A Quasi-moment-method Modelling of Energy Demand Forecasting

This paper presents a novel approach to the modelling of electrical energy demand forecasting, based on Quasi-Moment-Method (QMM). The technique, using historical consumption/demand data, essentially calibrates nominated ‘base’ models (in this case, nominal Harvey and Autoregressive models) provide significantly better performing models. In addition novelty use QMM, identifies hitherto unreported singularities generic / logistic model, through which simple, but remarkably pivotal modification is proposed, prior model’s as base model in QMM calibration schemes. treatment ‘Harvey singularities’ informed similar equally significant utilized paper. For purposes validation performance evaluation, computational results due are compared with corresponding reported three different journal publications, conventional regression And terms usual metrics (including Mean Absolute Percentage Error (MAPE) Root Square (RMSPE)), very clearly demonstrate superiority for both prediction forecasting. As representative examples, QMM-calibrated recorded an RMSE value 495.45dB total consumption prediction, against 618.60dB obtained model: was 131.35dB peak load demand, 173.40dB model.