|
|
|||||||||
|
||||||||||
|
As electronic devices continue to increase in power and decrease in size, the problem of heat management has become more complex. When designing electronic enclosures for harsh environments every aspect of cooling must be optimized. The advanced Product Development Group at Miltope Corporation has turned to Computational Fluid Dynamics (CFD) to help achieve proper cooling requirements in new product designs. Coolit is a Windows 95/NT based program that models fluid flow and heat transfer of electronic systems using CFD methods. Unlike many analysis programs, Coolit is specifically designed to model electronics and electronic enclosures. This allows the user to easily create system level models using various standard electronic components such as cabinets, printed circuit cards, fans, and heat sinks. Coolit is based on a graphical user interface (GUI) which allows the user to create models by first beginning with the enclosure, creating components and complex components, and then positioning the components within the enclosure. Characteristics, such as size, position, and properties, of the enclosure and components can be easily changed with the use of pull-down menus or by pop-up menus through the GUI. Once the model is complete, Coolit will automatically generate a grid. This grid can then be modified if required. Finally the model is solved and the computed results are graphically displayed. Graphical results include temperature, velocity vectors and total velocity, heat flux vectors and total heat flux, pressure, and turbulence intensity. Figure 1 - Model of Electronic Enclosure
The GUI provides three orthogonal projections and one isometric view of the model in either wire frame or solid perspective. All views can be zoomed in or out and the isometric view can be rotated to allow easy access to all components. The intuitive menu makes it easy to step through the creation of a model from the enclosure, to adding components, solving the problem, and displaying the results. Coolit was used to model the electronic enclosure shown below in figure 1. This was a relatively complex model, which included an enclosure with 1 fan, 2 vents, and 15 various components. Also include were 2 volume resistance models used to simulate heat sinks. The grid was generated and modified to produce approximately 115,000 nodes. This model was solved with an algebraic turbulence model for convection and conduction through the solid components. The solution was obtained in 261 iterations and about 5 hours on a 233 MHz Pentium PC with 96 Meg of memory. The model was first created with the size and placement of the cooling vents based on intuitive thinking, which was to use vents as large as possible. It was determined by the CFD model and later verified with lab testing that intuitive thinking can be very distant from the optimal design. By using the graphical display to view plots similar to that shown in figure 2, the initial results showed that one particular integrated circuit chip was reaching very high temperatures and would require a heat sink for normal operation. This was easily added to the model by first creating a volume resistance and then placing it on top of the integrated circuit. Figure 2 - Component Surface Temperature Distribution
More importantly, Coolit provided a graphical display of the flow field within the enclosure, showing the high velocity areas versus the stagnate areas. Velocity vectors and total velocity can be displayed on either of the XY, XZ, and YZ planes as shown in Figure 3. It was found that by reducing the size of the vents and placing them in the correct positions, the majority of airflow could be concentrated over the hot components. By using the pop-up menus, both the size and positions of the vents were varied until the optimal layout was determined. The final configuration was found more quickly and at a minimum cost when compared to iterative testing on physical models. Figure 3 - Temperature and Velocity Distribution
Coolit has other features that make it simple and easy to use. Changes to the model, which do not effect the grid generation, can be made and the solution found based on the previous results. This can drastically reduce computation times when testing several configurations. The software is also capable of handling any combination of predefined units or custom units when creating models. For example, length can be in inches, temperature in Kelvin, and power in watts. The results can also be displayed and re-displayed using different units or custom units. In addition, the program can solve problems while using your PC to do other work. Although computation times will increase by multitasking, Coolit will not completely tie-up your PC. Coolit comes with a user manual. It is fairly easy to read and includes tutorials to help you learn to create and solve models very quickly. An on-line help menu compliments the user manual. I was able to use the software to create models like the one shown above in about a week. Although I have gained some experience with CFD while in college, I did not attend a Coolit training course prior to using the software.
Mike Staiano is a Mechanical Engineer with Miltope Corporation. Miltope specializes in the ruggedization of computer equipment and peripherals for commercial, industrial, and military use. Coolit software allows Miltope to optimize the thermal design aspects of electronic enclosures without the extra time and cost of physical testing.
|
||||||||||
|
||||||||||
){kind=link}
){kind=link}
){kind=link}