Table 22 Simulation and analysis process information and specification of the system used.
Software and the system used | Parameters and description | |
|---|---|---|
DIgSILENT P.F. 15.1 | Workspaces | Graphical and DPL |
Simulation Type | Electro-Magnetic Transients (EMT) | |
Output Parameters for Each Event | 6 Parameters of the Current (Three Phases + Three Sequence Components) | |
Samples Number for Each Fault and Non-Fault Case Waveform | 64 Samples | |
Sampling Window Frequency | 3.2 kHZ | |
Output Data Type | Time Series | |
Recorded Output Data Environment | Microsoft-Excel | |
Simulation Duration of the Faults and Non-faults Cases (Average of all scenarios) | 18.5Â h | |
MATLAB | Workspace | Script File (m file) |
The Used Tools | Discrete Wavelet Transform (DWT) and Statistical Features Extraction | |
The Used Wavelet Family and Types | Db (2,4,6,8) and Sym (2,4,6,8) | |
Statistical Features of DWT Coefficients | Six Features (Max, Min, Average, Standard Deviation, Skewness, and Energy) | |
DWT Process and Feature Extraction Duration (Average of all scenarios) | 5.1Â h | |
Python 3.8 | Workspaces | TensorFlow Platform and Keras Library |
Epochs for the Scenarios | From 400 to 9500 (variable for each training) | |
Learning Method | Supervised | |
DNNs Structure Specifications | Table 6 | |
DNNs Training Duration (Average of all scenarios for different epochs) | From 40Â min to 3.8Â h | |
The System Used Specifications | CPU | Intel Core i5-6400 CPU @ 2.7Â GHz |
GPU | NVidia GeForce GTX 750 Ti | |
