ABSTRACT
On the 330kV Nigerian Network modelled with Matlab R2014a, this work proposes a better protection solution based on the usage of artificial neural networks. The neural network receives inputs from measured fault voltages and currents that have been decomposed using the discrete Fourier transform implemented via fast Fourier transform. The neural network’s output plots demonstrate its successful application in defect diagnostics (fault detection, fault classification and fault location). The neural network’s application to fault location has a mean square error of 3.5331 and a regression value of 0.99976, indicating that the output and target values provided to the neural network have a very close relationship. Unlike traditional protection techniques, the neural network can be adjusted to distances that span the entire length of the network. This research also looks into an artificial neural network-based adaptive auto reclosure system. To discern between transitory and persistent problems, the adaptive reclosure system has been successfully modified for usage in the Nigerian Network. The adaptive reclosure technique was able to detect a line-to-ground transient fault and clear it in 0.1s, while the line-to-ground permanent fault was cleared in 0.14s, according to simulation data. The auto reclosure technique is built utilizing two distinct neural networks, one to distinguish defects as transient or permanent, and the other to classify decisions as’safe to reclose’ represented by logic ‘1’ or ‘do not reclose’ represented by logic ‘0.’ The Fault Diagnosis Algorithm was developed by
CHAPTER ONE
INTRODUCTION
BACKGROUND OF THE STUDY
The demand for steady electricity supply in Nigeria is growing, yet it is being met with many constraints. System flaws are one of them. Faults on transmission lines, in particular, are of major interest to Nigeria’s power holding company, since more money is invested in restructuring and extending existing infrastructure.
Nigeria’s power sector is organized into four categories: policy, regulations, consumers, and operations. The operations division sheds light on the activities of Nigeria’s transmission firm, which is in charge of high-voltage power transmission from producing facilities to substations for distribution to distribution stations. T.C.N. manages a 330kv system with a capacity of 6870MW over a distance of 5650 kilometers[1], with the goal of maintaining power system stability. Distance protection, over current protection, differential protection, and other protection schemes are being used. Distance protection, which is the most common, suffers from inaccuracy because to relay constraints on protection systems, such as reach settings. The relay is unable to properly adapt to changes in power system circumstances, particularly on parallel lines, or to discern between temporary and permanent faults in the event of a short circuit.
This paper examines the use of artificial neural networks for improved power system protection in terms of fault detection, fault location, and adaptive auto reclosure techniques, as opposed to traditional approaches such as travelling wave approaches and synchronous compensators, to mention a few.
STATEMENT OF THE PROBLEM
Transmission lines are one of the most significant components of the power system network, and they are frequently affected by many sorts of problems. The transmission line, as well as the rest of the substation equipment and bus bar, account for 80 percent to 90 percent of the faults. All power systems must maintain operational dependability, which can be accomplished by detecting, categorizing, and isolating various problems that occur in the system. The protective device must make a remedial decision in order to shorten the length of trouble and limit outage time, damage, and other issues. If any faults or disruptions in the transmission are not rapidly discovered, located, and eliminated, It may induce power system instability and considerable variations in system quantities such as overcurrent, undervoltage, power factor, impedance, frequency, and power. Single line to ground faults occur 70-80 percent of the time, line to line to ground faults occur 10-17 percent of the time, line to line faults occur 8-10 percent of the time, and three phase faults occur 3 percent of the time. The three flaws are uncommon, but if they do exist in a system, they can be highly costly.
Leave a Comment