Stochastic Modeling and Forecast of Hydrological Contributions in the Colombian Electric System

TL;DR

This paper models and forecasts the periodic and mean-reverting hydrological contributions to Colombia's electric system using stochastic processes, providing insights into their long-term behavior and short-term predictions.

Contribution

It introduces a novel stochastic modeling approach with periodic functional tendencies for hydrological contributions in Colombia's electric system.

Findings

Hydrological contributions exhibit a three-year periodicity.

Mean reversion stochastic processes effectively model the data.

Forecasts for three-year periods are successfully generated.

Abstract

In this paper as show that hydrological contributions in the colombian electrical system during the period between 2004 and 2016 have a periodic dynamic, with fundamental periods that are repeated every three years and that tend to oscillate around a long term average, in this context, such contributions can be characterized by mean reversion stochastic processes with periodic functional tendency. The objective of this paper is modeling and forecast the dynamic behavioral of hydrological contributions in the colombian electric system. A description of climate and hydrology in Colombia is carried out, as well as an analysis of periodicity and distributional properties of the data. A Gaussian estimation is performed, which allows to find all the constant and functional parameters from the historical data with daily frequency. The forecasts of the hydrological contributions are graphically illustrated for a period of three years and two alternatives to construct forecast fringes are proposed.