Experimental study and steady-state modelling of a Commercial diffusion-absorption refrigerator

Transcription

1 Revue des Energies Renouvelables, Spécial ICT3-MENA, Bou Ismail, (2015) 7 14 Experimental study and steady-state modelling of a Commercial diffusion-absorption refrigerator R. Mansouri 1,2*, S. Mazouz 1, M. Bourouis 2 and A. Bellagi 1 1 U.R Thermique & Thermodynamique des Procédés Industriels Ecole Nationale d Ingénieurs de Monastir, ENIM Avenue Ibn El Jazzar, 5019 Monastir Université de Monastir, Tunisie 2 Universitat Rovira i Virgili, Department of Mechanical Engineering Av. Països Catalan N 26, Tarragona, Spain Abstract Interest in diffussion-absorption technique for production of cold is noticed nowadays. The mono-pressure refrigeration system are heat powered and use a threecomponent working fluids mixture: ammonia-water-hydrogen. As there are no moving parts, these systems are both quiet and reliable and are therefore often used in hotel rooms and offices. We report in this paper on the results of an experimental study in stationary mode carried out on a low capacity commercial unit. The test are performed under different ambient air conditions and variable heat supplies. All the essential features of the machine (temperatures at various points, heat duty of different heat exchangers) are determined. Further, the experimental results are used as input data for a theoretical simulation model using the flow-sheeting software Aspen-Plus platform. Before running the simulation model a choice of the appropriate thermodynamic properties model has to be made. In our case, the Peng-Robinson equation of state with the modification of Boston-Mathias is adopted and used to regress LV water/ammonia equilibrium data. The final theoretical model is validated by comparing its predictions with the experimental data. Keywords: Absorption-diffusion, Refrigeration, Modeling, Ammonia-water, Aspen-Plus, Performance. 1. INTRODUCTION The absorption refrigeration was discovered by Naim in 1777 [1]. Carré [1] has built the the first absorption refrigeration system in Since these times only two working fluid systems are largely used: ammonia-water and water-lithium bromide. The ideal of eliminating the pump and making the pressure uniform in all components of the machine occurred first to Geppert in To allow the refrigerant to evaporate at low temperature in the evaporator, a third compound-an inert gas- is introduced. The first machine of this type was developed by the swedish engineers von Platen and Munters in 1928 [2]. This diffusion absorption refrigeration, DAR system uses ammonia as refrigerant, water as absorbent and hydrogen as inert gas. The working principle of the triple fluid Platen-Munters system is presented in figure 1. The generator contains a rich ammonia water solution (5). When sufficient heat is supplied, ammonia vapour is generated and moves up through the bubble pump carrying the liquid solution. At the top of the pump weak solution (6) and vapour (1) are separated. The vapour flow to the condenser, where it condenses, and further to the evaporator where the liquid refrigerant (2) meets hydrogen (4) coming back from the absorber. * rami.mansouri@estudiants.urv.cat, souhamazouz@yahoo.fr mahmoud.bourouis@urv.cat, a.bellagi@enim.rnu.tn 7

