프로세스 제어 시스템의 PID 게인 자동 튜닝을 통한 강인한 제어

Abstract

Proportional-Integral-Derivative (PID) controllers are extensively utilized in industrial process control due to their simplistic design and exceptional control performance. Typically, PID gains are devised based on transfer function models of plants. However, deriving a low-dimensional linear approximation model for a variable speed refrigeration system (VSRS) becomes highly challenging due to its inherent nonlinearity and high-order dynamic characteristics. While there are fuzzy control and AI methods that do not rely on models, their control logic is extensive, necessitating expensive microprocessors and relying heavily on expertise and experience, thus making maintenance very difficult. Furthermore, model-independent PID gain tuning methods are cumbersome, making it difficult to achieve the desired control performance and apply them to real systems. Therefore, robust auto-tuning and PID control of VSRS, which is multi-variable system with strong inherent nonlinearity, are described in this paper. This method is a simple tuning method through ON/OFF operation, and it is an integrated system that automatically switches to PID control after tuning is completed. In addition, this method was applied to the actual system to conduct simulations and experiments, and the tuning method was modified and proposed due to the influence of heat noise and disturbance during the experiment. Moreover, in the case of a multi-variable system, the operating conditions of the remaining variables were clearly revealed when auto-tuning one variable. Experimental validation corroborates the efficacy of the proposed method. Stability and reproducibility are confirmed through analyses involving the pole position of the closed-loop control system and the bode diagram of the open-loop control system. additionally applied to the EHP air handling unit, and the validity of the proposed method is verified through simulation and experiment. These results substantiate the effectiveness of the proposed approach in ensuring stable control despite the inherent complexities of the system, thereby solidifying its potential for real-world applications.