High discharging performance regulation of phase change materials based thermal energy storage using an advanced controller tuned by artificial intelligence approaches: A comparative study

Thermal energy storage based on phase change material (PCM) is recognised as a promising technology for efficient waste heat recovery and carbon emission reduction. However, achieving high-performance thermal discharging remains a key challenge for practical deployment. The complex nonlinear dynamics of PCM storage systems and the strong dependence on high-quality data limit the reliability and industrial applicability of model predictive control and machine learning–based control approaches. This study proposed an industry-oriented control solution based on an advanced Proportional–Integral (PI) controller tuned using nature-inspired Artificial Intelligence (AI) algorithms. A 200 kW-level distributed energy system integrated with PCM thermal energy storage was developed and experimentally tested to validate and compare the proposed control methods. The results showed that the PI controller tuned by Artificial Rabbit Optimisation (ARO) achieved superior temperature and power regulation performance compared with the empirical Ziegler-Nichols method and the conventional evolutionary search of Genetic Algorithm approach. The ARO-based framework provided a more adaptive exploration-exploitation balance while directly minimising the closed-loop tracking error under nonlinear thermal dynamics and actuator constraints. Based on the validated control results, a comprehensive techno-economic analysis was conducted to evaluate the influence of control strategies and operating setpoints on system performance. The results indicated that constant power regulation maintained a higher exergy-to-energy ratio for comparable thermal energy output, making it suitable for power generation applications. Meanwhile, the constant temperature strategy exhibited an optimal setpoint that maximised net profit and minimised payback period by balancing economic gains, capital investment, and operating expenditure. Overall, this study provided an AI-enhanced and industrially viable control strategy for improving the discharging performance of PCM thermal storage systems and offered valuable insights for their large-scale deployment in waste heat recovery and low-carbon energy systems.