Thermal Efficiency Enhancement of an A/C System Using Computer Simulations

Effects of airflow maldistribution in a prototype V-shaped evaporator (Model CAD501) for an air-conditioning (A/C) system were investigated using computer simulations in this study. The configuration of the fin-and-tube heat exchanger was first simplified into a 30-channel air flow passages, and velocity profiles of air flowing through the simplified heat exchanger passages were obtained using Computational Fluid Dynamics (CFD) software ABAQUS. The performance of the A/C system was then evaluated using thermodynamics software Engineering Equation Solver (EES) based on the velocity distribution provided by the CFD simulation. Methodologies are proposed to solve the problems associated with the airflow maldistribution. Preliminary experiments were conducted to qualitatively verify some of the computer simulations and results. It was found that the airflow maldistribution would reduce the Coefficient of Performance (COP) and cooling capacity of the A/C system and this reduction could be recovered by the proposed solutions.

Introduction The air conditioning system (A/C system) is crucial to modern life. It is widely used in both domestic and industrial worlds. As the old saying goes, every coin has two sides; while the A/C system brings us comfort, usage is associated with problems of energy consumption and environmental pollution. Although it is not feasible to eliminate the consumption of energy, an enhancement of energy efficiency is viable.

World Energy Problem
The Statistical Review of World Energy 2015 Report produced by British Petroleum (BP plc.) gives us a good overview of the current energy problem in the world. The review of 2014 in the report concludes, "Global primary energy consumption decelerated sharply in 2014, even though global economic growth was similar to 2013." While this seems to be good news, it actually is not. On the one hand, consumption increased for all fuels, reaching record levels for every fuel type except nuclear power. On the other hand, production increased for all fuels except coal. For oil and natural gas, global consumption growth was weaker than production. Therefore, the deceleration of primary energy consumptions means we are consuming more energy and doing it faster than ever. The emerging economies continued to dominate the growth in global energy consumption. China (+2.6%) and India (+7.1%) recorded the largest national increments to global energy consumption. A second consecutive year of robust US growth (+1.2%) was also reported.
Energy consumption of A/C systems plays a big role in this worldwide problem. Industrially, with the rapid development of cloud computing technology, the energy consumption of the global data centers is getting more critical. A study (Jonathan et al. 2011 [1]) shows that electricity used in global data centers in 2010 likely accounted for between 1.1% and 1.5% of total electricity usage. Specifically for the US, that number was between 1.7% and 2.2%. The A/C system accounts for around 37% of the data center energy consumption. As you can see, the data centers consist of just one small piece of this puzzle affecting the whole industry that uses A/C systems. Domestically, air conditioning is also very common in the United States. A survey conducted by the US Census Bureau says that in the United States, 88% of new single-family homes constructed in 2011 included air conditioning, ranging from 99% in the South to 62% in the West.

