JavaScript is disabled for your browser. Some features of this site may not work without it.
| dc.contributor.author | ADDISU, MENGESHA ADAL | |
| dc.date.accessioned | 2023-12-20T11:42:45Z | |
| dc.date.available | 2023-12-20T11:42:45Z | |
| dc.date.issued | 2023-01-30 | |
| dc.identifier.uri | http://ir.bdu.edu.et/handle/123456789/15535 | |
| dc.description.abstract | The overhead power transmission system is a network of conductors that are used to carry power from the generation station to the load side for most power plants around the world, including Ethiopia. Fault detection, identification, and location in transmission lines are critical issues to increasing the availability of power by reducing the time of interruption for maintenance in electric utility companies. In this thesis, fault detection, classification, and location for power transmission systems using artificial neural networks are developed for line-to-ground, line-to-line, line-to-line-to-ground threephase, and three-phase-to-ground faults in transmission systems. To develop this method, one of the rural transmission lines in Ethiopia (Oromia, Gefersa substation to Fitche transmission line) is used as a test line. This line is simulated using Matlab software to generate data for different fau lt conditions with different fault locations and types, which are the fault phase voltage and current. The generated data is preprocessed and put as an input for the neural network to be trained. MATLAB R2018a's neural network toolbox was employed to train artificial neural networks (ANN). The feed-forward multilayer network topologies of neural networks with improved back propagation and the LevenbergMarquardt learning algorithm were used to train the network. After the network (8-10-1) was trained for fault detection, the mean square error performance, regression plot, and error histogram analyses were made and found to have excellent performance with a regression coefficient of 1, validation performance of 1.6191e-23, and an error histogram range of -0.025 to 0.025. And the network (8-10-5) was trained for fault identification and location, followed by analyses of the mean square error performance, regression plot, and error histogram analysis the results showed to excellent performance with a regression coefficient of 0.98094, a validation performance of 0.010732, and an error histogram range of -0.056 to -0.006. Finally, it was found that ANN is one of the alternate options in fault detection, identification, and location design for transmission systems where sufficient transmission network data are available with a narrow fault location distance range from the substation. This provides significant benefits in assisting with the maintenance plan, saving efforts in fault location finding, and having economic benefits. Keywords: Fault detection, fault classification, fault location, neural network, transmission line, Matlab/Simulink | en_US |
| dc.language.iso | en_US | en_US |
| dc.subject | Electrical and Computer Engineering | en_US |
| dc.title | ANN-BASED FAULT DETECTION, CLASSIFICATION AND LOCATION FOR GEFERSA – FITCHE 66KV TRANSMISSION LINE | en_US |
| dc.type | Thesis | en_US |
