Академический Документы
Профессиональный Документы
Культура Документы
I have just completed my first time history analysis using Peng & Peng, Pipe Stress Engineering. As a check, I
thought I'd run the Fluid Transient option to have AutoPIPE automatically generate the TIH and TIL files so I
could make sure I had them structured correctly.
It appears to me (and being my first foray into dynamic analysis I could well be wrong) that the results output
from the Fluid Transient option are to be applied as a quasi-static analysis according tothe method
inEN13480 Annex A.2.2.2 and in Peng & Peng Section 12.6.2 "Closed Discharge System" and Section 12.7
"Steam Turbine Trip Load". For a time history analysis, the method in Peng & Peng Section 13.5.4 "Example
Time-History Analysis" gives completely dierent force-time histories and locations. However, it makes
moresense to me.
Has anyone come across this before? Thanks in advance for any insight that you can pass on.
I have pasted my methodology belowif anyone wants more detail on where I'm coming from. (Probably too
much detail for most, but my client's on-site piping engineer will take me to task otherwise.)
John Phelan
In addition to report 290677-DR01, Pressure Safety Valve discharge piping will be in accordance with;
ASME B31.1-2010, Power Piping, Non-mandatory Appendix II, Rules for the Design of Safety Valve Installations.
EN13480-3-2002 Amdt 4-2010, Metallic Industrial Piping, Part 3: Design and Calculation, (informative) Annex A, Dynamic
Analysis.
Peng, L.C. & Peng, T.L. 2009, Pipe Stress Engineering, ASME New York.
"II-2.3.2 Reaction Forces With Closed Discharge Systems. . . .Relief valves discharging into an enclosed piping
system create momentary unbalanced forces which act on the piping system during the first few milliseconds
following relief valve li . The pressure waves traveling (sic) through the piping system following the rapid
opening of the safety valve will cause bending moments in the safety valve discharge piping and throughout the
remainder of the piping system. In such a case, the designer must compute the magnitude of the loads, and
perform appropriate evaluation of their eects."
Under Paragraph II-3.5, Analysis for Reaction Forces Due to Valve Discharge, B31.1 goes on to say;
"II-3.5.2 Closed Discharge Systems. Closed discharge systems do not easily lend themselves to simplified
analysis techniques. The discussions on pressure in para. II-2.2.2 and on forces in para. II-2.3.2 indicate that a
time-history analysis of the piping system may be required to achieve realistic values of moments."
"There are a number of methods for the calculation of the eect of dynamic events, such as:
https://communities.bentley.com/products/pipe_stress_analysis/f/autopipe-forum/65534/fluid-transient-water-hammer-time-history-example 2/7
10/27/2017 Fluid Transient & Water Hammer (Time History) Example - AutoPIPE Forum - AutoPIPE - Bentley Communities
b) quasi-static equivalent;
Experience has shown that for properly supported piping, the use of simplified methods generally leads to
acceptable engineering solutions for the prevention of damage during dynamic events. Where complex analysis
is to be undertaken, care should be exercised in the selection of suitable programmes and consistent data for the
derivation of forces and allowable loads."
For rapid valve closure, and also rapid valve opening, EN13480 suggests three methods.
The first method, pressure rise assessment, ensures "that the minimum design pipe wall thickness can
withstand the operating pressure plus the maximum dynamic pressure rise dP" (EN13480, p.165). This is
already covered by determining the discharge piping design conditions using the method in Section 2.2. The
method in Section 2.2 uses AS4041 Appendix J and this method is specifically for a safety valve discharge
piping system. The method in EN13480 Annex A2.2.2 uses the Joukowsky's formula; it applies specifically to
water hammer problems and doesn't properly account for the jetting at the exit.
The second method, static assessment of dynamic loads, looks at the dynamic loads and determines
equivalent static loads to be applied to each leg.
"The eects of imbalance or surges on the piping system may be assessed by applying a calculated pressure
dierential to the ends of straight runs of pipe or at changes in direction. The dierential pressure is the
proportion of the peak pressure developed over the piping length under consideration and it is assumed to act
over the internal area of the pipe. In calculating the resulting forces, factors should be applied which makes
allowance for the variation in closure rate throughout the valve stroke and the dynamic nature of the actual
loadings."
This method is very similar to the method detailed in Peng & Peng.
https://communities.bentley.com/products/pipe_stress_analysis/f/autopipe-forum/65534/fluid-transient-water-hammer-time-history-example 3/7
10/27/2017 Fluid Transient & Water Hammer (Time History) Example - AutoPIPE Forum - AutoPIPE - Bentley Communities
"The net shaking force at each piping leg is taken as the dierence of the pressure existing at both ends of the
pipe leg under consideration. The maximum shaking force, Fmax, is determined by the length of the leg. . . If the
leg length is shorter than the sonic velocity multiplied by the valve [opening or] closing time, the maximum
shaking force is determined by direct proportion as
where lij is the length of the pipe leg located between point i and j.