Air-conditioning (A/C) System
In general, air conditioning (A/C) is used in order to control the temperature of a room or space to be more comfortable for those residing in it, particularly in a sweltering environment. The drop in temperature is achieved through a refrigeration cycle, wherein four separate processes combine in order to draw heat out of the occupied space and displace it to an outside area. Figure 1 shows the T-s (temperature-entropy) diagram of an ideal vaporcompression refrigeration cycle. The most commonly used cycle for A/C system, the vapor-compression refrigeration cycle, consists of four processes: 1-2 Isentropic (constant entropy) compression in a compressor; 2-3 Isobaric (constant pressure) heat rejection in a condenser; 3-4 Irreversible isenthalpic (constant enthalpy) throttling in an expansion valve;  heat absorption in an evaporator.
As shown in Figure 2, there are four main components comprising a typical vapor-compression refrigeration cycle: compressor, condenser, expansion valve and evaporator. Typically, the evaporator is placed in an indoor area to be refrigerated  (Ahamed et al. 2011[3]) and the condenser is placed in an outdoor space so that the refrigeration cycle takes away heat from the indoor area and rejects the heat to the outdoor space. The compressor is the only component which does work on the A/C system. In an ideal vapor-compression cycle, the refrigerant enters the compressor at state 1 and leaves at state 2. Refrigerant at the saturated-vapor-state is compressed isentropically from a low evaporator pressure to a relatively high condenser pressure. After the compression, the temperature of the refrigerant increases from the low evaporating temperature to a much higher temperature than the surrounding space. Then the refrigerant enters the condenser at state 2 and exits at state 3. The refrigerant condenses from a saturated vapor to a saturated liquid by releasing heat to the surroundings. After that, the refrigerant is throttled through the expansion valve from state 3 to state 4. The irreversible process takes its temperature down from above the outdoor temperature to below the indoor temperature. After all, the refrigerant enters the evaporator at state 4 and exits at state 1. The refrigerant completely evaporates from a low-quality saturated mixture to a saturated vapor by absorbing heat from the refrigerated indoor space. The cycle is completed as the refrigerant reenters the compressor from the evaporator. The heat transfer for internally reversible processes is expressed as the area under the process curve on the T-s diagram. Consequently, the area under the process curve 4-1 represents the heat absorbed by the refrigerant in the evaporator and the area under the process curve 2-3 represents the heat rejected in the condenser. It is shown that the evaporating and condensing processes appear as two horizontal lines (constant pressure lines) and the heat transfer in the condenser and  [2]) the evaporator is proportional to the lengths of the corresponding process curves 2-3 and 4-1. The throttling process appears as a vertical line on the diagram. The ideal vapor-compression refrigeration cycle is an internally irreversible cycle due to the involvement of the irreversible throttling process. The vapor-compression refrigeration can be analyzed as steady-flow processes since all four components constructing the refrigeration cycles are steady-flow devices. The kinetic and potential energy changes of the refrigerant are negligible since they are relatively small to the work and potential energy changes of the refrigerant. Then the steady flow energy equation on a unit-mass basis reduces to Equation (1), where q in is input heat transfer per unit mass; q out is output heat transfer per unit mass; w in is work input; w out is work output; h e is the specific enthalpy output; h i is the specific enthalpy input.
No work is involved in the condenser and the evaporator since only heat transfers are conducted in them. In addition, the compressor can be approximated as adiabatic. Then the Coefficient of performance (COP) of refrigerators operating on the vapor-compression refrigeration cycle can be expressed as Equation (2), where q L is heat transfer per unit mass with low-temperature body; win in is work input; h 1 is the enthalpy at state 1; h 2 is the enthalpy at state 2; h 4 is the enthalpy at state 4. Due to the irreversibility that occurs in various components of the cycle, an actual vapor-refrigeration cycle differs from the ideal one in several ways. The irreversibility is normally triggered by two factors: the fluid friction which causes pressure drops and heat transfer to or from the surroundings. Figure 4 shows the T-s (Temperature-entropy) diagram of an actual vapor-compression refrigeration cycle.
In the ideal cycle, the state 1 of the refrigerant leaves the evaporator and enters the compressor as saturated vapor. In practical use, however, it is a challenge to precisely control this state of the refrigerant. To simplify the problem, an overdesign is used in order to slightly superheat the refrigerant to prevent damage to the compressor. The design of superheating makes sure that the refrigerant is completely evaporated before it enters the compressor. In practical use, there could be a considerable pressure drop caused by fluid friction and heat transfer from the surroundings to the refrigerant due to the fact that the pipeline, which connects the evaporator to the compressor, is normally very long. The specific volume of the refrigerant is increased by the combination of superheating, heat gained in the connecting pipeline, and pressure drops in the evaporator and the connecting line. As a result, the total cooling capacity of the A/C system decreases since the mass flow rate of refrigerant is inversely proportional to the specific volume.
In the ideal cycle, the compression process is isentropic. However, the entropy of the refrigerant may be increased or decreased due to the involvement of frictional losses which increase the entropy, and heat transfer, which may increase or decrease the entropy depending on the direction. On the T-s diagram, process 1-2 shows the increase and process 1-2' shows the decrease of the entropy. In the case that the entropy is decreased, the compression process 1-2' is more desirable than the ideal isentropic compression process due to the reduction of the specific volume of the refrigerant which cause a decrease in work input requirements. In the ideal vapor-compression refrigeration cycle, the state 3 where the refrigerant leaves the condenser is assumed to be saturated liquid at the compressor exit pressure. In practical use, the pressure drops in the condenser and the long pipelines connecting the condenser to the compressor and to the expansion valve are inevitable. Similar to the case of the evaporator, the desired saturated liquid status of the refrigerant is difficult to ensure at the end of the condensation process. The sub-cooling design has been applied to the condenser to make sure that the refrigerant is completely Figure 5. A Fin-and-tube Heat Exchanger condensed before it enters the expansion valve.

