Chemical Engineering - June 1997 - Pipe Design for Robust Systems

8/13/2019 Chemical Engineering - June 1997 - Pipe Design for Robust Systems

1/9

COVER STORY

Stress analysis makes for asolid pipe system design.Software can help

After the piping and instru-mentation diagram (P&ID) ofa process design has beencompleted, the next step is

stress analysis of the piping network.In past years, engineering or operatingcompanies had staffs of layout andstress designers. Today, both thesefunctions might be in the hands of theindividual engineer. Either way, a solidunderstanding of pipe stress, and howto handle it, will lead to better plant de-sign and more reliable operation.

Engineers may perform pipe stressanalyses on a daily or occasional basis,or review the pipe stress analysis ofothers. Whatever their role, most ofthese individuals have only a basic un-derstanding of the topic. And manymay be unaware of the software toolsthat can automate many of the analyti-cal steps (Box, p. 92).

FIGURE 2. Thermal expansion affects the modeling of pipe joints such as this bend.Actual support points (directly under A; at C) produce spurious results. The modelercan substitute a square bend, or a rigid support at E to complete the modeling work,then specify the actual component afterwards

When to analyze

One of the most difficult decisions plantpersonnel face is whether or not to ana-lyze an existing or new piping system.It is often hard to assess the point atwhich a piping system should be field-routed or when a full analytical solu-tion is required. Although there are nosimple answers to these questions, hereare some conditions under which pip-ing analysis is advisable: Piping is attached to load-sensitiveequipment or is carrying Category M

Michael Bussler,Algor, Inc.,and Tony Paulin,Paulin Research Group

fluids (hazardous chemicals, as definedby ASME B31.3 rules), and the design

temperature is > 250F The pressure exceeds the maximumpressure for an ANSI Class 2500 B16.5fitting The system temperature is > 400F The system carries gas that hascooled to a liquid state The product of the pipe outside diam-eter (in in.) times the pressure (in psi)is 1,157 The system pressure is > 3,000 psi The system uses Glass ReinforcedEpoxy (GRE) pipe

The piping connects to rotatingequipment The system uses one or more expan-sion joints

Most piping codes, and perhaps 90cof all pipe stress analyses, involve three

principal loading types: (1) sustainedloads, such as pressure and weight; (2)expansion loads (i.e., from thermal con-ditions); and (3) occasional loads, suchas from wind and earthquakes.

Other types of loads include thosecaused by transient fluids, ice andsnow, ship or platform pitch and roll,explosion loadings, pressure loads,frost heave, fault movement, fluidsloshing and through-wall thermal ef-fects. These can all be analyzed, but aretypically reserved for more experienced

pipe-stress analysts.In general, an analysis involves sat-

isfying code requirements for stress inthe pipe, and manufacturers' require-

8 4 CHEMICAL ENGINEERING / JUNE 1997


2/9

ROBUST SYSTEMSments for loads on equip-ment, flanges, vessel noz-

es and the like. Basicode compliance involvesunning a computer pro-ram to compute a stress,nd making sure that it isss than a code calculatedlowable, i.e.,

Calculated Stress < Al-owable StressStresses in a piping sys-m are calculated by the

ollowing general-purpose

P = design pressure, psig

D = pipe diameter, in.

t =pipe thickness, in.

i = Stress Intensification Factor (aken from the piping code for the ap-licable geometry, i.e., bends or tees)M = square root of the sum of the

quares of the three moments acting atny point in the piping system. Sus-ained expansion or occasional loadsan cause the moment

Z = section modulus of the pipe [ap-

roximated by rr(r2)t; (r) is the midsur-

ace radius of the pipe]If pressure and weight cause the cal-

ulated bending moments (M), the al-owable stress is Sh, where Sh is the hotllowable stress from the applicableiping code. If the calculated bending

moments are due to thermal expansion,

he allowable stress is f i 1.25(Sc + Sh) -), where:

f = cyclic reduction factor (approxi-

mated by 6N-0-2) and N = the number of

ull-range thermal loading cyclesSc = the cold allowable stress from

he piping codeSh = the hot allowable stress from theiping codeSl = the weight and pressure stress in

he piping system at the point underudy

If the calculated bending momentsre due to pressure, weight, wind orarthquake loads, the allowable stress isqual to k(Sh), where k is a value giveny the piping code. For example,

BLOCK BY BLOCK WITH FEA

Finite element analysis (FEA) is amethod for testing how a part, product

or system will behave under real-world

conditions. FEA software performs theanalysis by breaking a system down intotiny blocks called finite elements. For ex-

ample, the grid on the pipe in Figure 9shows the breakdown of evaporator pip-

ing into finite elements.

FEA software tests how the pipe willen-dure heat, tensile or other types ofstress that occur during normal operationby applying mathematical computations,based on engineering theory, to each of

the finite elements. Once the software hasperformed all of the computations, it pro-vides numerical data and creates a colordiagram of the stress, such as Figure 1,which shows the amount of stress in the

different areas of the pipe. Red representshigh stress and dark blue represents low

stress. Similarly, FEA software tests thefluids within a piping system to predicthow chemicals of various viscosities will

behave in the real world. O

in ASME B31.3, a common piping codereference, k = 1.33.