2 8 R. Mansouri et al. Fig. 1: Working principle of the Platen-Munters system The ammonia partial pressure decreases, the liquid ammonia begins to evaporate by taking heat from the refrigerated space. The gaseous mixture of refrigerant and hydrogen (3) flows down to the absorber inlet. In the absorber, the ammonia vapour is absorbed by the weak solution (6), returning from the generator, transforming it to rich solution (5) which flows back to the generator. Hydrogen and residual not absorbed ammonia (4) continue to the evaporator. In diffusion-absorption refrigerators, hydrogen is commonly used as inert gas. Its flammability however limits its utilization. Helium is proposed as ideal substitute for hydrogen (Kouremenos et al., [3]). High temperature is needed to activate the ammoniawater bubble pump. Alternative working fluid mixtures have been proposed. NH 3/NaSCN/H 2 was used by Bourseau et al., [4]. Zohar et al., [5] examined the possibility of using R22, R32, R124, R125 and R134a as refrigerants, DMAC as absorbent and helium as inert gas. Ben Ezzine et al., [6] tested the R134a/DMAC/He and C 4H 10/C 9H 2O/He systems (Ben Ezzine et al., [7]). Thermodynamic analysis and modeling of the diffusion-absorption refrigeration systems were presented in several papers: Al-Shemmeri et al., [8]; Zohar et al., [9]; Startace et al., [10]; Sayadi et al., [11]. Also, experimental investigations were carried out on these systems (Chen et al., [12]). Vicatos et al., [13] modified the generator of the DAR to improve the response of the unit. A 2.5 kw prototype was constructed and studied by Jakob et al., [14]. Mazouz et al., [15] studied and experimentally investigated a commercial DAR. The aim of this paper is to present an experimental study on a diffusion absorption refrigerator and develop a steady-state simulation model of the machine using the flowsheeting software Aspen-Plus. 2. EXPERIMENTAL INVESTIGATIONS AND MEASUREMENT RESULTS The refrigerator studied is a small capacity refrigerator designed for hotel rooms presented in figure 2. The electric power input to the refrigerator is regulated by a controller. 11 K-thermocouples are installed at the inlet and the exit of the different components of the machine. The locations of the thermocouples are indicated by their ID in Table 1. The thermocouples are connected to a dedicated PC for data storage via a data acquisition system. Further the temperature of the refrigerator cabinet is also

3 Experimental study and steady-state modelling of a commercial diffusion 9 continuously measured as well as the ambient temperature. The tests are repeated for 12 different electric power supplies: 27, 35, 39, 44, 46, 48, 51, 53, 56, 58, 61 and 67 W. Fig. 2: Refrigerator studied Table 1: The locations of the thermocouples All the essential features of the machine were experimentally determined: the global heat transfer coefficients ( UA) cab an ( UA) int are found respectively equal to W/K and 0.3 W/K. It is noted also, figure 3, that the minimum heat power to be supplied to the generator to ensure a stable functioning of th refrigerator after 2h: 30min is 39 W. The best performance of the refrigerator is obtained experimentally, figures 4 and 5, with a heat supply of 46 W by a generator temperature of 167 C. The corresponding COP is of and the cooling capacity, 7.3 W. Table 2 presents some important measured data for this heat power supply. Fig. 3: Evolution of the temperature of various components of the machine (39 W) Fig. 4: COP vs. generator power supply

4 10 R. Mansouri et al. Fig. 5: Cooling capacity vs. generator power supply Table 2: Measured data for Q gen 46W 3. THERMODYNAMIC PROPERTY MODEL We proceed now to the steady-state modelling of the commercial diffusionabsorption refrigerator using the flow-sheeting software Aspen-Plus platform. Aspen- Plus includes a large databank of thermodynamic property for estimating the mixture properties. The selection of a proper method for estimating the thermodynamic properties of the working fluid is one of the most important steps that can affect the simulation of the refrigeration cycles. Therefore, it is important to choose carefully an appropriate method to estimate the different properties of the working fluid. In this refrigerator ammonia is used as refrigerant, water as absorbent and hydrogen as inert gas to equalize the pressure. The predictions of the nine property models available in Aspen-Plus for this working fluid (Table 3) were evaluated and compared with regressed experimental vapour-liquid equilibrium data reported by Mejbri et al., [16] in the pressure range from 2 to 25 bars and the temperature range -19 to 220 ºC. In a second step, the Aspen-Plus Data Regression System facility was used to fit the interaction parameters of the property models to the VLE data of reference [16]. The results are presented in the form of T-x-y diagrams. Figures 6 and 7 illustrate this comparison for 25 bars for both steps. It is observed in the first step, simple comparison (figure 6), that the maximum deviation in the vapour phase temperature between the values predicted by the considered Aspen-Plus models and the data reported by Mejbri et al., [16] is about 10 ºC. It is concluded that none of these property models reproduces the ammonia/water vapour-liquid equilibrium data with sufficient accuracy. In a second step, the data regression system of Aspen-Plus was used to fit the parameters of the property models by using the VLE data reported in [16]. As can be noted (figure 7), all values of the ammonia mass fraction in the vapour phase predicted by the Aspen-Plus property models are now close to the data. Figure 7 makes also clear that the PR-BM model (Peng-Robinson eos with Boston-Mathias alpha function) is in better agreement with the considered vapour-liquid equilibrium data. The sum of squared errors for the nine Aspen-Plus property models tested and summarized in Table 4 confirms this conclusion. The simulations will be performed using the regressed PR-BM eos for the calculation of the properties of the binary ammonia/water mixture as well as for the properties of the ternary ammonia/water/hydrogen mixture.