Evaporator
The evaporator is the focus of this study. It is a key component in an A/C system. The fin-and-tube heat exchanger structure is used in this evaporator design.

Fin-and-tube Heat Exchanger
Fin-and-tube heat exchangers consist of mechanically expanded round tubes in a block of parallel continuous fins. An example is shown in Figure 5. A schematic diagram of a simplified heat exchanger is shown in Figure 6. Fin-and-tube heat exchangers are designed to maximize heat transfer between two fluids, in this study the air and the refrigerant, with a minimum pressure drop associated with each fluid. The air flows through the fins and the refrigerant flows through the tubes, completing the heat transfer between the two fluids. The design of fin-and-tube heat exchangers requires specification of more than a dozen parameters, including but not limited to the following: transverse tube spacing, longitudinal tube spacing, tube diameter, number of tube rows, fin spacing, fin thickness, and fin type etc.
The foci of this study are the problems associated with air maldistribution, thus Figure 6. Schematic Diagram of a Simplified Fin-and-tube Heat Exchanger (Liu et al. 2004[4]) the design and configuration of the heat exchanger are simplified. Under the ideal conditions, mass flow rates at the refrigerant side and the airside in the evaporator are uniform and match well at every channel, and the A/C system works as designed. However, in practical use the performance of the system Figure 7. Three Channel Design of the Evaporator (Groll et al. 2011[5])

Multi-Channel Design
degrades due to problems of airside maldistribution, which is the focus of this study.

