Journal of Power and Energy Engineering, 2014, 2, 288-296
Published Online April 2014 in SciRes. http://www.scirp.org/journal/jpee
http://dx.doi.org/10.4236/jpee.2014.24040
How to cite this paper: Han, S., Gao, L.Y., Liu, Y.Q. and Yang, W. (2014) Post Evaluation of Wind Resource Assessment and
Micro-Siting. Journal of Power and Energy Engineering, 2, 288-296. http://dx.doi.org/10.4236/jpee.2014.24040
Post Evaluation of Wind Resource
Assessment and Micro-Siting
Shuang Han1*, Linyue Gao1, Yongqian Liu1, Wei Yang2
1State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources,
North China Electric Power University, B e ij ing , China
2North China Power Engineering Co., Ltd. of China Power Engineering Consulting Group,
Xicheng District, Beijing, China
Received February 2014
Abstract
The design energy productions deviate from the actual situa ti on, which are affected by the accu-
racy of two significant factors - the wind resource assessment and wind farm micro-siting. A run-
ning wind farm in northern China was taken as the object in this investigation. The measured data
obtained in operation phase and the relevant information in design phase were integrated and a
post evaluation of wind resource assessment, micro-siting and generating capacity reduction fac-
tors of the wind farm in design phase was provided. The results indicate that the representative
year wind regimes of the wind farm in design phase can basically reflect the wind conditions in the
wind farm without the consideration of the trends of long-term changes in wind speed; micro-sit-
ing project in design phase is superior to that in practical; generating capacity reduction factors,
overall on the high side, should be further optimized considering 20-year operation period.
Keywords
Post Evaluation; Wind Resource Assessment; Micro Siting; Reduction Factor
1. Introduction
Wind power has entered the scale-development stage, but there are still some problems and obstacles. One of
them is the design generating capacity of a wind farm vary greatly from the practical one. The investor’s profits
and strateg ic decision are affected by the deviation, no matter it is positive or negative. The project post evalua-
tion work can be performed through the analysis and evaluation of the implemented part, which will improve the
subsequent decision-making and implementation of the project. With an increasing number of wind farms, the
post evaluation of wind farms is more significant than ever, and some scholars have conducted researches [1-4].
The above studies are mainly focused on the general methods of post evaluation, the project's overall economic
post-evaluation studies and relevant fields. Post-evaluation investigations for the wind resource assessment and
micro-sitin g of wind farms are comparatively rare.
*
Corresponding author.

S. Han et al.
289
The object of the study is a certain wind farm in northern China, which has been put into practical operation. In
this work, the measured data of the operational phase were compared with the index of design phase [5,6]. The
differences between them and the reasons for these differences were uncovered to improve and optimize the
wind farm design.
2. Research Object
A wind farm in the northe rn China was chosen as the research object. Wind meters were installed at the height
of 10 m, 55 m, 65 m and 80 m of the anemometer towers respectively during the process of wind measurement.
This wind fa rm shows great development value as well as abundant wind resources with the annual average
wind speeds of 7.1 m/, 8.7 m/s, 8.9 m/s and 9.1 m/s. The average power densities are 364 W/m2, 674 W/m2, 716
W/m2 and 772 W/m2, annual average wind power density in Class 6 for many years [7].
The feasibility research of this wind farm was completed in April, 2005, twenty WTG1500D wind turbines
were recommended in the feasibility study and corresponding calculation of energy production had been con-
ducted. But for various reasons, 36 WTG850C wind turbines were practically constructed by the owners, which
was inconsistent with the feasibility report.
The data of anemometer towers in the feasibility report were collected from the 1# anemometer tower. In fact,
all measurements were gathered from new l# and new 2# tower because the l# tower had been dismantled. After
testing, a seriou s lack of the measurements, which didn’t meet the standard of effective data rate of 90%, led to
useless data from new l# tower. New 2# tower has complete data records, but wind measurement data are rela-
tively lower at 70m and 60m due to the equipment failure. Therefore, the data obtained at 50 m of new 2# tower
was adopted in the following post evaluation in this investigation.
To post-evaluate design indicators of this wind farm, the actual operation data from October 1, 2008 to Sep-
tember 30, 2009 were selected to be compared with the relevant data and indicators designed before.
3. Data Collection
In order to conduct post evaluation more accurately, details in design phase as well as operational phase are
needed. The data collection included the wind farm work logs during operational phase, statistical reports of the
energy production during operational phase, data from wind monitoring system, the availabilities of wind tur-
bines, failure registration forms, wind measurement data both in design phase and operational phase, the feasi-
bility research report of the wind farm, the topographic maps and relevant data of wind turbines.
