S.S. Halilčević and C. Moraga
Automatic generation control, synchronizing equipment, frequency
This article presents a model of an automatic starter of the synchronizer and additional power source. As it is known that new stable state after a generator outage depends on the timing of the countermeasures, we propose an automatic starter that can lead to a faster and more secure decision on the issue of interconnecting power sources. The soft part of the proposed starter is based on an extended feed-forward, multilayer neural network. The multilayer neural network consists of a three-layered neural network using neurons with sigmoid activation and is extended by a perceptron with a biased hard limit transfer function. The adopted extended neural network is excited by signals of the generator’s operating status and by results of the logical decision-making process that take into account the power system load. The automatic starter reduces the time spent in making out whether the rescue action will involve resorting to other power sources. Such a reduction is a contribution to efforts at preserving a power system’s integrity during critical situations (e.g., generating unit outages). The extended neural network is able immediately to recognize crisis symptoms and suggest introduction of other power sources through the synchronizing equipment.
Important Links:
Go Back
