Long Short Term Memory Recurrent Neural Network and Statistical Time Series Analysis Forecast Models for Wind Farm Power Output and Electricity Price

Boland, John, Chiera, Belinda, Cirocco, Lui, Chopin, Josh

ISES Solar World Congress 2025 · Fortaleza, Brazil · 2025-11-03
Published by International Solar Energy Society (ISES)

Abstract

There is a plethora of modelling methods for forecasting weather and renewable energy generation utilising both statistical time series analysis and artificial neural networks.   Autoregressive Moving Average (ARMA) and Long Short-Term Memory (LSTM) neural networks are both versatile and easy to train. In this paper we construct one step ahead forecast models for wind farm output and regional electricity price data using both traditional time series methods and hyper parameter tuned LSTM models to compare and contrast both the model performance and the modelling effort.

Keywords

Wind farm, Electricity price, Forecasting, ARMA, LSTM.

← All proceedings papers