- Email: [email protected]
Experimental evaluation of a small-capacity, waste-heat driven ammonia-water absorption chiller
international journal of refrigeration 79 (2017) 89–100
Available online at www.sciencedirect.com
ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g
Experimental evaluation of a small-capacity, waste-heat driven ammonia-water absorption chiller Anurag Goyal, Marcel A. Staedter, Dhruv C. Hoysall, Mikko J. Ponkala, Srinivas Garimella * Sustainable Thermal Systems Laboratory, George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA
A R T I C L E
I N F O
A B S T R A C T
Article history:
This paper presents the results from an experimental evaluation of a small-scale, waste-
Received 4 February 2017
heat driven ammonia-water absorption chiller. The cycle is thermally driven, using the waste
Received in revised form 9 April
heat from diesel generator exhaust to desorb the refrigerant solution. The absorber and con-
2017
denser are directly-coupled to the ambient air. The remaining heat exchangers are packaged
Accepted 10 April 2017
in a compact microchannel monolithic structure for enhancing heat transfer. The system
Available online 13 April 2017
is designed to deliver 2.71-kW of cooling at extreme ambient temperature of 51.7 °C at a coefficient of performance of 0.55. Experiments on a heat pump breadboard system are con-
Keywords:
ducted at ambient conditions of 29.7–44.2 °C, with delivered cooling duties of 2.54–1.91 kW.
Vapor absorption chiller
System and individual component performance is analyzed and compared with cycle model
Ammonia-water
predictions, and deviations are explained. The absorber and desorber were identified to be
Waste heat recovery
the limiting components in the system. Effects of variation in ambient temperature are studied to characterize the performance at off-design conditions. © 2017 Elsevier Ltd and IIR. All rights reserved.
Évaluation expérimentale d’un refroidisseur à absorption à ammoniac-eau de faible puissance alimenté par de la chaleur perdue Mots clés : Refroidisseur à diffusion-absorption ; Ammoniac-eau ; Récupération de chaleur perdue
1.
Introduction
Vapor absorption based heating, ventilation and air-conditioning (HVAC) systems can utilize low-grade heat sources such as waste
heat to provide heating and cooling.These systems also provide an alternative to conventional vapor compression systems in terms of reducing the peak demand for electricity (Ziegler, 1999). Specifically, waste-heat or solar energy driven systems can provide more economical cooling/heating solutions in
* Corresponding author. Love Building, Room 340, Ferst Drive, Atlanta, GA 30332. Fax: +1 404 894 8496. E-mail address: [email protected] (S. Garimella). http://dx.doi.org/10.1016/j.ijrefrig.2017.04.006 0140-7007/© 2017 Elsevier Ltd and IIR. All rights reserved.
90
international journal of refrigeration 79 (2017) 89–100
Nomenclature specific heat capacity [J kg−1 °C−1] specific enthalpy [J kg−1] heat exchanger inside diameter [m] mass flow rate [kg s−1] OD outside diameter [m] P pressure [kPa] PWM pulse-width modulation q vapor mass quality [−] RPC refrigerant pre-cooler SHX solution heat exchanger T temperature [°C] x ammonia concentration [–]
Cp h HX ID m
Subscripts abs absorber cf coupling fluid conc concentrated solution cond condenser des desorber dil dilute solution evap evaporator in inlet liq liquid out outlet rec rectifier ref refrigerant vap vapor
conjunction with power generation systems for combined heat and power (CHP) applications as shown by Ziegler and Riesch (1993) and Filipe Mendes et al. (1998). These systems also use environmentally benign working fluids, such as ammoniawater, with reduced global warming potential (GWP) (Lorentzen, 1995). Vapor absorption systems typically require more heat and mass exchange components compared to conventional vapor compression systems. The mechanical compressor is replaced by a set of heat and mass exchangers that enable heat source utilization (Herold et al., 1996; Srikhirin et al., 2001).This leads to larger system size and high capital cost, which has historically limited the use of such systems for residential and light commercial applications. To improve the efficiency of the single-effect absorption cycle, advanced cycles with internal heat recovery and multi-staged heat transfer mechanisms have been proposed by researchers in the past (Kang et al., 2000; Srikhirin et al., 2001; Ziegler, 1999; Ziegler and Riesch, 1993). However, in such systems, the increase in system complexity typically does not yield commensurate benefits in efficiency. Some past research efforts experimentally evaluated absorption systems and advanced cycles for improved efficiency. Didion and Radermacher (1984) presented results on partload performance and ambient temperature variation for a 10.5 kW cooling capacity gas-fired ammonia-water absorption chiller. It was observed that the COP and cooling duty decreased as the ambient temperature increased. Treffinger (1996) presented a direct gas-fired ammonia-water system for water heating
applications. The COP of the system varied between 1.4 and 1.7, with the heating COP defined as the ratio of the total heat transfer rate from condenser and absorber to the heat input in the desorber. Erickson et al. (1996) demonstrated an ammoniawater generator-absorber heat exchange (GAX) cycle with cooling duty of 14.6 kW at a COP of 1.05. Mendes and Collares-Pereira (1999) presented results from the design and testing of a smallcapacity solar absorption heat pump.The system demonstrated a cooling capacity of 5 kW (COP 0.56) and a heating capacity of 9 kW (COP 1.42). Sözen et al. (2002) presented results from testing of an ammonia-water absorption heat pump using solar energy. De Francisco et al. (2002) also presented experimental analysis of a solar powered ammonia-water system for small-scale refrigeration applications. The system was designed to provide a cooling duty of 2 kW, but the experiments showed limited performance and low COP due to using only natural convection for condenser and absorber heat rejection. This also led to very large component size. These studies report a need for more compact heat and mass exchangers to reduce the overall system footprint. Additionally, the vast majority of reported work did not consider extreme ambient conditions, which introduce further development challenges. Reduction of the size of vapor absorption system components has been the focus of active research for some time. Garimella (1999) and Garimella (2004) presented a design for a microscale heat and mass exchange component for binary fluid mixtures. The falling-film type component utilized 1.58 mm OD tubes, and can be used as any component in the absorption cycle with minor modifications. Meacham and Garimella (2004) demonstrated the performance of the aforementioned device as an absorber and addressed issues with fluid distribution and techniques to improve the component performance. Later, Determan and Garimella (2011) and Garimella et al. (2011) demonstrated the use of the same device for desorption with minor modifications in fluid routing. However, all these designs were still less suitable for a compact heat pump packaged design with small fluid inventory. Hu and Chao (2008) demonstrated the performance of a very small-scale water-lithium bromide absorption system with 40 W cooling capacity. The components were fabricated using silicon wafers with photo-chemically etched micro-channels of hydraulic diameter of 131 µm. Pence (2010) investigated micro-scale fractal flow networks for design of heat and mass exchangers to increase the vapor flow area and reduce the pressure drop in two-phase components. Determan and Garimella (2012) presented a micro-scale ammonia-water absorption heat pump with all the components in a single monolithic structure. The system delivered 300 W of cooling at a COP of 0.375. It demonstrated the potential of component integration and efficient packaging to achieve small system sizes, and the potential for scaling the systems to higher capacities. Keinath (2015) presented a compact absorption heat pump water heater for residential applications. The prototype system utilized a monolithic fabrication approach to integrate multiple heat and mass exchange components. It delivered 2.79 kW of heating at a COP of 1.74. Garimella et al. (2016) demonstrated the performance of a microchannel absorption chiller packaged unit. The natural gas fired system delivered a cooling duty of 3.3 kW at a COP of 0.47. Their work demonstrated the scalability of microscale designs.
international journal of refrigeration 79 (2017) 89–100
The present study extends the applicability of microscale heat and mass exchange devices for the development of compact absorption heat pumps. The component design focuses on direct integration of the heat source, thereby avoiding the use of a secondary heat transfer loop. The heat source investigated in this work is exhaust from a diesel generator. However, the design of the heat coupling ensures flexibility of using in a direct-fired application as well. The absorber and condenser also reject heat directly to the ambient to eliminate hydronic loops, which require additional components and larger system size. In addition, the proposed system is designed for diverse applications such as mobile refrigeration units, residential heat pump systems, and military applications with severe ambient conditions such as those seen at forward operating bases. Results from cycle modeling and experiments are presented, and the deviations from predicted performance are discussed. Potential changes in component design to improve system performance are also proposed.
2.
System description and modeling
A single-effect ammonia-water absorption chiller is investigated. Ammonia-water is selected as the working fluid pair for the current system because ammonia has high latent heat of vaporization, which is necessary for efficient performance. Also, the high vapor-phase density allows for smaller component sizes. For the intended high-ambient temperature conditions, ammonia-water system does not present a crystallization risk, which would be the case with water-lithium bromide if
91
direct air-cooling is implemented. Moreover, the freezing point of ammonia is 196.15 K, which allows for low-temperature refrigeration as opposed to water-lithium bromide systems (Srikhirin et al., 2001). The desorber, rectifier, solution heat exchanger, and condenser operate at the high-side pressure, while the absorber and evaporator operate at the low-side pressure. A schematic of the different system components is shown in Fig. 1. The desorber uses heat supplied from a hightemperature source, the condenser and the absorber reject heat to the ambient, and the evaporator provides cooling by receiving heat from a low-temperature source. Internal heat recovery through recuperative heat exchangers on the solution and refrigerant circuits helps in improving the COP and decreasing the desorber and absorber heat duties. As shown in Fig. 1, the concentrated solution is pumped from low-pressure (1) to high-pressure (2). The concentrated solution cools the rectifier component condensing residual water in the generated refrigerant vapor leaving the desorber. This also recovers energy internal to the system by preheating the concentrated solution. The concentrated solution at (3) enters the solution heat exchanger and exchanges heat with the dilute solution exiting the desorber. The recuperative heat exchange further heats up the concentrated solution before entering the desorber at (4). The desorber receives heat from high temperature source and separates ammoniarich vapor from the liquid mixture. The ammonia-rich vapor exits the desorber (8), and is further purified in the rectifier to remove trace quantities of water. The purified refrigerant vapor leaves the rectifier and enters the condenser (9), while the condensed solution from the vapor is remixed into the desorber (10). The refrigerant vapor is condensed and the
Fig. 1 – Schematic of single-effect absorption cycle.
