jump to navigation

I shake my little tush on the…computer July 20, 2010

Posted by mareserinitatis in electromagnetics, engineering, science.

When people ask me and some of my friends about our work, we will mention that we do a lot of modeling. Of course, the first thing they think of is something like this:

I guess it’s an understandable mistake. For most people, computers are used to surf the net, keep in touch with friends, and maybe balance your checkbook. Most people don’t understand what scientists or engineers do with computers. If they do know anything, they will say they’ve heard of climate modeling, and then inform you of the fact that they don’t think those computers know anything about how the climate works. (I have to admit that I’ve wanted to tell them how ignorant they sound as they obviously have no idea how modeling works, but I usually refrain.)

If you happen to have this conversation with a scientist or engineer, they usually have a better idea of what you’re doing. Sometimes.

One problem is that scientists and engineers who aren’t familiar with it will often confuse modeling with it’s cousin, imaging. This has actually been a pain for me because I’ve run into a few scientists who think that “modeling” is something that engineers do, and engineers think it’s only something that scientists do. The truth is that there are groups of computational scientists who do modeling, and other groups who do imaging. Likewise, some engineers do modeling, while some do imaging. The end goal for the scientist is usually to find something out about an existing system, while the engineer is often more interested in designing a system that works the way it’s supposed to.

What’s the difference?

Modeling examines how something works while imaging figures out what’s there, roughly speaking.

More specifically, someone who does modeling will look at a system or device. They will take the laws of physics and put them into a computational form. Then they will try to constrain the computational form to match the real system. Finally, they will examine how the computer system behaves and compare that to a real system with the same or similar constraints. They use the computer to “enforce” the laws of physics for a system and then hope to use the computer to see how complicated systems will behave within the constraints of those laws.

How do they know their answers are right? They usually begin with problems which have known answers, called “canonical problems”, to verify that they’ve modeled the physics correctly. In some cases there are several problems people will try before using their models on new problems. In other cases, there are “benchmarks”: in the absence of canonical problems, one can make their models (or simulations) do certain things to show that it’s working properly. Climate modeling, for instance, would have to reproduce many qualities of the historical climate record before one can assume that the model will produce an accurate estimate of what the the climate may be like in the future.

Imaging is basically signal processing. One takes data from a system and attempts to figure out how it has behaved or what’s there. It involves having a good understanding of statistics and how changing the system will result in a different signal or output. Imaging involves a lot of number crunching. Imaging works a bit differently. You don’t “validate” images the same way you do models. You have to have an understanding or knowledge (sometimes from previous studies) of how changes in a system will provide different signals. Sometimes people find things that are unexpected when imaging a system. In this case, these are often interesting results that are confirmed or denied through other studies or other methods.

These two areas, however, are usually closely intertwined from a research perspective. Usually modelers use the results provided by images to test their theories of how systems behave. A modeler will try to recreate a signal or result that matches what the images have observed.


Maybe that clarified things for you. If not, sometimes examples help, so I’m going to mention a few areas where I am familiar with modeling and imaging.

Solar physics: People who model the sun or other stars try to create models of the physics of the inside of the sun to look at effects of different forces or phenomena. For instance, they may look at how fluids move in the convective region of the sun. However, these results are constrained by information about the sun’s structure and behavior that has been provided by helioseismologists. Helioseismologists slug through tons of data to find how pulses move through the sun because changes in the way these pulses move will tell us things about what’s inside the sun and how it operates.

This is really not much different than how things operate for geophysics. Seismologists and minerologists provide information about what is inside the Earth, which provides information about how the Earth works. Several types of earth scientists (some seismologists or geodynamists, for example) will try to create models of the physical behavior and see if this behavior can explain how the Earth operates to create the structure we see. Or they will model things like earthquakes and see how they move through the Earth and interact with the structure.

The whole process is a ‘closed loop’: the data comes in; something is found: the modelers try to recreate this tidbit; once recreated, the modelers may see something different than expected; the imagers try to get more data or comb through the old data to find more evidence…and then back to the beginning.

Engineering: In engineering, modeling and imaging are usually more distantly related.

If you’re modeling in engineering, you often know what your system looks like and you want to see how it will behave. For instance, my friend, who is a mechanical engineer, and I will model a device. I will be looking for information about things like signal strength, how much it radiates, etc. and try to determine if it will operate correctly. My friend looks how flexing or shaking or heating and cooling an object will affect its operation. We usually give this information back to the person who designed the device and then he or she can change it so that it will work better.

The “imagers” in the engineering world are often in the signal processing realm, and they are the ones developing ways of looking at signals, filtering, etc. One great example are the engineers who work on satellite imaging. They work on methods for figuring out how much water vapor or carbon dioxide is in the air. They try to develop new and more accurate methods to take this data and calibrate the equipment. And, of course, they may be involved in processing the data to figure out if they can see what they want to and how to filter out what they don’t.

I realize that this is a bit of a simplification and that there are people who do both. I guess the point is that using a computer to do science and engineering not simply sitting down and programming a computer. There are different ways to use computers to work on problems.

My personal preference is modeling. I like making the computer do cool things. I would go nuts slogging through data all day. I have the greatest respect for those who do that sort of thing.

Of course, my modeling is not nearly as cool as this modeling (warning: nudity at the end of the clip):



1. Fluxor - July 20, 2010

Funny, when I tell people I do modeling as part of my job, no one ever confuses me with being sexy. By the way, is there a model for how your little tush shakes on the computer?


mareserinitatis - July 22, 2010

Really? No one?!


I don’t have a model, but I’m sure I could develop one if I had the free time to do so.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: