Peak load forecast via neural network / Abu Khir Abdul Hamid

This thesis presents a new signal processing technique calted aruficial neural networkfANN) to peak load forecast based on short term basis. The uncertain elements such as weather conditions, holiday's effect, and historica! load is taken into account. Experiments were canied out to choose the most suitable ANN model for peak load forecasting. The resuits acquired through the experiments shows that the proposed methodologies produce accurate peak load prediction under a wide variety of power systems operating conditions.