Development of load profile prediction using tcbm and arima hybrid-modeling

For effective energy use and level off electricity demand throughout the day, thermal energy storage system is getting popular in HVAC systems in Japan. Load profile prediction can provide information to plan optimal operation, and now, various prediction approaches have been developed, for example, Autoregressive Integrated Moving Average (ARIMA) model, Topological Case Based Modeling (TCBM), Artificial Neural Network (ANN), and other approaches. To be considered characteristics of these approaches, hybrid prediction has have been applied. In this paper, load profile prediction approach using TCBM and ARIMA hybrid-modeling that has been developed for use in the optimal operation of thermal energy storage system is proposed. Performance of this approach was assessed by benchmark test held by the technical committee for optimization of thermal energy storage systems (TC-OTES) of heating, air-conditioning and sanitary engineers in Japan (SHASE). Results show good agreement between actual and predicted load.