Hand Calculations 10

[KevinDeSmet]

I would consider myself possessing an ample knowledge of finite element analysis, applying proper LBC, material properties and so on. If I would want to double-check a simulation study, I would opt for solving a simpler or complexer study by changing mesh types.

I know it is common practice to double-check by means of a hand calc.
However I just can not wrap my head around hand calculations and tend to want to avoid them. Does anybody have suggestions on how to get better at them?

And most important, are hand calculations essential to be taken serious as an employee involved with finite element analysis or is that a bridge to far?

Certified SolidWorks Professional

 

[rb1957]

"how to get better at them" ... doing them

 

[DanStro]

I agree wit rb1957, the more you do the better and more confident you'll get. I, and I think most of the people on this forum, think they are very, very, important. Even if you don't do a hand calculation every time after a while you should get a decent feel for what the answers should look like. Even if the problem needs to be simplified.

If the type of analyses that you are talking about is something like static, dynamic, or thermal then you can get books of problems (Schaum's or Problem Solving are two that come to mind).

 

[ESPcomposites]

Earlier in my career I did more FEM and less hand calcs. It was partially due to nature of space applications. Later I became better at hand calcs, to the point where I rely on them far more than FEM. What helped was getting the right reference books for your industry and making friends with the more senior engineers. You can help them with FEM and they can help with your goals.

The hand calcs will keep you in check and tremendously improve your ability to think about the right way to set up the FEM and greatly improve the result.

I know someone who prides himself on his FEM ability, but had poor ability translating theory into practice. He has built some of the most poor, ineffective, and costly FEMs that I have seen. In some cases I overrode the results via napkin scribble and a calculator. The other thing is that learning hand calcs is a more valuable skill than running a FEM program. This will help you in career opportunity and gain you respect. Not to mention that you can take those skills with you and don't have to worry about what software the company runs.

Brian

http://www.espcomposites.com

 

[ThomasH]

I also agree with rb1957. As always, you need to practice to get better. Avoiding them will get you nowhere.

I would say that if you "know" what to expect in terms of result before you run the analysis and than get the expected result, it is probably correct. At least there is a good chance. Then you have probably understood the problem and modelled it properly.

Otherwise you are, worst case, just pushing buttons if I put it harsh .

Regards

Thomas

 

[Twoballcane]

FEM is one data point. You will need to do hand calc to create the other data point to make sure your assumptions are going in the right direction. I will actually do the hand calcs first and then do the FEM just because the hand calcs will can explain all of the physics that you are trying to accomplish and have a better idea on what type of FEM model I would like to do. Just as simple as creating a free body diagram to know where the forces are happing, what moments you may encounter, in your structure so you can calculate out the stress. Just from this insight, your FEM mode may be smaller and easier to solve than throwing in the whole model that may take longer to build and solve. FEM is garbage in and garbage out. How do you justify your model to your boss/management/customer/other Engineers when they ask “how do I know this is right?”.

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”

 

[flash3780]

So... if you don't check your FEA with hand calculations... how do you know your FEA isn't out in left field?

Get a copy of Roark's and a copy of Peterson's. It helps to have good references.

 

[Andries]

A very interesting and topical thread. Amen ESPcomposites!

Andries

 

[corus]

Doing any kind of analysis involves a level of confidence in the results. In the first instance people may say they're not confident in the results because 'they don't look right' or because common sense tells you that can't be right. This may be because of the assumptions you've made or because you've simply made a mistake somewhere.

Once you're past that stage the level of confidence will increase if the results are similar to other results obtained in the past or they're 'in the right ball park' of what you expect, to use an americanism. That 'ball park' figure is best obtained from hand calculations.

A higher level of confidence is reached if the results from the model agree with measurements taken from an existing case, showing that your assumptions, material properties, etc., are validated. I wouldn't be necessarily highly confident that the model results were correct because they agreed with some 'noddy' calculations someone had made as the assumptions in them are generally very simple compared to a FE model, and of course mistakes could have been made in the hand calc itself.

