AbstractView references

This paper presents a novel procedure for short-term load forecasting in distribu-tion management systems. The load is forecasted for feeders that can be of a pri-marily residential, commercial, industrial or combined type. Each feeder has var-ious amounts of distributed energy resources installed, which accounts for multi-ple different load patterns. Hence, the distribution management systems cannot use a single short-term load forecasting model for all forecasts. The proposed procedure addresses the specificity of each particular feeder type, by creating customized short-term load forecasting models. It uses a genetic algorithm to se-lect the best inputs for different multiple linear regression models. The genetic algorithm chooses variables from a dataset constructed using load and tempera-ture measurements. The dataset is extended by adding non-linear transformations and mutual interaction effects of the measurements, as well as calendar varia-bles. This extension enables for the modelling of non-linear influences and ex-tracts the non-linearity to the domain of input variables. The models’ perfor-mance is assessed by the mean absolute percentage error. The proposed proce-dure is applied to a set of measurements collected from an US electric power util-ity, which operates in the city of Burbank, Cal., USA. The obtained multiple line-ar regression model is compared with a previously proposed naïve benchmark, and a special comparison model, developed by correlation analysis. The pro-posed method is extendable to suit distribution management systems with differ-ent types of electricity consumers. © 2021 Society of Thermal Engineers of Serbia.