The above equations are simplifica-tions of the actual rules given in thepiping codes, and serve to illustrate the

intent of the stress-evaluation proce-dure. When performing or reviewing astress analysis, an engineer shouldcheck each of the following allowablesto be sure of compliance:

Sustained (weight andpressure) loads toprevent the system fromcollapsing or from experi-encing excessive distortion (stress < Sh)Expansion (thermal)

load To avoid fatiguefailure, when cracks form,grow and then fail cata-strophically (stress < f[1.25(Sc + Shl - SI)

Occasional (wind and

earthquake) loads toprevent the system fromcollapsing or from experi-encing excessive distortion (stress < 1.33Sh)

Markl's landmark paper,

"Piping Flexibility Analysis," [1] estab-lishes most of these rules. It covers anumber of the assumptions used in theformulation of the code rules that arenot found in any of the code documentsthemselves, and is an excellent basic

reference on the subject.

Now, address pressure factors Mostpiping system designs start well beforethe plant or unit is laid out. Theoperator and the engineering designfirm agree on a piping specification togovern the basic pressure and materialdesign of the piping system. Pipingspecifications are primarily designed toensure that the piping system willwithstand the expected pressure and issuited for the intended process. Piping

specifications are typically writtenaround an applicable piping code. Themost common piping codes, arrangedby nation, are:

United States (ASME):

B31.1 - Power

B31.3 - Petrochemical

B31.4 - Oil and Slurry Pipelines

B31.5 - Refrigeration

B31.8 - Gas PipelinesB31.9 - Building Services PipingSection III -Nuclear Piping: NB (Class

1); NC (Class 2); ND (Class 3)Canada:

Z183 - Canadian Oil PipelinesZ184 - Canadian Gas Pipelines (

Continues)

CHEMICAL ENGINEERING / JUNE1997 85


3/9

TABLE 1.

Adapted fromASME B16.5, thistable providesguidance onthe allowablepressure ratingfor flanges andfittings based onmetal type (material group),temperatureand size

Europe:

BS806 British Power Piping

FDBR German Power PipingOther relevant standards include the

National Electrical Mfrs. Assn. SM 23;Standards of the Tubular ExchangerMfrs. Assn.; and Bulletins of the Amer-ican Petroleum Institute (API). Themost widely used piping code is ASME

(or American National Standards Insti-tute, ANSI) B31.3, Chemical Plant andPetroleum Refinery Piping.

Chapter 2, Part 2 of B31.3 gives theminimum requirements for pressure in

piping components. Piping specifica-tions are designed around these Chap-ter 2, Part 2 pressure requirements.Typically, the guide's requirements are

COVER STORY

satisfied well before any stress analysis(Chapter 2, Part 5) is performed. How-ever, analysts should not assume that asafe pipe-f lexibility analysis means a

safepressure analysis.Piping flexibility (or stress) analysistests pressure to a small extent but isprimarily concerned with weight, ther-mal and occasional loads. Most pipestress programs do not satisfy thesePart 2 requirements, since the pro-grams are primarily based on a differ-ent part: Chapter 2, Part 5.

A novice analyst might incorrectlyassume that a piping system satisfiesB31.3 allowables for thermal weightand pressure because a pipe stress pro-

gram indicates that B31.3 is satisfied.For shop- or field-fabricated intersec-tions, engineers must design these forpressure in addition to performing thepiping flexibility analyses.

This situation arises most often whena plant makes minor modifications toexisting pipe work. The plant engineerresponsible for the modificationaccepts the design of the tie-ins basedonly on a flexibility analysis, withoutchecking Part 2 pressure requirements.This is a significant mis-take. The most

dangerous problems occur when thepipes being connected have largediameters with small wall thickness, orwhere the branch pipe centerline is not90 deg to the run pipe. The engineershould perform simple handcalculations to ensure that the Part 2requirements are met.

Software speeds analysisPiping analysis software is available ina range of capabilities, from completingsimple drawings to performing complex

flow and stress studies (Table 2). Tech-niques such as finite element analysishave proven useful. The software cansave engineer's time and effort on thesystem's analysis.

Among the software's analytic capa-bilities are: static and dynamic analysisto calculate displacement or support re-actions; loads caused by time-depen-dent phenomena such as water ham-mer; code compliance; nonlinearreactions; and the integration of struc-tural components in the piping design.

