Publication

Abstract : This study investigates bio-inspired metaheuristic algorithms for robust proportional integral derivative (PID) controller tuning in nonlinear boiler-turbine systems (BTS), critical for efficient thermal power generation. Conventional control strategies often falter against BTS nonlinearities, multivariable couplings, and time-varying dynamics. Conventional PID controllers are the most commonly used control approaches for industrial BTS due to its simplicity, reliability, and ease of implementation. PID controllers function best when parameters are properly tuned, especially for complicated and nonlinear processes like the BTS. The Bell and Åström boiler model features nonlinearities, multivariable interactions, non-minimum phase characteristics, and operational limitations, making traditional tuning strategies insufficient for dynamic performance under diverse conditions. By balancing transient performance, steady-state accuracy, and control effort, objective function allows fair and consistent evaluation of optimization strategies. This research allows for systematic evaluation of bio-inspired algorithms’ performance in addressing BTS tuning issues. Five bio-inspired optimizers—particle swarm optimization (PSO), Grey Wolf Optimization (GWO), golden eagle optimization (GEO), whale optimization algorithm (WOA), and artificial rabbits optimization (ARO) are evaluated for PID parameter tuning, minimizing integral performance criteria (IAE, ISE, ITAE) to enhance transient response and tracking accuracy. These algorithms emulate natural behaviors for balanced exploration- exploitation: PSO via social cognition, GWO through hierarchical hunting, WOA with bubble-net foraging, ARO via evasion tactics, and GEO mimicking predatory soaring. The comparison of the bio-inspired algorithms signifies the improvement of controller tuning by reducing overshoot, rising time, and settling time. For the control of drum pressure, the bio-inspired algorithms showed nearly identical results, with rise times of ≈ 0.80 s and settling times of ≈ 0.98 s. Overshoot values are less (<0.0003%), but WOA technique achieved the lowest overshoot at 0.00022%, a 12% reduction compared to others bio-inspired algorithms (0.00025%). Bio-inspired algorithms for electric power control showed comparable performance, with rising times of 0.75 s and settling times of 0.92 s. While overshoot values were mod- est (<0.0004%), the GEO approach had the lowest overshoot (0.00028%), a 15% reduction compared to other bio-inspired algorithms (0.00033%). For the control of the drum water level, PSO, WOA, and ARO settled with a settling time of 3.54 s, significantly faster than GEO (12.47 s) and GWO (16.69 s), corresponding to an ≈ 79% reduction in the settling time com- pared to GEO and ≈ 79% faster than GWO’s response. Overshoot for GEO techniques (1.39%) was higher than the overshoot of the other bio-inspired algorithms (<0.0001%), indicating stability issues. A thorough analysis of the optimization algorithms showed that all evaluated approaches performed well for electric power and drum pressure control, with GWO and PSO showing the lowest variance and highest consistency. Notably, ARO, WOA, and PSO performed well in limiting overshoot and attaining faster settling times, while GWO performed exceptionally well for drum water level in terms of Integral of Absolute Error (IAE)/ Integral of Squared Error (ISE)/ Integral of Time-weighted Absolute Error (ITAE) measures. These findings emphasize the specific advantages of ARO and WOA for performance metrics.