Parameter identification of Hammerstein model for heat pumps based on SABO-CR algorithm

• Published in: Data Driven Control and Learning Systems Conference (Online) pp. 1809 - 1814
• Main Authors: Yang, Xiaohan, Zhao, Yihan, Lu, Meng, Li, Junhong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 09.05.2025
• Subjects: Atmospheric modeling; Computational complexity; Computational modeling; Control systems; Convergence; Data models; differential variance strategy; Hammerstein model; Heat pumps; Mathematical models; Nonlinear system; Parameter estimation; SABO-CR algorithm; Valves
• ISSN: 2767-9861
• DOI: 10.1109/DDCLS66240.2025.11065408

This paper discusses the identification of Hammerstein model for nonlinear air source heat pump control system exhibiting dead-zone characteristics. The control valve in the air source heat pump control system presents nonlinear characteristics. The Hammerstein model consists of static nonlinear block and dynamic linear block. In this paper, we establish the mathematical model of each series link of the heat pump control system based on the Hammerstein model. A single-input single-output Hammerstein model for air source heat pump control system with dead-zone characteristics is proposed. To overcome the limitations of the Subtraction-Average-Based Optimizer (SABO) algorithm, such as escalating computational complexity with iterations and susceptibility to local optima, we introduced differential variance factors and proposed the SABO-CR algorithm. The SABO-CR algorithm is employed to identify the model parameters using actual sampling data. Experimental results shows that SABO-CR algorithm has higher convergence rates, reflecting its superior performance.