Once engineers decide to perform for-mal stress calculations, they should re-alize that they will spend much of their

TABLE 1. PRESSURE-TEMPERATURE RATINGSFOR CLASS 150 PIPING

MaterialOrbup No. 1.2. 1.4 1.5 2.1 2.4 2.7

Carbon steels Alloy steels Austenitic steels

C-1/2Mo Type 304 Type 321 Type 310

Temp., F pressure, psig

-20 to100 290 235 265 275 275 260

200 260 215 260 235 235 230

500 170 170 170 170 170 170

750 95 95 95 95 95 95

1,000 20 20 20 20 20 20

FIGURE 5. A stressanalysis of a plasticpiping network showsdisplacements atelbows and the ends ofruns. Typically, suchnetworks areconstrained withanchors at both endsto limit thermal andpressure expansion

86 CHEMICAL ENGINEERING / JUNE 1997


4/9

TABLE 2. READING PIPE STRESS PROGRAM OUTPUT REPORTS

Output Discussion: The three most common types of reports printed from pipestress program are shown above. The deflection report shows the displacementsof the system in the selected light-handed cartesian coordinate system. Y or Z istypically vertical, and the other axes are selected by the analyst to ease the inputof data. The forces and moments on restraints show the loads the piping exerts onsupports. The stress report shows how close the system stresses are to the applica-ble Code allowable.Load : Dead Weight + Pressure 1 + Thermal 1

System Deflections

PointName

Displacements/in.X Y Z

Rotations/degreeX Y Z

5810a10b15

0.000

0.000

0.000

0.039

0.039

0.000

0.001

0.001

-0.016

-0.016

0.000

0.013

0.013

-0.001

-0.001

0.000

-0.001

-0.001

0.152

0.152

0.000

0.003

0.003

0.210

0.210

0.000

0.005

0.005

0.118

0.118

Deflection:

1) Verity displacements in the vicinity of all supports, particularly around supportswith friction and gaps, and directional characteristics, i.e., +Y supports.2) Check large horizontal displacements and be sure that they are in the correct di-rection and of approximately the right magnitude.

Forces and Moments acting on Restraints

PointName X

Forces/lbY Z

Moments/in.-lbX Y Z

54075

-4180

-222

3233-6,075

435

-2,9090

2,736

-21,459 19,6090 0

-30,408 -26,186

32,9370

4,824

Force on Restraints:

1) Make sure that loads on terminal points are reasonable:a) Will the support carry this much load without failing?b) Will the support deflect under load, distributing the load somewhere else?c) Is the support "rigid," or does it have a realistic stiffness?

2) Check large horizontal loads and be sure that steel supports will carry large

loads.3) At locations of large loads, make sure that the inclusion of friction will not signifi-cantly alter the solution.

System Stresses (ASME B31.1)

In-Point PlaneName SIF

Out-Plane

SIF

Section Stresses/

Longitu. Princ. Code Allow.Modulus psi

5 1.00 1.00 22.03 1,401 1,589 2,248 3,164 43,75(.:

8 1.00 1.00 22.03 1,401 1,389 2,143 2,670 43,750,

8 6.25 6.25 22.03 1,401 7,267 7,360 12,831 43,750

System Stresses:

1) Check the stresses at terminal points. There, the stress intensification

are 1.0. If a terminal point is another pipe or a vessel the SIF will not be 1.0.2) If stresses are close to allowables, be sure that other safety concerns area) Does the system cycle more than 7,000 times in its life?b) Is the system heavily corroded?c) Is the system temperature greater than 700`F. fd ) I s t he D / T f o r t he s y s t em g r ea t e r t han

80? e ) I s t he r e any pos s i b i l i t y o f dy nam i cl oad i ng due t o f l u i ds ?

active graphical capabilities so thatusers can verify that the physical rep-resentation matches the input data, itwill not identify all user errors. Themost common types of errors made are: Omitting data about weight of insu-

lation, valves and operators ncorrectly entering data about howpipes intersect

Oversimplifying or incorrectly enter-

ime inputting data or adjusting com-uter models of the piping system. En-ineers often rush to meet a deadline orring a quick end to the tedious task ofnputting data. Errors made duringnput account for a majority of analysis

roblems. Often the software will pro-uce results even if the engineer hasnput inaccurate data.

Although software usually has inter-

ing piping boundary conditions Ignoring or oversimplifying pipeshoes, guides, anchors and trunnionsupports

Using program algorithms withoutquestioning underlying assumptions,such as spring hanger design

To reduce errors, the pipe stress pro-gram user should follow these guide-lines:List and check the input. Don't as-

sume the input is correct just becausethe program does not report fatal er-rors. The first time that you list andcheck an input, and don't find an erroris when you can stop checking inputsfor errors

Don't just review code stresses inthe output report. Look at printeddisplacements (one of the factors lead-ing to the stress determination), mak-ing sure that they are reasonable, andthat the system is deflecting in an ex-pected manner. Frictionless supportswill occasionally allow for ridiculoushorizontal movementsCheck restraint reports closely.

