25 October 2019
SIMULATION: TOO MUCH OF A GOOD THING?
Ryan Vann, Chief Engineer, CAE, Simulation and Modelling, Prodrive, believes that aerospace and military engineering projects can be accelerated by more considered use of simulation, using techniques used in the motorsport and automotive sectors.
The aerospace and defence industries are beginning to recognise an uncomfortable truth about simulation and predictive modelling; one that much of the automotive industry has already embraced. For any particular engineering application there is an optimum level of simulation accuracy. While high accuracy is required for certain types of research projects and academic proof, such as verification and validation of numerical methods too much emphasis on accuracy for commercial engineering programmes can be as ineffective as too little.
How can this be? Results which are insufficiently accurate are self-evidently of no value in steering engineering decisions, but how can you have too much? We need look no further than the law of diminishing returns.
There is a threshold of accuracy, below which the modelling results are worthless; they are too coarse to distinguish between the performance of a good and a bad design. Once this threshold is passed, useful output is obtained, which can inform the design process. However, increasing accuracy demands greater levels of input data and the preparation of more sophisticated models. The quality of the input data itself can potentially limit the accuracy of modelling, so this must also be compiled to the necessary standards. These activities slow down the process significantly and delay the availability of the results.
Unsurprisingly, the 80-20 rule applies; we can get 80% of the way in 20% of the time but the last 20% will take 80% of the time. Ultimately there is no escape; we must get the final product 100% right or risk the consequences. But during the very early stages of the design process, where important trade-offs are made and alternative solutions are ranked in order of merit, we do not need the last 20% in order to make the right decisions. In fact, we haven’t got the time to wait for the last 20%. We need to confirm the design direction before major commitments are made to long lead items, such as tooling and facilities. Clearly there is an optimum level of modelling accuracy for a given stage in an engineering programme; one that will deliver the right design decisions as early as possible.
It is time pressure that drove Prodrive to find smarter ways of applying simulation and predictive modelling. Initially in the motorsport sector, where no team ever had sufficient time or budget to achieve perfection, success came to those who applied their resources optimally within the time constraints of the racing calendar. These lessons can be applied in other sectors, such as automotive, aerospace and defence. In automotive, this is now enabling new model introduction schedules to be compressed at the same time as greater rigour is applied to development programmes, in order to supply a wider range of products while satisfying consumers with higher expectations of faultless operation. It is important to note that rigour can be maintained without the need for high analysis accuracy by applying appropriate engineering processes. Prodrive employs functional analysis to support earlier concept DFMEA generation, which enables a more robust virtual DVP.
With typical automotive programme timing reduced to as little as three years while defence and aerospace programmes still take 10-20 years, there seems to be great potential for closing the gap, at least partially. Prodrive has been using the principle of ‘optimum modelling accuracy’ since the late 1990s and has recently applied the technique to several military and defence projects.
One programme, to investigate and develop a chassis/suspension system for military vehicles which require a paradigm shift in capability to meet future generation platform targets, was completed in just 78 days. Engineering feasibility was determined, issues were identified and consideration was given to factors such as occupant comfort and hardware scalability. The level of robustness achieved in the results was enough to justify the allocation of additional funding to progress the concept further. It is precisely this kind of funding go:no go milestone that demands the optimum combination of accurate yet timely information.
Also completed in a short timeframe, especially when compared to traditional timings in the aerospace industry, was the fast-track simulation and subsequent manufacture of a transmission used for jet engine power take-off. Presented with the idea by the client, Prodrive was able to use its simulation experience to simplify the concept – making it less prone to breakage whilst reducing cost – and progress it in-house to the manufacture of a physical prototype. Feedback confirmed that the speed in which the task was completed more than compensated for lower accuracy enabling the project to be quickly escalated to prototype status.
Procurement is another area where time pressures demand a prompt decision, but enormous quantities of data must be evaluated with the necessary accuracy in order to identify the optimum selection. Evaluating the performance of candidate vehicles in a virtual environment under a range of appropriate scenarios would be unrealistic if attempted with very high accuracy. By choosing the optimum simulation methods and accuracy levels, useful results can be obtained quickly, allowing a greater number of options to be assessed, or a wider range of scenarios to be included.
We should never lose sight of the fact that excessively elaborate modelling, however reassuring it may feel, adds to the cost and timescales of a programme and can contribute little to making the right engineering decisions.
For more information on Prodrive's CAE and modelling capabilities click here