A comparison of two extremum seeking control strategies based on simulation and laboratory tests for heat pump air conditioning

Abstract

Extremum seeking control (ESC) strategies are a class of model-free real-time optimization techniques. Because conventional ESC (CON-ESC) based on dither demodulation has relatively slow convergence, various alternative forms of ESC have been proposed to improve the transient performance. The purpose of this study is to compare a recently proposed input–output correlation ESC (IOC-ESC) strategy with that of CON-ESC for an air-source heat pump (ASHP) application. The ESC performance is evaluated under different tuning parameters, initial conditions, and measurement noise using a Modelica simulation model of the ASHP. Simulation results show that the IOC-ESC is simpler to tune, has less sensitivity to tuning parameter and initial condition, and converges more quickly than the CON-ESC. The performances of the two ESC algorithms are affected by measurement noise, but the IOC-ESC still achieves faster convergence under the measurement noise levels being evaluated. An experimental comparison is also made between the IOC-ESC and CON-ESC using a mini-split air conditioning system. The experimental results of both single-input and two-input ESC show that the IOC-ESC converges more quickly than the CON-ESC.

Acknowledgments

The authors are grateful for financial support from Johnson Controls International plc, and to TLK-Thermo for its permission to access the TIL Suite product for control simulation.