92
international journal of refrigeration 79 (2017) 89–100
Table 1 – Steady-state cycle model design inputs and calculated heat transfer rates. Component
UA (W K−1)
Coupling Fluid −1
Absorber Condenser Desorber Evaporator Rectifier Pre-cooler Solution heat exchanger
379.6 187.5 55 846 43.9 25 74.1
Baseline Duty (W)
Flow rate (kg s )
Inlet Temperature (°C)
0.383 0.383 0.0255 0.0957
51.67 51.67 398.9 19.8
heat of condensation is rejected to the ambient. The liquid refrigerant at high-pressure (11) exchanges heat with colder low-pressure refrigerant vapor in the refrigerant pre-cooler. The liquid refrigerant (12) is throttled to the low-pressure side through an expansion valve. The refrigerant enters the evaporator (13) and exchanges heat with low-temperature source, exiting in a high quality two-phase state at (14). The refrigerant vapor gains heat in the refrigerant pre-cooler, leaving at (15), and mixes with the dilute solution at the inlet of the absorber. The dilute solution leaving the desorber (5) exchanges heat with concentrated solution in the solution heat exchanger and cools to a high-pressure liquid state at (6). The dilute solution is throttled to the low-side pressure (7) across an expansion valve. The absorber rejects the heat of absorption to the ambient, and regenerates the concentrated solution, thus completing the cycle. A steady-state cycle model is developed to predict the baseline heat exchanger sizes for different components and conduct parametric analyses to study the effects of variations in operating conditions such as ambient temperature, solution mass flow rate, and heat source temperature and mass flow rate. The model is developed on the Engineering Equation Solver (EES) platform (Klein, 2014). Three independent thermodynamic properties are required to completely define the state of the binary working fluid. Initial overall heat conductance values (UA) of the heat exchangers are determined by making appropriate assumptions about their closest approach temperatures or effectiveness. The baseline model is provided with ambient conditions, heat source inlet temperature, coolant inlet temperature, and flow rates of coupling fluids and the concentrated solution. To achieve optimal system performance, the mass fraction of ammonia in the refrigerant vapor exiting the rectifier was specified at 0.998. The high purity of the refrigerant vapor is achievable with compact desorber and rectifier designs as shown by Keinath et al. (2015). Mass, species, and energy conservation equations are setup for each component and the simultaneous system of equations is solved iteratively. The baseline UA values are then optimized to maximize the system COP for specified operating conditions. The air inlet temperature to the condenser and absorber was specified at 51.67 °C, with a relative humidity of 18%. An air mass flow rate of 0.383 kg s−1 for both heat exchangers was specified. Hot air was supplied to the desorber at 398.9 °C and a mass flow rate of 0.025 kg s−1. The concentrated solution flow rate was specified at 0.008 kg s−1. Temperature in-equilibrium values for rectifier vapor inlet and reflux outlet, and desorber
5147 2529 4937 2713 1496 345 2196
solution inlet and vapor outlet were also specified. The trace amount of water in the refrigerant leads to a rise in temperature of the refrigerant along the length of the evaporator. To optimize the heat transfer performance of the evaporator, temperature glide in the evaporator is specified at 2.5 K. Table 1 presents the inputs to the model and the heat transfer rates in different components. The high-side pressure at design conditions was calculated to be 2928 kPa, while the low-side pressure was 585.3 kPa. The refrigerant flow rate and dilute solution flow rate were calculated to be 0.0025 kg s −1 and 0.0055 kg s−1, respectively. The ammonia mass fractions of the dilute and concentrated solution were calculated to be 0.163, and 0.422, respectively. The cooling COP of the system, as defined in Eq. (1), was calculated to be 0.55.
Q evap COPcooling = Q des
(1)
The overall COP of the system, accounting for electrical input for pumps, fans, electronic valves, and control system, is calculated as defined in Eq. (2). Using an estimated total electrical load of 375 W for pump, fan and control system electricity consumption, and a primary energy load factor of 3.18 to obtain source-based fuel energy consumption, the overall COP of the system is calculated to be 0.44.
Q evap COPoverall = Q des + E total
(2)
Fig. 2 shows the variation in COP and cooling duty as the ambient temperature changes. A lower ambient temperature allows improved heat rejection from the absorber and the condenser, and thereby leads to an increase in the overall system cooling capacity. Improved absorber heat rejection also leads to reduced low-side pressure, and thus lower refrigerant inlet temperature at the evaporator and greater cooling capacity.
3.
Experimental setup
Fig. 3 shows a schematic of the breadboard test facility, which allows the evaluation of the entire system, as well as individual component performance. The key features of each component are listed in Table 2.
international journal of refrigeration 79 (2017) 89–100
Fig. 2 – Variation in ambient temperature and predicted cooling capacity and COP.
3.1.
Heat exchangers
Detailed, segmented heat and mass transfer models were developed for individual heat exchangers using operating conditions computed by the thermodynamic model of the system. The condenser and absorber are custom-designed multi-pass finned-tube heat exchangers. The heat of
93
absorption and condensation is rejected to the ambient air, which is circulated by dedicated fans. A novel desorber based on a diabatic distillation column concept that integrates heat transfer stages with purification stages was developed. This provides a compact design with very small liquid-vapor approach temperature differences, and consequently, high-purity vapor leaving the desorber. A small, integrated analyzer section in the column further improves ammonia concentration in the vapor stream and decreases the rectification requirement. Due to the direct coupling to engine exhaust, the design allows for a maximum pressure drop of 2.5 kPa on the gas side. The rectifier is a compact shell-andtube heat exchanger with partial condensation of vapor on the shell-side and concentrated solution flowing on the tube-side. The refrigerant pre-cooler, evaporator, and solution heat exchanger are all microchannel plate type heat exchangers, integrated in one monolithic block, as shown in Fig. 4. The bonded monolithic heat exchanger consists of stainless steel plates with microscale features. The two plate designs are stacked in an alternating pattern for working fluid and coupling fluid sides of the heat exchanger, and enclosed by a cover plate and an endplate. The fabrication process involved photochemical etching of microscale flow passages on the shims, followed by bonding the assembly using diffusion bonding or vacuum brazing. The microchannel fluid passages have a hydraulic diameter of 440 µm. The experimental test setup is shown in Fig. 5.
Fig. 3 – Breadboard test facility schematic.
94
international journal of refrigeration 79 (2017) 89–100
Table 2 – Component geometry details. Component
Type
Absorber
Air-cooled, cross-flow, finned-tube
Condenser
Air-cooled, cross-flow, finned-tube
Desorber
Evaporator (Coolant coupled)
Direct-coupled, shell-and-tube with finned trays on shell-side, counter-flow liquid and vapor Microchannel array, counter-flow, liquid coupled
Rectifier (Solutioncooled) Pre-cooler
Solution-cooled shell-and-tube, counterflow Microchannel array, counter-flow
Solution heat exchanger
Microchannel array, counter-flow
3.2. Heat source coupling loops and auxiliary components The evaporator coolant loop circulates a 25% mixture of propylene-glycol and water, and contains a pulse-width modulation (PWM) controlled electrical heater, which maintains the coolant inlet temperature at the set-point.
Fig. 4 – Monolith shim (top); bonded monolith (bottom).
Key Dimensions Tube ID: 0.0141 m, Fin density: 551 and 393 m−1 H × W: 0.419 × 0.61 m, 12 tube passes Tube ID: 0.0141 m, Fin density: 393 m−1 H × W: 0.419 × 0.31 m, 9 tube passes Two-columns, Shell ID: 0.102 m Exhaust Tube bundle: ID 0.011 m, 22 tubes Height of column: 0.416 m Channel Hydraulic Diameter: 440 µm 66 plates: 22 refrigerant, 44 coolant flow 62 channels/shim, Refrigerant channels: 0.0995 m, Coolant channels: 0.0795 m, 0.154 × 0.126 m (L × W) Exergy LLC 35 Series, Shell-side: OD 0.038 m Tube-side: OD 0.0032 m, 55 tubes, 0.254 m Channel Hydraulic Diameter: 440 µm, 66 plates: 33 pairs, 42 channels/shim, Straight channels: 0.109 m, Diagonal Channels: 0.0746 m, 0.227 × 0.085 m (L × W) Channel Hydraulic Diameter: 440 µm, 66 plates: 33 pairs, 42 channels/shim, Straight channels: 0.109 m, Diagonal Channels: 0.0746 m, 0.227 × 0.085 m (L × W)
An electrically heated air loop serves as the waste heat source and simulates a waste heat stream such as exhaust gas from a diesel generator. A PWM controller for the heaters regulates the temperature set-point of the air supplied to the desorber. The air-heater assembly simulates a waste heat source, such as exhaust gas from an electric generator accurately, and provides precise control over the flow rate and temperature. Storage tanks are installed at the outlet of absorber, condenser, rectifier, and dilute solution outlet of the desorber. These storage tanks ensure appropriate fluid inventory in different parts of the system, allowing for smoother steady-state operation, and accurate measurement of flow rates by ensuring that the different fluids streams are in single phase. At the outlet of the rectifier, refrigerant vapor is separated from the
Fig. 5 – Experimental setup.
95
international journal of refrigeration 79 (2017) 89–100
Table 3 – Equipment and measurement probe specifications. Instrument Thermocouple Pressure Transducer Mass Flow Sensor Coolant Mass Flow Sensor Level Measurement Probe Electronic Expansion Valves Data Acquisition and Control
Manufacturer/Model
−250–300 °C 0–750 °C 0–1034.2 kPa 0–3447.3 kPa 0–0.01 kg s−1 0–0.1 kg s−1 0–0.6056 kg s−1 0–0.152 m
Omega® T-type Omega® J-type Omega® 4–20 mA Output® Absolute Rheonik® Coriolis Sensor MicroMotion® Coriolis Sensor Endress+Hauser® Liquicap M FMI51 Carel® E2V Series National Instruments® c-RIO 9024
condensate in a storage tank before it enters the condenser. The reflux combines with the solution inlet stream to the desorber. The dilute solution storage tank ensures that the solution heat exchanger receives liquid-phase solution. The dilute solution and refrigerant vapor are mixed in a tube-in-tube type injection port at the inlet of the absorber. The regenerated solution is stored in a solution storage tank downstream of the absorber, which provides liquid-phase solution to the pump. A positive-displacement piston pump circulates the ammoniawater mixture.
3.3.
Range
Instrumentation and controls
As shown in Fig. 3, the temperature at the inlet and outlet of each stream of individual heat exchangers is measured using T-type thermocouples. Pressures are measured using electronic pressure transducers at the outlet of the desorber, the outlet of the condenser, inlet to the evaporator, and outlet of the absorber. The two pressure measurements on the lowpressure side also help in characterizing the total pressure drop on the low-pressure side components. Four high accuracy (±0.1%) Coriolis mass flow sensors are used to measure the mass flow rates of the concentrated solution, dilute solution, refrigerant solution, and evaporator coolant. It should be noted that this leads to very small uncertainties in calculated variables due to mass flow rate measurement, and the corresponding error bars are not visible on the plots. In addition, an analog air-flow meter is used to measure the volumetric flow rate of hot air supplied to the desorber. A capacitance based level measurement probe is used on the refrigerant tank to measure the liquid level. This is used to monitor and regulate appropriate liquid storage level in the tank during transients and change in operating conditions. The speed of the pump and the fans for the condenser and absorber are controlled using variable speed drives. Both the expansion valves in the system are variable-opening electronic expansion valves positioned by a stepper motor. A proportional control based feedback loop, as described by Goyal et al. (2015), is used to regulate the valve position. Data acquisition and control is done using National Instruments® LabVIEWTM. Steady-state data are recorded for a 20-minute duration at a scan rate of 100 ms. Table 3 provides specifications of all the measurement probes.