Literature Review
This chapter briefly reviews some of the researches published in the literature and intends to show an overview of the work done prior to this study. Lee et al. 1997 [1]  Payne and Domanski et al. 2002 [2] investigated the potential benefits of controlling refrigerant distribution to achieve equal exit superheats. In the experimental part of their study, they found that if the overall superheat was held constant while the individual channel superheat was allowed to vary, then the cooling capacity decreased by up to 41% and 32% for the wavy and wavy lanced fin evaporator, respectively. For non-uniform air flow tests, it was found that the capacity was recovered to within 2% of the original value if superheats were controlled to the original value. In parallel to the experiments, they used their evaporator model EVAP5 to predict possible savings in evaporator core volume if a smart distributer was used. They found that in extreme cases, the savings in core volume can be up to 40%.
In a numerical study by Lee et al. 2003 [3], non-uniform airflow profiles reduced the cooling capacity of the evaporator up to 6%. Domanski et al. 2004 [4] proposed a computational model which used an evolutionary algorithm to optimize evaporator circuitry for uniform and non-uniform airflow, termed as ISHED (intelligent system for heat exchanger design). It was found that for a 3 bank evaporator with 12 tubes each, the capacity of the design for uniform airflow operated at uniform airflow was slightly smaller than the one obtained for non-uniform airflow when operating in non-uniform airflow (both approximately 5.25 kW). The capacity of the design obtained for uniform airflow under non-uniform airflow dropped to 4.82 kW. Domanski and Yashar et al. 2007 [5] used a newer version of ISHED to optimize evaporator and condenser circuitries for different refrigerants. For the evaporator designs, it was found that the ISHED generated design outperformed the best manually generated design for all refrigerants. For the condenser designs, it was found that the ISHED generated design performed equally well or better than the best manually generated design. They furthermore found, that for an Atype evaporator coil, as found in residential AC systems, the capacity improved by 4.2%, if the ISHED optimized circuitry was compared to the original circuitry. For this simulation, the velocity profile was based on particle velocity measurements and CFD simulation. TJoen et al. 2006 [6] investigated the effects of airflow maldistribution on the performance of a 3 row air to water finned tube heat exchanger. Using different velocity profiles, he found that the overall heat transfer coefficient dropped by up to 8%, if a quadratic velocity profile was applied. AbdelAziz et al. 2008 [7], pointed out that the airflow might also create a recirculation zone in the lower part of the coil. They carried out simulations of the airflow through an A-shaped evaporator using CFD (computational fluid dynamics) simulations. These recirculation zones in the coil led to a reduction in the cooling capacity since the recirculated airflow was not exchanged. Kim et al. 2008 [8] introduced the hybrid control scheme for evaporators, sug-gested possible balancing valves and evaluated possible performance benefits using a numerical model. Kim et al. 2008 [9] describes the modeling approach and results in a more concise form. balancing method, they found that most of the performance degradation caused by uneven airside flow was recovered when individual circuit superheats were balanced. Specifically, for an air distribution factor of 0.6, capacity degraded by about 6% and COP by about 4% without balancing superheats. If the upstream flow balancing method was applied, most of the losses were recovered.  [10] explain the model used in Kim et al. 2008 [13] in greater detail. Kim et al. 2009 [11] investigates additional cases of maldistribution for the same model as previously described, considering R410A, R134A and R22 as refrigerants. For an air flow maldistribution factor of 0.6, corresponding to a flow distribution factor V C1 /V C2 of 0.4, they found a cooling capacity degradation of 16% and a COP degradation of 11%. They considered combinations of airside and refrigerant maldistribution. They found that for all considered cases, most of the capacity and COP degradation was recovered, if the upstream flow balancing method is applied. Brix et al. 2009 [12] studied maldistribution in an R134a mini-channel evaporator for an automotive air-conditioning system. Both inlet vapor quality and airflow non-uniformities were investigated numerically with simplified two-channel geometry. When only liquid entered into channel 2 and the remaining mixture entered channel 1, the cooling capacity was reduced by 23%. When the air velocity across channels 1 and 2 were 2.24 m/s and 0.96 m/s, the cooling capacity decreased by 19%. Kaern et al. 2009 [13] described a simulation model for air conditioning systems, which uses a two-pass evaporator to study the effects of maldistribution.
They considered airside flow velocity distribution using a factor F air = V f r,1 /V m = 0.1 and found the COP degradation was 38%. If the individual circuit superheats were controlled with applied airside distribution factor of 0.1, the COP degradation decreased from 38% to 7%. Their overall conclusion was that the conductance and COP reduction become significant when full evaporation is not reached in one of the circuits.
To compensate for maldistribution, a new method was evaluated in Mader et al. 2010 [14] with respect to cooling capacity and the coefficient of performance (COP). This method involved a coupled expansion and distributor device that was able to control the individual channel superheat by measuring only the overall superheat. Kaern et al. 2011 [15] found that non-uniform airflow distribution leads to a COP reduction of up to 43.2% and quality maldistribution in the distributor leads to a COP reduction of up to 13%. Kaern et al. 2011 [21] confirmed their previous result that most of the capacity and COP degradation, for a wide range of airside maldistribution, can be recovered if the individual channel exit superheat is uniform.
In conclusion, based on the basic literature search, there has been a considerable amount of research done in the past to greatly advance our knowledge on the problems of air flow maldistribution in the A/C system. The focus of this work is based on the particular problem of airflow maldistribution within the V-shaped evaporator designs and on the solutions to remedy this problem.
[10] J.-H. Kim  The focus of the study is on the system with V-shaped evaporators.
Since the refrigerating process is dynamic and interactive, many factors work simultaneously in practical use. It is difficult to isolate the individual factor that influences system performance during refrigerant maldistribution using experimental methods. With computational simulations, we can theoretically isolate these individual factors and focus on the dominant factors that affect system performance.
The computational simulations consist of three parts : 1)

