Maybe my insight is not new.  But I thought it was interesting nonetheless.

I have been working on a project in the solar energy arena.  My company is developing a two axis tracking system that will improve the energy harvest of flat roof and residential solar installations.  Its going really well.  A real mechatronic challenge.  The magic is in creating an electromechanical solution that moves a single solar panel cost effectively.  Not an easy project.

We modeled the initial solution and found some limitations in the range of motion that would have negative effect on the energy harvest.  Its easy to get the solar panel to rotate 45 degrees to the east from a flat storage position.  But that isn’t always enough to maximize the amount of solar energy converted to electricity.  And when the elevation angle is combined with the rotation of the azimuth, there were significant limits to the motion, and therefore limits the amount of energy from the solar panels.  So we started working on an improved solution.

This, by the way, is the great thing about software simulation.  You can study something without making any hardware.

And this lead me to think about the nature of solid model and finite element analysis software.  I think there may be a mis-impression about the nature of these products.  The goal of solid modeling is not to eliminate the prototyping process, the goal is to make prototyping faster and less expensive.  Which goes hand in hand with the explosion in rapid prototyping technology.  More on that later.

Maybe its a poor analogy, but I can’t help recalling Thomas Edison’s search for the improved electric light.  And by the way, Edison didn’t patent the electric light.  He improved it after purchasing a Canadian patent from Dr. Henry Woodward and Matthew Evans.  The electric light was first demonstrated by Sir Humphry Davy in 1809, but it was a platinum filament powered by a huge array of batteries.  And the race was on for the balance of the 1800’s with people all over the world trying to come up with a solution that would make electric light practical.

Anyway, Edison is said to have conducted ten thousand experiments with the goal of creating a light source that would have sufficient life expectancy to be commercially viable.  Solid modeling would not have helped Edison since the work was primarily based on material science.    But modeling software has become so widespread and so powerful, that it is very widely used in product development.

The key thing to remeber is that software is a tool.  It has its uses, and it has its flaws.  We’ve found several flaws along the way.  But it allows us to work with the design and understand the system’s physical properties before we cut metal.  This saves huge amounts of time and money.  But it does not replace the fabrication step.  Even in a system as simple as the one I am working with, it is clearly necessary to build hardware and test performance.  There are too many areas that software models are not sufficiently detailed to be able to analyze reliably.

Nothing replaces hardware and testing.  Not yet.