lidar is positioned to minimise obscuration by the mast of its
scan disk in the expected upwind and downwind directions.

The measurement cone rapidly moves clear of any mast wake effects as the focus height increases.
Positioning the lidar close to the mast minimises any risk of variation arising from terrain effects. It is
sometimes necessary to position the lidar a greater distance from the mast (e.g. because of difficult access
or availability of power). If this is unavoidable, it is worth bearing in mind that even in flat terrain the
correlation will deteriorate with increased separation.

The ZephIR measurement height is taken from the top window of the unit and the terrain around the mast
may not be flat, so any height differences between the ground level at the mast and the ZephIR window
must be taken into consideration. Quite small height differences can have an impact on the correlation: for
a typical shear profile at 30m, the bias in mean wind speed can be of order 1% per metre of height
difference. This ignores any possible terrain effects that might lead to further difference due to speed
changes over obstacles.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

2
 Figure 1 below illustrates best practice, and shows a ZephIR unit undergoing verification tests against
ZephIR Ltd’s 90m mast at Pershore, U.K, with all parts of the scan well clear of any cups or vanes. With
cups at 20m such as the case of the Pershore mast, it is necessary to reduce the lidar-mast separation to
7.5m or below.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

3
 2. COMPARISON OF LIDAR AND MET MAST DATA

At present it is common for the performance of remote sensing equipment to be verified against a
traditional met mast. This practice is likely to reduce in the future as more confidence in lidar is gained by
the wind industry. Data comparisons usually consist of velocity and direction analysis, with possible
extension to include parameters such as shear, inflow angle and turbulence intensity. This section outlines
the method used for velocity comparison of 10-minute averaged data at ZephIR Ltd’s test site at Pershore,
U.K.

It is assumed that care has been taken to filter mast data affected by shadowing or icing. Ensure analysis
periods are concurrent, and that there is no relative time offset (e.g. 10-minute data can be period
beginning or ending). For good comparisons, high-quality cups must be used in conjunction with an IEC-
compliant mast arrangement. The following steps are used to perform a comparison between the lidar unit
under test and the mast (used as reference):

 Apply a calm filter of 3m/s to both test and reference data sets

o This eliminates data that is of little consequence for the wind industry, and where either
sensor may be prone to larger error or offset
 Perform a least-squares fit, forced through the origin (0,0)

o Ideal agreement will yield a gradient of exactly 1.00 and a coefficient of determination (R^2)
that also approaches unity

 It is often sensible (particularly with Excel) to reduce the point size to the smallest on any
comparison plots; otherwise the points can form a solid mass, and the impact of outliers can be
greatly over-emphasised.

Appendix 1 gives an example of analysis of a ZephIR 300 lidar at Pershore, based on this approach.
Our preference for the forced fit is based on two arguments:

 When the wind speed does not occupy a large range, the fit parameters can be subject to larger
error when the intercept is allowed to float.

 For wind resource assessment, the forced fit gives the best indication of any bias between the test
and reference systems.

o The gradient of the forced fit indicates whether the test unit will over- or under-estimate the
wind resource with respect to the reference unit.

o Note that the forced fit can lead to a slight reduction in the value of R^2
The value of R^2 is a useful indicator, but it must be used with caution. During a period of relatively
constant wind speed, the value of R^2 is inevitably reduced but this should not be taken to imply a
reduction in the accuracy of the lidar.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

4
 3. DIAGNOSIS OF PROBLEMS IN ZEPHIR/MAST COMPARISON DATA SETS

3.1 Introduction

This section provides simple guidelines of what to look for when faced with correlation data that exhibit a
poor gradient or R^2. When diagnosing such a problem, there are several types of cause to consider. The
summary below is intended to be used as a quick check list to ensure all the most likely causes have been
considered.

3.2 Mast problems

Met mast anemometry is prone to a number of well-known problems that can lead to apparently poor
agreement with lidar.

 Cup/vane icing can cause serious under-reading. Filtering of data below 2 degrees C is a common
precaution.
 Shadowing effects: perform basic directional analysis to check whether directional filtering has
been correctly applied. Check paired cups where possible to verify any assessment. This can be
valuable even when the cups are not at the same height. Note that flow distortion by the mast can
be significant at the 1-2% level even for an IEC-compliant set-up.
 Malfunctioning cup: check mean shear profile to identify any obvious anomalies via a clear
discontinuity. Have the wrong cup calibrations been applied?

