PDelta(large)_1(Converge).stdPDelta(large)_1(Iteration).std
Hello,
Firstly, thanks to all of you reading this.
I tried to understand what they(number of iteration, converge) do of P-delta analysis.
As per the help of STAAD(C:\Program Files\Bentley\Engineering\STAAD.Pro CONNECT Edition\Help\std_dls_P-Delta_Analysis.html)
- "Number of iteration" make perform analysis for the specified number of iteation whether or not converges.
- "Converge" make perform analsysis for the speicfied number of iteration. But if the convergence made, then the analysis will be stoped.
I thought that there should be no difference if there is no difference of numbers.
As you could find out the attached files, the only difference of command is that usage of "converge" or "number of iteration".
The numbers are same "3". and the they both did 3 times iteration.
So, I assume that the results will be same.
However the some results are different as follows:
1. Support reaction at Node 1.
1) Iteration : Fx=-2 kN
2) Converge : Fx=-7.712 kN
( It is obvious that, the "converge" makes the moment equlibrium of support reaction from the second order moment(77.12kN-m) )
2. All Relative Displacement at 5m height of Beam 1
1) Iteration : y = 282.148 mm
2) Converge : y= 75.106 mm
(I could not find out what this means)
3. Node Displacement at Node 2
1) Iteration : x = 1504.790 mm
(if, this function do not care about convergence, it shoud be 390.24mm = PL^3 / 3EI = 2.0 kN * 10m^3/ (3*2.05*10^8kN/m^2 * 8.33*10^-6m^4) )
2) Converge : x = 1504.790 mm
(From simple manual calculation - 1504.790mm = PL^3 / 3EI = 7.71 kN*10m^3/ (3*2.05*10^8kN/m^2*8.33*10^-6m^4)
As you could see above mention, not like the STAAD help explanation,
Not only the "convergence check" is added by the "converge function".
However there are no "detailed explanation" of them.
To sum up, Could anyone kindly tell me
1) What "number of iteration" and "converge" are ?
2) What "number of iteration" and "converge" do differently ?
I have the same questions. Can anyone answer this question?
No body? Really no one knows?
For PDELTA CONVERGE, the program compare the displacements between 2 successive iterations against a tolerance. The default value of that tolerance is based on the span of the structure. (See Section TR.37.2 P-Delta Analysis Options of the Technical Reference manual). Your model is a single column, due to which the structure has no span dimensions along X or Z.
Hence, the other method based on the number of iterations is the right one for your model. For this method, instead of a convergence check, the check is whether displacements are diverging from one iteration to the next, since excessive displacements can lead to a crash. If you specify N iterations, the large displacement D which is calculated at the node, and the small displacement d which is calculated within the span, are calculated N times, with the load being adjusted in each iteration.. So the final value of D and d represents the total displacement at the nodes and at the span location at the end of N iterations.
Firstly, I really appreciate your reply.
However, your reply is essentially explanation of P-delta analysis and STAAD's explanation.
1) Converge will stop when the convergence made(in other word "tolerance check") or specified number performed.
2) Iteration will stop when the specified number performed.
My question is then, the "two functions" results are different.
If, only diffrence of two method is the way of stop.
Then if the number of iteration is same(3) then, the results should be same doesn't it?
What I mean, the results below should be same, but it is not.
the Result of - "convergence function" analyze "3 times of iteration"
the Reuslt of - "nubmer of iteration function" analyze "3 times of iteration"
The question directly regarding the "function of staad"(number of iteration vs. converge)
As you could look into the attached staad file, the two model does not included small delta effect.
Because it could be obvious comparison analysis when it is done only with Large delat effect.
As per the results,
1. The support reactions are different(Fx=-2kN vs Fx=-7.712kN), even though two models are analyzed at "same" N=3 times.
2. However node displacements at the top nodes are same(x=1504.790mm), Surely the displacement is "correct"
But, what happened about support reactions?
As increase the numbers of iteration(3 to 10) in both model, the support reaction at "STAAD-Converge" is increased.
But the support reaction at "STAAD-Iteration" is not change(Fx=-2kN fixed value).
Both top node displacements are increased.
If these results are intend/purpose of the function, then the explanation should be provided in STAAD, I think.
Our developers are looking into this issue. We will provide the answer as soon as we have their findings.