CAD Modeling of the Evaporator
The evaporator is the key module in the A/C system which provides the cooling capacity by evaporating of refrigerant at a low predetermined pressure. It is the very same component in which the problems of refrigerant maldistribution induced by airflow maldistribution occurs. As discussed in the introduction section, the V-shaped evaporator configuration provides the maximum heat transfer surface area within a given space. However, this configuration causes the uneven distribution of airflow.
Since the subject of CFD simulations is the air flow patterns over the evaporator, the sketch to be generated from the evaporator design is the air space design are dense, small, detailed and direct computer modeling. Therefore, they require a considerable amount of computer memory, CPU time and software resources, which would cause problems in our CFD simulations. Therefore, the complicated 3-D air flow passage within the heat exchanger made of numerous finned coils is simplified to 30 2-D parallel placed straight passages. Figure 10 shows the 2-D sketch of the evaporator air space generated from the original design. Each block inside the frame represents the simplified and integrated groups of evaporator coils with fins.

CFD Simulations on Airflow through the Evaporator
In this study, the steady state CFD simulations are conducted using CFD software ABAQUS which allows users to define all the dimensions, properties and boundary conditions in the visual interfaces. Using ABAQUS, the geometry is meshed into pieces of elements, and the CFD simulations are run under two governing equations as follows.
Equation (3) is the conservation of mass equation, indicating that the reduction rate of the mass in the control body is equal to the rate of mass flowing out of the control body. Since the simulations are conducted in a steady state, the differential of time can be eliminated. Since the velocities of airflow inside the evaporator air space are relatively low, the Reynolds number in each air passages are small and the airflow can be considered as incompressible flow, thus we can get equation (5).
Substitute Equation (5) into Equation (3), then we can get the conservation of mass equation demonstrated by Equation (6).
Equation (4) is the Navier-Stokes Equation which was developed from the conservation of momentum equation, indicating that the flowing of fluids is determined by mass force, viscous shearing stress and pressure together. In this study the air is a Newtonian fluid so that the shearing stress can be expressed as Equation (7).
To solve a specific problem in fluid mechanics, several initial and boundary conditions must be defined. These conditions could be the velocity or pressure. Boundary conditions define the parameters at the boundaries of the flow field. In this study, the CFD simulations are run on the air space of the working evaporator.
The energy equation is not solved in the simulation.
In ABAQUS, the air space was meshed into small pieces. It solves the pressure field and velocity distribution using these two governing equations, and the initial and boundary conditions by using numerical iteration from piece to piece.
Results of air pressure and velocity distribution over the heat exchanger are displayed using pressure contours and velocity vectors, respectively using ABAQUS software. Figure 11 shows the meshing of the evaporator air space. The mesh size of the Figure 11. Meshing of the Evaporator Air Space elements inside the air passages is about 10 times smaller than the size of the ones outside to ensure the accuracy of the velocity iterations. Grid tests were conducted to ensure that the velocity profiles obtained from the simulation are independent of the grid size.