4. Content of Evaluation
Evaluation mainly consists of the post evaluation of wind resources and the post evaluation of energy produc-
tion.
The post evaluation of wind resources investigated through the comparison between the measured wind re-
gime in operational phase and the revised data of wind regime during the feasibility study stage to verify the ac-
curacy and reasonableness of wind resource evaluation, including annual and monthly wind conditions. The an-
nual wind conditions chie fly contain the contrast of the annual average wind speed, the contrast of the curves of
the diurnal variations of wind speed all year round, the contrast of the curves of the diurnal variations of wind
energy all the year round, the contrast of the curves of annual variations of wind speed, the contrast of the curves
of annual variations of wind energy, the contrast of the wind speed and wind power frequency distribution his-
tograms all year round, the contrast of the wind direction roses all year round and the contrast of the wind ener-
gy roses all year round. Evaluation on monthly wind conditions is similar with the annual one.
The post evaluation of energy production mainly compared the actual power generation capacity in opera-
tional phase and the estimated power generation capacity in design phase and revealed the significant reasons
for variations. This work is focused on the investigation of rationality in reduction factors adopted in estimation.
5. Assessment Case Study
According to national standard GB/T18710-2002, data in two phases were processed, and a variety of wind con-
ditions parameters worked were graphed [7-10]. In this way, changes in wind speed, wind direction and wind
energy can be discussed more visually, and it is more conducive to compare the trend of the parameters before

S. Han et al.
290
and after.
5.1. Wind Resource Post Evaluation
Post evaluation is mainly based on annual wind conditions.
1) Annual average wind speed
The annual wind speeds of design phase and operational phase are 8.70 m/s and 8.32 m/s, and wind speed in
operational phase is 4.37% lower than that in design phase.
2) Comparison of wind speed diurnal variation curve of a year
After screening and averaging wind speeds in a certain period of time in a whole year using Equation (1),
24-hour diurnal variation laws were obtained.
1
1n
i ij
j
Vv
n=
=
∑
(1)
In Equation (1): Vi is the average wind speed at ith period of everyday in a whole year (m/s, i = 0 - 23); n is the
number of wind speeds at i period of everyday in a whole year (365 in common); vij is the jth of all wind speeds
in ith period in a whole year, m/s.
Figure 1 shows a comparison of wind speed diurnal variation curves between design phase and operational
phase.
The maximum deviation, the minimum deviation and the average deviation of annual wind speed diurnal var-
iations curves in design phase and operational phase respectively are 10.30%, 0.32%, and 4.47%.
As can be seen in Figure 1, the annual wind speed diurnal variations curves are consistent with each other in
two phases, but the actual wind speed is slightly lower than that in design phase, mainly between 4:00 am and
16:00 pm of a day. The annual average wind speeds in operational phase are lower than those in design phase.
Wind power density curves are similar to the wind speed curves above, no narrative further.
3) Comparison between annual wind speed variation curves
Figure 2 shows a comparison chart of annual wind speed variation curves in design phase and operational
phase, which depends on statistics calculation results.
From Figure 2, the regulation pattern in annual variations of wind speed in operation phase is similar to that
in design phase. The maximum relative deviation, the minimum relative deviation and the average relative devi-
ation respectively are 28.65%, 5.95% and 13.93%. The larger deviations occur in January (27. 81 %), February
(22.35%) and June (28.65%). Because the wind energy production data in the same period of wind speed data
are available, wind regime variation tendency could be confirmed by the trends behind changes of energy pro-
duction.
4) Annual frequency distribution of wind speed and Wind energy
The frequency distribution of wind speed and wind energy throughout the year is based on a wind speed in-
terval of 1m/s. The occurrence frequencies of wind speed and wind energy in an individual wind speed interval
were counted and the number of each wind speed section represents the median.
Figure 1. Annual wind speed diurnal variation curves.

S. Han et al.
291
Figure 2. Annual wind speed diurnal variation curves.
To calculate wind energy frequency, the wind energy density within each wind speed segment (i.e., a unit of
wind energy) is required depending on Equation (2).
3
1
1( )()
2
m
iij ij
j
D vt ρ
=
=∑
(2)
In Equation (2): Di is the ith interval of wind energy density (W∙h)/m2; m repres ents the number of wind speed
intervals; ρ is air density (kg/m3) ; vij is the jth wind speed (m/s) cubic values in the ith wind speed interval; tij is
the Occurrence time of the jth wind speed in the ith wind speed interval within a certain sector or all sectors (h).
