Designing a new refrigerator cabinet based on optimization of the insulation material injection

There are many advantages of using polyurethane materials, including low cost, light weight, enhanced thermal and electrical insulation, and high impact strength and, therefore, it has many applications, as for example, producing various kinds of complex parts including automobile interior furnitures, household appliances and housing industry. Use of polyurethane foam is also spreading rapidly in refrigerators as insulation material. Reaction injection molding of polyurethane foam is a process in that two or more chemical components mix, chemically react and finally form a foam which flows into the mold cavity where the polymerization is initiated. Energy efficiency of modern refrigerators demands a very effective insulation and, hence, the foam is required to fill complex cavities in very short time scales. In order to increase cavity filling efficiency while shorten the injection molding process, it is necessary to understand the phenomena of mixing, reaction kinetics, bubble nucleation and growth and two-phase flow behaviour during mold filling. Lack of control on reactive injection molding parameters results in variations from load to load and even from part to part, leading to serious quality degradation and, thus, the whole process becomes labor-intesive at elevated material costs.International competition between companies is rapidly growing. Hence companies race in the energy efficiency field and invest large amount of money on R&D activities for having more efficient products. For refrigerators; energy efficiency is highly relative to its insulation performance. Rigid polyurethane foam is used as insulation material between the refrigerators cabinet walls. Good insulation performance depends on the homogeneity of foam in the cabinet. Another function of the foam is the mechanical support to thin cabinet walls. Short injection time and minimized chemical costs are desired in mass production. To provide these requirements air outlets orientation, numbers, diameters and progress of chemical reactions, the temperature of the injected mixture and cabinet temperature are important parameters. If we want to find these optimum parameters we have to understand the nature of the polyurethane injection process. Thus polyurethane injection process become more controllable and we can predict process parameters for new designs before the manufacturing. In the light of these informations the aims of the thesis obtained as; Decreasing the usage of the foaming materials for single cabinet volume, Decreasing the costs due to change of the initial conditions of the process, Decreasing the foam filling time of the cabinet, Increasing the insulation and structural strength performance of the foam with minimizing the air traps in it, Optimization positions, sizes and the numbers of air outlets to make the polyurethane injection procces more efficient, Understanding the nature of polyurethane flow and getting a solid scientific knowledge about it. From this point of view a project started by ARCELIK Inc. and ITU Fluids Group with the support of the Ministry of Science Industry and Technology Commissary. A refrigerator with four doors chosen for the computational fluid dynamics (CFD) simulations. Flow of the rigid polyurethane foam in the cabinet is three dimensional, reactive, multiphase and time dependent. Also solidification occurs due to chemical reactions. Chemical reactions effect the flow very strongly. Except the CFD simulations, a chemical kinetics study has been done without flow to obtain optimum initial reaction conditions locally such as initial mass fractions of the species and initial temperature. ANSYS Meshing and ICEM CFD were used for mesh generation. CATIA and SOLIDWORKS were used for CAD and geometry refinements. On the other hand a FORTRAN code were used for modeling the chemical kinetics. Chemistry code was developed in the ITU Fluids Group. CFD simulations were done with the FLUENT software. Chemistry code was implemented in FLUENT by using the user defined functions (UDF) and they run simultaneously. Post processing had been done with the TECPLOT and MATLAB. Furthermore, EXCEL and some FORTRAN codes which were written during the project, are used for converting the physical, chemical and thermodynamic data for calculations. The validation of the chemical kinetics with real process is done by comparison between an experimental data. Gelling and blowing reaction rates are adjusted with the calibration of the Arrhenius parameters. Thus a similar behavior obtained with real reactions. Density, viscosity and enthalpies are calculated with sixth degree polynomials. Except from the CFD calculations a numerical study has been done with the chemical kinetics only to obtain the optimum process conditions such as; initial temperature, initial mass fraction of the species. Three different cases were investigated; in the first case isocyanate is decreased and the water is increased; in the second case polyol is decreased and water is increased and in the third case physical blowing agent is increased and all other species mass fractions decreased on condition that their ratios remain constant one to another. Temperature is changed 20 ℃ to 45 ℃ with a step of ∆t = 5℃. According to these chemical kinetics are investigated at 11763 different initial conditions. As a result a number of 1411560 element solution space is obtained. This numerical study shows that considerable reduction can be achieved without changing the polyurethane foam properties. Cutcell meshing method was used for creating computational grid and finite volume method was used to discretization of the governing equations on the grid. Unsteady Reynolds Averaged Navier-Stokes (URANS) equations and twelve scalar equations for enthalpy and mass fractions of isocyanate, polyol, polyurethane, water liquid and vapor, carbamic acid, amine and urea are solved for a mesh size of 1.2 million cells and at a average time step of ∆t=0.01 seconds. Complex turbulent flow and chemical reactions limit the time step. As the flow is turbulent, a conventional two-equation k-ε turbulent model was used with wall functions and standart model constants. There is an inlet at the bottom and fifteen air outlets at the top of the refrigerator cabinet. The injected mixture temperature is 21 ℃ and the cabinet wall temperature is 40 ℃. Adiabatic boundary condition assumed for all cabinet walls. Hence no heat transfer occurs through the walls. The injection time period is 12.99 s and the inlet mass flow rate is 1.5 kg/s. Generally 64 processors and 2 Gb memory was used for calculations. The results of CFD shows very detailed aspect of the reacting, turbulent, multiphase solidification flow of the polyurethane foam in the refrigerator cabinet. According to the CFD results an asymmetric flow is observed in the cabinet. At the left side of the cabinet foam flow slower than right side and this caused by the thin walls of icemaker. Also it is observed that foam strike to the air outlets and this shows the importance of optimization of locations, sizes and number of air outlets. Finally some suggestions were made with the results of the CFD simulations and chemical kinetics calculations. The future work plans are designing a new refrigerator cabinet geometry in line with these suggestions and make CFD simulations with new design. Afterward, new and previous results will compare and the improvements will be observed. According to observations some experiments will be done with the new and old designs if it will necessary. Thanks to this, we will have extensive knowledge about polyurethane injection process and manufacture competitive products for global market.