Real Time Investigation on Congestion Forecasting in De-regulated Power Markets using Artificial Neural Network

This paper demonstrates a short-range electricity load and cost forecasting for a system via Artificial Neural Network. Bagging regression tree models that were utilized to calibrate the forecast, day-ahead loads, given temperature prediction, holiday load data, and historical loads periodically. Bidding and trading activities among the market traders promote the accurate estimation in power system planning and operation. The models are shown to produce highly accurate day-ahead forecasts with average errors which in turn help’s in identifying the factors such as congestion and other abnormalities to be modified or adjusted to maintain system integrity. Operations involved in minimizing the estimated error and the data requirements are exploited substantially reduced thus the operation becomes much simpler and easier. Forecasting results, based on the information obtained from the New-England ISO validate the practicability and precision are explained using Artificial Neural Network (ANN).