Page 209 - Decision Making Applications in Modern Power Systems

172 Decision Making Applications in Modern Power Systems

TABLE 6.1 Classification results for empirical mode
decomposition Hilbert transform ANN.
Sl. no. Power quality Total no. No. of samples Classification
event of samples classified correctly accuracy (%)
1 S1 18 15 83.3
2 S2 18 6 33.3
3 S3 18 8 44.4
4 S4 18 13 72.2
5 S5 18 14 77.7
6 S6 18 10 55.5

7 S7 18 17 94.4
Overall classification accuracy 65.8

TABLE 6.2 Classification results for empirical mode
decomposition Hilbert transform probabilistic neural network.

Sl. no. Power quality Total no. No. of samples Classification
event of samples classified correctly accuracy (%)
1 S1 18 16 88.8
2 S2 18 15 83.3

3 S3 18 15 83.3
4 S4 18 9 50
5 S5 18 18 100
6 S6 18 11 61.1
7 S7 18 18 100
Overall classification accuracy 80.9

6.3.7.2 Classification of power quality events using support
vector machine
In this section, SVM has been used for fault classification. A detailed discus-
sion on SVM has already been done in Section 6.3.5. In this work, LIBSVM
[30] has been referred to for the parameters of SVM. Seven PQ events have