3.3 Analysis problems

There are a number of simple causes of poor agreement that can be easily checked:

 Check time synchronisation in the case of poor R^2.

o Even an offset of just a few minutes can lead to significantly degraded correlation, although
the gradient is not affected. If poor synchronisation is suspected, plot the lidar and mast
time series and look for a time offset.
o In particular, look out for 10-minute offset (data sets can be labelled by period beginning or
period ending), or integer number of hours in different time zones.
 Correlation analysis:
o Check for housekeeping errors - have correct heights been compared?
o Is there enough of a range of speeds during the analysis period for meaningful
comparison?
o Has the intercept been allowed to float (see previous section)? This can give large
uncertainty in gradient when speed range is low.

3.4 Terrain & deployment problems

For practical reasons, it is not always possible to follow precisely the best-practice guidelines as outlined in
the previous sections. The impact of this can be more severe in complex terrain.

 What is the likelihood of complex flow? Has the impact of trees, turbine wakes been assessed?
Has the Dynamics adjustment been applied correctly?
 What is the ZephIR – mast separation? In complex terrain, larger separation will have more impact
and lead to reduced agreement.
 Height difference between ZephIR and mast: it’s not always clear how this should be taken into
account. At low heights the flow follows the contours; at great heights it ignores them! This issue is
minimised when the ZephIR can be located very close to the mast, as recommended in section 1.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

5
 Appendix 1: Example performance verification of ZephIR against Pershore mast

PERFORMANCE VERIFICATION
Functional tests & full performance verification:

Velocity testing, system integrity check, focus mechanism and performance verification

Associated parties:

System supplied by:

Verification certificate:

System ZephIR laser anemometer

Supplier ZephIR Ltd., UK

Unit Number ZP300

Customer Anonymous

Order number -

Date of tests and verification -

Test Station A (397557 E 249426 N BNG), Separation = 5m

If fully complete and compliant this certificate documents the traceability of systems to a full verification test against a tall mast with
calibrated cups, and the subsequent functional tests at a UK test site operated by ZephIR Ltd. Users may request additional full
verification tests at their expense.

This document may not be reproduced other than in full without the written permission of ZephIR Ltd., and is invalid without
signatures.

Approved signatures:

Test engineer Quality Manager Date

PRINT:

SIGN:

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

6
 Functional tests:

Velocity testing and system integrity tests provide a check of several key components including wedge angle, wedge mounting, laser wavelength
and software configuration:

Test: Criteria: Result:

(delete as appropriate)

VELOCITY TEST* VELOCITY ERROR <0.5% PASS / FAIL

DIRECTION TEST* DIRECTION ERROR <0.5⁰ PASS / FAIL

SENSITIVITY* SNR >60 PASS / FAIL

FOCUS CALIBRATION** RANGE ERROR <1m PASS / FAIL

*Velocity, direction and SNR checked against calibrated moving belt at speeds up to 5m/s
**Focus calibrated at 68m focus range

Full performance verification against a tall mast:

Verification of horizontal wind speed performance against the certified tall mast operated by ZephIR Ltd’s Ledbury office. The comparison is
based on the slope of the forced regression line for 10-minute average values obtained over a timescale of at least 7 days. The ZephIR data is
processed using standard filters with an additional calm filter of 3m/s. Mast filters are also applied to eliminate invalid cup data. A minimum of
400 valid concurrent data points is required for the comparison.

Test: Criteria: Value: Result:

(delete as appropriate)

AVERAGE WIND SPEED Slope 1.00 ±2.0% 0.43% PASS / FAIL

AT 91.5 METRES R2>0.970 0.994 PASS / FAIL

AVERAGE WIND SPEED Slope 1.00 ±2.0% 0.03% PASS / FAIL

AT 70.5 METRES R2>0.970 0.995 PASS / FAIL

AVERAGE WIND SPEED Slope 1.00 ±2.0% 0.18% PASS / FAIL

AT 45.5 METRES R2>0.970 0.995 PASS / FAIL

AVERAGE WIND SPEED Slope 1.00 ±2.0% -0.18% PASS / FAIL

AT 20.5 METRES R2>0.970 0.994 PASS / FAIL

Additional comments:

The tall mast operated by ZephIR Ltd. has been constructed to conform fully with the recommendations for mast anemometry in IEC 61400-12-1 :
Power Performance Measurements of Electricity Producing Wind Turbines and has been approved for use by technical and engineering services
provider GL Garrad Hassan. Technical details of the test mast are presented in the annex to this document.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

7
 Correlation Plot Of Unit 300 v Tall Mast at Height 91.5m
Data Range: 16/11/2011 – 25/11/2011

Correlation Plot Of Unit 300 v Tall Mast at Height 70.5m

Data Range: 16/11/2011 – 25/11/2011

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

8
 Correlation Plot Of Unit 300 v Tall Mast at Height 45.5m
Data Range: 16/11/2011 – 25/11/2011

Correlation Plot Of Unit 300 v Tall Mast at Height 20.5m

Data Range: 16/11/2011 – 25/11/2011

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

9
 Annex A Test Mast Technical Specification
Annex A Test Mast Technical Specification

North West (300°) A B South East (120°)

C D
0.7m

Height AGL Dimensions

Notes
[m] [mm]

91.5 50.0 x 50.5 Square Section

Ø 48.3
E F 70.5 Round Section

45.5 Ø 48.3 Round Section

20.5 Ø 48.3 Round Section

Lattice Porosity = 0.67

G H
I J

K L

M, N

Lab Height Orientation (°) Mast to Calibration Cup to boom centre Instrument to mast centre
Type Manufacturer/Model Calibration*
el (m) Instrument Date height (mm) length (mm)

SOH
A 91.5 300 Cup Anemometer Risø P2546A 07/06/2010 1500 1025
MEASNET

3D Sonic
B
Lab 91.5
Height 120
Orientation (°) Mast to Metek USA1 - - 1500 Calibration 1025
Cup to boom centre height Instrument to mast centre length
Anemometer Type Manufacturer/Model Calibration*
el (m) Instrument Date (mm) (mm)
C 88 300 Direction Vane Vector W200P - - 920 3700
SOH
A 91.5 300 Cup Anemometer Risø P2546A 07/06/2010 1500 1025
D 88 120
Temperature/Hu Campbell Scientific
- - MEASNET- -
midity CS215

EB 70.591.5
3DRisø
Sonic SOH
300 120 Cup Anemometer P2546A Metek
MEASNET USA1
07/06/2010 - 960 - 3700 1500 1025
Anemometer
SOH
F 70.5 120 Cup Anemometer Vector A100LM 21/06/2010 1160 3700
MEASNET
C 88 300 Direction Vane Vector W200P - - 920 3700
SOH
G 45.5 300 Cup Anemometer Risø P2546A 07/06/2010 960 3700
MEASNET
Temperature/Humi Campbell Scientific
D 88 120 SOHCS215
- - - -
H 45.5 120 Cup Anemometer dity A100LM
Vector 21/06/2010 1160 3700
MEASNET

IE 43.570.5
SOH 920
300 300 Direction Vane Cup Anemometer
Vector W200P Risø
- P2546A - 07/06/2010 3700 960 3700
MEASNET
Temperature/Hu Campbell Scientific
J 43.5 120 - - - -
midity CS215 SOH
F 70.5 120 Cup Anemometer Vector A100LM 21/06/2010 1160 3700
SOH
MEASNET
K 20.5 300 Cup Anemometer Risø P2546A 07/06/2010 960 3700
MEASNET
SOH
G 45.5 300 Cup Anemometer Risø
SOH P2546A 07/06/2010 960 3700
L 20.5 120 Cup Anemometer Vector A100LM
MEASNET
21/06/2010 MEASNET
1160 3700

Campbell Scientific
M 6 - Pressure - - SOH - -
H 45.5 120 Cup Anemometer
CS1000 Vector A100LM 21/06/2010 1160 3700
MEASNET
Campbell Scientific
N 6 - Data Logger - - - -
CR1000
I 43.5 300 Direction Vane Vector W200P - - 920 3700

Temperature/Humi Campbell Scientific

J anemometer
*Cup 43.5 calibration 120
certificates are available - - - -
dityon request. CS215
Mast Installed 13/08/2010.
SOH
K 20.5 300 Cup Anemometer Risø P2546A 07/06/2010 960 3700
MEASNET

SOH
L 20.5 120 Cup Anemometer Vector A100LM 21/06/2010 1160 3700
Page 5 MEASNET

Campbell Scientific
M 6 - Pressure - - - -
CS1000

Campbell Scientific
N 6 - Data Logger - - - -
CR1000

*Cup anemometer calibration certificates are available on request.