Wind energy frequency in one wind speed interval was obtained by dividing wind energy densities in an indi-
vidual wind speed interval obtained from the Equation (2) by the sum of them (the total wind energy density).
Figures 3 and 4 are respective comparison of the annual frequency distribution of wind speed and wind
energy in design phase and operational phase.
Frequency distributions of the two phases are similar, indicating that the revised wind regimes in the repre-
sentative year in design phase corresponding to the actual conditions.
5) Comparison between two phases of annual wind direction roses and wind energy roses
The hourly data in design phase and operational phase were drawn into two-stage annual wind direction rose
diagrams, as shown in Figure 5.
Wind directions during design phase and operational phase are relatively concentrated. The predominant wind
directions in design phase and operational phase respectively are NW (21.2%) and WNW (23.9%).
According to the hourly observation data of wind speed and wind direction, statistical calculation of wind
energy in each sector was conducted, and wind energy frequency were obtained by dividing the wind energy in
individual sector with the total wind energy. Wind energy in each sector was calculated by Equation (2), but the
meaning of i in Equation (2) changed into ith sector. Wind energy frequency in each sector is the wind energy in
this sector divided by the total wind energy.
The wind energy roses were drawn according to the wind energy frequency in each sector. The annual wind
energy rose diagrams in design phase and operational phase were shown in Figure 6.
Wind energy are concentrated in design phase as well as in operational phase, consistent with wind direction
respectively, NW (36.6%) and WNW (33.7%). The dominant wind direction, northwest , is consistent in two
phases. The dominant wind direction in operational phase indicates a south deviation compared with that in de-
sign phase.
Comparison of wind regime monthly equals to compare the monthly wind speed and wind power density di-
urnal variation curves, and monthly wind direction roses and wind energy roses. Overall, the design phase of di-
urnal variation curves of wind speed and wind power density of each month are representative to describe diur-
nal variations of wind speed and wind power density in each month.
5.2. Post Evaluation of Energy Production Estimation
In order to compare the wind regime between design phase and operation phase of the wind farm, WAsP was

S. Han et al.
292
Figure 3. Annual wind speed and wind energy frequency distri-
bution histograms in design stage.
Figure 4. Annual wind speed and wind energy frequency distri-
bution histograms in operational phase.
Figure 5. Annual wind direction roses in design phase (left) and
operational phase (right).
Figure 6. Annual wind energy roses in design phase (left) and
operational phase (right).

S. Han et al.
293
utilized to calculate the wind regime data of two phases respectively. During the calculation, wind farm digital
maps, wind turbines arrangement and wind generator models were based on the documents in design phase
while standard power curves were adopted under standard density.
1) Comparison of equivalent full load hours
The average equivalent full load hours of program WTG850C from design phase are 2361, while program
WTG1500D are 2497, full load hours during the actual operation are 2711. However, some data are vacant and
required to be fixed before using due to the wind farm monitoring system communication problem. Moreover,
the recommended scheme was not adopted during the actual construction. Thus this value is only for reference.
2) Comparison of wind turbine availability
According to statistics in normal time of wind turbines in the wind farm production & technology department
record sheets, the wind turbine availability ratio can be calculated, as shown in Table 1.
As can be seen in Table 1, the efficiency of 28th unit wind turbine is too low, only 32.1%. The number of
normal hours is 0 by querying the data of 28th unit from January, 2009 to August, 2009. After consultation with
operation personnel, the 28th unit had been a long time downtime, but did not appear for several months long,
which indicated lack of data records. After removing the effects of the 28th unit, average utilization rate of wind
turbines in the wind farm was up to 95.81%. The utilization rate (95%) in the feasibility study phase is consis-
tent with the actual situation.
3) Dispatch of load limitation
During the operation of the wind farm, scheduling notification of load limitation generally had great influence
on the energy production of wind farm. According to commands on scheduling load limit and electricity loss
caused by them in the statistics records of the wind farm, major load limit appeared in January, February, March,
November, December and some other months with good wind conditions, which led to a certain impact on
energy production. The loss of generated capacity by dispatching load li mitation was accumulated to 16,964
MWh in sum from October, 2008 to September, 2009. But this record is the total power loss of three wind farms,
it is impossible to identify the specific issue of power loss in a certain wind farm, because there are not clearly
marked specific restrictions in the records. The loss of power due to the scheduling load limit in the first project
wind farm added up to about 4.3MWh if pro rata estimated. Because grid construction is lagging behind, load
limit of wind farms in Inner Mongolia and the Northeast is much more serious, especially during heating sea-
sons, and these months are just the time with the best wind conditions. Prior to requirements in power grid con-
struction, variation s of generating capacity caused by this factor should be noted in the feasibility study.