Accuracy ±0.5 °C ±1.1 °C ±0.03% Range ±0.1% ±0.1% ±0.5%
480 motor steps
–
–
–
4.
Results and discussion
4.1.
Data analysis
Experiments were first conducted to analyze the system performance at nominal ambient conditions of 30 °C, and followed by testing at increasing ambient temperatures approaching design conditions. Data analysis is conducted as described in Lee et al. (2012), using time-averaged values of different measured quantities to define inlet and outlet states of each component. Heat transfer rate in each component is then calculated and system performance analyzed. Key assumptions in the analysis are: • The refrigerant vapor leaves the desorber/analyzer section in the saturated state. • The vapor leaves the separator tank after the rectifier and enters the condenser in the saturated state. • The reflux from the rectifier leaves the bottom of the separator in the saturated state. These phase-quality assumptions are required to determine the concentration of the refrigerant, reflux, and concentrated solution streams, and hence set the third required state property along with the measured pressure and temperature. The dilute solution concentration is determined by performing mass and species conservation on the absorber component. Mass and energy conservation are also used to validate the steady-state nature of the recorded data set. Engineering Equation Solver (EES) (Klein, 2014) is used for data analysis. Using the measured heat transfer rate, which is obtained from the mass flow rates and measured temperature and pressure, the overall conductance can be expressed as shown in Eq. (3),
UA =
Q LMTD
(3)
where, Q is the measured heat transfer rate, and LMTD is the logarithmic mean temperature difference. The overall conductance calculated here is indicative of the spatially averaged conductance in the component; detailed in-situ measurements
96
international journal of refrigeration 79 (2017) 89–100
Fig. 7 – Condenser heat transfer rate and overall conductance. Fig. 6 – Absorber heat transfer rate and ambient conditions.
to understand local variations were not performed. These calculations provide an overall estimate of the component performance when integrated into the complete chiller system.
4.2.
Absorber
The heat and mass transfer processes inside an absorber take place across adverse concentration gradients and small driving temperature difference, making the role of the absorber very crucial for the overall system performance (Beutler et al., 1996). Fig. 6 shows the heat transfer rate for the absorber as a function of the concentrated solution flow rate. The heat transfer rate is calculated on the working fluid-side using measured flow rates and states of two-solution streams mixing at the inlet of the absorber. The dilute solution expansion valve is assumed to be isenthalpic for calculation purposes. Mass, species and energy conservation are used to establish the concentration and inlet enthalpy in the absorber. Overall mass balance for the dilute and the refrigerant solution mass flow rate differs from the measured concentrated solution flow rate by a maximum of 6.5%. The outlet temperature and pressure are measured, and using the absorber concentration, the outlet enthalpy is determined. The overall heat transfer is calculated as shown in Eq. 4,
conc (habs,in − habs,out ) Q abs = m
(4)
As shown in Fig. 6, the heat duty for the absorber is below the design values. The temperature values on the plot indicate the ambient condition of the test. The measured heat transfer rate varied between 4.47 and 5.16 kW. The design values for heat transfer rate and UA are 5.14 kW and 0.38 kW K−1, respectively. Overall, the heat transfer rate remains below design for most test cases indicating limitation in the performance of the component. The heat transfer rate for the absorber is below design at both ends of the ambient temperature. This can be attributed to limitation in the overall conductance and
variation in the driving temperature difference between the working fluid and ambient air. Near design ambient conditions, the limitation in heat transfer requires that the refrigerant flow rate be reduced to decrease the low-side pressure and absorber concentration, leading to an overall decrease in system performance. Modifications to the absorber, including additional tube passes with reduced tube diameter of 9.5 mm OD are expected to increase tube-side heat transfer coefficients as well as surface area. To increase liquid-vapor interfacial area, twistedtape type steel inserts can also be used. The overall pressure drop on the low-pressure side of the system, from evaporator inlet to absorber outlet, stayed below 27 kPa for all the test cases, which is significantly less than the design pressure drop of 15 kPa in any individual component. Therefore, the changes to incorporate additional interfacial area and reducing the tube diameter will not exceed the design pressure drop value.
4.3.
Condenser
Fig. 7 shows the measured overall conductance and heat transfer rate for the condenser. The heat transfer rate is calculated on the refrigerant side using measured refrigerant mass flow rate at steady state, and temperature, pressure and concentration to calculate the refrigerant enthalpy, as shown in Eqs. 5 and 6,
xref = f ( Pcond, Tcond,inlet , qvap )
(5)
ref (href ,cond,in − href ,cond,out ) Q cond = m
(6)
where, xref is refrigerant vapor concentration, Pcond is pressure measured at condenser outlet, Tcond,inlet is temperature measured at the inlet of the condenser, qvap is the saturated vapor quality (=1) at the inlet. The heat transfer rate is calculated using ref and the refrigerant inlet and the measured mass flow rate, m outlet enthalpies. The heat transfer rate varied between 2.15 and 2.901 kW as compared to the design value of 2.53 kW. The overall
international journal of refrigeration 79 (2017) 89–100
Fig. 8 – Desorber heat transfer rate and overall conductance.
Fig. 9 – Evaporator heat transfer rate and overall conductance.
conductance of the condenser varied between 0.145 and 0.197 kW K−1, as shown in Fig. 7. However, as a result of the limiting performance of the absorber leading to a decreased absorber concentration, the refrigerant flow rate is lower than the design value of 0.00247 kg s−1 except for the lowest ambient temperature case. The high uncertainty associated with certain data points is due to a near-saturation outlet state of the condenser, and a consequently large change in outlet enthalpy within the temperature measurement uncertainty. The UA of the condenser increases with an increase in refrigerant mass flow rate as the heat transfer coefficient increases. At neardesign refrigerant mass flow rate, the UA is close to the design value, while the heat transfer rate varies with LMTD depending on the ambient conditions. Overall, the condenser seems to perform near-design, especially when the design flow rates are provided, and slight modifications in tube diameter or number of tube passes can be made to further augment the performance.
measured heat duty and UA varied between 4.26 and 4.97 kW, and 0.027 and 0.037 kW K−1, respectively, while corresponding design values are 4.94 kW and 0.055 kW K−1, respectively. The heat transfer rate and UA are lower than design owing to the dilute solution stream leaving as a two-phase mixture, and consequently a lower refrigerant generation rate. At the lower concentration of the solution, the solution mass flow rate should be increased above design to achieve the same refrigerant mass flow rate and suppress the two-phase mixture leaving though the dilute solution outlet. However, in compact geometries this can often lead to flooding in the component as the liquid solution can be entrained in the exiting vapor stream, leading to a poor quality and concentration in the rectifier. Further experiments will focus on validating the quality at the desorber exit, as a small fraction of vapor in the dilute solution significantly affects the performance of the desorber and the absorber.
4.5. 4.4.
97
Evaporator
Desorber
The heat transfer rate in the desorber is measured on the solution-side using steady-state energy conservation equation as shown in Eq. 7:
vap,des,out hvap,des,out + m liq,des,out hliq,des,out − m conc,inhconc,in Q des = m reflux,inhreflux,in −m
(7)
vap,des,out is the vapor phase mass flow rate leaving the where, m desorber, and the enthalpy of vapor, hvap,des,out , is determined using measured temperature, pressure assuming a saturated liq,des,out is the liquid dilute vapor (quality =1) state. Similarly, m solution exiting the desorber and the enthalpy hliq,des,out is calculated using measured temperature, pressure, and concentration calculated from the absorber states. The inlet streams into the desorber are the concentrated solution from solution heat exchanger and reflux from the rectifier, with their states determined from the absorber and rectifier, respectively. Fig. 8 shows a plot of the desorber heat duty and overall UA as a function of the concentrated solution flow rate. The
Owing to the small temperature difference on the refrigerant side and strong dependence of refrigerant enthalpy on temperature, the heat transfer rate is calculated on the coupling fluid side of the evaporator. The heat transfer rate is determined using Eq. 8:
cf ,evapCpcf ,evap ( Tcf ,evap,in − Tcf ,evap,out ) Q evap = m
(8)
Fig. 9 shows a plot of the evaporator UA and heat duty as a function of refrigerant flow rate. The measured heat transfer rates and UA varied between 2.54 and 1.91 kW, and 0.486 and 0.724 kW K−1, respectively, while the corresponding design heat duty and UA are 2.71 kW and 0.850 kW K−1, respectively. The high uncertainty in the reported values is due to the relatively smaller temperature difference (~4–6 °C) on the coupling fluid side and the constant associated uncertainty of temperature measurement (±0.5 °C). The temperature glide, i.e., the difference between the refrigerant outlet and inlet temperature, varied between 1.55 and 8.44 °C as compared to the design value of 2.5 °C. Both the heat transfer rate and UA increase with
98
international journal of refrigeration 79 (2017) 89–100
Fig. 10 – Refrigerant pre-cooler heat transfer rate and overall conductance.
Fig. 11 – Solution heat exchanger heat transfer rate and overall conductance.
an increase in refrigerant mass flow rate, which is expected due to an increase in the heat transfer coefficient. The evaporator performance was below design as the refrigerant mass flow rate remained below design for most test cases. The temperature glide was higher than the design value for the same reason. It is expected that at design refrigerant flow rate, the evaporator performance will be near-design.
be used to calculate the overall conductance. In this work, an average heat transfer rate of the two streams is used.
4.6.
Refrigerant pre-cooler
The refrigerant pre-cooler allows heat recuperation between high-pressure refrigerant liquid and low-pressure refrigerant vapor, and reduces the liquid temperature at the inlet of the refrigerant expansion valve. Using the measured temperature and pressure on both fluid streams, and known refrigerant concentration and mass flow rate from the condenser, the heat transfer rates for both fluid streams are calculated using Eqs. 9 and 10:
ref (hliq,in − hliq,out ) Q RPC,liq = m ref (hvap,out − hvap,in ) Q RPC,vap = m
(9)
dil (hdil,in − hdil,out ) Q SHX ,hot = m
(11)
conc (hconc,out − hconc,in ) Q SHX ,cold = m
(12)
Fig. 11 shows the measured UA and heat transfer rate for the solution heat exchanger as a function of the concentrated solution flow rate. The measured heat transfer rate and UA varied from 3.49 to 4.47 kW and 0.141 to 0.199 kW K−1, respectively. The design heat duty and UA are 2.19 kW and 0.074 kW K−1, respectively. As expected, the UA and heat transfer rate increase with concentrated solution mass flow rate. The significantly higher heat transfer rate is attributed to the twophase state at the dilute solution inlet leading to condensation of that mixture followed by sensible cooling. The effectiveness of the heat exchanger varied between 0.88 and 0.94 as compared to the design effectiveness of 0.94. As the heat transfer rate for SHX is observed to be above design for all test cases, it is concluded that the component performs above design.
(10)
4.8. Fig. 10 shows a plot of the UA and heat transfer rate for the refrigerant pre-cooler as a function of the refrigerant flow rate. The measured heat transfer rates and UA varied between 0.163 and 0.385 kW, and 0.024 and 0.070 kW K−1, respectively. The design heat duty and UA are 0.345 kW and 0.025 kW K−1, respectively. As shown in the plot, the heat transfer rate and UA increase with refrigerant mass flow rate as expected. The effectiveness of the heat exchanger varied between 0.89 and 0.98 as compared to the design effectiveness of 0.92.
