14th December 2011 by Royston Jones
It seems a long time ago (showing my age – the early 80’s) when people at my university (Swansea) carried around their finite element programs in computer punch cards stacked so high you couldn’t see their faces and Professor Zienkiewicz, a pioneer of the finite element method, announced the arrival of a ‘real’ virtual reality where complex physical events would be simulated in real time.
As one of his disciples, released into industry in 1984, my first assignment was the impact simulation of a nuclear flask where I needed to design my mesh on graph paper because I could only afford 2,000 solid elements. Consequently, the ‘actual’ reality was still along way from achieving his vision.
Early Impact Simulation Containing Just 2,000 Elements
Throughout the 90’s simulation increased in industry but in a rapidly evolving design environment it was having limited impact. I remember asking a designer why had he placed a reinforcement rib in a certain location and being told “Why not”, in the absence of simulation input, he had a point because simulation wasn’t quick enough to give him the input. Depressingly, on occasions, I would stumble across simulation projects performed solely as a ‘tick in the box’ exercise to satisfy an OEM with the results never used At best simulation was a verification tool.
The start of the millennium witnessed the immergence of new computer architectures (clusters, CPU processors) and since simulation was always shackled to the power of the computer, we started to feel a sense of liberation. Importantly, new intelligent software technologies began to mature with the potential to inject design innovation, most notably the emergence of structural optimization technology.
My own epiphany occurred around 2000 where we used optimization technology to develop a process to automatically evolve the optimum geometry of a bracket required to absorb knee bolster impact. We worked with Jaguar to reduce a six month manual ‘trail and error’ process to two weeks. The newly defined process is still used today. Below are some pictures from an initial numerical experiment where the technology demonstrated it could automatically growth a simple single strut to an optimise crash performance geometry. The accompanying paper is also available: Optimisation of a Knee Bolster Loading for Euro & US NCAP.
In addition, we performed our first commercial aerospace study on a wing rib of the Airbus A380 and developed an optimum minimum mass / maximum strength design using optimization techniques which are routinely used today. Both these projects demonstrated how intelligent software technology could inject innovation into design (gold dust!). Again, the paper is available: Application of Topology, Sizing & Shape Optimization Methods to Optimal Design of Aircraft Components.
Free Form Optimization Used to Design Light Weight A380 Wing Ribs
By 2010, all OEMS acknowledge simulation as a key component of a modern product design process and are increasing their investments and ambition. Companies are thinking through how to introduce a more aggressive deployment of simulation technology and are developing execution models such as optimization centres (e.g. Airbus) or cells (e.g. Renault), in order to integrate the technology into established design processes to quickly deliver the business benefit.
And so to the future and the crystal ball. In our everyday lives, intelligent software and inexpensive computing has revolutionised our access to information and the immediacy with which it is communicated (e.g. smart phones, social media). This new era of digital technology will also transform the product development process. The rapid generation of engineering data and its ability to generate optimum geometry will be transformational. The general availability of digital data through the complete product development process will force a more collaborative way of working. The future of design will be more multidisciplinary which will require less attribute silos.
I hope that follow-on blog entries from my colleagues and friends in industry will give their thoughts on how they see simulation technology impacting the future product development process.
I feel we are closer to the Professor’s vision of the ‘real’ virtual reality.
Realistic Rendering of a Full Vehicle Crash Simulation