Performance Evaluation

In order to improve the system’s performance under varying operating conditions, a robust LQG controller is designed according to the Steps 1-5 described in Subsection 10.5.3. For the proposed approach, the smallest Иж norm-bound is found to be у = 1.15. A norm-bounded controller is designed to have an Ию norm less than 1 /у. The weighting matrices are chosen as

Qr = diag(1,1; 1; 12; 20,10), Rr = 1.5; Rf = 1.

Qr and Rr are chosen so as to give good regulator performance and Rf is chosen so as to satisfy (10.22). There is a considerable freedom in choosing the regulator weighting matrices Qr and Rr, but limited freedom in choosing the filter weighting matrices Qf and Rf. This is to be expected because the objective is to limit the overall controller gain which consists of the product of the LQR and KBF gains. A suitable balance between the LQR and KBF gains can be obtained by adjusting Qr and Rr in such a way as to permit a smaller Rf [22].

Fig. 10.16 Voltage at PCC (bus 13) for load composition 3 (80 % induction motor in composite load model)

The frequency response of the proposed controller in Fig. 10.15 signifies that it provides adequate bandwidth and margins to stabilize the system under different operating conditions. The closed-loop eigenvalue of the dominant mode is found at -4.512. To examine the robustness of the proposed controller, the following contingencies are applied:

• a symmetrical three-phase short-circuit fault close to the substation; and

• an asymmetrical fault close to the DG unit.

Updated: October 23, 2015 — 12:41 pm