Forecast of schedulable capacity for thermostatically controlled loads with big data analysis

Abstract

Demand Response (DR) is an important technique to realize the reliable and economic operation of power system. As a key demand side resource, thermostatically controlled load (TCL) can be used through various demand side project to interact with power system for maintaining power system stability. In this paper, we propose the schedulable capacity forecasting (SCF) methods for thermostatically controlled load by big data analysis method. Under the Hadoop and Spark platform, considering the users comfort constraint and using the indoor, outdoor temperatures and real-time TCLs status to forecast the real-time schedulable capacity (SC) of TCLs. Meanwhile, using the large amount of results of the real-time SCF and weather information as historical data, for training the parallel random forest (RF) and compared with the decision tree (DT) model to do the one day ahead SCF. The results provide the basis for the economic operation analysis and energy optimization of the active distribution network system. The simulation results show that the proposed methods can be used to predict the schedulable capacity of the TCLs in a fast and accurate manner.