Tara

 

[HankTexas]

As someone who started a career before FEA was readily available I used hand calculations, a slide rule and reference materials for every design. Over time you develop a ‘feel’ by just looking at the structure and the loads on the structure for what the results should be in a lot of cases. And today the first thing I do when starting a design is some hand calculations to verify I’ll be in the ballpark with the FEA.

I’m now often in a position of checking over some of the younger engineers work before it goes into fabrication. It’s not that unusual to find a design that’s not even close to meeting the loading requirements. When I consult with the engineer I will always find they didn’t have a clue as to what the results should look like. They just took the results of the FEA and put it on paper. Of course the results are accurate to ten decimal places but the structure will not handle the loading. That’s scary.

 

[johnhors]

Hank,

It is my belief that no "young engineer" should be allowed to use FEA until they are competent in hand calculations and the use of free body diagrams.

However what I find scary is the push by management in many companies to out-source FEA to cheap consultancies where the "analysts" know what buttons to press in the software to produce pretty pictures but have no understanding of stress analysis.

http://www.Roshaz.com

quality, cost effective FEA solutions

 

[rb1957]

john ... willing to bet you're older than 40

 

[johnhors]

54

FEA is a fantastic and phenomenally powerful tool but like powerful physical tools (hand held chainsaw) it can be a deadly weapon in untrained hands. Push button technology in every industry has made people mentally lazy (remember when staff in small shops used to have to add up the cost of goods in their heads? in £sd?). The situation is not helped by a culture (amongst ignorant managers) that loves seeing glossy contour plots and loathes any hand written calculations. So the hand calculations get put into excel or mathcad and you're back to the push button technology again where the spreadsheet just becomes a template for subsequent users to plug a few numbers into, again with little or no thought behind their actions.

http://www.Roshaz.com

quality, cost effective FEA solutions

 

[jagad5]

Just because you can model it with FEA or other fancy software doesn't mean you should. A true stress analyst can often provide useful answers on the back of an envelope faster than he/she can start ANSYS. I have seen months worth of computer simulations negated by a guy with pencil and paper. (Seriously. At one of the major jet engine companies, a new PhD engineer spent a few months doing some complex CFD simulations. During the project review, the chief engineer demonstrated using hand calculations that fit on a napkin that the complex analysis was wrong, and he included a sketch of what the results graph should have looked like.)

Get a copy of Roark & Young and an older version of Stress Concentration Factors by Peterson (the new ones are too small and don't lie flat on your desk.)

Doug

 

[corus]

You have to bear in mind that FE models make assumptions as to the loads/restraints etc. In general hand calculations make bigger assumptions. One can check against the other to make sure that the results from the FE model are sensible. Neither of the solutions will be 100% correct (if there is a correct solution), and all you can say is that the FE models may be more correct, if carried out responsibly.

Tara

http://tinyurl.com/4ydjg7m

 

[gerbii]

Hi all

As someone stated, FEA can be dangerous.. "FEA will make great engineer from a good one, but it will make dangerous one from a bad engineer" that's how we describe this in firm

To be better at FEA ? expierence is essential (from my point of view).
As for the hand calc.. if it isn't too time costy - why not. It's better to sleep well (sometimes checking critical points of problem it's enought).

"If an experiment works, something has gone wrong"

 

[Qshake]

Permit me $0.02 on this matter. I am also of the age where hand calculations were dominant before widely avaialble FEA became the norm.

I don't know if you are mechanical or structural or otherwise in your studies. I rely on statics, mechanics of materials and a variety of structural analysis methods (energy and numerical) to determine overall responses and or stresses.

Other good sources for this include Roarks Stress and Strain as well as other such as Practical Stress Analysis in Engineering Design by Blake.

I agree with the others, it's practice, practice, practice.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[KevinDeSmet]

Whether I am mechanical or structural in my studies, well I quit school to work on video game level designs. But there's not a lot of work in that sector so... one thing led to another, and now I'm interested in CAD and CAE!

Certified SolidWorks Professional