Fault Location Method in Nonsolid-Earthed Network Based on Spatial Domain Image Fusion and Convolution Neural Network

Table 2

Parameters of fault location model based on GAF and CNN.

Layer name	Structural parameters	Output size
Input layer
Convolution lay	, , “Valid”
Activation layer 1	ReLu
Normalization layer 1	Local response norm
Pooling layer 1	Max pooling
Convolution layer 2	2 groups of 128 , , “Same”
Activation layer 2	ReLu
Normalization layer 2	Local response norm
Pooling layer 2	Max pooling
Convolution layer 3	, , “Same”
Activation layer 3	ReLu
Convolution layer 4	2 groups of 192 , , “Same”
Activation layer 4	ReLu
Convolution layer 5	2 groups of 1, , “Same”
Activation layer 5	ReLu
Fully connected layer	1 neuron
Regression layer	Distance-to-fault	1