ANALYSIS & OPINION
16 August 2012Tweet
The cloud will accelerate the use of simulation-driven design, says Scott Reese of Autodesk
When early versions of the games console Xbox overheated and then stopped functioning, Microsoft spent more than $1 million extending warranties to restore customer faith in the product. A sensible investment as the Xbox went on to become one of the best-selling consumer products of all time. However, not many other companies could afford to do the same. Most products made, marketed and sold with design faults become nothing more than expensive – or even dangerous – mistakes.
Computers have been used to carry out physics-based simulations to predict the performance of a product for more than 40 years. Yet during this time, simulation has generally been seen as complex, expensive and the work of experts. Only very large companies have been able to afford to hire full-time simulation specialists to develop methods to apply to their particular products.
There have been attempts to introduce simulation to non-experts, but CAD operators are not trained to calculate the engineering evaluations involved and the tools were only able to carry out simple linear analyses on single components.
Consequently, simulation experts rejected them as over-simplified and their view was backed by the fact that sometimes parts that had been verified by these tools failed the final validation tests. There were examples of success, but only for those organisations that were prepared to invest in training, standardise processes and generally work hard to force the technology to fit their needs.
Many manufacturers and product designers have, therefore, continued to use specialists and handle simulation in an inefficient serial design-prototype-test-redesign cycle. At the same time, the search has continued for a way to enable designers and engineers to perform these tasks as part of their day-to-day work.
Recently, several factors have come together to create a breakthrough. The integration of intelligent 3D modelling technology with multi-physics simulation makes simulation-driven design early in a product development process a reality. This can have an enormous impact on product cost, quality and development time. For example, by bringing finite element analysis (FEA) and mechanical event simulation (MES) in house as part of its Autodesk digital prototyping solution, snow vehicle manufacturer, the Tucker Sno-Cat Corporation slashed both the time and cost for design validation by between 50 and 60 per cent.
Primera Plastics, which produces custom plastic parts for international customers, including furniture manufacturers and automotive OEMs, has used injection moulding simulation software to produce more accurate quotes and optimise injection mould design prior to manufacturing. Using Autodesk Moldflow, engineers run a range of simulations to study the flow of melted plastic, evaluate various runner systems and gate configurations, determine the most efficient cooling systems and predict volumetric shrinkage. It can also use simulations to predict how long it will take for a plastic part to reach ejection temperature and how long it will take for a part to warp.
The cost of creating injection moulds often plays a major role in the total cost of manufacturing – in some cases up to 75 per cent. ‘If the mould isn’t designed correctly, it can wreak havoc on the entire production process,’ said project engineer Scott Leatherman. ‘A great deal of money is wasted if we have to go back to them after the mould is produced.’ The simulation software also helps find and fix potential defects such as short shots, air traps and weld lines well before manufacturing begins.
However, as product complexity continues to grow, more and more computing power is needed to perform multi-physics analyses and this can exceed the scope of the desktops and workstations used by most engineers and designers. Yet, fortunately, this is a shortfall that new cloud technology can fill, enabling users to maximise the use of simulation-driven design.
Engineers no longer have to spend time simplifying geometry prior to analysis, the infinite computing power available in the cloud means they can now perform complex multiple pre-processing or multiple simulation tasks in parallel. The software automatically acquires all the appropriate geometry and inputs, such as limits and parameters, and then transfers the information to the high-performance computing in the cloud. This performs the meshing, executes the necessary solvers, runs the appropriate post processor and then returns the results to the user who has been able to continue with other tasks on their own computer in the meantime.
In conclusion, cloud-based simulation is a game-changer as it can significantly lower the costs, making it available to smaller businesses unable to invest in local high-performance IT infrastructure.