Antaki Design for Analysis

8/2/2019 Antaki Design for Analysis

1/12

r,A/SRC-_S--91-075IDE92 009837

DESIGN-FOR-ANALYSIS OR THE UNINTENDED ROLE OF -._ANALYSIS IN THE DESIGN OF PIPING SYSTEMS (U)

/_t/_ f

by f .

G.A. Antaki " ? _!9_ _ _Westinghouse Savannah River Company _ _ _._ = _-_ _ #.Savannah River Site = _ ,, _ _, ,. -Aiken, South Carolina 29808 _" _ ""

F .-. _'d'_ -.--_ _ =

_= _ _ _..__. _ _''_'_ _ _ ''" _- oa

, _,,_ OJ "_ _ _ 0

A paper proposed for presentation and publication at the _ _==_,_ _-_ -_ _'-_-=-"_ASME Pressure Vessel and Piping Conference =__"_ _= _ "__=_ _"--oan_-t>mgo,alifornia .== ___ _, = _ _ ,, _, ._ ,,,,. _ _ _ ,.,. _ _, ii, c_June 23-27, 1991 _-_.__-_o_-=" = "=_""

_N.._ _. ta ,._ t,._ '_ e'_ m

_ ," _ _,_, _ o m '_, '__

This paper was prepared in connection with work done under Contract No. DE-AC09-89SR18035with the U.S. Department of E_ergy. By acceptance of this paper, the publisher and/or recipientacknowledges the U.S. Government's right to retain a nonexclusive, royalty-free license in and toany copyright covering this paper, along with the right to reproduce and to authorize others toreproduce all or part of the copyrighted paper.

gtSTR{SUTION.FTHISDOCUMENTS UilILlr,/IITED


2/12


3/12

j

DESIGN-BY-RULEThe original editions of the B31 standards, strongly relied on the use of "design-by-rule",which consisted of standard practices established and confm'ned through experience.The introduction of A.R.C. Markrs thermal stress equations into the 1955 B31.1 pipingstandard, had a limited intent [1, page 141]:

"Formal analysis or model tests shall be required for pipe lines whichsimultaneously satisfy the following conditions: Maximum normal operating metal temperature greater than 800F Nominal pipe diameter greater than 6 inches Rated service pressure greater than 15 psi"

Further, he warned [1, page 148]""... against too imp_cit a reliance on the rules to the exclusion of good judgment"

An example of good judgment is provided in Markl's paper [1, page 141]""The requirements for analysis shall be considered satisfied for duplicate units ofsuccessfully operating installations or for replacements of piping systems with arecord of satisfactory service"

To a certain extent, this reliance on performance record for design is maintained in thecurrent B31.1 [2] Section 104.7.2 which allows pressure design to be established through"extensive, successful service experience under comparable conditions with similarlyproportioned components of the same or like material".The reliance on experienced judgment has been recently recognized and formalized formechanical and electrical equipment by the Seismic Qualification Utilities Group [3]. Whileexperience was originally widely used in the field of piping design, today's practice islargely reliant on analysis, despite recent calls for:

"... more independent reviews of piping designs by experienced engineers ratherthan by the current emphasis on design by stress analysis" [4].

This trend to quantify design-by-rule was recognized early on by the ASA B31.1 task force[1, page 148]:"We further agree that the vague guidance [rules] which the [B31.1 ] Task Force feltmore or less compelled to retain is quite sufficient for a Standard of Good Practice.The salient point, however, is that the Piping code may be considered no Iongersuch a standard since it is rapidly being adopted as a Safety Cgde, its rulesbecoming mandatory and enforceable by law".

2


4/12

r

The B31 committee correspondence of the late 19,40s reflects the apparent difficulties ofreconciling safety with simplicity, in design [5]:

.... "The charge of the committee is not a simple one: always remember that this is asat'ety code and not a design code".

DE SIGN-B Y-ANAL YSISStress analysis of piping evolved from graphic methods for deadloads in the 1930s, tostatic quantitative thermal expansion in the 1950s, closed form dynamic solutions forsimple piping configurations and static seismic (uniform building code approach) in the late1950s, limited computerized static and dynamic stress analysis in the late 1960s and f'mally"production" computerized analysis in the 1970s and 1980s [6].