4.7.
Rectifier
The heat transfer rate is calculated using the measured temperature and pressure on both fluid streams, and used to calculate the ammonia fraction of the concentrated solution using Eq. 13–15:
ref = qrec m des,vap,out m qrec,out = f ( Prec, Trec,out , xdes )
xdes = f ( Pdes, Tdes,out, qvap,des = 1) reflux = m des,vap,out − m ref m
(13)
Solution heat exchanger
The heat transfer rate is calculated as shown in Eq. 11 and 12, using the measured temperature and pressure on the dilute solution and concentrated solution streams, and known concentrations and mass flow rates from the absorber and desorber. At steady state, both sides will indicate almost identical heat transfer rates. Therefore, either of these heat transfer rates can
vap,des,out (hvap,rec,in − hvap,rec,out ) Q rec,vap = m
(14)
conc (hconc,out − hconc,in ) Q rec,coolant = m
(15)
Eq. 13 presents the mass balance and the associated thermodynamic state equations to determine the mass flow rate
international journal of refrigeration 79 (2017) 89–100
99
Fig. 12 – Rectifier heat transfer rate and overall conductance.
Fig. 14 – Variation in system COP and cooling duty with ambient temperature.
of vapor leaving the desorber and mass flow rate of the reflux based on the outlet quality of the rectifier. The concentration of vapor leaving the desorber xdes is determined using measured temperature, pressure at desorber outlet assuming saturated vapor quality ( qdes = 1). Fig. 12 shows the measured UA and heat transfer rate for the rectifier as a function of the concentrated solution flow rate. The measured heat transfer rate and UA varied between 0.776 and 1.91 kW and 0.028 and 0.077 kW K−1, respectively. The design heat duty and UA are 1.49 kW and 0.044 kW K−1, respectively. As shown in the plot, the UA and heat transfer rate increase with concentrated solution mass flow rate, which is expected, and the rectifier is appropriately sized to perform at design conditions.
to 2522 kPa and 568.4 to 665.7 kPa, respectively. The design values are 2928 kPa and 609.4 kPa, respectively. Both, the lowand high-side pressures increase with an increase in ambient temperature. The high-side pressure is determined by the temperature at which complete condensation occurs in the condenser. As the ambient temperature increases, the refrigerant outlet temperature in the condenser increases, and consequently increases the high-side pressure. The low-side pressure is affected by the performance of the evaporator and the absorber. As the absorber performed below design and the mass flow rate of the refrigerant was adjusted below design to achieve reasonable cooling performance from the system, the change in the low-side pressure is less significant with ambient conditions. However, if the absorber performs as intended, a similar increase in the low-side pressure with ambient conditions is expected. Fig. 14 shows the variation in cooling duty and COP with ambient temperature. The cooling duty varied from 1.91 to 2.54 kW for a variation in ambient temperature from 44.2 to 29.7 °C. As the ambient temperature increases, it becomes increasingly difficult for absorber and condenser to reject the heat, and therefore, the overall system performance settles at a lower value. The cooling COP, calculated using Eq. 1, followed a similar trend and varied between 0.417 and 0.547. The minor anomalies in the trends of cooling duty and COP, such as duty and COP increasing from 36 °C to 38 °C, are due to changes in experimental parameters such as the overall charge amount and concentration, solution circulation rate, and using two absorber configurations of different fin densities, leading to a change in overall system performance. A total charge of 2–2.5 kg ammonia-water mixture, with the overall concentration varying between 0.45 and 0.6, was used in the system.
4.9.
System performance
Fig. 13 shows the variation in system pressures with ambient temperature. The high- and low-side pressure varied from 1826
5.
Fig. 13 – Variation in system pressures with ambient temperature.
Conclusions
An experimental evaluation of a small-capacity, waste-heat driven ammonia-water absorption chiller for operation at extreme ambient conditions was conducted, with
100
international journal of refrigeration 79 (2017) 89–100
performance documented at the component and system levels. The system is intended to deliver a cooling capacity of 2.71 kW at an ambient temperature of 51.67 °C and a cooling COP of 0.55. In these experiments, it delivered 2.54 kW to 1.91 kW of cooling as the ambient temperature varied between 29.7 and 44.2 °C. The absorber and desorber are identified as the “bottleneck” for system performance. Most other components performed near design and were affected by the performance of the absorber and desorber. Design modifications for the air-coupled absorber have been identified to increase interfacial area and overall fluid-side heat transfer area. For an accurate assessment of assumptions of quality of fluid leaving the desorber, flow visualization capability has been added to the test setup as well. Future work will involve modifying the heat exchangers to approach design performance more closely. A packaged unit system incorporating the modified components will be fabricated and tested at design conditions and coupled with a heat source such as the exhaust from a diesel generator.
REFERENCES
Beutler, A., Hoffmann, L., Ziegler, F., Alefeld, G., Gommed, K., Grossman, G., et al., 1996. Experimental investigation of heat and mass transfer on horizontal and vertical tubes. In: Proceedings of the International Ab-Sorption Heat Pump Conference, Montreal, Canada. pp. 409–419. De Francisco, A., Illanes, R., Torres, J.L., Castillo, M., De Blas, M., Prieto, E., et al., 2002. Development and testing of a prototype of low-power water–ammonia absorption equipment for solar energy applications. Renew Energy 25 (4), 537–544. doi:10.1016/S0960-1481(01)00093-3. Determan, M.D., Garimella, S., 2011. Ammonia–water desorption heat and mass transfer in microchannel devices. Int. J. Refrigeration 34 (5), 1197–1208. Determan, M.D., Garimella, S., 2012. Design, fabrication, and experimental demonstration of a microscale monolithic modular absorption heat pump. Appl. Therm. Eng. 47, 119– 125. doi:10.1016/j.applthermaleng.2011.10.043. Didion, D., Radermacher, R., 1984. Part-load performance characteristics of residential absorption chillers and heat pump. Int. J. Refrigeration 7 (6), 393–398. doi:10.1016/01407007(84)90011-2. Erickson, D.C., Anand, G., Papar, R.A., 1996. Branched GAX cycle gas fired heat pump. In: Energy Conversion Engineering Conference, 1996. IECEC 96. Proceedings of the 31st Intersociety. IEEE, pp. 1078–1083. Filipe Mendes, L., Collares-Pereira, M., Ziegler, F., 1998. Supply of cooling and heating with solar assisted absorption heat pumps: an energetic approach. Int. J. Refrigeration 21 (2), 116– 125. doi:10.1016/S0140-7007(97)00076-5. Garimella, S., 1999. Miniaturized heat and mass transfer technology for absorption heat pumps. In: Proceedings of the International Sorption Heat Pump Conference, Munich, Germany. pp. 661–670. Garimella, S., 2004. Method and Means for Miniaturization of BinaryFluid Heat and Mass Exchangers. U. Patent. US. Garimella, S., Determan, M.D., Meacham, J.M., Lee, S., Ernst, T.C., 2011. Microchannel component technology for system-wide
application in ammonia/water absorption heat pumps. Int. J. Refrigeration 34 (5), 1184–1196. doi:10.1016/ j.ijrefrig.2011.03.005. Garimella, S., Keinath, C.M., Delahanty, J.C., Hoysall, D.C., Staedter, M.A., Goyal, A., et al., 2016. Development and demonstration of a compact ammonia-water absorption heat pump prototype with microscale features for spaceconditioning applications. Appl. Therm. Eng. doi:10.1016/ j.applthermaleng.2016.03.169. Goyal, A., Rattner, A.S., Garimella, S., 2015. Model-based feedback control of an ammonia-water absorption chiller. Sci. Technol. Built Environ. 21 (3), 357–364. Herold, K.E., Radermacher, R., Klein, S.A., 1996. Absorption Chillers and Heat Pumps. CRC Press. Hu, J., Chao, C.Y., 2008. Study of a micro absorption heat pump system. Int. J. Refrigeration 31 (7), 1198–1206. Kang, Y.T., Kunugi, Y., Kashiwagi, T., 2000. Review of advanced absorption cycles: performance improvement and temperature lift enhancement. Int. J. Refrigeration 23 (5), 388– 401. doi:10.1016/S0140-7007(99)00064-X. Keinath, C.M., 2015. Direct-Fired Heat Pump for Multi-Pass Water Heating Using Microchannel Heat and Mass Exchangers. Mechanical Engineering. Atlanta, GA, Georgia Insitute of Technology, Vol. PhD. Keinath, C.M., Hoysall, D., Delahanty, J.C., Determan, M.D., Garimella, S., 2015. Experimental assessment of a compact branched tray generator for ammonia–water desorption. Sci. Technol. Built Environ. 21 (3), 348–356. doi:10.1080/ 23744731.2014.1000797. Klein, S.A., 2014. Engineering Equation Solver, F-Chart Software. Lee, S., Bohra, L.K., Garimella, S., Nagavarapu, A.K., 2012. Measurement of absorption rates in horizontal-tube fallingfilm ammonia-water absorbers. Int. J. Refrigeration 35 (3), 613– 632. doi:10.1016/j.ijrefrig.2011.08.011. Lorentzen, G., 1995. The use of natural refrigerants: a complete solution to the Cfc/Hcfc predicament. Int. J. Refrigeration 18 (3), 190–197. doi:10.1016/0140-7007(94)00001-E. Meacham, J.M., Garimella, S., 2004. Ammonia-water absorption heat and mass transfer in microchannel absorbers with visual confirmation. ASHRAE Trans. 110 (1), 525–532. Mendes, L., Collares-Pereira, M., 1999. A solar assisted and air cooled absorption machine to provide small power heating and cooling. In: International Sorption Heat Pump Conference, Munich, Germany. pp. 129–136. Pence, D., 2010. The simplicity of fractal-like flow networks for effective heat and mass transport. Exp. Therm. Fluid Sci. 34 (4), 474–486. Sözen, A., Altıparmak, D., Usta, H., 2002. Development and testing of a prototype of absorption heat pump system operated by solar energy. Appl. Therm. Eng. 22 (16), 1847–1859. doi:10.1016/ S1359-4311(02)00109-6. Srikhirin, P., Aphornratana, S., Chungpaibulpatana, S., 2001. A review of absorption refrigeration technologies. Renew. Sustain. Energy Rev. 5 (4), 343–372. doi:10.1016/S13640321(01)00003-X. Treffinger, P., 1996. Development and performance of a single stage high efficient absorption heat pump for domestic heatings. In: Proc. Int. Absorption Heat Pump Conference, Montreal. Ziegler, F., 1999. Recent developments and future prospects of sorption heat pump systems. Int. J. Therm. Sci. 38 (3), 191–208. doi:10.1016/S1290-0729(99)80083-0. Ziegler, F., Riesch, P., 1993. Absorption cycles. A review with regard to energetic efficiency. Heat Recov. Syst. CHP 13 (2), 147–159. doi:10.1016/0890-4332(93)90034-S.
