Most mathematical field in Engineering? 1

[gman89]

Throughout my brief professional career so far as a mechanical engineer I've noticed that a lot of what gets done seems to be based purely on experience and the notion of "I know this is the best answer because it just is". I then find myself doing the same thing with tasks suitable to my small experience. Work to some extent becomes a mindless process where most thought and energy goes towards either thinking how to write a report (which is fine, I guess), communicating with clients or superiors or using some software via a black box (which is not fine).

The one thing I miss about uni is the application of theory to problems, now I find anything that is fundamental enough to be solved by theory has probably just been scripted and makes me feel intellectually dissatisfied. I take charge of some small jobs that have just become a routine use of scripts and software and no real difficult analysis.

What I would like to know is which field in engineering would you say requires the most mathematical analysis on a day to day basis to the extent that experience alone won't guarantee an answer? I'm thinking maybe some form of signal processing at the moment but I say that with hesitation, too.

 

[IRstuff]

When was the last time you did a fuel injector, or carburetor, rebuild?

Every engineering discipline has, at its core, a mathematical basis, which is why you, or at least I, were required to take classes that covered PDEs and ODEs. And like the insides of cars, there was a time when people spent large chunks of time cranking on calculations because there was no other way to do it. Now, most of our math is done with software programs, be it NASTRAN, SINDA, Mathcad, MATLAB, Maple, etc. If you don't do any, then it's simply due to the fact that your company isn't on the leading edge. However, those that are on the leading edge are doing mostly computer-based analysis. If that's still enough, then you need to be looking at academia.

TTFN
faq731-376

 

[Twoballcane]

What do you mean by mathematical analysis, as in using math from arithmetic up to calculus? If you are then, within Mechanical Engineering, you should look into the specialty of being an analyst in the field of structure, dynamics, and thermal. And when I say analyst, I don’t mean a FEM jockey, but a person who truly understands the math and physics in those fields. The trick is to meld theory with practice, but you have to be in a situation where you can do this type of work. If you are more interested in just using math to find new things, use calculus in your everyday calculation of anything. Just come up with an equation to define the situation (i.e. finance, physics, yearly temperature trends…) and then find the derivative and see what it gives you or take the integral see what the area under the curve means to you. A great book to start with is Calculus An Intuitive and Physical Approach by Kline. I picked this book up a year ago and it has great ways to see things differently or how other theories where first developed using calculus. Its not a math text book per sea but more on how to tackle subject base problems.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”
"People get promoted when they provide value and when they build great relationships"

 

[TheTick]

The mechanical design field is overrun with hammer-monkeys who do little more than throw more steel at problems they have already solved a hundred times over. Lots of progress to be made by applying even a modicum of analysis to even the most mundane products.

 

[zdas04]

I don't often derive new relationships, but I've done it when the "black box" gave me answers that I was unwilling to accept (its happened 6-8 times in the last 30 years). I do verify the output of the black box with real world data before I'll use a program that someone else wrote (and I find that about half of the time simple calculations have problems and nearly 100% of the time complex programs like pipeline simulators have serious problems). After evaluation I have to decide if: (1) the program is too broke to use: (2) if I avoid certain features it is OK; or (3) it seems to work well. The only program I can point to that I use that is in the last category is Uconeer and RefProp (the first from member Katmar, and the second is from NIST).

My peers in Field Facility Engineering for Oil & Gas tend to be Project Managers and mostly work in a Microsoft Office level of arithmetic. My point is that in any field you can find reasons to do as much or as little original thought as you want to. You can get by with black boxes and Excel. Or you can personalize your toolbox with relationships you've developed and black boxes you've confirmed to (at least sometimes) reflect reality. Either one will usually get a "good enough" answer, but I find that the later gives me more job satisfaction.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[MiketheEngineer]

Aero, Chem or EE

 

[Latexman]

I would add Nuclear too.

If you desire a more math related career, get a PhD and go into academia or research.

