A Good Engineer Always Does the Math Before the Test!

Most of one's training as an engineer, is theoretical.  How do you predict how many times a coil spring can be compressed before it breaks?  At the university, theory is important, and it's also all one can afford since testing is expensive.  When you enter the working world, you have access to design guidelines and test equipment.  This can make an engineer lazy.  Worse yet, we invented Finite Element Analysis.  This is a very valuable, but very misleading tool.  Here is an example:

This a simple I-beam like holds up most buildings.   Red is high strain and blue is low.  When you calculate that beam in college, you get one peak stress.  Here, you have a picture of stress at every point on the beam!  It's amazing!  It's beautiful!  It must be correct!  Sadly, it is only as good as the input assumptions.  So, do you give up and just go to testing?

No, you use the potentially inaccurate FEA and hand calculations to improve your engineering judgement!  This only works if you work the loop of design-calculate-test.  Any step on its own is useless.

I have totally violated this in building BB.  Oh, I did enough math to make sure nothing would break, which has largely been correct. (side bar, my few regular readers will remember frequent tow truck rides.  These were not due to engineering errors, but to poor judgement of the situation.  In 6 of 10, I was sure my fuel pump was broken but I had actually run BB out of fuel.)

I am now working in reverse to correct this omission.  It turns out there is a very accurate racing simulator called Assetto Corsa.  Remarkably, it is shareware and developed by volunteers.  You can configure your own car and it goes into great detail; including the location of each suspension pivot point, spring and damper rates, sway bar stiffness, and tire properties.

So, I used the Ohio winter to recreate BB in Assetto Corsa.  Here is her home page: 

And here is some of the detail behind that.  It's around 20 text files overall for engine, chassis, and body variables.  That took some time with the tape measure and straightedge under the car but more time getting it all entered in the correct format.  For example, one enters the sway bar stiffness as Newtons/meter of travel as if the sway bar was linear and not rotary.  That took a few spreadsheet pages to calculate!

I haven't had much time for virtual racing but I did learn that I need to reduce my camber setting.  Tire temperatures also support this so the simulation and reality roughly coincide.  Hopefully that works out in real life!  In the meantime, here is how BB looks in Assetto Corsa:

Spring is here!  We will soon see how simulation helps reality!  Keep on having #funwithcars !