Available online at www.sciencedirect.com
ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g
Experimental evaluation of a small-capacity, waste-heat driven ammonia-water absorption chiller Anurag Goyal, Marcel A. Staedter, Dhruv C. Hoysall, Mikko J. Ponkala, Srinivas Garimella * Sustainable Thermal Systems Laboratory, George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA
A R T I C L E
I N F O
A B S T R A C T
Article history:
This paper presents the results from an experimental evaluation of a small-scale, waste-
Received 4 February 2017
heat driven ammonia-water absorption chiller. The cycle is thermally driven, using the waste
Received in revised form 9 April
heat from diesel generator exhaust to desorb the refrigerant solution. The absorber and con-
2017
denser are directly-coupled to the ambient air. The remaining heat exchangers are packaged
Accepted 10 April 2017
in a compact microchannel monolithic structure for enhancing heat transfer. The system
Available online 13 April 2017
is designed to deliver 2.71-kW of cooling at extreme ambient temperature of 51.7 °C at a coefficient of performance of 0.55. Experiments on a heat pump breadboard system are con-
Keywords:
ducted at ambient conditions of 29.7–44.2 °C, with delivered cooling duties of 2.54–1.91 kW.
Vapor absorption chiller
System and individual component performance is analyzed and compared with cycle model
Ammonia-water
predictions, and deviations are explained. The absorber and desorber were identified to be
Waste heat recovery
the limiting components in the system. Effects of variation in ambient temperature are studied to characterize the performance at off-design conditions. © 2017 Elsevier Ltd and IIR. All rights reserved.
Évaluation expérimentale d’un refroidisseur à absorption à ammoniac-eau de faible puissance alimenté par de la chaleur perdue Mots clés : Refroidisseur à diffusion-absorption ; Ammoniac-eau ; Récupération de chaleur perdue
1.
Introduction
Vapor absorption based heating, ventilation and air-conditioning (HVAC) systems can utilize low-grade heat sources such as waste
heat to provide heating and cooling.These systems also provide an alternative to conventional vapor compression systems in terms of reducing the peak demand for electricity (Ziegler, 1999). Specifically, waste-heat or solar energy driven systems can provide more economical cooling/heating solutions in
* Corresponding author. Love Building, Room 340, Ferst Drive, Atlanta, GA 30332. Fax: +1 404 894 8496. E-mail address: [email protected] (S. Garimella). http://dx.doi.org/10.1016/j.ijrefrig.2017.04.006 0140-7007/© 2017 Elsevier Ltd and IIR. All rights reserved.
90
international journal of refrigeration 79 (2017) 89–100
Nomenclature specific heat capacity [J kg−1 °C−1] specific enthalpy [J kg−1] heat exchanger inside diameter [m] mass flow rate [kg s−1] OD outside diameter [m] P pressure [kPa] PWM pulse-width modulation q vapor mass quality [−] RPC refrigerant pre-cooler SHX solution heat exchanger T temperature [°C] x ammonia concentration [–]
Cp h HX ID m
Subscripts abs absorber cf coupling fluid conc concentrated solution cond condenser des desorber dil dilute solution evap evaporator in inlet liq liquid out outlet rec rectifier ref refrigerant vap vapor
conjunction with power generation systems for combined heat and power (CHP) applications as shown by Ziegler and Riesch (1993) and Filipe Mendes et al. (1998). These systems also use environmentally benign working fluids, such as ammoniawater, with reduced global warming potential (GWP) (Lorentzen, 1995). Vapor absorption systems typically require more heat and mass exchange components compared to conventional vapor compression systems. The mechanical compressor is replaced by a set of heat and mass exchangers that enable heat source utilization (Herold et al., 1996; Srikhirin et al., 2001).This leads to larger system size and high capital cost, which has historically limited the use of such systems for residential and light commercial applications. To improve the efficiency of the single-effect absorption cycle, advanced cycles with internal heat recovery and multi-staged heat transfer mechanisms have been proposed by researchers in the past (Kang et al., 2000; Srikhirin et al., 2001; Ziegler, 1999; Ziegler and Riesch, 1993). However, in such systems, the increase in system complexity typically does not yield commensurate benefits in efficiency. Some past research efforts experimentally evaluated absorption systems and advanced cycles for improved efficiency. Didion and Radermacher (1984) presented results on partload performance and ambient temperature variation for a 10.5 kW cooling capacity gas-fired ammonia-water absorption chiller. It was observed that the COP and cooling duty decreased as the ambient temperature increased. Treffinger (1996) presented a direct gas-fired ammonia-water system for water heating
applications. The COP of the system varied between 1.4 and 1.7, with the heating COP defined as the ratio of the total heat transfer rate from condenser and absorber to the heat input in the desorber. Erickson et al. (1996) demonstrated an ammoniawater generator-absorber heat exchange (GAX) cycle with cooling duty of 14.6 kW at a COP of 1.05. Mendes and Collares-Pereira (1999) presented results from the design and testing of a smallcapacity solar absorption heat pump.The system demonstrated a cooling capacity of 5 kW (COP 0.56) and a heating capacity of 9 kW (COP 1.42). Sözen et al. (2002) presented results from testing of an ammonia-water absorption heat pump using solar energy. De Francisco et al. (2002) also presented experimental analysis of a solar powered ammonia-water system for small-scale refrigeration applications. The system was designed to provide a cooling duty of 2 kW, but the experiments showed limited performance and low COP due to using only natural convection for condenser and absorber heat rejection. This also led to very large component size. These studies report a need for more compact heat and mass exchangers to reduce the overall system footprint. Additionally, the vast majority of reported work did not consider extreme ambient conditions, which introduce further development challenges. Reduction of the size of vapor absorption system components has been the focus of active research for some time. Garimella (1999) and Garimella (2004) presented a design for a microscale heat and mass exchange component for binary fluid mixtures. The falling-film type component utilized 1.58 mm OD tubes, and can be used as any component in the absorption cycle with minor modifications. Meacham and Garimella (2004) demonstrated the performance of the aforementioned device as an absorber and addressed issues with fluid distribution and techniques to improve the component performance. Later, Determan and Garimella (2011) and Garimella et al. (2011) demonstrated the use of the same device for desorption with minor modifications in fluid routing. However, all these designs were still less suitable for a compact heat pump packaged design with small fluid inventory. Hu and Chao (2008) demonstrated the performance of a very small-scale water-lithium bromide absorption system with 40 W cooling capacity. The components were fabricated using silicon wafers with photo-chemically etched micro-channels of hydraulic diameter of 131 µm. Pence (2010) investigated micro-scale fractal flow networks for design of heat and mass exchangers to increase the vapor flow area and reduce the pressure drop in two-phase components. Determan and Garimella (2012) presented a micro-scale ammonia-water absorption heat pump with all the components in a single monolithic structure. The system delivered 300 W of cooling at a COP of 0.375. It demonstrated the potential of component integration and efficient packaging to achieve small system sizes, and the potential for scaling the systems to higher capacities. Keinath (2015) presented a compact absorption heat pump water heater for residential applications. The prototype system utilized a monolithic fabrication approach to integrate multiple heat and mass exchange components. It delivered 2.79 kW of heating at a COP of 1.74. Garimella et al. (2016) demonstrated the performance of a microchannel absorption chiller packaged unit. The natural gas fired system delivered a cooling duty of 3.3 kW at a COP of 0.47. Their work demonstrated the scalability of microscale designs.
international journal of refrigeration 79 (2017) 89–100
The present study extends the applicability of microscale heat and mass exchange devices for the development of compact absorption heat pumps. The component design focuses on direct integration of the heat source, thereby avoiding the use of a secondary heat transfer loop. The heat source investigated in this work is exhaust from a diesel generator. However, the design of the heat coupling ensures flexibility of using in a direct-fired application as well. The absorber and condenser also reject heat directly to the ambient to eliminate hydronic loops, which require additional components and larger system size. In addition, the proposed system is designed for diverse applications such as mobile refrigeration units, residential heat pump systems, and military applications with severe ambient conditions such as those seen at forward operating bases. Results from cycle modeling and experiments are presented, and the deviations from predicted performance are discussed. Potential changes in component design to improve system performance are also proposed.
2.
System description and modeling
A single-effect ammonia-water absorption chiller is investigated. Ammonia-water is selected as the working fluid pair for the current system because ammonia has high latent heat of vaporization, which is necessary for efficient performance. Also, the high vapor-phase density allows for smaller component sizes. For the intended high-ambient temperature conditions, ammonia-water system does not present a crystallization risk, which would be the case with water-lithium bromide if
91
direct air-cooling is implemented. Moreover, the freezing point of ammonia is 196.15 K, which allows for low-temperature refrigeration as opposed to water-lithium bromide systems (Srikhirin et al., 2001). The desorber, rectifier, solution heat exchanger, and condenser operate at the high-side pressure, while the absorber and evaporator operate at the low-side pressure. A schematic of the different system components is shown in Fig. 1. The desorber uses heat supplied from a hightemperature source, the condenser and the absorber reject heat to the ambient, and the evaporator provides cooling by receiving heat from a low-temperature source. Internal heat recovery through recuperative heat exchangers on the solution and refrigerant circuits helps in improving the COP and decreasing the desorber and absorber heat duties. As shown in Fig. 1, the concentrated solution is pumped from low-pressure (1) to high-pressure (2). The concentrated solution cools the rectifier component condensing residual water in the generated refrigerant vapor leaving the desorber. This also recovers energy internal to the system by preheating the concentrated solution. The concentrated solution at (3) enters the solution heat exchanger and exchanges heat with the dilute solution exiting the desorber. The recuperative heat exchange further heats up the concentrated solution before entering the desorber at (4). The desorber receives heat from high temperature source and separates ammoniarich vapor from the liquid mixture. The ammonia-rich vapor exits the desorber (8), and is further purified in the rectifier to remove trace quantities of water. The purified refrigerant vapor leaves the rectifier and enters the condenser (9), while the condensed solution from the vapor is remixed into the desorber (10). The refrigerant vapor is condensed and the
Fig. 1 – Schematic of single-effect absorption cycle.