Make sure that excessive loads do notexist at steel support locations. Wherelarge loads exist, decide if friction

should be considered in the analysis, orif an error in the input is causing theloads. A close review of displacementand restraint load reports often revealsareas of the model that are incorrectCheck loads on all rotating

equipment nozzles. To verify these,use the manufacturers' guidelines, orindustry accepted guidelinesCheck displacements and rota-

tions at expansion joint ends. Makesure these displacements do not exceedmanufacturers' limits for the number of

design cyclesReview the graphic results

closely. Look at displaced shapes to besure that they make sense. Make surethat high stresses occur in the correctplaces.

Most stress-analysis programs fol-low an easily manipulated, trial-and-error procedure. Usually, engineersneed only make a few iterations to ad-

just supports or design springs so thatstresses are within allowables.

Linking to rotating equipmentUsually the most difficult work for thestress analyst is meeting the require-ments of rotating equipment and end

CHEMICAL ENGINEERING / JUNE1997 8 7


5/9

GURE 6. (above) The effects of watermmer can be seen In the bend andsplacement of this field pipe

FIGURE 7. A calculation of waterhammer's pressure wave shows Its

amplitude and duration

nnections. Pumps, turbines and com-essors are very sensitive to loadsom attached pipe, and a single appli-

tion of an overload on a pump or tur-ne can result in leaking seals, failedafts, excessive wear and vibration. Ape having only a 4,000-psi stress (nding stress, not fluid pressure) cansily overload a pump nozzle.To ensure compliance with rotating-uipment or end-connection codes,fer to NEMA SM23 (for steam tur-nes), API 610 (for pumps), and API7 (for compressors). Each of thesecuments gives allowable loads for dif-

ent sizes of piping inlets and outlets.ost computer programs include theseuipment specifications as part ofeir standard evaluation procedures.addition, some manufacturers of ro-ing equipment provide their ownidelines or use a multiplier of NEMA

RULES OF THUMB FORPIPE DESIGNERS

Steel pipe is very forgiving. At a

low number of cycles it will almostnever fail due to thermal expan-sion. Plastic pipe is certainly not asforgiving as steel pipe, so exerciseextra care Rotating equipment and expan-sion joints are not forgiving at all!When there is little room for error,contact the expertsOnly about 10% of the pipinganalyses require an expert's re-view. Competent chemical or me-chanical engineers can handle theremaining 90% as long as theyexercise care and prudence in theuse of the computer results anddesigns

COVER STORY

or API values. Use the manufacturer'sinformation whenever it is available.

It is more difficult to find a line rout-ing or support configuration that will

satisfy rotating-equipment allowablesas the temperature increases. Most ro-tating equipment codes have two typesof allowables:Individual nozzle limits provide

standards for each individual productnozzle on the rotating equipment. Forexample, the inlet to a pump may havean axial allowable of 2,000 lb, while theoutlet has an axial allowable load of 1,750 lb. Another approach to thisanalysis is to evaluate the suction side

of the pump independently from thedischarge side.Resolved load limits ensure that

forces and moments on the product noz-zles do not act in combination to createexcessive cumulative loads or momentson the equipment casing or base. Theanalyst must resolve all loads acting onall product nozzles attached to theequipment about a single point on theequipment body.

There are limits to the resolution ofthese loads. Since pipe analysts do nottypically study inlet and outlet pipingsystems together, they must rememberthat loads from both lines must satisfymost equipment requirements. For ex-ample, use suction loads and dischargeloads to evaluate resolved load limitsfor pumps.

When adjusting supports in thevicinity of rotating equipment it is im-portant to remember that:

1. Supports should not require con-struction or installation tolerances be-

yond the capability of the shop or field.Do not add supports where steel is notreadily accessible. Avoid moment re-straints, and translational restraintsrequiring precisely set gaps

2. Properly designed, horizontal loadsshould not cause deflection in theattached steel, which negates the sup-port's usefulness. Steel constructionsdesigned to keep piping loads awayfrom rotating equipment will transferthe load back to the rotating equipment

if excessive deflection occurs. The pipestress analyst must determine a tolera-ble amount of deflection of the steelconstruction

3. Evaluate the effect of friction atsupports on large lines, or on highly


6/9

A CAUTIONARY CHECKLIST FOR SOURCES OF PIPE FAILURES

1) The majority of catastrophic failures are due primarily to operational failures andhave little to do with pipe stress analysis2) Most catastrophic failures are due to a sequence of unfortunate events, which, if oc-curring individually, would have not caused a major failure3) The most common consequences of a poorly executed piping system analysis are:

a) Local overstressed areas in the piping system (usually has no further implicationsor decrease in system safety)

b) Supports that will be lifted off, or springs that are outside of their intended oper-

ating range (also causing little real decrease in system safety)c) Poorly functioning rotating equipment, resulting in leaking seals, alignment prob-

lems and other equipment difficulties (this situation tends to be the most common andmost expensive result of a poorly executed pipe stress analysis)

d) Leaking flangese) Failed pipe (rare)

4) Most catastrophic failures are due to a leak of dangerous material. Leaks commonlyoccur where:

a) Process conditions are not properly evaluated and controlledb) Thermal or pressure cycling of the system is not properly evaluated or analyzed (

the system has major load changes more than 7,000 times during its life)c) Welding procedures and material selection are applied improperly, e.g., high-

hardness welds in sour gas service

d) Incorrect pre- or post-weld heat treatment that creates brittle or chemically weakwelds

e) Corrosion is not adequately controlledf) Workmanship on welded or FRP joints is poorg) Large-diameter flanges are placed at locations in the system where there are high

bending momentsh) Bolted joints exist where the bolts are 1.5 in. or smaller and might be overtight-

ened, or where they are 1.75 in. or larger and might be undertightenedj) Erosion occurs at changes in directionk) The piping system has an improper design and contains an expansion joint I)High-temperature piping loads are not properly supportedm) Fluid loads are not properly evaluated and vibration occursn) Improperly balanced rotating equipment produces high-frequency vibration at

stress-intensified components, such as small temperature or pressure takeoffs orblinded valves

o) Rapid changes in temperature are not evaluated at discontinuities

During the design phase, the following situations should be monitored closely,and the following practices maintained:

1) D/T (pipe diameter divided by its thickness) greater than 100. Piping systems in thisD/T range are very thin and may perform more like thin-walled pressure vessels thanpiping systems. Some piping code rules are not well suited for piping systems in thisD/T range and extra caution is warranted2) If the highest design pressure is negative (vacuum), extra caution should be takenso that a buckling situation, not evaluated by most pipe stress programs, is avoided

3) Solve dynamic problems quickly.A vibrating system can easily undergo millions ofcycles in hours or days. Failures in overstressed dynamic systems can occur veryrapidly

4) Review old design drawings before making modifications to a piping system. Initialangular offsets at hanger locations, cold spring, or initial restraint-gap settings can beoverlooked and not properly reinstalled after work has been completed5) Look carefully for corroded areas, cracks or poor workmanship during maintenanceopportunities. It may be the leak from a small crack that sets off a chain reactionthatresults in catastrophic failure, fire or loss of life6)Consult experts when any decision is at all questionable. Developing a good work-ing relationship with a local pipe-stress expert can be invaluable for finding practical,cost-effective, safe solutions to most plant piping problems7) If there is a question about why a result from a pipe stress program looks a certainway, make sure to ask the question and get an answer. Piping program vendors can

often yield useful tips and insights O

oaded supports. Where friction is un-desirable, the designer might use poly-meric friction plates or metal slidingplates to reduce friction. Some analystsalways use friction in their analyses;

others never use it4. Take the millwright's practices into

account when designing springs,tanchions, and other supports near ro-ating equipment. Some spring designs

produce unbalanced loads on equip-ment when the system is in a cold con-dition. Millwrights may improperly ad-ust or locate these springs in the cold

condition so that the system is perfectlybalanced. Typically, the designer did notntend for the balance to occur in the

cold condition

Meeting at the end pointsf a piping system does not attach to ro-ating equipment, it probably connects to

a pressure vessel, a heat exchanger, oranother piping system. In these cases, atress increase will occur at the end orerminal point in the system. There, a

geometric discontinuity exists due to thentersection of two cylindrical shapes

meeting to form an opening.While some analytical software takes

hese terminal points into account; oth-ers do not, so the designer should exer-cise care. Stresses at terminal points areoften the governing forces in a de-sign.Some software programs use standardsfrom the Welding Research Council (WRC) Bulletins 107 or 297 to determinehe terminal point stresses. Analysts

whose software doesn't per-form eitherof these calculations can use the WRCtandards to manually compute stress aterminal points.WRC 107 and 297 methods for stress

and stiffness at nozzle connections arebased on simplified deformation theo-ies, adjusted by tests and experience.

Because of their ready availability inmany computer programs, these meth-ods have been used for wide varieties ofgeometries, well outside of their origi-nal intended scope. Users should becautious using them when the d/D ratios greater than 0.5, or when the t/T ratios one or less. In general, the WRC

documents will be conservative fortress, although possibly by hundreds of

percent.WRC 297 is the only WRC document

hat predicts stiffnesses of nozzles in

CHEMICAL ENGINEERING / JUNE1997 8 9


7/9

COVER STORY

TABLE 3. NOZZLE LOADINGS FROM ROTATING EQUIPMENTPumps and other devices put a stress (moment) on the interface between the device and piping.The following table, based on API Standard 610 (Table 2), provides stress estimates based on nozzleflange diameters

Nozzle flange nominal size, in.Force/moment 2 4 6 10 16top nozzle

FX 160 320 560 1,200 1,900

FY 200 400 700 1,500 2,300FZ 130 260 460 1,000 1,500FR 290 570 1,010 2,200 3,300

Each side nozzle

FX 160 320 560 1,200 1,900FY 130 260 460 1,000 1,500FZ 200 400 700 1,500 2,300FR 290 570 1,010 2,200 3,300

Each end nozzle

FX 200 400 700 1,500 2,300FY 130 260 460 1,000 1,500FZ 160 320 560 1,200 1,900FR 290 570 1,010 2,200 3,300

Each nozzle

MX 340 980 1,700 3,700 5,400MY 260 740 1,300 2,800 4,000MZ 170 500 870 1,800 2,700MR 460 1,330 2,310 5,000 7,200

F = force, in Ibr; M = moment, in ft-Ibr; R= resultant. X, Y and Z refer to the orientation of a nozzle;

Z is into or out of a nozzle; X is perpendicular to Z but in the same horizontal plane; and Y is in thevertical plane. FR is the square root of the sum of the squares of the individual forces.The "each nozzle" category refers to bending moments on a nozzle; These are not dependent onnozzle location, i.e., top or side. (Data from API Standard 610, Table 2, p. 10.)

linders. WRC 297 can be nonconserv-ve for stiffness. Some computer im-ementations of WRC 297 may predictffness values hundreds of percent toow. The worst errors occur for out-of-ane stiffnesses, and when the para-eter (d/D)v(D/T) is greater than one.e WRC 297 user is encouraged torify stiffnesses generated by compar-g them to a finite element result, or toedictions given in ASME Sec. III

B3685.7.Piping systems are attached to ves-s either by welding or by flanges.

pe designers use welding throughoutentire piping system when leaks

nnot be tolerated under any circum-nces. When small leakage can be tol-

ated, most pipe designers use flangents because maintenance is easier.

Most flanged joints are as strong ase attached pipe. A good software pro-am will enable the user to describe

e properties of these attachments inses where the joints are not as strongthe pipe. Some of these exceptions

e systems that have:

Fiberglass-reinforced plastic flangesSoft gaskets made with synthetic fiber

Ring-type joints

Flange components made of cast iron Larger (>12 in.) flanges (which arenotoriously susceptible to leakagecaused by external moments)

Flanges should be located in the pip-ing system at points of small bendingmoments, when possible. ASME B16.5gives pressure and temperature ratingsfor flanges of various materials. It isnot uncommon to see bending mo-ments and axial forces converted toequivalent pressure and a comparisonto B16.5 made. This comparison pro-vides a safety factor of approximatelyfive to eight.

How plastic pipe differsThe use of plastic piping systems is in-creasing because of declining pricesand improved corrosion resistance.Plastic piping systems differ from othersystems when pressure or temperatureis a concern. Plastic systems often ex-pand as much or more due to pressureinstead of temperature. Piping analystsshould not assume that knowledge ofsteel piping systems transfers to a plas-tic system.

Plastic piping systems fail most often

at flanges or at connections of pipe tothicker components, such as bends andtees. Many of these failures occur be-cause of incorrect support or because ofimproper joint makeup. Others occurbecause engineers do not properly eval-uate water-hammer events. Manymanufacturers are developing theirown solutions to these problems asmore and more plastic systems arebeing used.

What fluids can doMost pipe stress analysts are only con-cerned with a fluid's temperature andits specific gravity. These analysts usu-ally leave issues like flow velocities,valve closure rates, pump trips, chemi-cal decompositions or resonant acousticvibrations to the specialist, who iscalled in only when an extraordinaryproblem is discovered.

As new software becomes available,

the pipe stress analyst is becomingmore involved in transient as well assteady-state fluid-flow problems.Newer fluid programs use the 3D struc-tural mesh of the piping system to con-struct a 1D mesh of the fluid system.

Before the software's advent, the


8/9

process engineer usuallyperformed the fluid cal-culations. The mechani-cal engineer (pipe stress

analyst) was rarely con-cerned with flowrates orpressure drops, sincemost piping systems are

not sensitive to the

slight pressure drops caused by the ad-dition of an expansion loop or extrapipe and bends. Now, piping engineerscan involve themselves in the varietiesof fluid problems because they havetools that will minimize the timeneeded.

Pipe stress engineers should be most

concerned when fluid pressure pro-duces mechanical loads on the pipingsystems. This classical situation canhappen when:

A valve opens or closes quickly

Two-phase flow occurs

Notable flow oscillations exist

There are long liquid lines There is piping in and around com-pressors and pressure-relieving devices

Spring hangers hot and cold

Spring hangers are important in pipingsystems because they support theweight of hot piping as it thermally ex-pands. If a rigid support is used in placeof a spring hanger, it will exert a forceback on the pipe, creating a stress onthe pipe or a load on an equipment noz-zle when the pipe expands.

Because the spring hanger is flexible,it moves with the pipe as it expands.The spring load will change slightlyfrom this movement, but this loadchange will not be detrimental to the

piping system, and creates signifi-cantly less load than a rigid support.

Although a design with spring hang-ers may be accurate on paper, improperinstallation can defeat its usefulness. Itis not at all unusual to walk through afacility noticing springs that are notcarrying loads or that are bottomed out.Piping engineers need to provide spe-cific instructions with any field-ad-justed equipment, such as a load flangeor a turnbuckle on a spring support.

Engineers often misapply constant-

effort springs and place an excess num-ber of these springs where they are notneeded. They often make these mis-takes when they are concerned about

temperature, and thus support the pip-ing system over a large part of itslength using springs.

In these cases, engineers don't realizethat once a line turns in the horizontaldirection, considerable flexibility isgained by each pipe run between sup-ports. There is a point in a horizontalrun of pipe where entering horizontalruns will absorb any entry or exit verti-cal thermal expansion. Intermediate lo-cations simply do not require springs.

Even though springs are designed tocarry weight through a specified dis-tance, the proper use of springs imme-diately adjacent to equipment nozzlessolves many rotating equipment prob-lems. In this case, the spring carries thespecified weight exactly, even thoughthere is little movement in the support.Rigid supports around hot rotatingequipment if not properly in-stalled can cause significant problems. Thedesigner should exercise considerable

caution when analyzing and designingthese installations.If a designer fails to incorporate sup-

ports with axial or horizontal stops forhot piping systems, the pipes maysometimes move slowly around on thepipe rack, contacting other pipes orsteel. Between two fixed anchors, mostpiping systems have a location wherezero thermal movements occur in oneor more directions. Engineers shouldplace supports at these thermal nodes inthe computer model to provide fixed

stability to the system and to preventthe piping from moving.

For a hot pipe that cycles, it is best toinstall guides and limit stops wherever

possible. Use long pipe shoes or othersupports wherever excessive axial dis-placement is expected. (It is very em-barrassing to have a 12-in. pipe shoeslide off a steel support!)

Cold spring is the practice of pur-posely fabricating certain sections ofpiping to be either too short or too long.Cold spring will never improve a pipingsystem's response to thermal stress (forB31 code applications), but it can re-duce operating loads on rotating equip-ment, and can reduce creep problemscaused by hot stresses.

Some operating companies do notpermit cold spring to be used in a pip-ing system design. Cold spring canoverload equipment in the cold condi-tion and engineers can easily forgetabout it years later when they makemodifications. In general, it is best toleave cold spring out of a piping systemcalculation unless nozzle or supportloads cannot be reduced any other way.

Nozzle stiffness is the degree of flexi-bility at terminal points in vessels andheat exchangers. It permits reducedpiping loads at these locations becauseof an inherent flexibility in the shell ofthe vessel or heat exchanger. It can alsosignificantly affect force and momentdistributions in large-diameter, thin-walled systems or in short, stiff sys-tems that have no inherent flexibility.

Users employing nozzle stiffnesses toreduce end-connection loadings shouldbe aware that some implementations

can produce inaccurate results, compa-rable to the situation described on p.90. If engineers use nozzle stiffness toproduce a significant change in final re-

CHEMICAL ENGINEERING / JUNE1997 91


9/9

COVER STORY

suits, they shouldcompare the stiffnessvalues to an alterna-tive source. The mostreadily available infor-mation on nozzle flexi-bilities comes from:WRC 297, NB3685.7,and some Imite-ele-ment analysis pro-grams.

Expansion joints areflexible portions of apiping system for re-ducing loads or ab-sorbing thermal ex-pansion. Engineers can

design expansion

joints ,in a variety of -10ways and can changetheir function to take lateral or axialloads. Expansion joints are often themost critical part of a piping system.Extra care should be taken whenusing expansion joints in impor-tant piping applications. Most ex-pansion joint manufacturers are veryhelpful when it comes to providing de-sign guidance and recommendations

for use of their products.

Hot rustyThe effects of corrosion and high tem-perature on pipe stress analysis isamong the least quantitatively under-stood of all stress related effects. Testson piping fittings usually are done onessentially noncorroded fittings. Theeffect of corrosion on fatigue varieswidely and may be significant. There isno quantitative way to evaluate theseeffects today.

When a system is subject to intenselycorrosive conditions, and undergoes asignificant number (several thousand)of thermal or pressure loading cycles,extreme caution should be used in thedesign. If a particular pipe material andchemical medium has not beenpreviously used, metallurgical study ofthe interaction of corrosion and fatiguefor the particular combination of fluidand medium may be warranted.

In most pipe stress programs todaythe corrosion allowance specified is re-moved uniformly from the wall thick-ness of the pipe before stress calcula-tions are made. Depending on thepiping code employed and the particu-

lar program op-tions used, cor-rosion may not beremoved, may beremoved fromsustainedstress cases only,operating casesonly, thermalcases only, orfrom anycombination ofthe load cases. Itis a wise user whoquestions theapplication ofcorrosion in a

particular soft-ware program. Itis the prophet

who knows what to do with the answer.High temperature in piping implies

systems that operate in the creep rangefor the material. Material creep in-volves a complicated nonlinear interac-tion between temperature, stress andtime. High-temperature rules for pip-ing and pressure vessels are based ononly the simplest of material laws.

Creep phenomena usually becomesignificant at about 750F for steel andits alloys. High-temperature pipingshould be well supported and well con-trolled. Springs should be used care-fully, and travel stops provided to limitexcessive displacement.

