I model the building on Etabs easily, but I can never get the exact results of moments, shear or forces that Etabs give me. I still can't exactly know how the load tranfers, yes it's slab to beams to columns but don't know the exact value that each element will have. Can you please tell me what to do to get the right values of each element ?
You will never get the exact correct answer for the stress in a beam/column/slab from either computer analysis or hand calculations. And you will never get a hand calculation to match a complex computer analysis, nor will you get two complex computer analyses using different software to match exactly. There are too many approximations involved.
So structural design instead requires that your analysis give results that are close enough. Software is a tool to achieve that.
How close is close enough? That depends on how redundant the structure is (how many different paths a load could travel to the support/foundation), how ductile the structure is (ductility is what allows the structure to use those different load paths), and whether it is acceptable to take advantage of the redundancy (eg it could cause unacceptable cracking in concrete).
As an example, the concrete design code in my country allows 30% redistribution of bending moment for very ductile beams. Looking at that the other way, my analysis of bending moments only needs to be within 30% accuracy if I'm going to design for the bending moment straight from analysis without redistribution.
You are correct though that your analysis (hand or computer) needs to be checked somehow, and it obviously helps if you understand approximately how the structure will behave. Playing around with a computer model can be very helpful with this. Vary the size of members, vary their span, change the pattern of the loading etc and see how the results change. I think you'll learn more quickly by doing this than doing complex hand calculations trying to match a computer analysis.
Check small sections of the results wherever possible, eg the variation of bending moment and shear on a single span of a beam are related only to the loads on that span, even though the beam may be part of a complex building. Look for little 'tricks' like that when studying - some textbooks have sections on how to verify your analysis results which are very useful. Always check that your reactions are equal and opposite to the applied loads. Check that you don't have any unexpected large reactions. Check that the reactions are where you expect them to be. Check that the deflected shape is what you expected. All of these can be done with no calculations or only very basic calculations.
At university, I spent a semester doing stiffness matrix calculations by hand. I've never done this since graduating because computers are so much better at it - it's a waste of time when you're working for profit. However it was still useful as a learning exercise since it taught me what the analysis software is doing, and therefore how to input the right data to make to software do what I need. I suggest you study with that in mind - learn what the software is doing but don't try to beat the software at its own game.
My summary: hand calculations are first choice for simple structures only (or if you can make acceptable simplifications). Complex structures need software, and hand calculations are to check the software analysis, not the other way around.
I know it's ageneral question, but seriously I think it would e better to know how to the sense of transfering of loads.
I think you already have an idea about this. Load tends to go to the closest structural members. If it has a choice, it goes to the stiffer members (or splits according to the stiffness of the nearest members). Stiffness is very important in structural analysis. Playing with simple models before working up to complex models will help with understanding this.
