Conference Year



Flash tank, Air-source heat pump, Extremum seeking control


Vapor injection (VI) techniques have been well received as an effective technology for improving the performance of air-source heat pump (ASHP) under very low ambient temperature, for which the flash tank cycle (FTC) and the internal heat exchanger cycle (IHXC) are the two majaor configurations. In principle, FTC has higher achievable performance than IHXC because that the saturated vapor from the flash tank has a lower temperature which helps reduce the compressor discharge temperature and thus power consumption [1]. However, development of the FTC technology has been hampered by the lack of proper control/operational strategy that can optimize the thermodynamic characteristics of the vapor injection channel under variable ambient and load conditions. In the flash tank, the refrigerant is separated into the liquid and saturated vapor phase. The liquid refrigerant enters the lower-stage expansion valve and then circulates through the evaporator before entering the suction side of compressor, while the saturated-vapor refrigerant is injected into the intermediate pressure port of compressor. As saturated vapor is in principle the best choice for the vapor injection channel, superheat adjustment via the upper electronic expansion valve (EEV) is no longer viable. A liquid level measurement for the flash tank has been considered as feedback for the EEV control, however, determination of the optimum liquid level is rather difficult for practical operation due to the complexity of the underlying process and diversity in operating condition. In this paper, we propose an extremum seeking control (ESC) based strategy for efficient operation of the FTC-VI based ASHP heating systems [3]. ESC is a model-free real-optimization strategy, which is a dynamic gradient search with the online gradient estimation realized by a dither-demodulation scheme. For this problem, the setpoint for the intermediate pressure of injected saturated vapor is adopted as the manipulated input of the ESC, which is adjusted by the opening of the upper EEV via an inner-loop proportional-integral (PI) controller; while the total power consumption of the system is the only feedback needed. The heating load is regulated by the compressor capacity. To evaluate the proposed ESC strategy, a Modelica based dynamic simulation model of an FTC-VI based ASHP water heater is developed with Dymola and TIL Library. The hot-water outlet temperature is regulated by the compressor capacity, while the upper-EEV opening is used to regulate the intermediate pressure and liquid level of the flash tank. Simulation study is performed under different scenarios of ambient and thermal load conditions. The results show that the ESC is able to find the optimum intermediate pressure (corresponding to the optimum flash tank liquid level) by adjusting the upper EEV, which minimizes the total power consumption without sacrifice of heating load regulation and thus maximizes the system COP. To the authors’ best knowledge, the proposed strategy is a novel control solution to the optimal operation of FTC-VI ASHP systems, which does not require plant models or sensor measurements beyond power consumption. The presented results promises a great potential for the proposed strategy to facilitate the adoption of FTC technology.