5 Experimental study and steady-state modelling of a commercial diffusion 11 Fig. 6: T-x-y VLE diagram at P 25Bars for ammonia/water mixture Table 3: Thermodynamic property models used in the Aspen-Plus simulations Fig. 7: T-x-y VLE diagram at P 25Bars for ammonia/water mixture with regressed parameter Table 4: Sum of squared errors for the thermodynamic property models tested 4. MACHINE MODELLING The Aspen-Plus model for the diffusion-absorption machine is presented in figure 8. The simulations are started using the standard sequential approach. After convergence and synchronization, the Equation-Oriented method is used. The generator is modeled using the Radfrac block, the bubble pump by combining a Flash and a Heater block connected with a heat stream, the condenser and the evaporator using a single Heater block. An Absorber block is used for the absorber and a combination of a Flash and Mixer for the tank. An Lnghx three-flow exchanger simulates the gas heat exchanger and twoflow heat exchanger Heatx for the solution heat exchanger. Material streams connecting the various components of this machine are identified with their IDs given in Table 1. Two "Break points" are considered in the model for the cycle input conditions: The "first break" is inserted at state point 1 (rich solution exiting the tank), and the "second break" at state point 2 (inert gas flowing to the gas heat exchanger). The identity of stream1 and 1A on one side and stream2 and 2A on the other side constitutes the convergence criterion for the simulations. 5. ASSUMPTIONS AND SIMULATION RESULTS The main assumptions and operating conditions used as input for the simulations are based on our experimental study and summarized in Table 5. The model results and experimental data are compared in Table 6 and figure 8. Table 7 presents the convergence test at "break points" of the cycle, and Table 8 shows the comparison of model predicted and experimental performances. In view of all these

6 12 R. Mansouri et al. results it can be concluded that the predictions of the proposed Aspen-Plus model for the diffusion-absorption machine are in excellent agreement with the experimental data. Fig. 8: The Diffusion-Absorption machine model in Aspen-Plus Table 5: Assumptions and input for the simulations Fig. 9: Comparison between calculated and experimental temperatures Table 6: Calculated and experimental temperatures

7 Experimental study and steady-state modelling of a commercial diffusion 13 Table 7: Convergence test at the cycle break points Table 8: Comparisons in terms of performances between calculated and experimental data 6. CONCLUSION A commercial diffusion-absorption refrigerator is tested under different heat inputs to the generator. The temperature at the inlet and outlet of every component is continuously recorded. All the essential features of the refrigerator are determined experimentally, especially the overall heat transfer coefficients ( UA) cab 0.554W/ K and ( UA) int 0.3W/ K. The best performance of the refrigerator is obtained experimentally with an electric power supply of 46 W by a generator temperature of 167 C. The COP of the machine is found equal to An Aspen-Plus model for the diffusion-absorption machine is then developed. An appropriate property model for the working fluid mixture used has been selected: Peng-Robinson equation of state with the modification of Boston- Mathias. The main assumptions and operating conditions based on our experimental study have been used as input for the simulations. The calculated results show a good agreement with the measured data indicating that the model represents fairly well the machine. NOMENCLATURE ( UA), Global heat transfer coefficient COP, Coefficient of performance cab, Refrigerator cabinet evap, Evaporator gen, Generator abs, Absorber Q, Heat duty, W REFERENCES [1] K. Rajasuthan and M. PalPandi, Fabrication and Performance Evaluation of Electrolux Refrigeration System Coupled with IC Engine, International Journal of Engineering Sciences & Research Technology, Vol. 3, N 12, pp , [2] B.C. Von Platen and C.G. Munters, Refrigerator, US Patent N 1,609,34, [3] D.A. Kouremenos and A. Stegou-Sagia, Use of Helium Instead of Hydrogen in Inert Gas Absorption Refrigeration, International Journal of Refrigeration, Vol. 11, N 5, pp , 1988.

