Date of Award
2026-05-01
Degree Name
Doctor of Philosophy
Department
Computational Science
Advisor(s)
Tzu-Liang (Bill) Tseng
Abstract
The global energy landscape is undergoing a profound transformation, driven by the urgent need to decarbonize power systems, enhance energy security, and meet growing electricity demands. Solar photovoltaic (PV) power has emerged as a critical component of future energy infrastructure due to its abundance, scalability, and cost-effectiveness. However, PV generation is inherently variable and weather-dependent, introducing significant uncertainty into grid operations and complicating the task of balancing supply and demand. Accurate forecasting of PV power generation-particularly on day-ahead and hour-ahead horizons-has become a strategic necessity for grid stability, economic efficiency, and environmental sustainability. PV output is influenced by numerous factors, including panel orientation, geographic location, meteorological conditions, shading, equipment health, and system design. Understanding these factors and their interactions is essential for predicting grid load and optimizing energy dispatch. This dissertation addresses these challenges by proposing a novel Bayesian deep learning framework for uncertainty-aware PV power forecasting. The approach integrates hierarchical Bayesian neural architectures with spatial-temporal modeling to capture complex dependencies in PV generation data. Unlike conventional deterministic models, the proposed method provides probabilistic forecasts that are explicitly calibrated to achieve statistically reliable uncertainty estimates, enabling explicit quantification of epistemic and aleatoric uncertainty. Beyond methodological innovation, this research introduces a decision-theoretic operational framework that leverages uncertainty estimates to optimize reserve allocation and mitigate grid reliability risks. Using real-world datasets from the Texas Community Solar Assets that are part of the El Paso Electric Facility. The study evaluates predictive performance through advanced probabilistic metrics and simulates operational scenarios to assess economic and reliability impacts. By bridging innovative machine learning techniques with practical grid management strategies, this work contributes to the development of robust forecasting systems that support renewable integration and inform policy for sustainable energy futures.
Language
en
Provenance
Received from ProQuest
Copyright Date
2026-05
File Size
173 p.
File Format
application/pdf
Rights Holder
Pablo Abraham Bustamante
Recommended Citation
Bustamante, Pablo Abraham, "Bayesian Deep Learning For Photovoltaic Power Forecasting: A Probabilistic Framework For Uncertainty Quantification And Grid Reliability Optimization" (2026). Open Access Theses & Dissertations. 4636.
https://scholarworks.utep.edu/open_etd/4636