4) Rough calculation of energy production of the wind farm
There are differences in the location of wind towers and siting of wind turbines between the actual design
phase and the operational phase. The wind farm energy productions were calculated by WAsP, considering four
cases in the evaluation: the wind regime and the site election both in design phase; the wind regime in design
Table 1. Availability of various units of wind farms (%).
Unit NO. 1 2 3 4 5 6
Use Ratio 97.52 99.24 86.93 98.24 99.38 97.26
Unit NO. 7 8 9 10 11 12
Use Ratio 96.24 99.41 99.81 99.91 99.21 83.07
Unit NO. 13 14 15 16 17 18
Use Ratio 98.62 99.28 96.66 70.35 99.22 98.74
Unit NO. 19 20 21 22 23 24
Use Ratio 94.32 98.68 99.94 95.54 97.68 94.46
Unit NO. 25 26 27 28 29 30
Use Ratio 92.25 98.00 95.31 32.10 99.42 99.09
Unit NO. 31 32 33 34 35 36
Use Ratio 97.59 96.68 90.53 97.11 96.53 91.14

S. Han et al.
294
phase and the site election in operational phase; the wind regime in operational phase and the site election in de-
sign phase; the wind regime and the site election both in operational phase.
Case NO.1: the wind regime and the site election both in design phase. The calculation shows that the annual
net energy production is 110,854 MWh of 36 wind turbines totally.
Case NO.2: the wind regime in design phase and the site election in operational phase. The results ind icate
that the total net energy production is 87 ,456 MWh of 29 wind turbines, and the net energy production of 36
wind turbines is 118,566 MWh calculated in proportion. The energy production is slightly lower than that in
case NO.1, indicating that the siting in design phase is superior to the actual siting.
Case NO.3: the wind regime in operational phase and the site election in design phase. The calculation even-
tuated that the annual net energy production was 113,311 MWh of 36 wind turbines.
Case NO.4: the wind regime and the site election both in operational phase. The results indicate that the total
net energy production is to 89,174 MWh of 29 wind turbines, and the net energy production of 36 wind turbines
is 110,699 MWh calculated in proportion. Layouts of wind turbines in design phase were further proved better
than the actual layout scheme.
Through the calculation of four cases, the net total energy production calculated by WAsP is between 108,566
MWh to 113,311 MWh, and discrepancy between the maximum and minimum values is 4745 MWh.
5) Energy production and energy consumption within the wind farm
According to report statistics posted by the Electricity Production Technology Department, the monthly
energy production and on-grid energy production in the first project of the wind farm are listed in Ta bles 2 and
3 (from October, 2008 to September, 2009).
The generating capacity and wind speed were presented to do comparison in Figure 7.
Through the comparison, the wind speed trend is consistent with energy production exactly. Assessment of
wind regimes reflects the average wind speeds in January and February are lower than that in representative year
of design phase, as shown in Figure 2. The uniformity of energy production with monthly wind speeds, the
wind measurement data during operational phase are credible.
The annual generation capacity of the wind farm is 87,892 MWh and the on-grid energy production is 86,831
MWh by calculating monthly generating capacity available. On-grid energy production accounts for 76.63% -
79.98% of the rough calculation power generation. In the feasibility report, total generating capacity reduction
rate, 32%, is higher than the actual one 20.02% - 23.37%. Therefore, appropriate adjustment of the reduction
factor should be considered in the future feasibility study.
The energy consumption inside the wind farm was shown in Figure 8 using the data of energy production and
the on-grid energy production.
Table 2. First project energy production of the wind farm (10 MWh).
Month 10 11 12 1
Energy Production 806.86 1078.00 1170.96 854.17
Month 2 3 4 5
Energy Production 860.18 947.50 627.29 761.12
Month 6 7 8 9
Energy Production 654.79 276.39 232.36 519.64
Table 3. First project on-grid energy production of the wind farm (10 MWh).
Month 10 11 12 1
Energy Production 799.13 1064.13 1156.34 841.81
Month 2 3 4 5
Energy Production 849.35 935.21 618.93 753.83
Month 6 7 8 9
Energy Production 648.10 273.20 229.20 513.94

S. Han et al.
295
Figure 7. Comparison between wind speed and energy produc-
tion.
Figure 8. Energy consumption in the wind farm.
Total energy consumption and loss venue are 1061.100 MWh, accounting for 1.21% of the actual energy
production of the wind farm, and accounting between 0.94% - 0.97% rough-estimated generating capacities.
Energy consumption reduction, in feasibility report, is 5%, which is relatively higher than the actual one.