EES (Engineering Equation Solver) is a general equation-solving program that
can numerically solve nonlinear algebraic and differential equations. EES has a high accuracy thermodynamic and transport property database for hundreds of substances in a manner that allows it to be used with equation-solving capability.
The detailed introduction on EES software can be seen in Klein et al. 2015 [2]. the thermodynamic properties of the refrigerant are typically described by the following commonly used thermodynamics parameters: temperature T, pressure P, entropy s, enthalpy h, specific volume v and saturated liquid-vapor mixture quality x. According to the state principle, in a simple compressible system such as the A/C system studied in this work, two independent values will determine the values of the rest. Some key temperature values are determined by the operating condition of the A/C system, such as evaporating temperature T evap , condensing temperature T cond , low pressure vapor superheating temperature T suph and high pressure liquid subcooling temperature T subc . In an ideal thermodynamic A/C cycle, all the parameters can be obtained by combining the A/C operating temperature conditions and using state principles. The calculation of the performance of the A/C system includes total cooling capacity, the input work of the compressor and the coefficient of performance. These parameters can be defined as: The total cooling capacity QQ The input power of the compressoṙ And the coefficient of performance of the A/C system The mass flow rate,ṁ, of the A/C system is calculated as density (at the exit of the evaporator) multiplied by the volumetric flow rate of the compressor. The volumetric rate of the compressor is kept as a constant in EES modeling, while the density of refrigerant varies with the A/C operating conditions.
It is assumed in EES modeling that the maldistributed air flow and refrigerant flow mainly affect the performance of the evaporator; so that the refrigerant status in other components of the A/C system is not affected.

EES Modeling of the Evaporator
Since the rest of the components are assumed to work in ideal conditions in this study, the performance of the evaporator will determine the overall performance of the A/C system.
A typical evaporator of an A/C system consists of a large number of finned tubes. These tubes are usually evenly divided into many groups, connected one by one with 180 degree bends to form many back-and-forth flow channels. Figure 13 below shows a simplified evaporator model used by Kaern et al. 2009 [4]. In this model, the evaporator has only two flow channels, with each channel consisting of only one pipe. Liquid refrigerant passes through the distributor, gets evenly distributed into two flow channels, and evaporates inside the heat exchanger (releasing heat to the passing air stream). The vapor refrigerant then merges at the end of the evaporator (evaporator manifold) and the refrigerant is returned to the compressor. where V f is the fractional flow velocity and V m is the mean flow velocity. In their study, the refrigerant flow was limited to two channels. When F air is unity, the air-flow is indicated to distribute equally across the two tubes. When F air is zero, the air flows only across channel 1. The two-channel model represents an extreme case of air maldistribution. Figure 14 illustrates the concept of our strategies to optimize the performance of the A/C system. Panel A of Figure 14 shows the idealized heat transfer design that mass flow rates of air and refrigerant distribute evenly along the evaporator.
The rate of heat rejected from the air matches the rate of heat absorption by the evaporating refrigerant everywhere over the evaporator surface.
Panel B of Figure 14  We propose two types of basic strategies to solve the problems of air maldistribution. The first type is to improve the performance of the A/C system by matching the refrigerant distribution with the existing mal-distributed air flows.
The second type of strategy is to minimize the maldistribution of air flow from the source, before the air flow reaches the passage of the evaporator.  In this strategy, we will improve the air velocity profiles over the heat exchanger surface. This can also be achieved by several methods, such as by adding a flowresistant filter structure or by adding a flow guide device.
As stated in chapter 3.1, in this study, the complicated, many tubes of the whereṁ is the mass flow rate, C p is the specific heat capacity and ∆T is the mean temperature change of the air. The mass flow rate of the air can be expressed aṡ m = ρAv , where ρ is the density, A is a cross-section of the space the air flows through and v is the mean velocity of the air. The mass flow rate of air is directly proportional to velocity and the change of the temperature is small enough that it can be neglected. Therefore, for each coil, the heat loss of the air flows can be approximated as the linear function of mean flow velocity: where k and b are constant values to be determined by energy conservation principles. The determination of the constants k and b can be conducted by presuming that the state of the refrigerant exiting from the evaporator achieves the ideal sate under the mean velocity of the velocity profile. For each tube, the absorption of heat can be expressed as:Q whereṁ is the mass flow rate of the refrigerant in each tube, h 4 is the enthalpy at the entry of the coil and h 1i is the enthalpy at the exit of each channel. Without a control theme on the distributor, the mass flow rate distributed into each circuit is considered to be the same.
With the velocity profiles obtained from the CFD simulations, the equations can be solved and the state of the refrigerant at the exit of the evaporator, which is marked as state 1, can be determined.

