ML20114C808
| ML20114C808 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 01/18/1985 |
| From: | Williams N CYGNA ENERGY SERVICES |
| To: | George J TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC) |
| References | |
| 84056.040, NUDOCS 8501300530 | |
| Download: ML20114C808 (5) | |
Text
.
50- W S i
pd
$0 - W(/
101 California Street, Suite 1000, San Francisco, CA 941115894 415 397 5600 January 18, 1985 84056.040 Mr. J. B. George Project General Manager Texas Utilities Generating Company Comanche Peak Steam Electric Station Highway FM 201 Glen Rose, Texas 76043
Subject:
Cygna Study of Unistrut Torsional Capacity Comanche Peak Steam Electric Station Independent Assessment Program - Phase 4 Job No. 84056
Dear Mr. George:
Cygna's review of conduit support designs has noted instances where torsional
)
effects due to longitudinal loads were not considered. Cygna has perforned a study to evaluate these effects on the integrity of the longitudinal conduit supports. The study and its conclusions, as discussed below, indicates a need for further evaluation, not only of the Unistrut support designs, but also of conduit stresses and anchor bolts used in connection detail CSD-la.
During the conduit support reviews Cygna found that for longitudinally loaded supports, no consideration was given to torsional effects induced in the support due to either the eccentricity between the conduit and the main member (P1001C3) or the eccentricity at the connection (P2815) between the brace (P1001) and the main member. To assess the behavior of a conduit /Unistrut support system due to longitudinal loading in the presence of member eccentricities, Cygna constructed a finite element beam model and performed a conservative preliminary analysis of a conduit supported by three multi-directional supports (Type CSM-6a). The analysis clearly demonstrated that since the bending stiffness of the conduit is so much greater than the torsional stiffness of the main member, virtually all torsional moment due to conduit eccentricity was absorbed by the conduit and very little torque was transmitted to the torsionally flexible Unistrut main member and its cantilever support.
Pwas
&Q Q l
San Francisco Boston Chicago Richland
6 Ekz Mr. J. B. George January 18, 1985 Page 2 In contrast to the low torque developed in the main member between the conduit i
and the cantilever support, considerable torque was developed in the main member between the conduit and the connection to the brace due to the brace eccentric-ity about the center of rotation of the main member. p ending upon the assumed location of the center of rotation of the main member,\\p/ the rotations at the end of the main member ranged from 10' to 15*.
However, in reality these large rotations will not occur since the P2815 connections can develop moment resist-ance about an axis perpendicular to the pin axis. The combined normal stresses (bending plus warping normal) at the most highly stressed location on the cross section exceeded the maximum allowable stresses _llowed by AISC and AISI. AISC allowables are governed by Section 1.5.1.4.5(0.g33,000=19,800 psi) and AISI by Section 3.1 (0.6 x 45,900 = 27,540 psi).
Stresses at the mid-span of the member, which was the critical cross-section, as well as the member-end rotations are sumarized for three cases in Table 1.
Regarding this support, Cygna has the following concerns:
1.
The manufacturer has stated that Unistrut combination members have not been designed to resist torsion.
2.
A P1001C3 Unistrut member is a combination of three P1000 members joined by intermittent msistance spot welds on 2 to 3 inch centers such that its cross section possesses no axis of symmetry. As such, the evaluation of the torsional properties and the internal distribu-tion of stresses due to torsional loading of a P1001C3 member would require an extremely detailed analysis.
3.
Torsional moments and conduit longitudinal forces are applied through one of the P1000 members and resisted by another P1000 member to which the brace is connected. Thus, the spot welds joining the members are subjected to a combined state of shear and tension of an unquantifi-able magnitude.
(1) For twisting only, the center of rotation of the P1001C3 was assumed to occur at either the C.G. of the entire section or the shear center of the single P1000 member to which the conduit was attached. Torsional stiffness and torsiona) stresses were conservatively based on this same single P1000 member to which the conduit was attached.
(2) These are not necessarily the correct allowable stresses. They are the maximum allowable stresses allowed by the code for members satisfying very specific conditions.
ssZ Z_
Mr. J. B. George January 18, 1985 Page 3 4.
Resistance spot welds are designed to resist only shear and not the tearing type tension loads which must be developed to transmit from one P1000 member to another both torque and tensile reactions forces at the brace connection.
5.
To resist the rotation of the P1001C3 member, the brace connection P2815 must resist bending about an axis perpendicular to the pin axis. This connection was not designed or load-rated for this moment.
The results of the analyses performed by Cygna are not to be construed as li-mited to the support type analyzed (CSM-6a). The studies were performed to evaluate and illustrate the effects of torsional loadings on Unistrut members and connection components. The results should be considered in all cases where such loads occur and where spot welds are used to transfer loads other than shear. Given thr Ntent of the analysis, i.e., to evaluate the effect of tor-sion on Unistrut mponents, Cygna has not examined the effects of the torsion on the conduit stresses and the behavior of the support-conduit system. Such behavior was referenced (Gibbs & Hill letter GTN-69371, dated August 23,1984) in response to question A-2b of Cygna letter 84056.015. This response indicated that the anchor bolts used for CSD-la type connection detail were acceptable due t
to the behavior of the support conduit system.
In light of the above results, Cygna believes that these areas should be investigated further and the response should be revised accordingly.
Based on the analysis of the P1001C3 n:mber and the configuration of the connections Cygna believes that the calculation of a safe load carrying capacity cannot be made in accordance with the provisions of either AISC or AISI. Therefore, a safe load carrying capacity must be established by testing of the actual support configurations and their variations. Cygna recommends l
that tests be conducted in accordance with the (ISI Specification, Section 6.
l l
,
Hr. J. B.-George January lb,1985 Page 4 If TUGC0 elects this plan of action, Cygna recommends some further discussion between Cygna and TUGC0 to ensure the test specimens and applied loadings envelop the support configurations at CPSES. Please call if you have any questions on the above matter.
Very truly yours, o
e Nancy H. Williams Project Manager NHW/ajb Enclosure cc: Mr. R.M. Kissinger Mr. J.B. George Mr. D.H. Wade f
Mr. S. Burwell Mr. V. Noonan Mr. S. Treby Mrs. J. Ellis Mr. R. Ballard i
l l
l l
l
W TABLE 1 Results of Cygna's Scoping Analysis for I
Unistrut Conduit Supports Conduit Location Size & of Center of 2
Number Rotation 3(in)
Afin) 8(PSI)
O(DEG) 3-3/4" A
36 24 32,029 12.6 3-3/4" B
36 24 23,427 9.6 1-1-1/4" A
55 24 33,834 14.8 l
1.
Analysis was based on a 2-span conduit with support spac-ing, S.
The P1001C3 member of the CSti-6a type support was modeled as a cantilever with the brace attached at a dis-tance, i from ground. All moments at the P1001C3-brace connection were released.
2.
The analyses considered two center-of-rotation locations:
A-The center of rotation was located at the shear center of the P1001 member to which the conduit was attached.
B-The center of rotation was located at the center of gravity of the P1001C3 member.
3.
Summation of the flexural and warping normal stresses at mid-span of P1001C3 member (= 1/2).
4.
Member-end rotations about longitudinal centerline of P1001C3 member at P1001C3-brace joint.