GRU-based Power Consumption Prediction for Energy Management in Smart Building

Currently, the smart grid business is increasing worldwide. The energy storage system (ESS), which is essential for the smart grid business, is an ESS that charges the battery in advance and supplies power stably when needed. If you use ESS to charge during early morning hours when electricity rates are low and use it during peak-time hours when electricity rates are high, electricity usage charges can be very low. Currently, most ESSs are scheduled to be scheduled semi-automatically by predicting power consumption. However, if the ESS is automatically scheduled by predicting the power consumption through deep learning, the efficiency of electricity use of the ESS can be increased, and the electricity usage fee can also be lower than that of the existing semi-automatic ESS. In addition, Korea's existing electricity structure has a problem in that more than 15 % of the existing demand must be produced in preparation for peak times. However, if ESS and gated recurrent unit (GRU) are combined and distributed, it will be possible to gradually reduce the amount of electricity that is created in advance because it is produced according to an accurate forecast amount.