Load Forecasting based on Deep Long Short-term Memory with Consideration of Costing Correlated Factor

Abstract

In Day-ahead Power Market (DAM), Load Serving Entities (LSEs) needs to submit their load schedule to market operator beforehand. For reduction of the total cost, the disparity of the price of DAM and the price of RDM (Real Day Market) should be considered by the LSEs. Therefore, the problem is that a more accurate load-forecasting model sometimes provide a price that has an interspace will lead to a lower cost. Facing this issue, this paper initiates a load forecasting model considering the Costing Correlated Factor (CCF) with deep Long Short-term Memory (LSTM). The target of the forecast model contains both accuracy section and power cost section. At the same time, the construct of LSTM can of fset the sacrificed accuracy. Also, this paper uses an Adaptive Moment Estimation algorithm for network training and the type of neuron is Rectified Linear Unit (ReLU). A numerical study based on practical data is presented and the result shows that LSTM with CCF can reduce energy cost with acceptable accuracy level.