Modelling of sanitary hot water energy consumption using adaptive neuro-fuzzy inference systems

The electrical energy consumption associated with sanitary water heating makes up a large part of the total load associated with residential energy consumption, and therefore load models thereof could find use in various Energy Management (EM) applications. This paper presents the results of an investigation to model the electrical load associated with the combined sanitary hot water heating systems of 21 university residences using Adaptive Neuro-Fuzzy Inference Systems (ANFIS) within the MATLAB platform. The desired prediction horizon is defined as medium term, i.e. up to a year ahead forecasting. The training inputs considered include temperature, day of year, day of week, and daily time interval. The effects of compartmentalising the dataset into subsets representing different characteristics, thereby deriving different models representing different cyclic periods, are explored. K-fold cross validation is used in conjuncture with Mean Average Percentage Error (MAPE) calculations to provide a comprehensive breakdown of model performance.