Application of ANFIS and ANN for partial discharge localization in oil through acoustic emission
This article presents an examination on the acoustic partial discharge (PD) localization in oil based on adaptive neuro-fuzzy inference system (ANFIS) and artificial neural network (ANN) approaches. Impedance matching circuit (IMC) was used to measure the electrical PD. The acoustic PD was obtained...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers
2023
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/106638/1/Application%20of%20ANFIS%20and%20ANN.pdf http://psasir.upm.edu.my/id/eprint/106638/ https://ieeexplore.ieee.org/document/10092794/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This article presents an examination on the acoustic partial discharge (PD) localization in oil based on adaptive neuro-fuzzy inference system (ANFIS) and artificial neural network (ANN) approaches. Impedance matching circuit (IMC) was used to measure the electrical PD. The acoustic PD was obtained through an acoustic emission (AE) sensor and preamplifier gain unit. In total, 112 coordinates for each of the AE sensors were utilized to evaluate the location of the PD. Once the voltage reached 30 kV, the electrical and acoustic PDs were recorded. Next, the data were preprocessed by moving average (MA) and analyzed by time of arrival (TOA), ANFIS, and ANN. The distance between PD and AE sensor was calculated based on TOA to determine the PD location. These information were used as an input to train the network by optimizing epoch and neuron for ANFIS and ANN in order to locate PD. ANFIS has the best percentage of PD source prediction based on root mean square error (RMSE) and coefficient of determination ( R2) as compared to ANN. Meanwhile, the computation time for ANN is 1.75 s faster than ANFIS to perform PD localization based on AE PD signals. |
|---|