Hello everyone, I have this problem which always bother me a lot! whenever I do FEA on any part and compare the results against the hand calculations what I notice is huge anomaly which is clearly not in the range of 0 to 15% of FEA results. I have even showed those calculations to my supervisors which they think are correct but I just can't understand the reason for such deviation! Has it happened to anyone else? What is the solution for this?

There could be so many reasons for this. If your supervisors agree your calcs are correct then your model must be wrong. Have you done convergence checks? Most problems (assuming loads are corectly applied) are insufficient mesh density and incorrect assumptions at boundary conditions. Without a specific example I can't really help.

Like joninstjohn state first I would look at your mesh and boundary conditions this is the most common mistake that new users of FEA or any computational software make when setting up a problem. You can not just input some numbers and geometry and get the right results. I have spent months and nearly a year setting up and developing code to solve certain situations. I am assuming that your problem is simple if you are solving it by hand and depending on the equations you may not be holding enough decimals sometimes gives an error not as large as you a presenting. If you want more help provide us with some more details. Since it is simple enough to solve by hand you should get exactly the same answers if both methods are done properly with a good software package.

to achieve a true answer in a FEA software you need experience more than any knowledge.do you have enough experience to be sure about your results in software?

Make sure the boundary conditions are correct. In hand calculations of a simple beam with simple supports with a uniform load you should have a smile shape in regard to deflections. If in FEA you restrict the face of the edges you should notice that the deflection will not start at the edge but a short distance in towards the center. If the shape of the deflection is correct you should check the material properties used in the FEA to make sure that it matches the hand calculations. I normally find hand calculations will be off by about 10-30% depending on the loading and how complex the geometry is. Hope this helps.