_ Five factors have contributed to the widespread use of simplified (and admittedlyconservative) analysis methods for piping systems:1. Piping systems are composed of an assembly of standard components and fittings,with established flexibilities, stress intensification and allowables.2. The state of loading in piping follows a rather uniform pattern with predominant

bending and torsional moments and usually negligible axial or shear forces.3. Piping systems encompass a large volume of the structural systems and mechanical

components of a power or process plant. This volume calls for simplicity inanalysis methods to reduce costs, schedules, and risks of error.

4. Unlike vessels, piping systems considered individually do not constitute a largeinvestment. The added costs of a detailed analysis of each system, however wouldcontitute a significant investment compared to its material and fabrication costs.

5. Nucleaa"piping systems are designed to postulated severe loading (earthquakes,pipe breaks) for which little experimental or experience data has been assembledprior te,the mid-1980s. Analytical predictions filled the void created by this lack ofdocumented experience data.

Fueled by these factors, simplified "production" pipe stress analysis and pipe supportdesign software flourished in the 1970s and 1980s, with their supporting cast of qualified"stress analysts"; to the point where in the late 1980s piping analysis and design constitutedclose to 2 million of the approximately 11 million man-hours of engineering design for anuclear power plant [7].


5/12

DESIGN-FOR-AN ALYSISInterestingly, while the engineering costs of "production" analysis are commonlyrecognized, another aspect of the emphasis on analysis has not been clearly recognized yet:the appearance of "design--for-analysis"."Design-for-analysis" can be defined as the selection of design features tofacilitate stress analysis. Conversely, it is avoidance of good design features whenthey do not lend themselves to simplified "production" analysis.Following are a few practical examples of some designs introduced, and others avoided, inorder to facilitate analysis.The Snubber - complex and of limited reliability (hence the periodic inspections andtesting), the snubber has however been very often selected (upwards of 2000 in certainnuclear power plants) because it is seemingly easy to "analyze": inactive (zero stiffness) forthermal and deadweight mns and active (finite stiffness) for dynamic run.Solutions to the excessive use of snubbers have been sparked by the issue of new pipingdamping factors in the mid-1980s [9]. Several nuclear plants implemented successfulsnubber reduction programs, with snubber reduction rates upward of 50%. As we lookback at the snubber reduction programs, we recognize that together with the increase inpiping damping, the success of snubber reduction was also due to a careful, deliberateeffort to avoid the use of snubbers as analytical expedients to resolve overstress.The Energy Absorbing Devices - unlike the snubber, promising design concepts(such as viscous dampers or energy absorbing springs) are avoided because their non-linear behavior does not lend itself to simplified "production" analysis.The Rigid Piping System - pipe supports which are added to a system to simplify itsanalysis (using span, equivalent static or dynamic methods) increases its operationalthermal stresses, as well as the potential seismic excitation transmitted from the building.In the same vein, lateral gaps between pipe and pipe support frames have beensystematically shimmed to less than 1/16 inch. This was accomplished to bring thehardware in line with the linear analysis which does not account for gap effects.Interestingly, the tight shims have in tyurn led to a concern, mostly analytical, with theradial compressive contact stress created by the thermal radial expansion of the pipe.

A better treatment of this question would Mlow, and indeed encourage in many cases, aradial gap betweens the pipe and a suppoi_, frame, A I/4 inch gap would allow forimproved thermal performance, while remaining sufficiently small to avoid significantimpact effects. This gap could be shimmed to zero only where necessary, such as atsupports adjacent to pump nozzles.


6/12

i

The Rod Hanger - an efficient and simple support, proven in tests and earthquakes, therod hanger provides pendulum restoring force and energy absorption by compressivebuckling (f_gure 1). However rod hangers are ali too often avoided and at times replacedbecause they are difficult to analyze using linear methods.

'.." 4 "

FIGURE 1. Example of seismic ruggedness of flexible rod hung piping.(San Francisco Earthquake, 1971, courtesy of EQE Engr.)

A more realistic treatment of rod hung piping would account for the lateral restoring forceof the rod. This effect, introduced in several software codes for static loading, can also beapproximated in linear response spectra analysis, as illustrated in figure 2.

.....

leKv

-._ .......

I

Dr:t:Sv

FIGURE 2. Equivalent stiffness for a rod hangerFor small lateral swing angles, it can be shown that +Sv (the seismic vertical load) and Th(the static horizontal load, due for example to normal operation thermal expansion) can beneglected. The equivalent horizontal stiffness provided by the rod hanger restoring capacitycan then be approxinmted by KH= D,,/Ie.

