Residential Lighting Load Profile Predictor Using Computational Intelligence

Abstract This study presents the development, analysis and assessment of residential lighting load profile using computational intelligence based modelling Adaptive Neuro Fuzzy Inference System (ANFIS) and Neural network (NN) models for prediction (forecasting) and evaluation of lighting load and initiatives. Factors considered in the development of the models include natural lighting, occupancy (active) and income level. Trapezoidal membership and sigmoid transfer function were applied during the training process of the ANFIS-based and NN-based model respectively. Using computational and different validation approaches, ANFIS gave better correlation and error level results in comparison with the NN-based method analyses notably morning standard, morning / evening peak and daily TOU (time of use) periods. The inference attribute of the ANFIS model based on characterization factors and its reflection of occupants’ complexity on lighting loads in residential buildings makes it a better lighting predictor especially in demand side management & residential lighting load energy efficiency project initiatives.