Many high-pressure pipe spools aredelivered in thicknesses that vary con-siderably from the nominal specified.This altered thickness affects the totalweight of the system which is not in-cluded properly in a spring hanger de-

sign. The spring then pushes too muchor too little on the pipe resulting increep concentrations at critical bendingsections. Supports on high-tempera-tures piping systems should be de-signed very carefully.

Piping stress analysis can be a com-plicated, dangerous business, andshould be approached with great cau-tion. Like any other technology, pipingstress analysis software is only as goodas the person who uses it. While theprograms perform calculations and give

guidelines, engineers must applycommon sense and judgement whenusing these tools.

Edited by Nicholas Basta

References

1. Mark]. A. C., et al., "Piping Flexibility Analy-sis," in Pressure Vessel and Piping Design:Collected Papers 19271959 (Am. Soc. of Me-chanical Engineers, New York), 1972.

2. Process Piping, ASME B31.31996 Edition,

an American National Standard (Am. Soc. ofMechanical Engineers, New York), 1996.3. British Standard Spec. for Design and Con-

struction of Ferrous Piping Installations forConnection with Land Boilers, BS 806:1994. (British Standards Inst., London) 1994.

4. Centrifugal Pumps for Petroluem, Heavy-DutyChemical, and Gas Industry Services, API610, 8th ed., Aug. 1995, (American PetroleumInst., Washington, D.C.), 1995.

5. Steam Turbines for Mechanical Drive Service,NEMA Standards Pub. No. SM 23, (NationalElectrical Mfrs. Assn., Rosslyn, Va.

6. Pipe Flanges and Flanged Fittings,ASME/ANSI B16.51988, ASME, New York.

7. Mershon, J. L., Mokhtarian, K., Ranjan, G. V.,and Rodabaugh, E. C., "Local Stresses inCylindrical Shells Due to External Loadingson Nozzles," Supplement to WRC Bulletin No.107 (Welding Research Council, New York),

Aug. 1984.8. Minimum Design Loads for Buildings and

Other Structrues, ANSI/ASCE 793, andASCE Standard (Am. Soc. of Civil Engineers,New York).

9. Oil Pipeline Systems, CAN3-Z183-M86, a Na-tional Standard of Canada (Canadian Stan-dards Assn., Ontario, Canada).

10. Wichman, K. R., Hopper, A. G., andMershon, J. L., "Local Stresses in Sphericaland Cylindrical Shells due to ExternalLoadings," in Pressure Vessels and Piping:Design and Analysis, (ASME, New York).

11. Burgreen, D., Design Methods for PowerPlant Structures, (Arcturus Publishers, CherryHill, N.J.).

12. "Rules for Construction of Nuclear PowerPlant Components, Div. 1, Subsec. NB: Class 1Components," in ASME Boiler & Pressure

Vessel Code (ASME, New York), 1992.

Authors

Michael Bussler, P.E., is thefounder and president ofAlgor, Inc. (150 Beta Dr.,Pittsburgh, PA 15238; Tel:412-967-2700), an interna-tional company providing abroad range of technology-based products and services tothe engineering community.Under Mr. Bussler's leader-ship, Algor has made 17 major

engineering software innova-tions, including the first PC-based software for fi-nite element analysis. Mr. Bussler was namedEntrepreneur of the Year by Ernst & Young, amanagement consulting company.

Anthony W. Paulin is presi-dent of Paulin Research Group (25211 Grogans Mill Rd., Ste.315, The Woodlands, TX77380; Phone: 281-363-3790),and coauthor of the finite-ele-ment program, FE/Pipe. Hecurrently serves on the B31Mechanical Design Code Com-mittee of ASME. In 1984, heco-founded Coade EngineeringSoftware, Inc., and is the orig-

inal author of the Caesar II pipe stress program.Previous work includes positions at Brown &Root Inc., G. H. Bettis Co. and Scientific Inter-comp, Inc. Paulin is the recipient of the 1991 Her-bert Allen Award for Outstanding TechnicalAchievement, presented by the South TexasASME Chapter. His B.S.M.E. degree was earnedat Texas A&M University, and he is a certifiedProfessional Engineer in Texas.

PIPE STRESS PROGRAMS FOR

THE CPI

Write in the Reader Service numbercard for more information

ADLPipe Research Engineers, Inc.

Yorba Linda, Calif. 201

AutoPipe Rebis, Inc.

Walnut Creek, Calif. 202

CAEPipe SST Systems, Inc.

San Jose, Calif 203

Caesar II Coade Engineering

Software, Houston, Tex 204

PipePlusAlgor, Inc.

Pittsburgh, Pa. 205

TriFlexAAA Technologies and

Specialties Co., Houston, Tex. 206

9 2 CHEMICAL ENGINEERING / JUNE 1997

Chemical Engineering - June 1997 - Pipe Design for Robust Systems

Documents