Leveraging Publicly Available Data and Production Cost Models for Risk Assessment and Enhanced Operational Strategies in Electric Power Systems

Abstract

Decarbonizing the power sector requires a significant increase in electricity generation from intermittent variable renewable energy resources such as wind and solar. Despite ongoing efforts to expand VRE penetration, the clean energy transition faces significant integration challenges, including limited data accessibility, inadequate risk management strategies, and interprovincial barriers. The dissertation aims to address these VRE integration challenges and improve power system operations to ensure the affordability, reliability, and sustainability of electricity. Chapters 2 and 3 establish a standardized protocol for collecting, assessing, and utilizing publicly available data. Chapter 2 identifies key data variables required for power system models, sets quality standards, and lists publicly accessible sources that meet these criteria. Chapter 3 outlines how to preprocess and harmonize data from multiple sources, transforming them into ready-to-use inputs for power system models. Together, these two chapters enhance the transparency of model inputs and foster collaborative discussions. Chapter 4 introduces a risk-scoring mechanism to assess reliability risk at the electricity generating unit (EGU) level. I employ GADS data, eGRID data, and Risk-Adjusted Stochastic UC models to estimate the probability and consequences associated with the sudden and prolonged failure of a dispatchable EGU or storage unit. The framework is applied to a U.S. power system using data from July and December 2022 to demonstrate its effectiveness. Our results reveal that forced outages among the originally available units are rare events. While failure probabilities vary among EGUs within the same technology group, the consequence assessment displays a high degree of intra-group consistency, especially when CVaR differences are expressed in $/MWh. The 10% solar derating scenario results in substantially higher risk scores, driven by both elevated failure probabilities and significant consequences. Chapter 5 evaluates the economic, environmental, and reliability benefits of different BA expansion strategies - considering both full and partial integration - within a Chinese power system. The analysis simulates full-year operations in 2023 using a deterministic Production Cost model under different VRE-dominant scenarios. I find that full BA integration reduces total system operational costs by 37% (saving 138.7 billion RMB in the modeled year), load shedding by 98% (6.6 TWh/yr), and carbon emissions by 8% (74 million tonnes/yr) compared to provincial dispatch, primarily driven by a 37% increase in hydro generation. Under a solar-dominated scenario, these benefits are even more pronounced, with cost savings reaching 40% (saving 164 billion RMB).