Abstract

This article presents a new method for selection and combination of weather stations based on Wavelet Squared Coherence (WSC) that seeks to improve the performance of forecast models for the electric load. An Average Coherence Index (ACI) is obtained from WSC and used as an indicator of the relationship between the temperature of a weather station and the load of an area served by a utility. The ACI is used to categorize stations and select the best combination of stations for a forecast model. The method is validated with data from 18 load zones and 11 weather stations, offering a lower Mean Absolute Percentage Error (MAPE) than similar methods.