Dr. Nyangon Co-Authors New Study Advancing Electricity Price Forecasting in Renewable-Dominated Markets

January 20, 2024. Dr. Nyangon has co-authored a new peer-reviewed publication in WIREs Energy & Environment titled “Principal component analysis of day-ahead electricity price forecasting in CAISO and its implications for highly integrated renewable energy markets.” The study provides timely insights into how advanced data analytics can enhance electricity price forecasting in power systems undergoing rapid renewable integration.

Accurate electricity price forecasting is essential for grid reliability, market efficiency, and renewable energy integration. Yet, utilities and system operators often face challenges stemming from complex generation mixes and heteroskedastic price behavior. Drawing on the California Independent System Operator (CAISO)—one of the world’s leading renewable energy markets—the study analyzes hourly electricity price and demand data from 2016 to 2021 to address these challenges.

The research evaluates a combined methodological framework that integrates traditional interquartile range (IQR) outlier detection with advanced principal component analysis (PCA) techniques, including a supervised robust PCA (RPCA) approach. This framework significantly improves data quality by reducing skewness and enhancing symmetry prior to model development. The authors then compare multiple linear regression models using both raw and PCA-transformed features.

Results show that models built on PCA-transformed data, particularly when paired with SAS Sparse Matrix outlier removal via proc RPCA, achieve the highest forecasting accuracy. The findings underscore the value of PCA-based methods in improving day-ahead electricity price forecasts, supporting more efficient grid operations and accelerating the integration of solar and wind resources in highly renewable power markets.