92
international journal of refrigeration 79 (2017) 89–100
Table 1 – Steady-state cycle model design inputs and calculated heat transfer rates. Component
UA (W K−1)
Coupling Fluid −1
Absorber Condenser Desorber Evaporator Rectifier Pre-cooler Solution heat exchanger
379.6 187.5 55 846 43.9 25 74.1
Baseline Duty (W)
Flow rate (kg s )
Inlet Temperature (°C)
0.383 0.383 0.0255 0.0957
51.67 51.67 398.9 19.8
heat of condensation is rejected to the ambient. The liquid refrigerant at high-pressure (11) exchanges heat with colder low-pressure refrigerant vapor in the refrigerant pre-cooler. The liquid refrigerant (12) is throttled to the low-pressure side through an expansion valve. The refrigerant enters the evaporator (13) and exchanges heat with low-temperature source, exiting in a high quality two-phase state at (14). The refrigerant vapor gains heat in the refrigerant pre-cooler, leaving at (15), and mixes with the dilute solution at the inlet of the absorber. The dilute solution leaving the desorber (5) exchanges heat with concentrated solution in the solution heat exchanger and cools to a high-pressure liquid state at (6). The dilute solution is throttled to the low-side pressure (7) across an expansion valve. The absorber rejects the heat of absorption to the ambient, and regenerates the concentrated solution, thus completing the cycle. A steady-state cycle model is developed to predict the baseline heat exchanger sizes for different components and conduct parametric analyses to study the effects of variations in operating conditions such as ambient temperature, solution mass flow rate, and heat source temperature and mass flow rate. The model is developed on the Engineering Equation Solver (EES) platform (Klein, 2014). Three independent thermodynamic properties are required to completely define the state of the binary working fluid. Initial overall heat conductance values (UA) of the heat exchangers are determined by making appropriate assumptions about their closest approach temperatures or effectiveness. The baseline model is provided with ambient conditions, heat source inlet temperature, coolant inlet temperature, and flow rates of coupling fluids and the concentrated solution. To achieve optimal system performance, the mass fraction of ammonia in the refrigerant vapor exiting the rectifier was specified at 0.998. The high purity of the refrigerant vapor is achievable with compact desorber and rectifier designs as shown by Keinath et al. (2015). Mass, species, and energy conservation equations are setup for each component and the simultaneous system of equations is solved iteratively. The baseline UA values are then optimized to maximize the system COP for specified operating conditions. The air inlet temperature to the condenser and absorber was specified at 51.67 °C, with a relative humidity of 18%. An air mass flow rate of 0.383 kg s−1 for both heat exchangers was specified. Hot air was supplied to the desorber at 398.9 °C and a mass flow rate of 0.025 kg s−1. The concentrated solution flow rate was specified at 0.008 kg s−1. Temperature in-equilibrium values for rectifier vapor inlet and reflux outlet, and desorber
5147 2529 4937 2713 1496 345 2196
solution inlet and vapor outlet were also specified. The trace amount of water in the refrigerant leads to a rise in temperature of the refrigerant along the length of the evaporator. To optimize the heat transfer performance of the evaporator, temperature glide in the evaporator is specified at 2.5 K. Table 1 presents the inputs to the model and the heat transfer rates in different components. The high-side pressure at design conditions was calculated to be 2928 kPa, while the low-side pressure was 585.3 kPa. The refrigerant flow rate and dilute solution flow rate were calculated to be 0.0025 kg s −1 and 0.0055 kg s−1, respectively. The ammonia mass fractions of the dilute and concentrated solution were calculated to be 0.163, and 0.422, respectively. The cooling COP of the system, as defined in Eq. (1), was calculated to be 0.55.
Q evap COPcooling = Q des
(1)
The overall COP of the system, accounting for electrical input for pumps, fans, electronic valves, and control system, is calculated as defined in Eq. (2). Using an estimated total electrical load of 375 W for pump, fan and control system electricity consumption, and a primary energy load factor of 3.18 to obtain source-based fuel energy consumption, the overall COP of the system is calculated to be 0.44.
Q evap COPoverall = Q des + E total
(2)
Fig. 2 shows the variation in COP and cooling duty as the ambient temperature changes. A lower ambient temperature allows improved heat rejection from the absorber and the condenser, and thereby leads to an increase in the overall system cooling capacity. Improved absorber heat rejection also leads to reduced low-side pressure, and thus lower refrigerant inlet temperature at the evaporator and greater cooling capacity.
3.
Experimental setup
Fig. 3 shows a schematic of the breadboard test facility, which allows the evaluation of the entire system, as well as individual component performance. The key features of each component are listed in Table 2.
international journal of refrigeration 79 (2017) 89–100
Fig. 2 – Variation in ambient temperature and predicted cooling capacity and COP.
3.1.
Heat exchangers
Detailed, segmented heat and mass transfer models were developed for individual heat exchangers using operating conditions computed by the thermodynamic model of the system. The condenser and absorber are custom-designed multi-pass finned-tube heat exchangers. The heat of
93
absorption and condensation is rejected to the ambient air, which is circulated by dedicated fans. A novel desorber based on a diabatic distillation column concept that integrates heat transfer stages with purification stages was developed. This provides a compact design with very small liquid-vapor approach temperature differences, and consequently, high-purity vapor leaving the desorber. A small, integrated analyzer section in the column further improves ammonia concentration in the vapor stream and decreases the rectification requirement. Due to the direct coupling to engine exhaust, the design allows for a maximum pressure drop of 2.5 kPa on the gas side. The rectifier is a compact shell-andtube heat exchanger with partial condensation of vapor on the shell-side and concentrated solution flowing on the tube-side. The refrigerant pre-cooler, evaporator, and solution heat exchanger are all microchannel plate type heat exchangers, integrated in one monolithic block, as shown in Fig. 4. The bonded monolithic heat exchanger consists of stainless steel plates with microscale features. The two plate designs are stacked in an alternating pattern for working fluid and coupling fluid sides of the heat exchanger, and enclosed by a cover plate and an endplate. The fabrication process involved photochemical etching of microscale flow passages on the shims, followed by bonding the assembly using diffusion bonding or vacuum brazing. The microchannel fluid passages have a hydraulic diameter of 440 µm. The experimental test setup is shown in Fig. 5.
Fig. 3 – Breadboard test facility schematic.
94
international journal of refrigeration 79 (2017) 89–100
Table 2 – Component geometry details. Component
Type
Absorber
Air-cooled, cross-flow, finned-tube
Condenser
Air-cooled, cross-flow, finned-tube
Desorber
Evaporator (Coolant coupled)
Direct-coupled, shell-and-tube with finned trays on shell-side, counter-flow liquid and vapor Microchannel array, counter-flow, liquid coupled
Rectifier (Solutioncooled) Pre-cooler
Solution-cooled shell-and-tube, counterflow Microchannel array, counter-flow
Solution heat exchanger
Microchannel array, counter-flow
3.2. Heat source coupling loops and auxiliary components The evaporator coolant loop circulates a 25% mixture of propylene-glycol and water, and contains a pulse-width modulation (PWM) controlled electrical heater, which maintains the coolant inlet temperature at the set-point.
Fig. 4 – Monolith shim (top); bonded monolith (bottom).
Key Dimensions Tube ID: 0.0141 m, Fin density: 551 and 393 m−1 H × W: 0.419 × 0.61 m, 12 tube passes Tube ID: 0.0141 m, Fin density: 393 m−1 H × W: 0.419 × 0.31 m, 9 tube passes Two-columns, Shell ID: 0.102 m Exhaust Tube bundle: ID 0.011 m, 22 tubes Height of column: 0.416 m Channel Hydraulic Diameter: 440 µm 66 plates: 22 refrigerant, 44 coolant flow 62 channels/shim, Refrigerant channels: 0.0995 m, Coolant channels: 0.0795 m, 0.154 × 0.126 m (L × W) Exergy LLC 35 Series, Shell-side: OD 0.038 m Tube-side: OD 0.0032 m, 55 tubes, 0.254 m Channel Hydraulic Diameter: 440 µm, 66 plates: 33 pairs, 42 channels/shim, Straight channels: 0.109 m, Diagonal Channels: 0.0746 m, 0.227 × 0.085 m (L × W) Channel Hydraulic Diameter: 440 µm, 66 plates: 33 pairs, 42 channels/shim, Straight channels: 0.109 m, Diagonal Channels: 0.0746 m, 0.227 × 0.085 m (L × W)
An electrically heated air loop serves as the waste heat source and simulates a waste heat stream such as exhaust gas from a diesel generator. A PWM controller for the heaters regulates the temperature set-point of the air supplied to the desorber. The air-heater assembly simulates a waste heat source, such as exhaust gas from an electric generator accurately, and provides precise control over the flow rate and temperature. Storage tanks are installed at the outlet of absorber, condenser, rectifier, and dilute solution outlet of the desorber. These storage tanks ensure appropriate fluid inventory in different parts of the system, allowing for smoother steady-state operation, and accurate measurement of flow rates by ensuring that the different fluids streams are in single phase. At the outlet of the rectifier, refrigerant vapor is separated from the
Fig. 5 – Experimental setup.
95
international journal of refrigeration 79 (2017) 89–100
Table 3 – Equipment and measurement probe specifications. Instrument Thermocouple Pressure Transducer Mass Flow Sensor Coolant Mass Flow Sensor Level Measurement Probe Electronic Expansion Valves Data Acquisition and Control
Manufacturer/Model
−250–300 °C 0–750 °C 0–1034.2 kPa 0–3447.3 kPa 0–0.01 kg s−1 0–0.1 kg s−1 0–0.6056 kg s−1 0–0.152 m
Omega® T-type Omega® J-type Omega® 4–20 mA Output® Absolute Rheonik® Coriolis Sensor MicroMotion® Coriolis Sensor Endress+Hauser® Liquicap M FMI51 Carel® E2V Series National Instruments® c-RIO 9024
condensate in a storage tank before it enters the condenser. The reflux combines with the solution inlet stream to the desorber. The dilute solution storage tank ensures that the solution heat exchanger receives liquid-phase solution. The dilute solution and refrigerant vapor are mixed in a tube-in-tube type injection port at the inlet of the absorber. The regenerated solution is stored in a solution storage tank downstream of the absorber, which provides liquid-phase solution to the pump. A positive-displacement piston pump circulates the ammoniawater mixture.
3.3.
Range
Instrumentation and controls
As shown in Fig. 3, the temperature at the inlet and outlet of each stream of individual heat exchangers is measured using T-type thermocouples. Pressures are measured using electronic pressure transducers at the outlet of the desorber, the outlet of the condenser, inlet to the evaporator, and outlet of the absorber. The two pressure measurements on the lowpressure side also help in characterizing the total pressure drop on the low-pressure side components. Four high accuracy (±0.1%) Coriolis mass flow sensors are used to measure the mass flow rates of the concentrated solution, dilute solution, refrigerant solution, and evaporator coolant. It should be noted that this leads to very small uncertainties in calculated variables due to mass flow rate measurement, and the corresponding error bars are not visible on the plots. In addition, an analog air-flow meter is used to measure the volumetric flow rate of hot air supplied to the desorber. A capacitance based level measurement probe is used on the refrigerant tank to measure the liquid level. This is used to monitor and regulate appropriate liquid storage level in the tank during transients and change in operating conditions. The speed of the pump and the fans for the condenser and absorber are controlled using variable speed drives. Both the expansion valves in the system are variable-opening electronic expansion valves positioned by a stepper motor. A proportional control based feedback loop, as described by Goyal et al. (2015), is used to regulate the valve position. Data acquisition and control is done using National Instruments® LabVIEWTM. Steady-state data are recorded for a 20-minute duration at a scan rate of 100 ms. Table 3 provides specifications of all the measurement probes.
Accuracy ±0.5 °C ±1.1 °C ±0.03% Range ±0.1% ±0.1% ±0.5%
480 motor steps
–
–
–
4.
