From: Mark Smith Sent: Thursday, July 28, 2011 11:46 AM To: Peter Titus; Ali Zolfaghari Cc: Lawrence E. Dudek; Bob Simmons Subject: RE: pedestal calc Pete, I've reread the calculation along with your responses. Ok on the technical content. Please update the calculation with the new analysis (flat lid info) along with the relevant responses to my emails. When this is done, I'll review it a final time. Barring some unforeseen issue, I will sign it. Mark Smith From: Peter Titus Sent: Wednesday, July 27, 2011 8:01 AM To: Mark Smith; Ali Zolfaghari Cc: Lawrence E. Dudek; Bob Simmons Subject: RE: pedestal calc Section 9 is from the global model that includes the TF OOP loading, and the TF torque. It is the same model and analysis used for the TF torsional shear. Global machine torque is included. Global torque effects are discussed in a couple of places in the calc - Torque reactions at the base and bolt circle and in the discussion of the different stresses in the Vee legs indicating that a torque was being reacted by the pedestal. If you look at the figure 9.2-5 you don't see a red contour. - the highest you see is a light green - around 135 level - stresses beyond this are very localized at the intersections and are an indication of the FEA capturing the stress concentration factor. - In the attachment I show the stress components in the pipes. the importat observation is that the primary load in the pipes is compressive which, regardless of the stress concentration, will not allow cracks to propagate. In the attachment I include plots from the latest global model run for a number of the equilibri - some go above the 135 MPa level, but not by much. I think the 135 MPa quote was a reasonable readin of the contours. with the higher values more indicative of the very local peaks at the intersections. The attachment is for the flat lower spoked lid. The plots in the calc are for the bent spoke and earlier modeling of the lower lid. The pedistal stresses are varying a bit based on the lower lid design. Sometimes tuburlar structure stress concentration factors are based on beam modeling of the truss. The results I am quoting are for the FEA solution which captures the stress concentration factor directly - in so far as the mesh density allows. I think this treatment is adequate for now - especially since the stress is biased towards compression. -Peter From: Mark Smith Sent: Tue 7/26/2011 3:07 PM To: Peter Titus; Ali Zolfaghari Cc: Lawrence E. Dudek; Bob Simmons Subject: RE: pedestal calc Pete, For question (1) below, your answer did not include the machine torque (effects of the TF legs). Is some percentage of the machine torque applied to the pedestal analysis? If so, how much and what model (figure)? Also, you mentioned (below) 233 MPa for the maximum stress. Executive summary page 6 shows 135 MPa, page 19 figure 9.2.-5 shows 233MPa. Please clarify. Lastly, I found various articles for tubular structure's weld stress concentration factors (SCF) based on steel, oil rigs, etc. The SCF's can be between 3-10, give or take a few, depending upon connection parameters. So, I think it prudent not to use the girth weld SCF =1. We need more margin. Mark Smith From: Peter Titus Sent: Monday, July 25, 2011 2:18 PM To: Mark Smith Cc: Ali Zolfaghari Subject: RE: pedestal calc From: Mark Smith Sent: Mon 7/25/2011 12:54 PM To: Peter Titus; Ali Zolfaghari Cc: Lawrence E. Dudek; Bob Simmons Subject: pedestal calc Hi Pete. I have a few questions for the pedestal calculation. (1) What figure has the results for the simultaneously applied loads? Section 9 has some of the results from the global model in which the vacuum loads, dead weight, Centerstack casing normal operating thermal expansion loads, TF Expansion Loads, and all loads from the PF coils for the given equilibrium, are applied. I have another global model run in which the bent spoke lower lid was straightened out. These may produce different pedestal stresses - will update as soon as I have access to Andrei's computer. Also, in the executive summary (calc page 6), no stress concentration factors are used for the welds. This seems generous. Typically, you have used a factor of 4. (2) Is no stress factor within AWS standards for full penetration welds? The factor of 4 is for fillet welds. In AISC and ASME a full penetration girth weld is assumed to have a stress concentration factor of 1.0. These pipe welds will have a stress concentration factor above 1.0, even though they are modeled as being full penetraition, because their intrersections are at some angle - but the FEA should include some measure of this - at least in so far as the mesh density can capture it. - This is why I keep emphasizing that the Thompson scattering port reinforcements should have full penetration welds, so that you can claim that the FEA represents the stress. Rigorously the mesh density should be fine enough to resolve the stress concentration. The peak loading (Lorentz-Deadweight) in the pedestal is compressive which doesn't propagate fatigue cracks. Note, scenario 79 has been used for the worst applied torque. Based on the design point sheet, there is about a factor of 1.5 between sc 79 and sc 21 torque. Assuming linear behavior, sc 79 would scale to about 202 MPa. I am not sure how much global torques appear in the pedestal - especially with the stiffer spoked lid. Figure 9.2.5 is for scenario 21 and is described as having the largest torque found. The local Tresca stress for this EQ, at the intersection of the Vee and middle flanges is 233 MPa. This includes the dead weight and all other loads - but is for the bent spoke lid. A modest weld stress factor (1.48 vs 4) puts the pedestal above the 300 MPa weld allowable. Given the reported stress is 233 Tresca, and much of this is compressive, I think weld fatigue will not be a problem for the Vee Pipe Pedestal. Mark Smith Mechanical Engineer Princeton Plasma Physics Laboratory P.O. Box 451 Princeton, NJ 08543-0451 (609) 243-2778