2 August, 2017
Gridding is perhaps the key phase in any thermal simulation - both in terms of determining the accuracy of your simulation and the speed at which your solver will arrive at a result.
Gridding has come a long way since the early days of CFD and thermal simulation software. Back then computers simply did not have the processing power to deal with more than a few thousand grid cells. This meant that models required drastic simplification, particularly when it came to complex shapes or experimental product designs. Despite this, engineers often had to wait days for any results.
Of course, as processing power improved, the resolution of the model (number of grid cells) also increased – dramatically improving the reliability and accuracy of simulation results. However, the use of structured Cartesian grids, which provided stable and efficient solving could produce a lot of unnecessary grid cells. This meant that the processing requirements were still extremely high - without the latest hardware solving could still take days.
To address this issue, unstructured grids offer the ability to alter the resolution of the grid to match different levels of complexity across the design. While this can bring significant time savings when it comes to solving, the need to manually amend grid resolutions across a model in many simulation packages can absorb most of those savings.
What 6SigmaET has developed is a unique “multi-level unstructured staggered grid Solver” that enables thermal engineers to achieve genuine time savings in both the gridding and solving phases and significantly reduces the hardware requirements for thermal simulation. Starting with a high resolution grid over the whole model, a hierarchy of increasingly coarse Cartesian grids is constructed and the Solver selects from these only those cells that are truly necessary to map out the geometry of the design.
In the 6SigmaET solver the generation of the unstructured grid is automated. Using in-built intelligence, the 6SigmaET Solver is able to ensure that areas which do not call for highly detailed grids are treated at a suitable resolution, while smaller, more complex objects are provided with a far finer simulation grid. The grid can still be modified manually using powerful grid controls if required.
In this way, a design that previously demanded nearly one billion grid cells can be simplified to just six million cells, providing an accurate solution in a matter of minutes rather than hours or even days.
By providing such a substantial saving on both time and required processing power, our unstructured solver helps to maximise the thermal efficiency of designs, providing unprecedented performance and simulation delivery times.
By: Tom Gregory, Product Manager