An Effective Short-Term Load Forecasting Methodology Using Convolutional Long Short Term Memory Network

In this research work, an improved short-term electrical load forecasting technique is developed. Short-term load forecasting plays a key role in scheduling of generating units and spinning reserve evaluation. It essentially estimates the near-future electricity demand, which helps to ensure seamless operation of all three stages of a power system - generation, transmission and distribution. There are a number of strategies for short-term load forecasting. However, conventional methods may not provide precise forecasting outcomes. Therefore, it is necessary to develop a more accurate technique to predict short-term electrical load. To this end, a convolutional long short term memory network (Conv-LSTM) is developed in this paper, which results in good forecasting precision. The developed methodology is applied to the electric load data of Bangladesh power system. Further, the performance of the proposed technique is compared to that of the conventional linear regression based forecasting algorithm in terms of different error metrics. Simulation results reveal that the proposed approach provides better forecasting and outperforms the conventional algorithm.