Design of Self-Tuning SISO Partial-Form Model-Free Adaptive Controller for Vapor-Compression Refrigeration System

Vapor-compression refrigeration systems (VCRS) are applied extensively in domestic, commercial and industrial refrigeration and are responsible for a high percentage of worldwide energy consumption. To achieve high energy efficiency, the application of advanced control methods in VCRS has increasingly attracted the attention of academia and industry. The model-free adaptive control (MFAC) strategy, as an important branch of advanced control research, encounters the problem of parameter tuning when applied to VCRS with strong nonlinearities. In this work, a parameter self-tuning methodology based on the self-learning and self-adapting properties of back propagation neural networks with the System Error set and/or Gradient Vector set as inputs is proposed to adjust the parameters utilized in SISO Partial-Form Model-Free Adaptive Control (SISO-PFMFAC). To test the performance of this novel methodology named SISO-PFMFAC-NNSEGV, qualitative and quantitative comparisons are carried out between the proposed method and the decentralized single PIDs given in the simulation platform provided by the benchmark PID 2018. The integral absolute error (IAE), the integral time-weighted absolute error (ITAE), the integral absolute variation of control signal (IAVU) and a combined index Jc are used to evaluate the performance. As a result, the proposed method shows the best performance with a higher tracking accuracy and less variation of the control signal with a combined index J
_c
= 0.7088, which represents a 29.1% improvement over the benchmark PID controller and a 9.6% improvement over the SISO-PFMFAC controller, making it a promising control method for vapor-compression refrigeration systems.