Research Article

Applying a Cerebellar Model Articulation Controller Neural Network to a Photovoltaic Power Generation System Fault Diagnosis

Table 9

Diagnosis results of the test data with additional errors and under 300–1000 W/m2 irradiation and PV module surface temperatures between 31 and 40°C.

Test no.Output weight for various fault types Known fault type Diagnosed results
PF1PF2PF3PF4PF5PF6PF7PF8PF9PF10

10.830.60.550.740.290.190.230.240.320.5PF1PF1
20.830.40.440.46000000PF1PF1
30.330.730.630.600000.240.17PF2PF2
40.770.830.60.50.250.270.20.250.140.19PF2PF2
50.330.610.670.360.270.1800.3600PF3PF3
60.360.630.660.420.3300.210.2500PF3PF3
70.350.450.620.8200.10.10.130.390.41PF4PF4
80.740.760.660.830.370.380.650.380.360.24PF4PF4
90.160.20.110.120.580.190.230.240.10.27PF5PF5
100.50.20.220.150.60.20.50.510.230.2PF5PF5
110.160.240.140.190.470.620.220.450.480.2PF6PF6
120.160.220.180.140.560.690.20.250.140.45PF6PF6
1300000.090.240.5000PF7PF7
140.090.220.130.180.130.170.520.080.110.08PF7PF7
1500.160.1400.5600.10.670.420PF8PF8
160.430.550.470.460.610.50.730.910.740.77PF8PF8
170000.140.2500.380.230.540.27PF9PF9
180.160.40.220.150.610.230.410.580.620.23PF9PF9
190.160.240.140.270.410.470.340.450.580.84PF10PF10
200000.50.34000.20.570.78PF10PF10