mechanical engineering should know

mechanical engineering should know

In my experience of 20+ years building and repairing machinery of many types, knowlege of manufacturing processes and tolerancing are highly valuable to a mechanical engineer.

For example, when considering tolerance callout or dimensioning on a drawing, 1" ,1.0" ,1.00" ,1.000", are not equivalent values and their use significantly determines the process to produce the required results. To describe this example further suppose the following general tolerance callout on any mechanical drawing;

Fractional = +/- 1/16"
             .0 = +/-  0.3"
           .00 = +/-  0.01"
         .000 = +/-  0.002"
Unless otherwise stated.

To produce a specific length of steel dimensioned at 6" means a conventional saw cut may suffice. Yet, to produce a length dimensioned at 6.000" means the use of a more accurate process( possibly face milling).

The cost to produce these two similar lengths are considerably different and without proper tolerance consideration, can substantially affect the cost of a design project.

Another commonly overlooked requirement for understanding is accumulative tolerancing. Consider the same piece of steel but with 4 holes in line and equally spaced at 1" increments. This common note in dimensioning can produce a fourth hole 1/4" out of true position (1/16" x 4 equal spaces). 

In short, knowlege of proper tolerancing (positional, accumulative, geometric etc,) and knowlege of manufacturing processes ( flame cutting, saw cutting, machining, shearing, stamping etc.,) are required to accurately convey your design requirements to a manufacturer. Accurate information provides accurate costing and minimizes error.

Mechanical engineered drawings are used to convey information from the engineer to others. The accuracy given in the drawing directly relates to what is produced from them.

OR

A lot of things, but I will recommend the following in order of importance..

1. Experience with prototyping in lathe machine, at least the knowledge of the process flow. Most important because modeling in software is easy, not making it in real life.

2. Software for solid modeling - Solidworks (easy to learn, lots of tutorials available)

3. Good knowledge of basic mechanics, strength of materials or fluid mechanics and mechanism design.

4. Presentation skills - Microsoft Excel, Powerpoint, Word, LaTeX

5. FEM or CFD and software - Ansys/Abaqus/Comsol

6. Numerical methods with Programming - MATLAB/C++/Mathematica/Python.

And, creative thinking. Edward de Bono recommended.


Im in agro-mech right now. I had a long running personal project of mine that started several years before I took thermo. It was to model the behaviour of this thing:

I took thermo and the teacher said, you can do a project. And so I knew immediately what I was going to do. The project exploded and I ended up spending 80hrs on it. I had to look up the pump inlet pressures and look up the nasa 7 term polynomials for the calculation of enthalpy and entropy of a specific cut of kerosene. My assumptions were extremely critical. I had to look up on 50yr old tables the enthalpy of liquid oxygen at multi mega Pascal pressures. I wanted to kick that dead horse so much that no person could ever hope to follow in my footsteps. I wanted that project to be an example to students much farther along, in the 400 level courses. I attribute my project to making me pass the class. C+ under the meanest teacher in the land. The moral of the story is, take your A's B's and C's home and actually try to use them in real life situations. You will soon find that your A means nothing at all. That you did not learn anything about the actual application of your tools. The fact that I take my grades home and actually try to use the information makes me the best engineering student around for miles. I am a true engineering student, a shark. I can solve real world problems because I make pure and rigorous mathematics come together to support and model real world situations. Have you ever tried to make a vector field of the spiral flow or a fluid through a centrifugal compressor?

F(x,y,z)=cos(x*pi/2r)(x-y)i-sin(x*pi/2r)y i + cos(y*pi/2r)(x+y)j+sin(y*pi/2r)xj -zk

 Do you know how to make a spiral vector field? Do you know how to do the flux integral so that you can calculate the convective heat transfer coefficients across the wall of a pump carrying liquid oxygen? If you are a C student in 300 level engineering you should be able to do this on your own and if you dont then you are a paper tiger. the projects that you do for classes don't really amount to much and you might as well not even try. Do you have the strength of will t reteach yourself how to do calc 3 even though you have not touched it in 4 years? You need to be able to do this. These are not high expectations, we are all nerds in engineering. I might call myself a genius, but there are bigger fish than me, far more powerful, but they dont choose to flex their muscle and that is a shame, one would wonder if all that they are capable is putting pointy letters on a report card. I am a C student, my GPA is just over 2.1, in my free time I make A students eat my dust, but its only because they dont try. I bet, that even in MIT you have people that float along and try to get the easiest teachers. they don't have their own personal projects. They dont use what they learned.   

Anyways, my professor in thermo,he taught me that any engineering problem is to be broken up into parts.
sketch
assumptions
concepts
approach
solutions
properties

sketch
     this needs to be a work of art. highly detailed and realistic, almost to scale. the neater it is and the more realistic, te faster te problem will be done. this is half the battle.

assumptions
     What you put down here is stuff like    frictionless, isothermo, liquid fluid, gaseous fluid,  V1 zero, no pump W= zero. pressure atm.  g=9.81m/s.
e=.7  yadda yadda yadda. It is very important that you assume correctly. Your assumptions can make the difference between spending 1.5hrs on a problem (which is kind of normal) vs. 2 weeks.

concepts:
     here you put things like the full off the shelf version of the extended bernoulli equation. The area of a circle. The relationship between radius and diameter ( i use that combination all the time). The relationship between pipe flow cross section area and flow average velocity with regards to flow rate.