A Novel Transformer-Based Approach for Reliability Evaluation of Composite Systems With Renewables and Plug-in Hybrid Electric Vehicles
This paper proposes a novel hybrid framework that integrates machine learning (ML) techniques with Sequential Monte Carlo Simulation (SMCS) to enhance the reliability assessment of modern power systems incorporating renewable energy resources (RER) and plug-in hybrid electric vehicle (PHEVs) integra...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10976708/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This paper proposes a novel hybrid framework that integrates machine learning (ML) techniques with Sequential Monte Carlo Simulation (SMCS) to enhance the reliability assessment of modern power systems incorporating renewable energy resources (RER) and plug-in hybrid electric vehicle (PHEVs) integration. While PHEVs can leverage RER to significantly reduce greenhouse gas emissions, the increased energy demand from large PHEVs fleets poses potential challenges to power system reliability. To address these issues, this research presents an advanced mixed-integer linear programming (MILP)-based algorithm for optimizing EV charging. The algorithm prioritizes clean energy utilization through intelligent power allocation strategies while considering cost-revenue trade-offs. A probabilistic model is developed to account for factors such as driving distance, charging times, locations, battery state of charge, and charging needs of PHEVs. The proposed approach is tested on the IEEE RTS-79 test system and evaluates multiple ML architectures, including Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and Transformer models, often combined with boosting algorithms, across three scenarios: base case, uncontrolled charging, and intelligent charging. Results highlight that ML-based approaches, particularly the Transformer model, achieve computational time reductions of up to 49% compared to traditional SMCS methods while maintaining comparable accuracy. The Transformer model identified 1,788 loss-of-load states compared to 1,510 actual instances, requiring only 176 minutes of computation. Among all models, the BiLSTM with Adaptive Boosting (BiLSTM+AB) achieved the lowest overestimation, exceeding actual instances by just 256 states. Performance metrics such as Loss of Load Probability (LOLP) and Expected Demand Not Supplied (EDNS) validate the effectiveness of the proposed ML approaches in balancing accuracy and computational efficiency. |
|---|---|
| ISSN: | 2169-3536 |