Approah 1:For modelling very long pipe line one has to divide it in to finite number of parts separated by anchors.Then one should perform analysis for each part separately.
Approach 2:Alternatively we can prepare the whole problem in one mathematical model and conduct analysis.

I have observed that there are differences in results at nodal points.

can anyone please clarify which method one should follow.

thanks

c srinivas

please clarify can we conduct analysis for the

I must confess I'm a bit confused. Are you talking about modeling a long straight run and having multiple anchors along it?

In "approach 1" above, you can only break the line into different models at TRUE anchors (where they really exist in the field). If you arbitrarily break a line, and analyze the two pieces in different runs, then you have lost the interaction between the two halves. I suspect this is your problem.

My quirie is simple.

Consider Instrument air line for Nulear Power Plant.The line starts from Air drier and terminates at various user points.We have to use ASME SEC III Subsection NC/ND.

I have two options

Option 1:I will break the Instrument air line at TRUE/PHYSICAL Anchors then conduct analysis for each part.(anchors I am locating generally at columns or at beams.I am welding the pipe to the Building through Structural members and Embeded Plate as the Building is made out of R.C.Concrete Floors.All degree of freedom are arrested at the point of Anchor.I take realistic stiffness for anchor/support by modelling the anchor/support in STAADPRO.This stiffness I am using in the program for conducting Resonse spectrum analysis).I have ANALYSED for DW,TH,RESPONSES IN EVENT OF EARTHQUAKE.Ihave noted down the results at various node points for Option 1.

Option 2:If we include all the parts in Option 1 then we get the Mathematical Model for Option 2.I can Model and perform the analysis for the problem in one run.In this method I can do so by including all the anchors considered in Option 1.(By introducing anchor the forces on either side of the anchor SHOULD NOT get transmitted to the other side).The results for NODES should be same as that we get in CORRESPONDING PART in Option 1.The method of analysis in same as Option 1.

At actual plant when the piping gets erected the results will be as per Option 2.

However I am finding variance in results at Corresponding nodal Points in Option 1 and Option 2.

I hope I could present the problem properly

thanking you

C srinivas

Check your natural frequency results for the two different conditions.

If you're using a frequency cut-off, you will get 1 mode past the cut-off in the combined model, and two modes over the cut-off in the split models (one for each). If these modes are significant contributors, then there will be a difference in the results.

If you're using a set number of modes, you'll get the specified number in the combined model, but 2 times that number in the split models.

Try using a frequency cut-off, and setting it to a larger frequency value (if you're running seismic, change the 33hz value to something like 70hz) and see if the results are closer.

Since your anchors have stiffness values computed and are not "infinitely rigid", they will not prevent loads and displacements from being transmitted across the "anchor" point. That is why you are seeing a difference between your continous and partial models. You may want to consider running your model with rigid anchors for your static analysis, assuming you are allowed to.

Truth be told, in the B31.3 and B31.1 world that I work in, I would be drug out and beaten for doing any kind of stress review on an instrument air line, much less the level of detail you are putting into your review. However, I understand from others who've suffered through it that stress engineering in the nuclear world requires that kind of detail.

Best of Luck,
--ED

I missed this important point in the post above. Ed is exactly right, if the anchors are not rigid, they will transmit forces and moments to the connecting pipe.

Even if you do change the stiffness to rigid, watch the frequencies as I previously mentioned.