Results and discussion
4.1.
Data analysis
Experiments were first conducted to analyze the system performance at nominal ambient conditions of 30 °C, and followed by testing at increasing ambient temperatures approaching design conditions. Data analysis is conducted as described in Lee et al. (2012), using time-averaged values of different measured quantities to define inlet and outlet states of each component. Heat transfer rate in each component is then calculated and system performance analyzed. Key assumptions in the analysis are: • The refrigerant vapor leaves the desorber/analyzer section in the saturated state. • The vapor leaves the separator tank after the rectifier and enters the condenser in the saturated state. • The reflux from the rectifier leaves the bottom of the separator in the saturated state. These phase-quality assumptions are required to determine the concentration of the refrigerant, reflux, and concentrated solution streams, and hence set the third required state property along with the measured pressure and temperature. The dilute solution concentration is determined by performing mass and species conservation on the absorber component. Mass and energy conservation are also used to validate the steady-state nature of the recorded data set. Engineering Equation Solver (EES) (Klein, 2014) is used for data analysis. Using the measured heat transfer rate, which is obtained from the mass flow rates and measured temperature and pressure, the overall conductance can be expressed as shown in Eq. (3),
UA =
Q LMTD
(3)
where, Q is the measured heat transfer rate, and LMTD is the logarithmic mean temperature difference. The overall conductance calculated here is indicative of the spatially averaged conductance in the component; detailed in-situ measurements
96
international journal of refrigeration 79 (2017) 89–100
Fig. 7 – Condenser heat transfer rate and overall conductance. Fig. 6 – Absorber heat transfer rate and ambient conditions.
to understand local variations were not performed. These calculations provide an overall estimate of the component performance when integrated into the complete chiller system.
4.2.
Absorber
The heat and mass transfer processes inside an absorber take place across adverse concentration gradients and small driving temperature difference, making the role of the absorber very crucial for the overall system performance (Beutler et al., 1996). Fig. 6 shows the heat transfer rate for the absorber as a function of the concentrated solution flow rate. The heat transfer rate is calculated on the working fluid-side using measured flow rates and states of two-solution streams mixing at the inlet of the absorber. The dilute solution expansion valve is assumed to be isenthalpic for calculation purposes. Mass, species and energy conservation are used to establish the concentration and inlet enthalpy in the absorber. Overall mass balance for the dilute and the refrigerant solution mass flow rate differs from the measured concentrated solution flow rate by a maximum of 6.5%. The outlet temperature and pressure are measured, and using the absorber concentration, the outlet enthalpy is determined. The overall heat transfer is calculated as shown in Eq. 4,
conc (habs,in − habs,out ) Q abs = m
(4)
As shown in Fig. 6, the heat duty for the absorber is below the design values. The temperature values on the plot indicate the ambient condition of the test. The measured heat transfer rate varied between 4.47 and 5.16 kW. The design values for heat transfer rate and UA are 5.14 kW and 0.38 kW K−1, respectively. Overall, the heat transfer rate remains below design for most test cases indicating limitation in the performance of the component. The heat transfer rate for the absorber is below design at both ends of the ambient temperature. This can be attributed to limitation in the overall conductance and
variation in the driving temperature difference between the working fluid and ambient air. Near design ambient conditions, the limitation in heat transfer requires that the refrigerant flow rate be reduced to decrease the low-side pressure and absorber concentration, leading to an overall decrease in system performance. Modifications to the absorber, including additional tube passes with reduced tube diameter of 9.5 mm OD are expected to increase tube-side heat transfer coefficients as well as surface area. To increase liquid-vapor interfacial area, twistedtape type steel inserts can also be used. The overall pressure drop on the low-pressure side of the system, from evaporator inlet to absorber outlet, stayed below 27 kPa for all the test cases, which is significantly less than the design pressure drop of 15 kPa in any individual component. Therefore, the changes to incorporate additional interfacial area and reducing the tube diameter will not exceed the design pressure drop value.
4.3.
Condenser
Fig. 7 shows the measured overall conductance and heat transfer rate for the condenser. The heat transfer rate is calculated on the refrigerant side using measured refrigerant mass flow rate at steady state, and temperature, pressure and concentration to calculate the refrigerant enthalpy, as shown in Eqs. 5 and 6,
xref = f ( Pcond, Tcond,inlet , qvap )
(5)
ref (href ,cond,in − href ,cond,out ) Q cond = m
(6)
where, xref is refrigerant vapor concentration, Pcond is pressure measured at condenser outlet, Tcond,inlet is temperature measured at the inlet of the condenser, qvap is the saturated vapor quality (=1) at the inlet. The heat transfer rate is calculated using ref and the refrigerant inlet and the measured mass flow rate, m outlet enthalpies. The heat transfer rate varied between 2.15 and 2.901 kW as compared to the design value of 2.53 kW. The overall
international journal of refrigeration 79 (2017) 89–100
Fig. 8 – Desorber heat transfer rate and overall conductance.
Fig. 9 – Evaporator heat transfer rate and overall conductance.
conductance of the condenser varied between 0.145 and 0.197 kW K−1, as shown in Fig. 7. However, as a result of the limiting performance of the absorber leading to a decreased absorber concentration, the refrigerant flow rate is lower than the design value of 0.00247 kg s−1 except for the lowest ambient temperature case. The high uncertainty associated with certain data points is due to a near-saturation outlet state of the condenser, and a consequently large change in outlet enthalpy within the temperature measurement uncertainty. The UA of the condenser increases with an increase in refrigerant mass flow rate as the heat transfer coefficient increases. At neardesign refrigerant mass flow rate, the UA is close to the design value, while the heat transfer rate varies with LMTD depending on the ambient conditions. Overall, the condenser seems to perform near-design, especially when the design flow rates are provided, and slight modifications in tube diameter or number of tube passes can be made to further augment the performance.
measured heat duty and UA varied between 4.26 and 4.97 kW, and 0.027 and 0.037 kW K−1, respectively, while corresponding design values are 4.94 kW and 0.055 kW K−1, respectively. The heat transfer rate and UA are lower than design owing to the dilute solution stream leaving as a two-phase mixture, and consequently a lower refrigerant generation rate. At the lower concentration of the solution, the solution mass flow rate should be increased above design to achieve the same refrigerant mass flow rate and suppress the two-phase mixture leaving though the dilute solution outlet. However, in compact geometries this can often lead to flooding in the component as the liquid solution can be entrained in the exiting vapor stream, leading to a poor quality and concentration in the rectifier. Further experiments will focus on validating the quality at the desorber exit, as a small fraction of vapor in the dilute solution significantly affects the performance of the desorber and the absorber.
4.5. 4.4.
97
Evaporator
Desorber
The heat transfer rate in the desorber is measured on the solution-side using steady-state energy conservation equation as shown in Eq. 7:
vap,des,out hvap,des,out + m liq,des,out hliq,des,out − m conc,inhconc,in Q des = m reflux,inhreflux,in −m
(7)
vap,des,out is the vapor phase mass flow rate leaving the where, m desorber, and the enthalpy of vapor, hvap,des,out , is determined using measured temperature, pressure assuming a saturated liq,des,out is the liquid dilute vapor (quality =1) state. Similarly, m solution exiting the desorber and the enthalpy hliq,des,out is calculated using measured temperature, pressure, and concentration calculated from the absorber states. The inlet streams into the desorber are the concentrated solution from solution heat exchanger and reflux from the rectifier, with their states determined from the absorber and rectifier, respectively. Fig. 8 shows a plot of the desorber heat duty and overall UA as a function of the concentrated solution flow rate. The
Owing to the small temperature difference on the refrigerant side and strong dependence of refrigerant enthalpy on temperature, the heat transfer rate is calculated on the coupling fluid side of the evaporator. The heat transfer rate is determined using Eq. 8:
cf ,evapCpcf ,evap ( Tcf ,evap,in − Tcf ,evap,out ) Q evap = m
(8)
Fig. 9 shows a plot of the evaporator UA and heat duty as a function of refrigerant flow rate. The measured heat transfer rates and UA varied between 2.54 and 1.91 kW, and 0.486 and 0.724 kW K−1, respectively, while the corresponding design heat duty and UA are 2.71 kW and 0.850 kW K−1, respectively. The high uncertainty in the reported values is due to the relatively smaller temperature difference (~4–6 °C) on the coupling fluid side and the constant associated uncertainty of temperature measurement (±0.5 °C). The temperature glide, i.e., the difference between the refrigerant outlet and inlet temperature, varied between 1.55 and 8.44 °C as compared to the design value of 2.5 °C. Both the heat transfer rate and UA increase with
98
international journal of refrigeration 79 (2017) 89–100
Fig. 10 – Refrigerant pre-cooler heat transfer rate and overall conductance.
Fig. 11 – Solution heat exchanger heat transfer rate and overall conductance.
an increase in refrigerant mass flow rate, which is expected due to an increase in the heat transfer coefficient. The evaporator performance was below design as the refrigerant mass flow rate remained below design for most test cases. The temperature glide was higher than the design value for the same reason. It is expected that at design refrigerant flow rate, the evaporator performance will be near-design.
be used to calculate the overall conductance. In this work, an average heat transfer rate of the two streams is used.
4.6.
Refrigerant pre-cooler
The refrigerant pre-cooler allows heat recuperation between high-pressure refrigerant liquid and low-pressure refrigerant vapor, and reduces the liquid temperature at the inlet of the refrigerant expansion valve. Using the measured temperature and pressure on both fluid streams, and known refrigerant concentration and mass flow rate from the condenser, the heat transfer rates for both fluid streams are calculated using Eqs. 9 and 10:
ref (hliq,in − hliq,out ) Q RPC,liq = m ref (hvap,out − hvap,in ) Q RPC,vap = m
(9)
dil (hdil,in − hdil,out ) Q SHX ,hot = m
(11)
conc (hconc,out − hconc,in ) Q SHX ,cold = m
(12)
Fig. 11 shows the measured UA and heat transfer rate for the solution heat exchanger as a function of the concentrated solution flow rate. The measured heat transfer rate and UA varied from 3.49 to 4.47 kW and 0.141 to 0.199 kW K−1, respectively. The design heat duty and UA are 2.19 kW and 0.074 kW K−1, respectively. As expected, the UA and heat transfer rate increase with concentrated solution mass flow rate. The significantly higher heat transfer rate is attributed to the twophase state at the dilute solution inlet leading to condensation of that mixture followed by sensible cooling. The effectiveness of the heat exchanger varied between 0.88 and 0.94 as compared to the design effectiveness of 0.94. As the heat transfer rate for SHX is observed to be above design for all test cases, it is concluded that the component performs above design.
(10)
4.8. Fig. 10 shows a plot of the UA and heat transfer rate for the refrigerant pre-cooler as a function of the refrigerant flow rate. The measured heat transfer rates and UA varied between 0.163 and 0.385 kW, and 0.024 and 0.070 kW K−1, respectively. The design heat duty and UA are 0.345 kW and 0.025 kW K−1, respectively. As shown in the plot, the heat transfer rate and UA increase with refrigerant mass flow rate as expected. The effectiveness of the heat exchanger varied between 0.89 and 0.98 as compared to the design effectiveness of 0.92.
