Air Flow Study inside the Supermarket Refrigeration System Sandeep Palaksha Senior CAE Engineer Hussmann Services India LLP #9, Anand s Corner, 2 nd Floor Market Road, Netkallappa Circle Basavanagudi, Bangalore 560004 p.sandeep@hussmann.com ABSTRACT CFD study is performed to understand the Physics of Airflow inside a Hussmann refrigeration system and to check how the gap between coil and wall is impacting the air flow distribution on the condensing coil. CFD is beneficial in achieving accurate results and reduce cycle time of conventional Lab testing. This was done by using HyperMesh to build mathematical model and AcuSolve to simulate the Lab conditions. Hussmann Proto-Aire refrigeration system is an outdoor, air-cooled refrigeration system based on the Protocol (refrigeration system) concept. It uses multiplexed scroll compressors and is available in a variety of configurations that can be tailored to specific installations. This system also has a condenser with FAN which throws out the heat from condenser coil. These are located on the roof top or outside the buildings. While performing CFD on newly designed Hussmann Proto-Aire refrigeration system it is found that the air flow distribution over the condensing coils is not uniform. Then we started investigation using AcuSolve CFD solver. CFD study leads us to redesign the airflow path, for which a prototype was built and tested in Lab. The result from Test lab and Simulation from AcuSolve have been correlated very well. Introduction: Supermarket refrigeration system are a parallel system, often referred to as a distributed system, is a multiple compressor refrigeration unit piped in parallel to yield smooth capacity control as compared to a single compressor unit. A parallel system can be located in a back room or on a roof in close proximity to refrigeration equipment for reduced piping. Supermarkets are one of the most energy-intensive types of commercial buildings. Significant electrical energy is used to maintain chilled and frozen food in both product display cases and walk-in storage coolers. Supermarkets have a wide range of sizes. In North America, store sizes vary from roughly 2,000 to 11,000 square meters. A typical supermarket consumes roughly 2 million kwh annually, and roughly half is for refrigeration. Thus, improvement in energy efficiency of supermarket refrigeration will affect the store s bottom line of profit margin. The most commonly used refrigeration system for supermarkets today is the parallel rack direct expansion system using a HFC refrigerant such as R404A. Figure 1 shows a diagram of a typical parallel rack system. Multiple compressors operating at the same saturated suction temperature (SST) are mounted on a skid, or rack, and are piped with common suction and discharge refrigeration lines. Using multiple compressors in parallel provides a means of capacity control, since compressors can be turned on and off to meet refrigeration load. All display cases and cold 1
store rooms use direct expansion (DX) air-refrigerant evaporator coils that are connected to compressor racks in a remote machine room typically located in the back or on the roof of the store. Heat rejection is usually done with air-cooled condensers because these are least costly to install and maintain. A typical supermarket requires 1400 to 2300 kg of refrigerant A wide range of compressors is available, including digital scrolls, to better match capacity and budget needs while providing the most energy efficient and environmentally friendly system. Parallel units, matched to an appropriate condenser, may be modular allowing you to add or change compressors as needed. The standard rack housing and frame is made of durable galvanized steel with other materials as an option. Figure 1: Parallel rack refrigeration system Advantages of Parallel Refrigeration system: a. Energy Savings: Parallel units can match refrigeration capacity to actual load. This amounts to a 20%+ savings over a single compressor unit. b. Optional subcooling increases refrigerant efficiency on low temp applications by approximately 17%. c. Optional heat reclaim increases energy efficiency by reclaiming waste heat from the condenser. Optional hot gas defrost increases energy efficiency by using waste heat for defrost. These are very huge refrigeration system which can ran 100s of refrigerated display cases. The capacity of the refrigerant is fixed, if customer wants to add few more refrigerated display cases then they cannot increase capacity of the refrigeration system. To solve this problem PROTOAIRE was evolved. PROTOAIRE is a complete refrigeration system which contains compressor along with condensing unit. PROTOAIRE design can be modified depend upon the required capacity. PROTOAIRE can also be used in a small stores and gas stations. 2
