Application example

Energy2green Wind And Solar Power System

Get Instant Access

The IEEE New England (10-generator-39-bus) system was employed as benchmark to test the proposed model and method. The single line diagram of the test system is given in Fig. 8. In this system, the classical generator model is applied to the synchronous generator G2. The 4th order generator model with a simplified 3rd order exciter model is applied to the remaining 9 synchronous generators. It should be noticed that there is no any power system stabilizer considered in the test system. All simulations were implemented on the MATLABTM environment.

Fig. 8. Single line diagram of IEEE New England test power system

A wind farm with 200x2MW DFIGs is integrated into the non-generator buses, i.e. busl-bus29. The corresponding parameters of wind turbine and DFIG are given in Table 1.

According to the procedure given in Fig. 6, the frequency distribution of wind speed by applying Monte Carlo method to the Weibull probability distribution of wind speed can be calculated as depicted in Fig. 9. The sample size is set to be 8000 during simulation.

 Parameters Values P 1.2235 kg/m3 R 45 m C 0.473 Vcut-in 3m/s Vcut-off 25m/s Vrated 10.28m/s Rs 0.00488 Xls 0.09241 Xlr 0.09955 Xm 3.95279 Rr 0.00549 H 3.5 Ki 0.1406 Ti 0.0133 K2 0.5491 T2 0.0096
Table 1. Parameters of wind turbine and DFIG with 2 MW capacity

800

700

600

n

e

500

tq

£

400

300

200

100

Fig. 9. Frequency distribution of wind speed

Next, in accordance Eq. (1) and the frequency distribution of wind speed as shown in Fig.9, for the wind farm with 200*2MW capacity, the probability distribution of wind farm power output can be finally obtained as shown in Fig. 10. From Fig.10, there exist two

concentrations of probability masses in the distribution: one corresponds to the value of zero, in which the wind farm is cut off; the other corresponds to the value of 400MW, in which the rated power output is generated by the wind farm.

3000

2500

fr 2000 S

1000

100 200 300

Wind farm power output (MW)

Fig. 10. Probability distribution of wind farm power output

Fig. 11 shows the frequency distribution of the real part of eigenvalues when the wind farm is connected to bus20.

3GGG

2SGG

2GGG

1SGG

1GGG

Small signal stable region

Real part value

Fig. 11. Probability distribution of real part of eigenvalues with wind farm integration into bus20

According to the statistical analysis based on Fig. 11, we found that there is a probability of roughly 39.1% (3128 out of 8000 in simulation) that the real part of the eigenvalue will be positive, in which situation the system is small signal unstable. Therefore we can conclude that the stability probability of the test system is 60.9% in current operating condition. Furthermore, there exist two concentrations of probability masses in the distribution: the left one corresponds to the situation that the wind farm is cut off; the right one corresponds to situation that wind turbine generates rated power.

Under the same wind speed condition, the wind farm is connected to the bus1-29, respectively. The corresponding results are given in Table 2.

 Bus No. of wind generator Stable Probability 20 200 60.9% Others 200 100%

Table 2. Small signal stable probability with different wind farm integration position

Table 2. Small signal stable probability with different wind farm integration position

Electro-mechanical oscillation mode can be picked out according to the electro-mechanical relative coefficient or the frequency of oscillation, i.e. p>1 or 0.1<f<2.5Hz (Wang et al., 2008). The mean and standard error of the mode properties: frequencies, Electro-mechanical relevant ratio, damping ratio, and the participating factors are given in Table 3. We found that the 9th EM oscillation mode is unstable with a probability of 39.1%, and it is actually the pair of eigenvalue that determines the small signal stability probability of the whole system. In summary, according to the simulation results discussed above, we can conclude that the deterministic small signal stability analysis can be considered as a special case study in the probabilistic small signal stability analysis. Especially, the probabilistic small signal stability based on the Monte Carlo method can evaluate the test power system more objectively and accurately.