The generation ready-reserve power determined in the interconnected power system by means of the artificial neural network

The paper presents the usage of artificial neural network (ANN) for the online determination of the amount of the ready reserve (RR) required for compensation of power deficit in power systems. Here the power deficit is a consequence of the outage of the synchronized generators and failure of the other forms of the operational reserve to act compensationally. RR consists of rapid start-up generating units such as gas turbines and hydro plants. The amount of the deficit-compensating RR, given the stochastic nature of the power system´s work, is information without which no serious outage-counteraction can be prepared. Such information, along with a reliable assessment of its contents, is obtained from a supervised, backpropagation trained ANN. Prompt in assessing the scope of the anti-outage exploitation of the RR generators, it both minimizes the setback in supplying the consumers with power and, consequently, maintains a high reliability of the whole system.