Objective of the project: To understand the air flow behavior inside the PROTOAIRE unit and its impact on the flow distribution over the condenser coil using CFD. Validate the CFD prediction with test data in terms of air flow distribution on the condenser coil. Propose better designs with respect to air flow management to attain uniform flow distribution over the condenser coil Process Methodology: As the PROTOAIRE unit is symmetrical, one fan section was considered for the CFD simulation. Figure 2: Four fans PROTOAIRE CAD model Figure 3: one fan section considered for CFD Fig 1 shows the preliminary design of the PROTOAIRE. Fig 2 shows one fan section considered for performing CFD simulation. This model has coil, fan motor assembly and air flow path. Before starting this project it s necessary to confirm whether given fan and condenser coil geometry is correct or not. To confirm that, numerical investigation on fan and condenser coil was performed and compared with test lab results. If the numerical results are promising and matches with manual test lab results then only these components can used for performing this project. Numerical investigation on given Fan: Given Fan CAD model was placed in the virtual wind tunnel and performed simulation to check numerical fan performance. 3
Figure 4: Fan inside wind tunnel Figure 5: Measured mass flow at outlet For the given rpm mass flow rate predicted from CFD = 2.8 m 3 /s. For the same rpm volume flow rate mesured from test lab = 4958 CFM. Simple calculation to cross check Known volume flow rate = 4958 CFM (from test lab) = 2.34 m3 /s so Mass flow rate = 2.34 m3 /s X 1.225 kg/ m3 (density of air) Mass flow rate = 2.87 kg/s (from calculation) Hence this shows that the actual and simulation results are matching and confirms given fan geometry is right. Numerical investigation on given condenser coil: Condenser coil is placed inside the wind tunnel (Figure6) to check the pressure drop at given mass flow rate. Figure6: Condenser coil inside wind tunnel Figure 7: Condenser coil pressure drop curve Using the above curve fitting equation (Figure7) CDarcy and CForch values are calculated and used as inputs in porous media modeling. As fan is delivering 2.8 kg/s mass flow same value is used as input and performed CFD simulation. CFD predicted 98N/m 2 as pressure drop across the coil and the pressure drop measured in test lab is 0.394 in-w g, after converting, the test lab pressure drop is 98.141N/m 2. CFD and test lab results are very well correlated. From these investigations it proves that the given fan and condenser coil data are correct and this information can be used for future CFD simulations. 4
Grid development: Altair HyperMesh used for surface mesh and then that surface mesh was imported in AcuConsole to generate solid tet. The dimensions of computations space is shown in figure 8. To capture boundary layer phenomena accurately, 5 layers were generated on all surfaces with first layer height as 0.1 mm and growth rate as 1.3. Fine and irregular grid at extended air domain was used for accurate commutation. Total mesh size was 9.8 Mn. tetrahedral elements. Boundary condition: Figure8: Volume mesh Condenser coil is modeled as porous media and fan is rotating at given rpm. Reference frame is created for the given rpm and the reference frame is assigned to fan surface. Air entry side bottom wall is considered as wall because air cannot enter from bottom, other all four surfaces are considered as inlet so that the air can enter freely. Air exit top surface is considered outlet. 5
Numerical methodology: In this work, the Navier-Stokes equations were solved using AcuSolve, a commercially available flow solver based on the Galerkin/Least-Squares (GLS) finite element method. AcuSolve is a general purpose CFD flow solver that is used in a wide variety of applications and industries. The flow solver provides fast and accurate transient and steady state solutions for standard unstructured element topologies. AcuSolve ensures local conservation for individual elements. Equal-order nodal interpolation is used for all working variables, including pressure and turbulence equations. The resultant system of equations is solved as a fully coupled pressure/velocity matrix system using a preconditioned iterative linear solver. The iterative solver yields robustness and rapid convergence on large unstructured meshes even when high aspect ratio and badly distorted elements are present. The following forms of the Navier-Stokes equations were solved by AcuSolve to simulate the flow inside the PROTOIRE refrigeration system: Due to low mach number involved in these simulation, the flow was assumed to be incompressible, and the density time derivative in Eq. (1) was set to zero. The three dimensional steady flow is simulated using RANS single equation Spalart-Allmaras turbulence model. The turbulence equation is solved using GLS formulation. The model equation is as follows: 6