4.7.
Rectifier
The heat transfer rate is calculated using the measured temperature and pressure on both fluid streams, and used to calculate the ammonia fraction of the concentrated solution using Eq. 13–15:
ref = qrec m des,vap,out m qrec,out = f ( Prec, Trec,out , xdes )
xdes = f ( Pdes, Tdes,out, qvap,des = 1) reflux = m des,vap,out − m ref m
(13)
Solution heat exchanger
The heat transfer rate is calculated as shown in Eq. 11 and 12, using the measured temperature and pressure on the dilute solution and concentrated solution streams, and known concentrations and mass flow rates from the absorber and desorber. At steady state, both sides will indicate almost identical heat transfer rates. Therefore, either of these heat transfer rates can
vap,des,out (hvap,rec,in − hvap,rec,out ) Q rec,vap = m
(14)
conc (hconc,out − hconc,in ) Q rec,coolant = m
(15)
Eq. 13 presents the mass balance and the associated thermodynamic state equations to determine the mass flow rate
international journal of refrigeration 79 (2017) 89–100
99
Fig. 12 – Rectifier heat transfer rate and overall conductance.
Fig. 14 – Variation in system COP and cooling duty with ambient temperature.
of vapor leaving the desorber and mass flow rate of the reflux based on the outlet quality of the rectifier. The concentration of vapor leaving the desorber xdes is determined using measured temperature, pressure at desorber outlet assuming saturated vapor quality ( qdes = 1). Fig. 12 shows the measured UA and heat transfer rate for the rectifier as a function of the concentrated solution flow rate. The measured heat transfer rate and UA varied between 0.776 and 1.91 kW and 0.028 and 0.077 kW K−1, respectively. The design heat duty and UA are 1.49 kW and 0.044 kW K−1, respectively. As shown in the plot, the UA and heat transfer rate increase with concentrated solution mass flow rate, which is expected, and the rectifier is appropriately sized to perform at design conditions.
to 2522 kPa and 568.4 to 665.7 kPa, respectively. The design values are 2928 kPa and 609.4 kPa, respectively. Both, the lowand high-side pressures increase with an increase in ambient temperature. The high-side pressure is determined by the temperature at which complete condensation occurs in the condenser. As the ambient temperature increases, the refrigerant outlet temperature in the condenser increases, and consequently increases the high-side pressure. The low-side pressure is affected by the performance of the evaporator and the absorber. As the absorber performed below design and the mass flow rate of the refrigerant was adjusted below design to achieve reasonable cooling performance from the system, the change in the low-side pressure is less significant with ambient conditions. However, if the absorber performs as intended, a similar increase in the low-side pressure with ambient conditions is expected. Fig. 14 shows the variation in cooling duty and COP with ambient temperature. The cooling duty varied from 1.91 to 2.54 kW for a variation in ambient temperature from 44.2 to 29.7 °C. As the ambient temperature increases, it becomes increasingly difficult for absorber and condenser to reject the heat, and therefore, the overall system performance settles at a lower value. The cooling COP, calculated using Eq. 1, followed a similar trend and varied between 0.417 and 0.547. The minor anomalies in the trends of cooling duty and COP, such as duty and COP increasing from 36 °C to 38 °C, are due to changes in experimental parameters such as the overall charge amount and concentration, solution circulation rate, and using two absorber configurations of different fin densities, leading to a change in overall system performance. A total charge of 2–2.5 kg ammonia-water mixture, with the overall concentration varying between 0.45 and 0.6, was used in the system.
4.9.
System performance
Fig. 13 shows the variation in system pressures with ambient temperature. The high- and low-side pressure varied from 1826
5.
Fig. 13 – Variation in system pressures with ambient temperature.
Conclusions
An experimental evaluation of a small-capacity, waste-heat driven ammonia-water absorption chiller for operation at extreme ambient conditions was conducted, with
100
international journal of refrigeration 79 (2017) 89–100
performance documented at the component and system levels. The system is intended to deliver a cooling capacity of 2.71 kW at an ambient temperature of 51.67 °C and a cooling COP of 0.55. In these experiments, it delivered 2.54 kW to 1.91 kW of cooling as the ambient temperature varied between 29.7 and 44.2 °C. The absorber and desorber are identified as the “bottleneck” for system performance. Most other components performed near design and were affected by the performance of the absorber and desorber. Design modifications for the air-coupled absorber have been identified to increase interfacial area and overall fluid-side heat transfer area. For an accurate assessment of assumptions of quality of fluid leaving the desorber, flow visualization capability has been added to the test setup as well. Future work will involve modifying the heat exchangers to approach design performance more closely. A packaged unit system incorporating the modified components will be fabricated and tested at design conditions and coupled with a heat source such as the exhaust from a diesel generator.
REFERENCES
Beutler, A., Hoffmann, L., Ziegler, F., Alefeld, G., Gommed, K., Grossman, G., et al., 1996. Experimental investigation of heat and mass transfer on horizontal and vertical tubes. In: Proceedings of the International Ab-Sorption Heat Pump Conference, Montreal, Canada. pp. 409–419. De Francisco, A., Illanes, R., Torres, J.L., Castillo, M., De Blas, M., Prieto, E., et al., 2002. Development and testing of a prototype of low-power water–ammonia absorption equipment for solar energy applications. Renew Energy 25 (4), 537–544. doi:10.1016/S0960-1481(01)00093-3. Determan, M.D., Garimella, S., 2011. Ammonia–water desorption heat and mass transfer in microchannel devices. Int. J. Refrigeration 34 (5), 1197–1208. Determan, M.D., Garimella, S., 2012. Design, fabrication, and experimental demonstration of a microscale monolithic modular absorption heat pump. Appl. Therm. Eng. 47, 119– 125. doi:10.1016/j.applthermaleng.2011.10.043. Didion, D., Radermacher, R., 1984. Part-load performance characteristics of residential absorption chillers and heat pump. Int. J. Refrigeration 7 (6), 393–398. doi:10.1016/01407007(84)90011-2. Erickson, D.C., Anand, G., Papar, R.A., 1996. Branched GAX cycle gas fired heat pump. In: Energy Conversion Engineering Conference, 1996. IECEC 96. Proceedings of the 31st Intersociety. IEEE, pp. 1078–1083. Filipe Mendes, L., Collares-Pereira, M., Ziegler, F., 1998. Supply of cooling and heating with solar assisted absorption heat pumps: an energetic approach. Int. J. Refrigeration 21 (2), 116– 125. doi:10.1016/S0140-7007(97)00076-5. Garimella, S., 1999. Miniaturized heat and mass transfer technology for absorption heat pumps. In: Proceedings of the International Sorption Heat Pump Conference, Munich, Germany. pp. 661–670. Garimella, S., 2004. Method and Means for Miniaturization of BinaryFluid Heat and Mass Exchangers. U. Patent. US. Garimella, S., Determan, M.D., Meacham, J.M., Lee, S., Ernst, T.C., 2011. Microchannel component technology for system-wide
application in ammonia/water absorption heat pumps. Int. J. Refrigeration 34 (5), 1184–1196. doi:10.1016/ j.ijrefrig.2011.03.005. Garimella, S., Keinath, C.M., Delahanty, J.C., Hoysall, D.C., Staedter, M.A., Goyal, A., et al., 2016. Development and demonstration of a compact ammonia-water absorption heat pump prototype with microscale features for spaceconditioning applications. Appl. Therm. Eng. doi:10.1016/ j.applthermaleng.2016.03.169. Goyal, A., Rattner, A.S., Garimella, S., 2015. Model-based feedback control of an ammonia-water absorption chiller. Sci. Technol. Built Environ. 21 (3), 357–364. Herold, K.E., Radermacher, R., Klein, S.A., 1996. Absorption Chillers and Heat Pumps. CRC Press. Hu, J., Chao, C.Y., 2008. Study of a micro absorption heat pump system. Int. J. Refrigeration 31 (7), 1198–1206. Kang, Y.T., Kunugi, Y., Kashiwagi, T., 2000. Review of advanced absorption cycles: performance improvement and temperature lift enhancement. Int. J. Refrigeration 23 (5), 388– 401. doi:10.1016/S0140-7007(99)00064-X. Keinath, C.M., 2015. Direct-Fired Heat Pump for Multi-Pass Water Heating Using Microchannel Heat and Mass Exchangers. Mechanical Engineering. Atlanta, GA, Georgia Insitute of Technology, Vol. PhD. Keinath, C.M., Hoysall, D., Delahanty, J.C., Determan, M.D., Garimella, S., 2015. Experimental assessment of a compact branched tray generator for ammonia–water desorption. Sci. Technol. Built Environ. 21 (3), 348–356. doi:10.1080/ 23744731.2014.1000797. Klein, S.A., 2014. Engineering Equation Solver, F-Chart Software. Lee, S., Bohra, L.K., Garimella, S., Nagavarapu, A.K., 2012. Measurement of absorption rates in horizontal-tube fallingfilm ammonia-water absorbers. Int. J. Refrigeration 35 (3), 613– 632. doi:10.1016/j.ijrefrig.2011.08.011. Lorentzen, G., 1995. The use of natural refrigerants: a complete solution to the Cfc/Hcfc predicament. Int. J. Refrigeration 18 (3), 190–197. doi:10.1016/0140-7007(94)00001-E. Meacham, J.M., Garimella, S., 2004. Ammonia-water absorption heat and mass transfer in microchannel absorbers with visual confirmation. ASHRAE Trans. 110 (1), 525–532. Mendes, L., Collares-Pereira, M., 1999. A solar assisted and air cooled absorption machine to provide small power heating and cooling. In: International Sorption Heat Pump Conference, Munich, Germany. pp. 129–136. Pence, D., 2010. The simplicity of fractal-like flow networks for effective heat and mass transport. Exp. Therm. Fluid Sci. 34 (4), 474–486. Sözen, A., Altıparmak, D., Usta, H., 2002. Development and testing of a prototype of absorption heat pump system operated by solar energy. Appl. Therm. Eng. 22 (16), 1847–1859. doi:10.1016/ S1359-4311(02)00109-6. Srikhirin, P., Aphornratana, S., Chungpaibulpatana, S., 2001. A review of absorption refrigeration technologies. Renew. Sustain. Energy Rev. 5 (4), 343–372. doi:10.1016/S13640321(01)00003-X. Treffinger, P., 1996. Development and performance of a single stage high efficient absorption heat pump for domestic heatings. In: Proc. Int. Absorption Heat Pump Conference, Montreal. Ziegler, F., 1999. Recent developments and future prospects of sorption heat pump systems. Int. J. Therm. Sci. 38 (3), 191–208. doi:10.1016/S1290-0729(99)80083-0. Ziegler, F., Riesch, P., 1993. Absorption cycles. A review with regard to energetic efficiency. Heat Recov. Syst. CHP 13 (2), 147–159. doi:10.1016/0890-4332(93)90034-S.