An adsorption refrigeration chiller has great potential for improving energy conservation. It uses adsorbing materials with low regeneration temperatures that can be regenerated by waste heat from a gas engine or hot water from a solar heat collector. To optimize the operation of an adsorption refrigeration chiller, a performance prediction method is required. In this study, a model of an adsorption refrigeration chiller was developed for annual building energy simulation. Models have been developed to calculate the thermodynamic properties of an adsorbent. Many of the adsorption refrigeration chillers currently available in the market use an AFI-type structure ferroaluminophosphate zeolite (FAPO-5) as the adsorbent. As the adsorption isotherms of FAPO-5 have sigmoidal characteristics in contrast to those of general zeolite or silica gel, they were approximated using a logistic function (Eq.(1)). Fig. 2 shows the adsorption isotherms of FAPO-5. To increase the calculation speed, a linear regression model (Fig. 3) was also proposed. The correlations between relative pressure [-], adsorbent temperature [K], and adsorption [kg/kg] were expressed in the model. The relative and absolute error rates of adsorbent temperature were 0.009% and 0.024%, respectively, in the range of applications. To calculate the temperature of saturated water from relative pressure and adsorbent temperature, a linear regression model of saturated water and vapor was also developed. The model of the adsorption refrigeration chiller consists of four components: a condenser, an evaporator, an adsorber, and a desorber. Fig. 4 shows the heat balance diagram of the adsorption refrigeration chiller. Heat flow was modeled using cooling water, chilled water, and hot water. Heat flow inside the chiller occurred by vaporization, condensation, desorption, and adsorption of refrigerant water. The heat exchange at each component was expressed with the NTU (number of transfer units) effective method as expressed in Eqs. (15), (22), (25), and (27). The changes in adsorption at the adsorber and desorber were expressed using differential equations as shown in Eq. (18) and Eq. (24). Fig. 5 shows the time series of the adsorption refrigeration cycle assumed in this model. Eq. (18) and Eq. (24) can be solved to Eq. (36) and Eq. (40) by using the linearity assumption of adsorption isotherms. Fig. 6 shows the calculation flow diagram of the model. Three iterative computations were required. The co-efficient of performance (COP) of the chiller was calculated iteratively in the outer loop. The hot water outlet temperature and desorb temperature were calculated iteratively in the inner loop. When the value of adsorption exceeded the linear range in the iterative calculations, the value was corrected as shown in Fig. 7. A sensitivity analysis was performed to evaluate the developed model. Fig. 8 shows the results of sensitivity analysis. The cooling capacity and COP were calculated by varying the temperature and flow rate of cooling water, hot water, and chilled water. The plot in Fig. 8 represents the actual measured data from a previous study. The lines in Fig. 8 are the calculated results for the developed model. The solid lines denote the results that have the same boundary conditions as the measured data. The results of the sensitivity analysis on flow rates were consistent with the measured data. Sensitivity analysis on the hot water inlet temperature shows slightly different trends compared to measured data. All source codes developed in this study are distributed under General Public License and can be downloaded from the website ( http://www.hvacsimulator.net ).
87