Experimental Setup
The experiments of this work are conducted in the E-cooling Laboratory of the Nanjing Canatal Air-conditioning Electrical & Mechanical Co., Ltd.  Due to time limitations of this Master's thesis work, the objective of the experimental work is to qualitatively verify the feasibility of some proposed solutions that attempt to overcome the problems associated with the maldistribution of airflow and refrigerant flow.
The pressures of the A/C system at each state were measured using a Testco differential pressure meter. The temperature of the V-shaped evaporator was monitored and recorded using a Testco infrared camera.

Results and Discussions
As discussed earlier, the advantage of the V-shaped evaporator design greatly increases the number of coils in a relatively small cross sectional area. However, its side effects of air maldistribution and refrigerant maldistribution is also significant. In

Benchmark Simulation: Thermodynamics simulations on an ideal A/C system (without air maldistribution)
The coefficient of performance, defined as COP as a function of evaporating temperature, T evap , and condensing temperature T cond is shown in Figure 16.
In all of these simulations, the subcooling and superheating temperatures are  Figure 16 shows that the COP monotonically increases with the evaporating temperature T evap ; and is inversely related to the condensing temperature T cond . The cooling capacity (kW), Q L , as a function of the evaporating temperature, T evap , and condensing temperature T cond is shown in Figure 17.  Figure 17 shows that cooling capacity Q L increased as the evaporator heated up, and it decreased as condensing temperature T cond increased. Figure 17 shows that cooling capacity (kW), Q L has a similar trend as well as similar magnitude of increase as COP shown in Figure 16.
The coefficient of performance, COP, as a function of the superheating, T suph , is shown in Figure 18. And the cooling capacity, Q L , as a function of the superheating, T suph , is shown in Figure 19.
The COP and the cooling capacity Q L were both plotted three times as a function of T suph with condensing temperature T cond 35 • C, 45 • C and 55 • C, respectively. Since the room temperature is usually around 25 • C, the subcooling temperature is set to T subc = 5 • C to ensure that the temperature of refrigerant exiting the condenser is above the room temperature. Similarly, the temperature of cooled air flowing out of the evaporator is about 15 • C so that the evaporating temperature is usually set to T evap = 5 • C to make sure the evaporating temperature is below the temperature of cooled air. Figures 18 and 19 show that the COP and Q L were both nearly independent of superheating temperature T suph . It is noted that the total cooling capacity is calculated as mass flow rate multiplied enthalpy changes across the evaporator. Although the increase of superheating temperature T suph increases the refrigerant cooling per unit mass flow, the superheating also increased the specific volume of the refrigerant. As a result, the mass flow rate from the compressor decreased. In other words, the effect of vapor superheating is offset by the effect of density decrease. The superheating at the exit of the evaporator is controlled to ensure that the refrigerant is fully evaporated and no wet refrigerant can enter and damage the compressor. Simulations show that effects of vapor superheating to COP and cooling capacity Q L is negligible. Table 1 shows the settings of the temperature parameters in the simulations of case 2 to case 6.
The calculation results under these temperature settings are as follows: Cooling The ideal A/C system under the specific temperature conditions shown in Table 1 is marked as case 1. The temperature conditions are also applied to the other cases to be discussed in the following paragraph.

Original V-shaped Evaporator Design Based Simulations
In Section 4.1, we discussed the performance of an A/C system under ideal conditions without problems associated with air maldistribution. In this Section the effects of air maldistribution caused by the original V-shaped evaporator design are studied.     Figure 23 shows the air velocity profile extracted from the CFD simulation.