5


7/12

0

As the lateral swing angle increases, the non-linear nature of Ka becomes more evident.For large displacements, a linear approximation KH(SH,le,Dv) must be calculated by aniterative process. However, it can be shown that the relationship KH= Dr/le remains a goodapproximation if the predicted rod lateral swing angle is less than approximately 30 , andprovided the connections at A and B (figure 2) can accomodate the swing.Boundary Anchors - anchors (six-way restraints)are at times introduced to divide apiping system into smaller subsystems, easier to analyze (small computer models). Thisanalytical expedient may adversely affect the system integrity by increasing thermalexp,'msion stresses and introducing welded attachments to the pipe, at the anchor point.The solution to this problem is two-fold:

Where dynamic analysis is required, the model size (dynamic degrees of freedom)should be minimized. The emphasis should be on the dominant first modes ofresponse. This can be accomplished by simplifying the model through the use ofsimple valve mass models, simplified representation of short protrusions such asvents or drains, etc..The rules for dynamic overlap and decoupling should not be arbitrary (such as twosupports irl each direction or 33 Hz overlap span requirements). Welding ResearchCouncil Bulletin 300 [9] provides effective guidance on this subject.

Saddles or One-Way Supports - simple and commonly used in the non-nuclear field,one-way (resting) supports are avoided, or converted to two-way restraints, to eliminate thecomplexities of analyzing uplifting effects.Two methods have been used in the past to attempt to account for "downward-only"supports: I) the s;upport is assumed inactive if the resultant upward load exceeds thedownward operating load; or 2) an average, reduced stiffness is assumed acting in bothdirections. A better approach would focus on the more practical effects of.one-waysupports as they could affec _.the system's structural integrity. To this end, the followingrules are proposed:

One-way support (or rod hangers) may be modeled and analyzed as two-wayrestraints provided the following four conditions are satisfied:1. Uplift, should it occur, would not result in the pipe falling off its supports.2. The pipe spans adjacent to the one-way support do not contain impact

sensitive components, instruments, or fittings.3. The one-way support is not adjacent to equipment sensitive to nozzle loads.4. The one-way support structure will not eail under possible impact load

resulting from uplift. The impactr load may be estimated to be equal to thenormal operating load plus twice the dynamic load, based on a two-wayacting restraint.

6


8/12

t

Large Support Members- Massive support frame sections are installed to react topostulated accident loads while maintaining elastic behavior. A better design would takeadvantage of inelastic design rules (such as Appendix F of the ASME B&PV Code) toallow, and indeed encourage, some plasticity under severe accident loading and hence moresimple and streamlined support structures (figure 3).

FIGURE 3. Example of ductile deformation of a pipe restraint subjectto large seismic impact loads. No damage occurred to thepipe. (Loma Prietta Earthquake, 1989, courtesy of EQEEngr.)

Gap Supports - an ideal configuration of nuclear piping design would consist of simpleconfigurations: deadweight supported, flexible piping with regularly spaced gap supports,lined with shock- absorbing material. The concept however introduces non-linear features(gg.pand clearances) difficult to analyze with simplified methods, and is currently avoided.

_PROMISING DEVELOPMENTSSeveral recent industry and regulatory initiatives are promising and may weil, if successful,reverse the trend towards "design-for-analysis":1. An important development has occurred in the 1980s with the adoption of

experience data in the evaluation of seismic adequacy of mechanical and electricalequipment [3]. The extension of this concept to piping systems should be pursuedto permit greater emphasis on actual failure modes and the use of experience basedrules [8].


9/12

w0

2. The Pressure Vessel Research Council has sponsored important studies to draw onthe lessons learned in the past 15 years of nuclear industry practice [9, 10]. TheWelding Research Council Bulletin 353 points out [10, page 4].

"... the piping analyst/support designer should not put excessive emphasison the analysis itself, but rather should focus on a good design of theoverall piping system. A detailed or rigorous analysis does not necessarilyinfer a good design"

3. The Pressure Vessel Research Council has also sponsored research on seismicdamping [9] which has resulted in the ASME Code Case N-411, increasing thedamping values, up to 5% in the low frequency range. There remains some debatein the engineering community, on whether damping should be 3%, 4%, or 5%. Thequestion pales however in comparison to the significant benefits of snubberremoval programs completed in the commercial nuclear industry.