For the steady state solutions presented in this work, a first order time integration approach with infinite time step size was used to iterate the solution to convergence. Steady state convergence was typically reached within 500 time steps. Results & Discussions: Base line model: Velocity distribution on inside surface of the condenser coil is measured and monitored. Proper air flow distribution across the condenser coli will help to remove or eject all the heat from the condenser coil. After the CFD simulation it s observed that there is no proper velocity distribution across the coil area. In Figure9 it s observed that more velocity at top due to no restriction so more air is entering from the top, in the same time less velocity is observed in the bottom area this is due to restriction or obstruction. In this scenario condenser coil will heat up in the bottom which will leads to shutting down the unit. In the base model fan is delivering 1.55kg/sec mass flow rate which is almost 50% less than the actual, this is due to static offered by the design shape. New design iterations: Figure9: Velocity vector plot in m/s To distribute uniform air flow across the coil 3 design iterations performed by increasing the area of restriction, so that more air can enter from the bottom to result in uniform distribution. Figure10 Figure11 Figure12 7
Iteration 1 (Figure10):-Added 8 volume in the cabinet to enhance the air flow distribution over the coil. Iteration 2 (Figure11):-Added 12 volume in the cabinet to enhance air flow distribution over the coil. Iteration 3 (Figure12):-Added full volume in the cabinet to enhance air flow distribution over the coil. All the 3 iterations were performed with the same boundary condition and the results are plotted below. Figure13: Velocity distribution across the condenser coil (Air side) From the above results it clearly indicates that iteration 3 velocity distribution is uniform across the coli inlet area. Conclusions: From the above CFD simulations it clearly shows that iteration 3 air flow is better than all the other design, and the flow distribution is almost uniform across the condenser coil. The result and conclusions obtained by the present simulation can be summarized as follows. New PROTOAIRE design is contributing some resistance so the air flow is not uniformly distributed across the condenser coil. New PROTOAIRE design with 8 added volume (iteration1) is contributing some resistance so the air flow is not uniformly distributed across the condenser coil. New PROTOAIRE design with 12 added volume (iteration2) is looks ok in terms of air distribution across the condenser coil compare to above 2 conditions. New PROTOAIRE design with full length added volume (iteration3) is looks good in terms of air distribution across the condenser coil. Benefits Summary: With the help of AcuSolve, Leading commercial finite element CFD code, we at HUSSMANN product development team quickly take a decision, whether we go for existing new design or suggested iteration3 8
design, without doing manual test which is expensive. And it was very useful for us to do parametric study by varying the restriction gap without spending much time as it was in physical test. Challenges: When we do parametric study in this project, every time when we modify restriction gap we should generate volume mesh again and again which is time consuming. It would be useful for Acusolve users, if automatic volume mesh updating option is there. ACKNOWLEDGEMENTS The authors would like to thank Mr. Kamleshwar Rajender, CFD Specialist, Altair Engineering India Pvt Ltd. and Mr. Balakrishna, Abhijith, Lead Engineer-Modeling&Simulation, Ingersoll Rand for their valuable support and contributions during this project. We also would like to thank Mr. Krishna, Mudumby, General Manager, HUSSMANN Services India LLP for allowing us to publish this paper. REFERENCES: 1. Refrigeration and Air Conditioning Technology by Bill Whitman (Author), Bill Johnson (Author), John Tomczyk (Author), Eugene Silberstein (Author), Whitman (Author), Larry Johnson (Author), Tomczyk (Author), Silberstein (Author) 2. Refrigeration System and Application by Ibrahim Dincer (Author), Mehmet Kanoglu (Author) 9