COP and Cooling Capacity of the A/C System
It clearly demonstrates the air maldistribution along the vertical direction. The velocity of air decreases monotonically as the distance further away from the entrance. The rate of flow slowing down is relatively quick for the first 10 air passages and then the decrease becomes gradual. This air velocity profile will be used as input for the EES simulation of an A/C system. As discussed in Chapter 3, the flowing of the refrigerant inside an A/C system is a dynamic and interactive process, and it is regulated by the expansion valve. In this study, the performance   Thus mass flow rate is reduced and consequently, cooling capacity is decreased.
For the condensing temperature T cond of 45 • C and targeted superheating temperature T suph of 10 • C, the reduction of the COP values is in the range of 6.98% at an evaporating temperature T evap of 5 • C to 7.21% with the evaporating temperature of 15 • C. Figure 25 shows the cooling capacity, Q L , of the ideal case 1, as compared with the results of case 2. Similar to Figure 24, the cooling capacity of the A/C systems has a decrease of about 3.45% to 3.91%. The designed A/C system under the certain temperature conditions shown in Table 1  Compared to case 1, COP drops 6.98% and cooling capacity drops 3.43% due to air maldistribution.

Effects of air velocity at the entrance
The inlet speed of air into the air space is controlled by the centrifugal fan   Figure 27 shows the velocity distribution inside the air space of the evaporator. Figure 28 shows the velocity distribution along the heat exchanger passages.
Compared to Figure 22 the speed of air through each channel increases and the air flow distribution parameter, F air , decreased from 1 to approximately 0.4 for each passage along the heat exchanger in Figure 28. In other words, the maldistribution of air becomes more severe by increasing the inlet speed.

Proposed solutions
The expansion valve will work incorrectly when the refrigerant in some coils is The second type of strategy is to minimize the maldistribution of air from the very beginning. The increase of resistance to air flow would affect the coils with higher air velocity more than those with lower air velocity. Therefore, the modifications in the air space of the evaporator aim to increase the resistance to airflow through each coil and improve performance, as 4.3.3 and 4.3.4 demonstrates.

Rearrangement of Evaporator Channels
The heat exchanger coils are rearranged into 30 new channels in this case.  Table 1

Installment of Filter Nets on the Heat Exchanger Surface
A practical approach to change air flow patterns along the evaporator surfaces is to install a layer of high flow resistance material, such as the filter nets over the surface of the heat exchangers. The higher the air velocity is, the greater the air flow resistance would be. Thus the filter structure creates smart and adaptive resistance to air flows, particularly at the entrance where air velocity is higher.
The designed A/C system with installment of filter nets on the heat exchangers is marked as case 5. Figure 34 presents the pressure field inside the air space of the evaporator. It is clearly shown that the pressure distribution became much more even than the original case design without the filter structure. Figure 35 shows the velocity distribution inside the air space of the evaporator with the installation of the filter structure. Figure 36 shows the velocity distribution along the heat exchanger air passage.
The air flow distribution parameter, F air , decreases from 1 to approximately 0.85 for each passage along the heat exchanger. The level of the maldistribution of air flow is greatly decreased.
It is noted that an air filter is a necessary and required component for an A/C system. Typical locations of the air filter are at the entrance. Case 5 shows that installing the filter at the surface of the evaporator is a practical and feasible approach to achieve two purposes with one design.

Installment of a Diverter in the Evaporator Air Space
Another design approach to even the airflow distribution is to install an air diverter in the entrance region of the air space. This design forces the incoming air to flow first through the pre-designed narrower paths of the diverter to achieve a more even distribution before it flows through the heat exchangers. The designed A/C system with installment of a diverter in the evaporator air space is marked as case 6. is clearly shown that the pressure drops across the coils are more even than the original case. Figure 38 shows the velocity distribution inside the air space of the evaporator.

Experimental Results
The temperature distribution of refrigerant can be measured by observing the infrared picture of the refrigerant inside the channel. The dark blue color   However, it is also noted that the temperature in the left heat exchanger is lower than the temperature in the right one, which indicates that the refrigerant is not uniformly distributed symmetrically. The result is possibly caused by the refrigerant maldistribution from the distributer and this phenomenon will be studied in the future.

Future Work
In general, our future work will include the following parts: