Every month that a project runs costs a car company money. In the ’80s a major car project could last 8 years, of which four was probably fully manned. These days the figure is 36 months but it could drop to 24. How?
One aspect of the drive to cut development time is in rapid prototyping. This has been going on for quite some time. In 2006 it was normal practice to mill quick models of car interior trim prior to tooling so as to see how the form looked in three real dimensions. Non-visible parts have also been made quickly so as to speed production. Even CAD modelling itself is a time-saver (or can be) which is now taken for granted.
The latest shift is in the simulation of dynamic aspects of the car. Wind-tunnelling is costly and time-consuming. Crash testing is violent, noisy and pricey. These are being made the subject of computer modelling (which assumes good models). Ansible Motion in Norwich provide driving simulators to allow designers and engineeers to experience how a car drives before even taking a mule out on the road. Sales of black and white sticky tape may be affected. Ford are among the first users of the systems, reports Automotive News. There are some interesting blog posts at Ansible’s website here.
Only the manufacturers and Ansible know how accurate their simulators are in recreating the sensory rush of driving a car, even at moderate speeds and in undemanding conditions. A question for maths modellers is whether or not the simulators can help find errors in the characteristics of the design or simply tell the designers what they think they know already. Put it this way – if the inputs are describing a set of characteristics a, b and c and the simulator reproduces that then it tells you that the simulator is giving one a correct output.
You can determine if you like that particular output. What must be tricky is assuring that the inputs a, b and c used on the simulator are translated fully to the real test car, assuming one is ever used. Think of it like musical score. The way it is played is dependent on the way the musician translates that notation into physical movements. I would be interested to know if Ansible can promise much more than to provide a representation of idealised driving dynamics.