ENERGY USE PREDICTION BY MULTIPLE-TIME SERIES AND STATE VARIABLE MODELLING BASED ON ENERGY MONITORING SYSTEM FOR INDUSTRIAL FACILITY IN INDONESIA

Measures for energy saving and de-carbonization in industrial sector is of great importance to realize a low-carbon society. In Indonesia, promoting “Green Industry” is a critical policy carried out by national government which focuses on reducing energy consumption in industrial sector. However, the current difficulty in accurately measuring energy consumption and energy-saving effects in Indonesia becomes an unavoidable barrier during policy implementation. This study conducts a survey on real-time energy consumption of industries in Indonesia based on an innovative energy monitoring system with high resolution at each device through public and private collaboration. An energy consumption prediction model is developed for each industrial process using the auto-regression exogeneous modeling and markov switching modeling. Results indicate the feasibility of developing an energy consumption prediction model in Indonesia, and reveal a high consistency between measured value and predicted value. This will also help in system design for demand response due to the high-accuracy measuring and prediction.