A Two-Stage Forecasting Approach for Day-Ahead Electricity Price Based on Improved Wavelet Neural Network With ELM Initialization

In deregulated electricity markets, reliable day-ahead electricity price forecasting (EPF) is the basis for developing bidding strategies, operating dispatch controls, and hedging volatility risks. However, electricity prices are highly volatile, non-stationary, and multi-seasonal, making it challen...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on industry applications 2024-05, Vol.60 (3), p.5061-5073
Main Authors: Qu, Ziyu, He, Li, Ge, Xinxin, Wang, Fei, Xu, Fei, Lu, Jinling
Format: Article
Language:English
Subjects:
Algorithms
Artificial neural networks
Classification
Day-ahead electricity price forecasting
Deregulation
Electricity pricing
extreme learning machine
Forecasting
improved wavelet neural network
Machine learning
Neural networks
Optimization
Pattern recognition
Power systems
Prediction algorithms
Prediction models
Predictive models
Prices
Training
Volatility
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In deregulated electricity markets, reliable day-ahead electricity price forecasting (EPF) is the basis for developing bidding strategies, operating dispatch controls, and hedging volatility risks. However, electricity prices are highly volatile, non-stationary, and multi-seasonal, making it challenging to estimate future trends. So the accuracy of most existing forecasting models is insufficient to meet real-world needs. To this end, a two-stage forecasting algorithm combining pattern recognition, machine learning, neural network models, and classification prediction is proposed. The algorithm operates in two stages, with the first stage focusing on predicting patterns of day-ahead electricity prices with machine learning algorithms. In stage two, an improved wavelet neural network (IWNN) model based on extreme learning machine (ELM) initialization is proposed to build classification prediction models for different electricity price patterns, which effectively solves the problem of slow or even non-convergence of traditional wavelet neural networks. Case studies based on PJM market data show that the proposed approach outperforms other state-of-the-art approaches, especially when the volatility of electricity prices is high. Moreover, our algorithm exhibits reliable generalization capabilities.
ISSN:0093-9994
1939-9367
DOI:10.1109/TIA.2024.3365456