Applicability of load forecasting techniques for customer energy storage control systems

Abstract

There is an opportunity for commercial customers to use energy storage to charge during low load periods and discharge during peak load periods to reduce demand charges. Energy storage control systems that incorporate load forecasts have an economic relationship with forecast error. The less the forecast error is, the more economically feasible energy storage will be. A range of time series forecast models and exponential smoothing forecast algorithms were compared to determine their applicability for use in these energy storage control systems. Model coefficients were estimated by regression and an optimization algorithm. The ARIMA model and double exponential smoothing algorithm performed the best out of the developed set of models.