Application of neural networks for load forecasting in a regional power company

The supervision of purchases from power generators is, for each regional power company (RPC), an important economy measure. A prerequisite for this measure is to have knowledge of short-term load development in the area to be supplied. In an automated load forecasting system, the accuracy of the forecasts is increased in order to give the experienced RPC a better safety margin in critical decision-making situations. This can only be achieved when the factors influencing the load are taken into account. Furthermore, an automated load forecasting system, which consists of hardware and software components, must be capable of being integrated in a functional manner into the existing grid control and load control system of the RPC, in order to achieve a successful upgrade. An upgrade, because of its lower cost when compared to a self-contained system which has its own process control, is especially suitable for local and regional power companies