What is the value of simulations in science and engineering?

[denjay]

I'm looking at some Materials Science graduate programs and some of the research areas involve simulation of material growth and properties. I'm kind of wondering what the value of this kind of research is as opposed to more practical research. This probably goes into the philosophy of science but really what is the practical use of computational research in something like materials science?

 

[SteamKing]

It might be quicker and cheaper than the real thing.

That's why the structure of a building or an airplane, for instance, is modeled numerically before construction begins. It's a lot cheaper to fail a model than it is to fail an actual structure.

Until now, material properties, by and large, could only be determined by experiment. If you wanted to test the properties of a new alloy, first you would have to manufacture the alloy, make a sample, and then test it. Lots of time and money to do that. But if you could predict by calculation the properties of a new alloy, you might eliminate the need to make trials of different alloy formulations and concentrate your limited resources on the most promising alloy formulations.

Leave philosophy of science to the philosophers.

 

[timthereaper]

In terms of research, I'd say simulations help you understand the difference between how you've modeled the material behaving and how it actually behaves. Being able to compare results between theory and reality is the main benefit to modeling for research. As well, if the model is accurate to real-world tests, other programs can be written to help designers simulate how that material will behave for their products. Like SteamKing said, it's way cheaper to break a model than fab up a part and break it.

Just for the record, computational research *is* practical research.