Research on Improving the Control Performance of the Fuzzy PID-Smith Predictive Fusion Algorithm in Large-Lag Temperature System

Abstract

During the vacuum thermal test process, the infrared thermal cage, serving as the temperature control object, typically exhibits characteristics such as significant inertia, substantial lag, nonlinearity, and strong disturbance interference. Traditional PID control algorithms face challenges in achieving high-precision and high-stability temperature regulation, and a single improved control algorithm is unable to simultaneously satisfy the requirements of lag compensation, parameter adaptation, and robustness. To address these technical challenges, this paper proposes an improved fuzzy PID-Smith predictive fusion control algorithm. By optimizing the structure of the Smith predictor model, the algorithm reduces reliance on precise mathematical models of the controlled object, effectively compensating for the system’s pure lag component. Furthermore, an adaptive fuzzy inference mechanism is constructed to optimize the membership function and control rules of the fuzzy PID controller, enabling online dynamic tuning of PID control parameters. Using the MATLAB/Simulink simulation platform, a simulation model of a large-lag temperature control system is established, and tests including step response test, anti-disturbance performance test, model mismatch adaptability test, and multi-condition comparison test are conducted to comprehensively evaluate the control performance of the proposed algorithm. Simulation results demonstrate that, compared to traditional PID, fuzzy PID, and conventional Smith predictive PID algorithms, the improved fusion algorithm achieves an overshoot of only 1.2%, reduces regulation time to 150 seconds, and maintains a steady-state error within ±0.05℃. Even under disturbance suppression and model mismatch conditions, the algorithm maintains excellent control quality, showcasing stronger robustness and engineering applicability.