Mast Installed 13/08/2010.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

10
 Annex B Location and Environment

Based upon Ordnance Survey data with the permission of the Controller of Her Majesty's Stationery Office, © Crown Copyright. Licence No. AL 100020693.

Test Mast 397557E 249426 N BNG

Figure B1: Test Mast Location and Environment

The terrain in the vicinity of the mast is flat and covered with sparse low growing vegetation. A free
standing lattice tower of approximately 40m in height exists on a bearing of 270° at 230m from the mast. 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 500m
to the North-East lies the small village of Throckmorton which consists of a few scattered farms and
houses. 700m 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.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

11
12
4L

8L 8L

1 km 1 km L = 100m
L = 100m
Mast Base Height = 35m AOD

Figure B2: Local Deviation from Plane Figure B3: Local Slope

Based upon Ordnance Survey data with the permission of the Controller of Her Majesty's Stationery Office, © Crown Copyright. Licence No. AL 100020693.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

Page 7
13
1 km
Figure B4: Regional Orography

Based upon Ordnance Survey data with the permission of the Controller of Her Majesty's Stationery Office, © Crown Copyright. Licence No. AL 100020693.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

Page 8
 The terrain surrounding the tall mast at Pershore has been assessed against IEC 61400-12 (Power
Performance Measurements of Electricity Producing Wind Turbines) [1] in order to determine whether any
sectors need to be screened from the mast data due to the orography of the site or local obstacles. Figure
B2 shows the difference in height between points within the local area of the mast and that of the mast
base. The colour bands in this plot have been set to match the criteria for maximum terrain variation from
the plane assuming a level plane through the mast base. Green represents variation within 0.04 times the
top instrument height (H). Yellow represents variation within 0.08 times H. The site meets the IEC
requirements for maximum terrain variation in all sectors. Figure B3 shows the slope of the terrain within
the local area of the mast. The colour bands in this plot have been set to match the criteria for maximum
slope in [1]. The coloured bands from blue to red represent slopes of < 3%, 3% - 5%, 5% - 10%, 10% -
30% and 30% + respectively. The site meets the IEC requirements for maximum slope in all sectors.
An obstacle assessment against [1] has been carried out for the lattice mast located to the South West of the test
mast. This analysis indicates that screening of data from directions ±14° about a bearing of 270° should be
considered. Detailed analysis of lidar versus mast measurements from this sector has shown negligible impact on
2
slopes, but potentially lower R values at 20.5m and 45.5m.

Annex C Instrument Selection and Data Filtering

Mast
Validation wind speed data is provided by the Risø P2546A cup anemometers on the North West side of the test
mast and the Vector A100LM cups on the South East side of the mast. Comparison of paired cups is used to provide
a robust method for identifying any problems with the mast instrumentation. Direction data is taken from the
Vector W200P wind vanes at the 88.0m and 43.5m levels.

Data has been screened where it may be affected by instrument or tower shadow. Direction measurements at
88.0m were used to screen wind speed data at 91.5m and 70.5m. Direction measurements at 43.5m were used to
screen wind speed data at 45.5m and 20.5m.

ZephIR
Wind measurements obtained with any method of anemometry are prone to increased levels of uncertainty in
certain conditions. ZephIR’s filtering software identifies rare conditions of reduced certainty and rejects the
corresponding wind measurements from the output data file. The standard set of filters supplied to ZephIR
customers was used to screen the wind speed data presented in this report.

Annex D References
[1] : IEC 61400-12-1 International Standard : “Part 12-1 : Power performance measurements of electricity
producing wind turbines,. Edition 1.0 2005-12, International Electrotechnical Commission.

11/28/12 | GUIDELINES FOR COMPARISON OF ZEPHIR AGAINST A METEOROLOGICAL MAST

14