SOYEMI, ADEBOLA OLUYEMISI and Covenant University, Theses
(2020)
*ARTIFICIAL NEURAL NETWORK BASED VOLTAGE COLLAPSE PREDICTION FOR POWER SYSTEM NETWORKS.*
Masters thesis, COVENANT UNIVERSITY.

PDF
Download (6MB) |

## Abstract

Unmitigated voltage instability often results in voltage collapse and or system blackout, which constitute a significant concern in power system networks across the globe but most especially in developing countries. This research work proposes an online voltage collapse prediction model through the use of a machine learning technique such as the artificial neural network and a voltage stability index called the new line stability index (NLSI_1). The proposed neural network approach uses a multilayer feed-forward neural network with the variables of the NLSI_1 as the inputs to the network. In the proposed approach, this is done in two phases; first load flow study is carried out for all load buses in the test-systems then voltage collapse prediction is carried out on the test-systems. Both phases are implemented using appropriate neural network models developed in MATLAB neural network toolbox. The proposed method was tested on the IEEE 14-bus system and the 28-bus, 330-kV Nigeria National Grid (NNG), for the two test systems two scenarios were considered; base case and contingency analysis (that is a variation of the reactive loads in the network). For the base case, all the buses and lines of the IEEE 14-bus system were stable with index values less than unity (<1). During contingency analysis simulations the 14th bus was ranked as the weakest bus in the system, having a reactive power margin of 74.6 MVAr and a percentage change in voltage magnitude of 38.30%. For the IEEE 14-bus system, the results obtained by the developed neural network and from the conventional method are approximately equal with negligible errors between them, which thereby validates the efficacy of the proposed method. For the NNG system, during base case simulation, the index values for all buses and lines were less than unity (<1). During contingency simulation, load bus 16 was ranked the weakest bus in the system with a reactive power margin of 139.5 MVAr and percentage change in voltage magnitude of 32.06%. The proposed method had an R-value of 0.9975 with a Mean Square Error (MSE) of 2.182415x10−5 for the IEEE 14-bus system and the NNG 28 bus system, an R- value of 0.9989 with an MSE of 1.2527x10−7. In this dissertation, the results presented indicate that artificial neural networks and voltage stability indices are highly capable of assessing the voltage stability of power system networks.

Item Type: | Thesis (Masters) |
---|---|

Uncontrolled Keywords: | ANN, Critical line, MATLAB, NNG, Power system networks, Voltage stability, Voltage stability indices. |

Subjects: | T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering |

Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |

Depositing User: | Mrs Hannah Akinwumi |

Date Deposited: | 24 May 2022 09:51 |

Last Modified: | 24 May 2022 09:51 |

URI: | http://eprints.covenantuniversity.edu.ng/id/eprint/15874 |

### Actions (login required)

View Item |