tion capacity of power grid is significantly lower in the winter heating period. The following Table 1 shows

minimum number of thermal power unit and minimum output in heating period and non-heating period in North

China [5].

3. Analysis of Optimization Model

In this paper, a power grid model was established based on the power balance of the whole grid to calculate

wind power consumption, with full consideration of spinning reserve, output range of thermal power and other

restrictions.

3.1. Object Function

The objective function of the time sequential production simulation is maximum consumption of wind power,

2.1 4.2

8.4

18.1

31.1

45.1

60.8

77.2

95.8

64

103 100

116

71

45

35

27 24

0

20

40

60

80

100

120

140

0.0

20.0

40.0

60.0

80.0

100.0

2006 2007 2008 2009 2010 2011 2012 2013 2014

Rate of rise (%)

Capacity (MW)

X. F. Li et al.

164

Characteristics of load

Unit commitment

Spinning capacity

Peaking

capacity

of power

grid

Transmission

capacity

Pump storage

Accommodation

capability of

wind power

Peaking capacity of power unit

Figure 2 . The influence factors of wind power accommodation.

Region 1

Region 1:

Region 2Region 3

Curtailment by peak-load capacity

Region 2:Curtailment by transportation

difficulties

Region 3:Curtailment which is convertible

Figure 3 . The relationship between peak-load regulation capacity of po-

wer grid and transmission capacity restriction.

Table 1. The operation situation of thermal power in North China.

Period The minimum number of thermal power The minimum output of thermal power (%)

Non-heating period 32 58

Initial and end of heating period 47 64

Middle of heating period 55 66

namely, total maximum wind power of each region in different periods, as shown in Equation (1).

( )

11

max ,

TN

tn

Ptn

= =

∑∑

(1)

Among them, N is the number of partitions of the power grid; T is the calculation step, for 1 hours;

( )

,Ptn

is wind power of the n partition in the t period [6].

3.2. Balance of Electric Power and Energy

( )( )( )

1

, 1,2,

N

i

i

ptTiePtL ttT

=

+= =

∑,

(2)

In Equation (2),

( )

i

pt

is the unit i output at time t; N is the number of unit (include wind power);

( )

TieP t

is the tie line power at time t (input is positive, output is negative);

( )

Lt

is the whole electricity load at time

t.

3.3. Spinning Reserve Constraints

In Equation (3), SP and SN are respectively positive/negative spinning reserve of power system;

t

N

C

is the

credible wind power capacity. Taking wind power into the thermal power boot program can minimize conven-

X. F. Li et al.

165

tional unit commitment and enhance peak-load regulation capacity of power grid [7]. Spinning reserve takes the

higher value of 5% of load or the largest single unit capacity.

max

1

min

1

j

Ntt

jNl P

j

Ntt

jNl N

j

PC PS

PCPS

=

=

+ ≥+

+ ≤−

∑

∑

(3)

3.4. Power Output Limit

( )

ii i

p ptp≤≤

(4)

In Equation (4),

i

p

is the lower limit and

i

p

is the upper limit of thermal power unit.

3.5. Pumped Storage Model

0

tt

ph

pp⋅=

(5)

In Equation (5),

t

p

p

is the pumping condition of pumped storage unit;

t

h

p

is the generating condition.

Pumping and generating can’t be conducted simultaneously [8].

4. Calculation Verificatio n

4.1. Calculation for Previous Years

Take a provincial power grid as example and calculate its annual wind power output. The wind power operation

of Power Grid A in 2013 is calculated with its wind power output sequence over different years. The calculation

result is compared with the actual operation data of Power Grid A in 2013, so as to verify the precision of this

model.

4.2. Data for Ca lcula ti on

1) Load curve

Take the predictive value of load as the annual load curve for calculation. The maximum value is 24.74 GW

and the minimum value is 14.85 GW. The maximum daily peak valley difference value is 5.10 GW and the

minimum difference is 1.71 GW, as shown in Figure 4.

