Residential load forecasting by a PSO-tuned ANFIS2 method considering the COVID-19 influence

The most important feature of load forecasting is enabling the building management system to control and manage its loads with available resources ahead of time. The electricity usage in residential buildings has increased during the COVID-19 period, as compared to normal times. Therefore, the performance of forecasting methods is impacted, and further tuning of parameters is required to cope with energy consumption changes due to COVID-19. This paper proposes a new adaptive neuro-fuzzy 2 inference system (ANFIS2) for energy usage forecasting in residential buildings for both normal and COVID-19 periods. The particle swarm optimization (PSO) method has been implemented for parameter optimization, and subtractive clustering is used for data training for the proposed ANFIS2 system. Two modifications in terms of input and parameters of ANFIS2 are made to cope with the change in the consumption pattern and reduce the prediction errors during the COVID-19 period. Simulation results obtained by MATLAB software validate the efficacy of the proposed ANFIS2 in residential load forecasting during both normal and COVID-19 periods. Moreover, the performance of the proposed method is better than that of the existing adaptive neuro-fuzzy inference system (ANFIS), long short-term memory (LSTM), and random forest (RF) approaches.