Good luck,
Latexman

 

[Venturini]

Well, from my short work experience I will tell you that you can actually use your knowledge and scientific basis to solve a lot of day to day real world problems. The thing is that lots of people that I came across didn't do this and didn't apreciated it either when someone else tried to do. They used to say that "we are not doing rocket science, just solving simpler problems".

Most of the time the problems were due lack of knowledge on the basis of mechanical design and could be solved just for aplying some old and well developed concepts that you get in college. There are some areas where math models are hard to work with, as modeling goes more complex - but the theory is valuable in every area of knowledge. Just remember, theory was developed through observation of real problems and phenomena.

You can do good engineering and thinking in any job, but it can be hard and you have to work your way through more experienced people with more time on the company but that are kind of skeptic with newer engineers. Of course this was based on personal experience.

 

[SNORGY]

gman89,

Nothing wrong with Mechanical Engineering for what you are looking for. Sure there are canned programs to crunch numbers for you, but in the absence of such things, when an answer "based on experience" is needed in one hour, I often state the required answer on a "derivation to follow" basis; for example: "How big is the treater? 8-D x 40-L? That'll be a 3" x 4" PSV with about a J orifice." Then Procurement looks for the PSV while I generate the back-up calculation for the Q/A file. On another recent project, I had to calculate the buoyant force on an inclined V-shaped "stealth bomber" fabricated plate assembly, and the ensuing use of equations for planes and their normal vectors resulted in a triple integral that, at the end of the day, caused me to derive from first principles the formula for the volume of a triangular pyramid - which, in retrospect, I suppose I could have looked up - and from that, I could quantify the shear force exerted on the bolts holding it down and, subsequently, size the bolts. On another occasion, I wanted to calculate a force on a pipe support going around a 90-degree bend but our CAESAR key was out of runs, so while I waited for them to reflash the ROM on the ESL, I derived equations from the free body diagram that, oddly enough, were none other than those for the guided cantilever method described in Peng (and others). To this day, if I want to, I can describe a loading function on a pipe or beam and integrate it four times and solve the boundary conditions to get the answers I need, or set up an iteration to solve some problem or other (e.g., Colebrook-White or Van Der Waal's).

If you really want to stick to number crunching, then in Mechanical, your best bet is likely in some application of FEA, machine dynamics, or computational fluid dynamics. Otherwise, especially in the energy sector, the guys who do the most math are the Process / Chemical guys (i.e., HYSYS).

The trick, as others have alluded to, is to use the various software tools out there, but know enough to recognize when the answers "don't make sense" or "can't possibly be right". To me, I found my Zen state when I started saying, "Here is my answer. Proof - or revision - to follow.". Then I forced myself to do the follow-up calculations just for a matter of record. Nobody in their right mind would trust me to integrate any "square root of ee to the eye theta" thingies any more, but for what I do, if I want the math, it's there.

 

[ScottyUK]

Electrical is pretty heavy on the maths, especially in the control and systems analysis fields. As noted above for other disciplines, software has taken a lot of the direct number-crunching away because most of us find it time-consuming and our employers would rather we used our skills to fix problems instead of crunching numbers at a tiny fraction of the speed that a computer can achieve.

 

[Gymmeh]

I think it really depends on what your company does and if they are "cutting edge".

During the economic poop-er, my last company revolutionized its product line, which put them in the front of the market. The great part was that all of the engineering disciplines had to go back to the basics because, no one had "done this before". However after the product was done being redesigned, I went back to being a glorified paper pusher, and trying to keep personalities from causing design flaws. (some would call this baby sitting)

The thing is this experience led me to another job which is involved with more math engineering. However, i find its almost impossible to avoid paper pushing at some point.

 

[CivilTom]

What about financial engineering? Creating powerful algorithims with lines and lines of code.

 

[SomptingGuy]

NVH diagnostics can be quite exciting if you like measurement and subsequent interactive time-series analysis, using tools like Matlab or equivalent. The only problem is that management see what you are doing and try to "productize" it to either sell or allow monkeys to do it.

- Steve