2) Capacity of wind power

Wind power installed sequence is total wind power installed for wind power production. Wind power in-

stalled capacity increases by steps. 1020 MW wind power is expected to be added annually and the cumulative

wind power installed capacity will reach 7.35 GW, as shown in Figure 5.

Figure 4 . Annual load curve.

X. F. Li et al.

166

Figure 5 . Installed capacity sequence of wind power.

3) Output of wind power

Wind power output data is normalized with consideration of wind power reduction. The full load hour of

wind power is 2050 hours. The highest of wind power normalization is 83% and the minimum is 0, as shown in

Figure 6.

4) Peaking capacity of thermal power

Thermal power unit is mainly composed of pumping thermal unit and air condensing thermal unit. Air con-

densing thermal unit is not heating units and its minimum output is unchanged; pumping thermal unit generates

both heat and electricity and its minimum output is different in heating period and non-heating period, as shown

in Table 2.

4.3. Calculation Resu lt s

Calculation results are shown in Table 3. The data for 2013 is the actual operation statistics, while those for

2010, 2011, 2012 and 2014 are calculation data. The deviation between calculated output and actual power gen-

eration in 2013 is no more than 5%, which indicates the credibility of calculation results.

5. Analysis on Wind Power Accommodation Capacity

This calculation model aims at maximizing wind power consumption and optimizing thermal power unit com-

mitment and output. Because variables of the model are huge and it is a typical mixed integer programming

problem, this paper adopts GAMS to solve it. The calculation results show that, on the premise of maximum ac-

cepted wind power, the whole year is expected to consume 11.6 TWh of wind power with 1.2 TWh of curtail-

ment and the full load hours of wind power is 1675 hours.

5.1. Impact of Spinning Reserve on Wind Power Accommodation

Spinning reserve capacity affects the unit's boot capacity, the bigger the spinning reserve, the higher the mini-

mum technical power and the more the wind power curtailment. When the power system spinning reserve is re-

spectively assumed as 5%, 10%, 15% and 20%, wind power accommodation conditions are detailed in Table 4,

the wind power generation quantity decreases while the wind power curtailment and its proportion increase.

5.2. Impact of Wind Resources on Wind Power Curtailment

Wind power curtailment varies between power grids with different wind resources, that is, more wind resources

lead to higher curtailment rate fewer wind power resources lead to lower curtailment rate. And the influence de-

gree varies between different provinces and regions (A1 to A6), as can be seen in Figure 7, where A1 and A6,

with curtailment rates standing at 4.6% and 6.0% respectively, are influenced more by wind resources than A2

and A3 whose curtailment rates are both 1.2%.

X. F. Li et al.

167

(a) (b)

(c) (d)

Figure 6. Normalization output sequence of wind power in different years.

Figure 7 . Influence of wind resources on wind power curtailment.

5.3. Impact of Load on Wind Power Curtailment

Wind power curtailment varies between power grids with different load capacity, that is, lower load leads to

higher curtailment rate while higher load leads to lower curtailment rate. And load influence degree varies be-

tween different provinces and regions, as can be seen in Figure 8, where A2 and A3, with curtailment rates

standing at 4.5% and 5.5% respectively, are influenced more by load capacity than A5 and A6 whose curtail-

ment rates are 2.7% and 2.1% respectively.

0

0.2

0.4

0.6

0.8

1

1

463

925

1387

1849

2311

2773

3235

3697

4159

4621

5083

5545

6007

6469

6931

7393

7855

8317

Output

Time (h)

0

0.2

0.4

0.6

0.8

1

1

439

877

1315

1753

2191

2629

3067

3505

3943

4381

4819

5257

5695

6133

6571

7009

7447

7885

8323

Output

Time (h)

0

0.2

0.4

0.6

0.8

1

1

463

925

1387

1849

2311

2773

3235

3697

4159

4621

5083

5545

6007

6469

6931

7393

7855

8317

Output

Time (h)

0

0.2

0.4

0.6

0.8

1

1

463

925

1387

1849

2311

2773

3235

3697

4159

4621

5083

5545

6007

6469

6931

7393

7855

8317

Output

Time (h)

0%

5%

10%

15%

20%

25%

30%

A1

A2

A3

A4

A5

A6 Normal wind

resources

Wind resources

reduced by 10%

Wind resources

increased by 10%

X. F. Li et al.

168

Table 2. P eak-load regulation capability of thermal power.

Capacity of pumping

thermal power

(MW)

Heating period Non-heating period

Minimum number of

operating unit Minimum output of

operating unit (%) Minimum number of

operating unit Minimum output of

operating unit (%)

100 4 75 2 70

200 14 64 9 58

300 18 69 10 57

330 13 67 7 52

600 6 50 4 47

Table 3. Calcuation results of 2013.

Year of normalization

output sequence of wind

power

Generated energy of

wind power (TWh) Curtailment energy of

wind power (TWh) Curtailment ratio of

wind power (%) Full load hour of wind

power (h)

2010 11.78 1.73 12.8 1703

2011 11.64 1.79 13.3 1682

2012 11.47 2.00 14.8 1658

2013 11.73 1.80 13.3 1692

2014 11.93 1.59 11.7 1725

Table 4. Impact of thermal power peak-load regulation capacity on wind power accommodation.

Spinning capacity

(%) Electricity generation of wind

power (TWh) Curtailment of wind power

(TWh) Curtailment ratio of wind

power (%)

5 11.23 1.82 13.96

10 11.11 1.94 14.85

15 10.93 2.12 16.28

20 10.60 2.45 18.77

Figure 8 . Influence of load capacity on wind power curtailment.

0%

5%

10%

15%

20%

25%

30%

A1

A2

A3

A4

A5

A6 Normal load

Load increased by

3%

Load increased by

5%

X. F. Li et al.

169

6. Conclusion

This paper establishes the time sequence simulation model with targeted maximum wind power consumption.

The impact of unit commitment, peaking capacity of thermal power units and spinning reserve on wind power

accommodation has been analyzed. Calculation of specific power grid operation statistics in 2013 verifies the

precision of the adopted simulation model. Spinning reserve, wind resources and load capacity have significant

impact on wind power curtailment, and the impact of each factor on wind power accommodation varies between

different power grids. Calculation results in this paper will provide reference to solving the problem of wind

power curtailment.

References

[1] Li u , C., Lv, Z.H., Huang, Y.H., et al. (2013) A New Method to Simulate Wind Power Time Series of Large Time Scale.

Power System Protection and Control, 41, 7-13.

[2] Kang, C.Q., Zhou, T.R., Chen, Q.X., et al. (2009 ) Assessment Model on Lo w-Carbon Effects of Power Grid an d Its

Application. Power System Technology, 33, 1-7.

[3] U mmels, B.C., Gibescu , M. and Pelgrum, E. (2007 ) Impacts of Wind Power on Thermal Generation Unit Commitment

and Dispatch. IEEE Transactions on Energy Conversion, 22, 44-51. http://dx.doi.org/10.1109/TEC.2006.889616

[4] Zhu, L.Z., Chen, N. and Han, H.L. (2011) Key Problems and Solutions of Wind Power Accommodation. Automation

of Electric Power Systems, 35, 29-34.

[5] Xue, B., Zhang, X. and Zheng, Y.X. (2014 ) Influence of Thermal Unit Spinning Reserve on Wind Power Accommoda-

tion. East China Electric Power, 42, 1550-1553.

[6] Cao, Y., Li, P. and Yuan, Y. (2014 ) Analysis on Accommodating Capability of Renewable Energy and Assessment on

Low-Carbon Benefits Based on Time Sequence Simulation. Automation of Electric Power Systems, 38, 60-66.

[7] Li u , D.W. and Huang, Y.H. (20 11 ) Research and Enlightenment of High Percentage of Wind Power in Spain. Renewa-

ble Energy, 33, 25-28.

[8] Xie, G.H. and Li, Q.H. (2011) Research on Calculating Model and Empirical Analysis of Wind Power Ancillary Ser-

vice Cost. Electric Power, 44 , 82-85.