WSEAS Transactions on Power Systems
Print ISSN: 1790-5060, E-ISSN: 2224-350X
Volume 20, 2025
Application of Shape Functions to the Calculation of an Annual Electricity Demand Forecast
Authors: ,
Abstract: This work proposes a methodology to construct an electricity power demand annual profile using a novel model to reproduce the demand behavior during weekends and holidays. These days have the common characteristic that the demand decreases during the day, or weekend, and then increases again. This behavior is represented by a simple deterministic model that is systematically applied to a normalized hourly demand profile based on similar days, allowing a relatively fast construction of an annual profile that reflects the actual demand characteristics and is useful for load demand forecasting, and as support for other medium or long term analysis, such as electrical expansion planning or fuel economics planning. The electricity demand profile construction starts with hourly measurements of demand as input and a base profile is prepared with historical data from previous years. It is based on the characterization of the weekdays by normalization and grouping into several time periods along the year. The base profile made with normalized days is then shaped by functions that allow the characterization of the demand behavior during weekends and holidays. In this work, a shape function is a one-dimensional vector that multiplies a demand vector and modifies its data for an interval of interest, leaving the rest of the vector unchanged. For the case of weekend modeling, the shape function spans 7 days, centering the modification on the weekend and leaving the initial and final days unchanged. The shape function for a public holiday spans two days and does not modify all the two-day interval, preserving the initial part of the first day and the last part of the second day. The objective is to generate shape functions with a simple model that systematically represents the real demand with low computational effort. In this work, the shape functions for weekends and holidays are based on the gamma probability distribution. The shape functions approach does not explicitly consider the weather, but it implicitly considers stationarity effects by dividing the yearly time data into segments, each one with its own characteristic properties, which vary along the year. The shape functions methodology is demonstrated with the construction of a power demand forecast for the Mexican National Interconnected System for the year 2022.
Search Articles
Keywords: Demand forecasting, load forecasting, holiday load, annual electricity forecast, hourly forecast
Pages: 42-53
DOI: 10.37394/232016.2025.20.4