6. Conclusions
With integration of the actual operating data associated with the relevant information in design phase of a certain
wind farm in northern China, wind resource assessment, micro-siting, power generation calculation were post-
evaluated and the following conclusions were obtained:
1) The revised annual wind regimes at the feasibility study phase and the actual operation phase (October,
2008 to September, 2009) are generally agreed, indicating that the wind regimes of representative year is a good
response to the characteristics of local wind conditions.
2) The wind speed and wind energy during actual operational phase are slightly lower than the wind condi-
tions of the representatives in the feasibility study stage, which are possibly caused by the impacts of the
long-term trend of the reduced wind speed or special wind regime of this year. Specifically, the high wind speed
segment reduced (annual wind speeds in January and February according to the measured data are lower than
those of representative years), and low wind speed segment unchanged or slightly increased. Due to the consid-
eration including these factors should be taken when determining the models of WTGS.
3) The number of wind speeds lower than the representative years’, in January and February during actual
operation, is large. Wind speeds were low indeed rather than measurement problems based on the observation of
energy production changes over the same period of the wind farm. There are two statuses may lead to pheno-
mena: Firstly, the revised wind speeds in January and February of the representative year have positive deviation;
Secondly, wind speed data collected in the January and February of the year in operational phase are relatively
low. The data of additional years are required with further comparison to confirm.
4) Through the comparison of different wind conditions and layouts of wind turbines, design of micro-siting
in the feasibility study is superior to the actual distribution scheme.
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9101112
0
200
400
600
800
1000
1200
1400
wind speedenergy production
m/s 10 MWh

S. Han et al.
296
5) The average wind turbine availability comes to 95.81% according to the records, while the estimated value
in the feasibility study stage is 95%, whose reduction is more accurate.
6) Energy consumption inside accounts for 0.94% - 0.97% of rough-estimated energy production of the wind
farm. The energy consumption reduction of the venue is 8% in the feasibility study, higher than the actual situa-
tion, recommending a cautious consideration for future design.
7) According to the annual energy production during actual operation of the wind farm and re-estimated
rough calculation of the power output, the actual value of energy production is lower than the rough calculation
about 20.02% - 23.37%. The reduction of total reduction in feasibility study stage is 32%, which is a little higher.
Therefore, considering the twenty-year life of the wind farm, selection of the reduction factors should be further
investigated.
To obtain more scientific results, the increasing number of wind farms will be further post evaluated in an ex-
tended period of time.
Acknowledgements
The authors gratefully acknowledge the support from National Nature Science Foundation of China for the Pro-
gram: Physical Method Study for Wind Power Prediction based on CFD Numerical Simulation Database
(NO.51206051) .
References
[1] Wang, R., Ji, H.S., Lou, H.Y., Zhu, R.Z., et al. (2013) Discussion on the Post-Evaluation Method Step of Wind Farm
Generation. Electric Power, 46, 54-59.
[2] Peng, H.W. and Wang, X.L. (2009) Disc ussion on Post Evalua tion of Wind Farm Design. Renewable Energy Re-
sources, 27, 97-99.
[3] Yang, Y.H. and Li, X.D. (2008) Study of Theory and Method for Post-Evaluating of Wind Power. Journal of North
China Electric Power University (Social Sciences), 3, 6-9.
[4] Lv, T., Zha ng , L.S., Li, Z. and Dong, L. (2006) Application of the Analytic Hierarchy Process to Economic Perfor-
mance Evaluation of Wind Farm Operation. Electric Power, 39, 42-44. http://dx.doi.org/10.1109/TSG.2013.2258048
[5] Yang, W. (2010) Study on Post-Evaluation of Wind farm Design. Ms.D. Thesis, North China Electric Power Universi-
ty, Beijing.
[6] Jiang, G.X. (2009) Study on Po st-E valu ation of Wind Resource and Generated Energy in Wind Farm Design. Ms.D.
Thesis, North China Electric Power University, Beijing.
[7] General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China (2002)
GB/T18710-2002. Methodology of Wind Energy Resource Assessment for Wind Farm. Standards Press of China, Bei-
jing.
[8] General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China (2005)
GB/T18709-2002. Methodology of Wind Energy Resource Measurement for Wind Farm. Standards Press of China,
Beijing.
[9] National Development and Reform Commission (1997) DL/T 5067-1996. Code on Compiling Feasibility Study Report
of Wind Power Projects. Standards Press of China, Beijing.
[10] Wei, G., Wang, F., Zhang, Y.J., et al. (2004) Methods of Processing Shortage and Damage of Historical Data in Load
Forecasting. Electric Power Science and Engineering, 1, 16-19.