Appliance Scheduling Optimisation Method Using Historical Data in Households with RES Generation and Battery Storage Systems

In recent years, the importance of reducing car-bon dioxide (CO2) emissions has increased. With the use of technologies such as artificial intelligence, we can improve the way households manage their energy use to decrease cost and carbon emissions. In this paper, we use the Spectral Entropy and Instantaneous Frequency-based Bidirectional Long Short Term Memory (SE-IF BiLSTM) method so the home energy manage-ment system (HEMS) can learn from historical data of energy usage, as well as the preferred energy consumption patterns for the user. With this data, a multi-objective optimisation problem (MOP) that considers cost, C02 emissions and discomfort is formulated to schedule appliances in different scenarios. These scenarios include households with battery storage systems and with or without renewable energy sources. We compared the results by using multi-objective immune algorithm where we found a 10.06% reduction in cost and 20.56% reduction in CO2 emissions by using the proposed method.