AI-driven multi-objective optimization of FACTS device placement using self-adaptive metaheuristics and neural networks

Voltage constraints and stability issues often limit the full utilization of transmission line capacity. While Flexible Alternating Current Transmission System (FACTS) devices have been used widely for power system performance enhancement, the existing optimization procedures suffer from a lack of coordination between optimization and intelligent decision-making techniques and early convergence in determining the optimal placement of FACTS. To overcome these drawbacks, this paper presents an Optimal FACTS Device Allocation with Wind Energy (OFDA-WE) scheme using Self-Adaptive Bat Optimization (SABO) algorithm combined with Chaotic Fuzzy Multi-Model Neural Network (CFNN). The SABO algorithm achieves the global optimization results of FACTS device placement, and the CFNN algorithm adds an angle-based chaotic search mechanism to improve the optimization process to achieve better results in terms of search capability and the challenge of premature convergence. The proposed framework is designed to achieve optimal power flow, voltage stability, transmission losses, and economic performance for different levels of wind penetration. The effectiveness of the proposed approach is tested on IEEE 30-bus and IEEE 118-bus benchmark systems. The results obtained from the simulation confirm that the proposed OFDA-WE framework can realize the maximum loadability of the system as 165%, which is higher than the loadability of the system when using the OFDA-HCI framework and the OFDA-DLR framework, while the transmission loss is 4.86 MW, lower than the transmission loss obtained by the OFDA-HCI framework and the OFDA-DLR framework. The proposed approach also yields better voltage regulation, better generator dispatch and optimal use of FACTS under high wind penetration.