8 14 R. Mansouri et al. [4] P. Bourseau and R. Bugarel, Absorption-Diffusion Machines: Comparison of the Performances of NH3/H2O and NH3/NaSCN, International Journal of Refrigeration, Vol. 9, N 4, pp , [5] A. Zohar, M. Jelinek, A. Levy and I. Borde, Performance d un Cycle Frigorifique à Absorption-Diffusion utilisant des Fluides Actifs Organiques, International Journal of Refrigeration, Vol. 32, N 6, pp , [6] N. Ben Ezzine, R. Garma and A. Bellagi, A Numerical Investigation of a Diffusion- Absorption Refrigeration Cycle Based on R124-Dmac Mixture for Solar Cooling, Energy, Vol. 35, N 5, pp , [7] N. Ben Ezzine, R. Garma, M. Bourouis and A. Bellagi, Experimental Studies on Bubble Pump Operated Diffusion Absorption Machine Based on Light Hydrocarbons for Solar Cooling, Renewable Energy, Vol. 35, N 2, pp , [8] T. Al-Shemmeri and Y. Wang, Theoretical, investigation and parameter study of a diffusionabsorption refrigeration system, Proceeding of the 21 st IIR International Congress of Refrigeration, August, International Institute of Refrigeration (IIR), Washington D.C., USA, [9] A. Zohar, M. Jelinek, A. Levy and I. Borde, Numerical Investigation of a Diffusion Absorption Refrigeration Cycle, International Journal of Refrigeration, Vol. 28, N 4, pp , [10] G. Starace and L. De Pascalis, An Advanced Analytical Model of the Diffusion Absorption Refrigerator Cycle, International Journal of Refrigeration, Vol. 35, N 3, pp , [11] Z. Sayadi, N. Ben Thameur, M. Bourouis and A. Bellagi, Performance Optimization of Solar Driven Small-Cooled Absorption Diffusion Chiller Working with Light Hydrocarbons, Energy Conversion and Management, Vol. 74, pp , [12] J. Chen, K.J. Kim and K.E. Herold, Performance Enhancement of a Diffusion Absorption Refrigerator, International Journal of Refrigeration, Vol. 19, N 3, pp , [13] G. Vicatos, Experimental Investigation on a Three-Fluid Absorption Refrigeration Machine, Journal of Process Mechanical Engineering, Vol. 214, N 3, pp , [14] U. Jakob, U. Eicker, D. Schneider, A.H. Taki and M.J. Cook, Simulation and Experimental Investigation into Diffusion Absorption Cooling Machines for Air-Conditioning Applications, Applied Thermal Engineering, Vol. 28, N 10, , [15] S. Mazouz, R. Mansouri and A. Bellagi, Experimental and Thermodynamic Investigation of an Ammonia/Water Diffusion Absorption Machine, International Journal of Refrigeration, Vol. 45, pp , [16] Kh. Mejbri and A. Bellagi, Modelling of Thermodynamic Properties of the Water-Ammonia by Three Different Approaches, International Journal of Refrigeration, Vol. 29, N 2, pp , 2006.