Diffusion absorption refrigeration systems for thermally-driven cooling from low-temperature heat sources

Abstract

Diffusion absorption refrigeration (DAR) is a small-scale cooling technology that can be driven purely by thermal energy without the need for electrical or mechanical inputs. In this work, a detailed experimental evaluation was undertaken of two newly-proposed DAR units, aimed at solar-driven cooling applications in warm climates, especially in arid and semi-arid regions. Firstly, a detailed experimental evaluation was undertaken of a DAR unit with a nominal cooling capacity of 100 W. Electrical heaters were used to provide the thermal input which was varied in the range 150-700 W, resulting in heat source temperatures of 175-215 C at the generator. Tests were performed with the DAR system configured with the default manufacturer's settings (22 bar charge pressure and 30 % ammonia concentration). The measured cooling output (to air) across the range of generator heat inputs was found to be in the range of 24-108 W, while the coefficient of performance (COP) was in the range of 0.11-0.26. The maximum COP was obtained at a generator heat input of 300 W. Results were compared to performance predictions from a steady-state thermodynamic model of the DAR cycle, showing a reasonable level of agreement at the nominal design point of the system, but considerable deviations at part-load/off-design conditions. Temperature measurements from the experimental apparatus were used to evaluate assumptions in the estimation of the model state point parameters and examine their influence on the predicted system performance. Temporal variations in the heat-input power and temperature strongly affect the performance of DAR chillers. In particular, the instantaneous cooling power delivered by a DAR unit powered by an intermittent heat source such as solar heat directly depends on the dynamics of start-up and shutdown processes. Therefore, the dynamic performance of an 80 W thermally-driven DAR system using ammonia-water (NH3-H2O) as a working fluid and charged with hydrogen (H2) as an auxiliary gas was investigated. The cooling temperature was set to 5 C. The (electrical) heat-supply power (200 W to 700 W), charge pressure (15 bar, 18 bar, and 21 bar at a 200 W heat input), and mixture concentration (25%, 30%, and 38% at a 200 W heat input) were varied and their effects on the system start-up time were investigated experimentally. The start-up time decreased with increasing heat-input power, and increased with increasing charge pressure, but in both cases, the system COP reduced by 30%. Furthermore, it was found that reducing the charge pressure and increasing the refrigerant concentration led to COP improvements. Further to this, a simple mathematical model was developed for the DAR performance when this receives a transient thermal input. Performance maps were generated according to the operating temperature of the system (5 C for refrigeration and 23 C for air-conditioning) at three charge-pressure levels (15 bar, 18 bar, and 21 bar) and three refrigerant concentration levels (25 %, 30 %, and 38 %) which can be used to select the desirable properties of a DAR system according to application. The best DAR performance for constant heat sources was achieved when the charge-pressure was set to 18 bar and the refrigerant concentration to 38 % (COP of 0.350 for HVAC and 0.22 for refrigeration purposes). For intermittent heat sources such as a solar input, a typical day with maximum solar irradiance of 650 W/m2 is selected as a case study and cooling demand of 452 Wh for air-conditioning purposes is considered between 11:45 to 16:15. The system uses evacuated tube collectors to heat up the generator. The adjustable parameters (21 bar of charge pressure and the refrigerant concentration of 38 %) are selected based on the availability of the solar energy. The system performance was best with a 3.7 m2 solar array, a COP of 0.25 at which point the system covers 420 Wh of the cooling demand. A comparison of a DAR system with a vapour-compression refrigeration (VCR) unit when both receive input energy from the sun. Whereas the DAR system has to store some part or all of the solar energy in the form of cold (ice) for the use overnight, some VCR scenarios store the PV electricity into a battery array. The analysis shows that the DAR system can compete with VCR systems in terms of the levelised cost of cooling, which is 0.32 $US and that is 0.16 $US less than the VCR best scenario. However, the DAR systems need more space to be installed because of their significant solar array areas. Apart from occupied space, DAR systems are more profitable than VCR systems when both cooling systems are totally driven by solar energy.