Fault type | Module 1 and module 2 | ANN architecture (multilayered feed-forward neural network) | Goal | Type of activation function in hidden/output layer | Learning rule |
---|---|---|---|---|---|
Single phase to ground | ANN-1L | 12-30-7 | 10−8 | Tansigmoid | Levenberg–Marquardt |
ANN-1D | 12-27-30-1 | ||||
Double phase to ground | ANN-2L | 12-30-7 | |||
ANN-2D | 12-35-35-1 | ||||
Three phase to ground | ANN-3L | 12-30-7 | |||
ANN-3D | 12-37-40-1 | ||||
Four phase to ground | ANN-4L | 12-39-7 | |||
ANN-4D | 12-35-40-1 | ||||
Five phase to ground | ANN-5L | 12-30-7 | |||
ANN-5D | 12-30-35-1 | ||||
Six phase to ground | ANN-6L | 12-35-7 | |||
ANN-6D | 12-40-40-1 | ||||
Double phase to phase | ANN-7L | 12-30-7 | |||
ANN-7D | 12-40-40-1 | ||||
Three phase to phase | ANN-8L | 12-30-7 | |||
ANN-8D | 12-20-20-1 | ||||
Four phase to phase | ANN-9L | 12-30-7 | |||
ANN-9D | 12-30-30-1 | ||||
Five phase to phase | ANN-10L | 12-30-7 | |||
ANN-10D | 12-30-30-1 | ||||
Six phase to phase | ANN-11L | 12-30-7 | |||
ANN-11D | 12-30-1 |