Enhancing power system reliability and carbon reduciton through fuzzy unit commitment with renewable energy and vehicle to grid system

Name: Seu, Sophany

keyword: Fuzzy mixed-integer linear programming (FMILP)

; Unit commitment

; Renewable energy integration

; Electric vehicle (EV) charging

; Vehicle-to-grid (V2G)

; Power system reliability

; Loss of load probability (LOLP)

; Expected energy not supplied (EENS)

; Energy storage systems (ESS)

; Pumped storage hydropower (PSH)

; Battery energy storage (BES)

Abstract: Driven by national policy initiatives aimed at decarbonization and long-term sustainability, the increasing penetration of renewable energy sources (RES) has introduced considerable uncertainty and operational complexity into large-scale power systems. As solar and wind energy continue to replace conventional dispatchable generation, the planning and scheduling of power system operations face significant reliability challenges. Simultaneously, the rapid adoption of electric vehicles (EVs) adds new layers of stochastic demand, while emerging technologies such as Vehicle-to-Grid (V2G) provide potential opportunities to treat EVs as both loads and mobile energy storage units. These developments necessitate advanced modeling frameworks that can ensure stable and efficient system operation in the face of high uncertainty. This thesis proposes a Fuzzy Mixed-Integer Linear Programming (FMMILP) model for day-ahead Unit Commitment (UC), designed to manage the combined uncertainties of RES generation, forecast load errors, and EV charging demand. The model explicitly incorporates Pumped Storage Hydropower (PSH) and Battery Energy Storage (BES) to enhance system flexibility and considers the potential role of V2G- (2) enabled EVs as dispatchable energy assets in future scenarios. The integration of fuzzy logic within the UC problem formulation allows imprecise and uncertain variables to be represented through membership functions, offering a more realistic and adaptive planning framework compared to traditional deterministic or stochastic approaches. To evaluate the model’s robustness, the research follows a two-phase validation strategy. First, a small-scale system is employed to demonstrate the theoretical soundness and operational performance of the proposed FMMILP models under various fuzzy logic aggregators. Second, the computational scalability and applicability of the model are tested using a detailed case study based on Thailand’s national power system in the year 2037. This large-scale simulation incorporates realistic data on renewable generation, EV adoption, and energy storage deployment across three seasonal scenarios: summer, rainy, and winter seasons. A key contribution of this research lies in the comprehensive reliability analysis, conducted using two established power system reliability metrics: the Loss of Load Probability (LOLP) and the Expected Energy Not Supplied (EENS). LOLP quantifies the probability that system demand will exceed available generation capacity within a given time frame, while EENS measures the expected amount of energy shortfall during these reliability breaches. The proposed FMMILP models, particularly those incorporating advanced fuzzy operators, demonstrate substantial improvements in both reliability and cost-efficiency when compared to the deterministic baseline. In the winter scenario characterized by high uncertainty and demand variabilitythe best-performing fuzzy model reduces LOLP by 51.44% and achieves a significant decrease in EENS, indicating enhanced system resilience under adverse operating conditions. In addition, a sensitivity analysis using Taguchi’s orthogonal array design is carried out to identify the most influential uncertainty parameters. The results show that spinning reserve requirements, load forecasting error, and solar power variability have the greatest impact on both operational cost and reliability. This sensitivity analysis provides valuable insights for power system planners in prioritizing investment and risk management strategies. The outcomes of this study have practical implications for policy makers and utility operators. By enabling more adaptive and robust generator scheduling, the proposed FMMILP model supports better alignment of operational decisions with (3) policy objectives such as carbon emission reduction and grid resilience. It facilitates effective planning under uncertain conditions, helping utilities to maintain high reliability while integrating higher shares of renewable and distributed energy resources. Future research directions include expanding the model to incorporate demand-side flexibility through dynamic pricing or demand response programs and modeling real-time operation adjustments. Additionally, the integration of spatially distributed renewable energy sources and region-specific V2G behaviors should be considered. Enhancing the computational efficiency of the model using decomposition techniques or parallel computing could further support its application in real-time or rolling-horizon dispatch environments. Finally, the development of a decision-support interface tailored for policy planners could translate these technical advancements into actionable insights for long-term energy strategy and infrastructure development.

Thammasat University. Thammasat University Library

Address: BANGKOK

Email: preserv@tu.ac.th

Name: Chawalit Jeenanunta

Role: advisor

Name: Sapa Chanyachatchawan

Role: co-advisor

Created: 2025

Modified: 2025-11-24

Issued: 2025-11-24

วิทยานิพนธ์/Thesis

application/pdf

DOI: 10.14457/TU.the.2024.524

eng

DegreeName: Master of Engineering

Level: Master's Degree

Descipline: Engineering Technology

Grantor: Thammasat University

©copyrights Thammasat University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	334734Seu.pdf	2.84 MB

ใช้เวลา

0.036005 วินาที