4. Through a better understanding of "leak-before-break" and the elimination of"arbitrary intermediate breaks", the industry and regulatory focus has turnedtowards greater attention for operating loads (erosion and corrosion, thermalstratification, fatigue crack growth) with a more reasonable approach towardshypothetical non-mechanistic behavior (guillotine breaks, whip and jet loads, etc.).

5. The Electric Power Research Institute (EPRI) has sponsored significant researchprograms for the understanding of piping failure modes under extreme dynamicloads [11, 12]. The EPRI work greatly contributes to our knowledge of failuremodes under severe, non-linear, dynamic loading, and helps explain and confirmthe earthquake experience knowledge [13].

6. "HaeNuclear Construction Industry Group and EPRI have set the foundation for areturn to design-by-rule, with the reintroduction of support spacing concepts andrules of good practice in a recent guide for the design of small bore piping [14].

CONCLUDING REMARKSIn the field of nuclear piping design, there has been, in my opinion, a trend in the 1970sand 1980s towards "design-for-analysis", the selection of design features on the basis ofsimplicity of analysis. Snubbers and tight fitting supports are but some examples of"design-for, analysis". Conversely, and as significant, optimum hardware designs havebeen avoided at times, when too co.rr.plexto analyze in a "production" mode.The price paid in efficiency and operational penalties has prompted a serious rethinking ofthe piping design process. Promising industry and regulatory initiatives are underwaywhich may reverse the trend and reintroduce a certain measure of "design-by-rule", basedon proven and simple design features, as intended by the original piping codes.


10/12

a

This paper_:wa_ prepared in connection with work done under Contract No. DE-AC09-89SR18035 with the U.S. Department of Energy. By acceptance of this paper, thepublisher and/or recipient acknowledges the U.S. Government's right to retain anonexclusive, lroyalty-free license in and to any copyright covering this paper, along withthe right lo reproduce and to authorize others to reproduce ali or part of the copyrightedpaper.

REFERENCES[1] A,R.C. Markl, "Piping Flexibility Analysis" Transactions of the ASME, volume77,1955[2] ASME Code for Pressure Piping, B31.1-1_,o9 Edition.[3] Seisrrtic Qualification Utilities Group, "Generic Implementation Procedure for

S;eisrnic Verification of Nuclear Plant Equipment", EPRI, Septembex '.990.[4] R.P. Kennedy, "Consulting Paper on Seismic Design of Piping", StructuralMechanics Associates, SMA 12209.08, December 1983.[5] ]Letter, F.S.G. Williams, ASA B31 committee chairman to committee members,

October 1, 1949. ASME Archives, United Engineering Center, NY.[6] G.C. Slagis, PVRC Project 89-6, "Basis of Current Dynamic Stress Criteria for

Piping", October 1990.[7] J.D. Stevenson, PVRC Subcommittee on General Topics Section III and XI.

Position Paper GT-03, "Economic Impact", January 1990.[8] G.A. Antaki, G. Hardy, and F. Rigamonti, "Screening Criteria for the Verification

of Seismic Adequacy of Pipint Systems", Pressure Vessel and Piping Conference1991.[9] WRC Bulletin 300, "Technical Position on Damping and on Industry Practice",December 1984.[10] WRC Bulletin 353, "Position Paper in Nuclear Plant Piping Supports", May 1990.[11] D.J. Guzy, "Piping _ d Fitting Dynamic Reliability Program", Proceedings of the

14th Water Reactor Safety Information Meeting, Washington, D.C., December1986.[12] K.L. Merz and Y.K. Tank, "Dynamic Testing of Eroded/Corroded Piping

Components", ASME PVP Conference, 1990.

9


11/12

.i,

[13] P.J. Yanev, et al., "Piping Seismic Adequacy Criteria Recommendations Based onPerformance During and After Earthquakes", EPR/Research Project RP-2635-1,1987.[14] J.D. Stevenson and Associates, EQE Engineering, "Procedure for Seismic

Evaluation and Design of Small Bore Piping (NCIG-14)",EPRI NP-6628, ElectricPower Research Institute, April 1990

10


12/12

Antaki Design for Analysis

Documents