. . . The piping is then analysed . . . statically by applying proper DLFs. The static method requires the application
of all forces at the same time."
Comparison of equation 12.58 in Peng & Peng with equations A2.2-5 and A2.2-6 in EN13480 shows that the
following two terms to be the same.
FT = dP..Di2/4
This leaves the EN13480 terms of 2M/A and 4M/A to equate to the "proper DLF" that Peng & Peng apply to
Fmax. Noting that Peng & Peng use a linear ramp function for the time history, then the EN13480 term M/A may
reduce to 1. This is a conservative maximum value of M/A per EN13480 Equation A.2.2-8. This implies that
EN13480 is using a maximum DLF of 2 for sti piping and 4 for flexible piping. However, these terms do not
strictly equate since the EN13480 M/A term is not a function of the natural period of the piping system,
whereas in Peng & Peng the DLF is a function of the natural period. Hence, EN13480 already has a DLF built in
and only distinguishes between sti and flexible piping. However, EN13480 provides no guidance on what is
sti and what is flexible. Since Peng & Peng provide more guidance and aligns more with the B31.1 Appendix II
approach, this method will be used.
The "proper DLF" that Peng & Peng apply to Fmax needs further examination. They say that "the DLF can be
taken from the appropriate shape or combination of shapes given in Figs. 12.5 and 12.6" (p.403). Using a ramp
impulse load, Fig. 12.5 is applicable. The theoretical dynamic load curve in this figure is repeated in Fig. 19
where a Design Curve is drawn to envelop the theoretical curve. This Design Curve is the same as the curve in
B31.1 Fig. II-3-2, Dynamic Load Factors for Open Discharge System, and EN13480 Fig A.2.4-2, Dynamic Load
Factor. The only remaining issue is what installation period is to be used to determine the DLF from B31.1 Fig.
II-3-2.
For this analysis, T will be taken as the fundamental (first) frequency of the piping system unless a higher
frequency can be shown to be the critical frequency. For example, a higher captured modal mass percentage
may indicate a more critical first frequency. For large piping models, the model should be reduced to only
include those lines influencing the modal frequency of the PSV inlet and discharge lines. (Larger models
general include more flexible piping than is around the PSV.) In fact the model should be reduced to only the
inlet and outlet lines with an anchor at the start of the inlet piping and an anchor at the end of the outlet
piping. Except for extremely sti systems, it is unlikely that the DLF will be less than 2.0 since the rapidly
opening PSV gives a very small numerator in the tO/T term. Hence the analysis eort to determine the
https://communities.bentley.com/products/pipe_stress_analysis/f/autopipe-forum/65534/fluid-transient-water-hammer-time-history-example 4/7
10/27/2017 Fluid Transient & Water Hammer (Time History) Example - AutoPIPE Forum - AutoPIPE - Bentley Communities
installation period will generally not be of benefit, particular for lines where higher temperatures mean
thermal expansion must be accommodated in the inlet and/or discharge piping.
The third method discussed in EN13480 is advanced methods of calculation. EN13480 provides a good over
view but provides no details for its execution.
"The characteristics and eects of the pressure wave created by rapid valve closure may also be assessed by time
history or modal analysis.
The development of the pressure pulse throughout the piping system can be idealised using mathematical
modelling of the events, and these pressures used at a large number of time intervals to determine the forces at
terminals, or changes of direction. The forces thus derived can be used as the driving factor in an analysis of the
vibrational response of the piping to these forces.
If modal analysis is used, the designer should check that the cut o frequency does not exclude any significant
higher modes resulting from the interaction of waves in the piping, as the system can be relatively sti for these
frequencies.
These advanced methods may incorporate coupling between the fluid and the piping and can thus incorporate
the damping of the pressure wave by the transfer of energy to relatively sti piping. For steam, or similar fluids
where the mass of the fluid is negligible relative to that of piping, the advantage of the use of the advanced
method is small.
Whilst these methods oer a potentially more accurate and less conservative solution to the problem of rapid
valve closure, the advanced techniques for rapid valve closure analysis can be very sensitive to the modelling of
the fluid source, the valve characteristics, the supports, and the fluid behaviour. The designer should be satisfied
that the mathematical representations of all aspects are suitable and accurate."
Peng & Peng provides a detailed, theoretical approach for a time history analysis. They show how to develop
the force-time histories and how to apply these to the piping system. Their example in Section 13.5.4 (p.451-
456) is the method that will be used.
The flexibility analysis so ware, AutoPIPE, has a feature to automate the development of the time history
profiles and the application of these profiles. This feature . . . it appears wrong???
The locations for the application of the force time histories in Peng & Peng (see fig 13.22), dier to the defaults
provided by AutoPIPE's automatic Fluid Transient load case setup. The Peng & Peng approach makes more
sense and is used in this calculation.
leg. The only real dierence I can see is that Peng & Peng recommend the length of the time history
cover the first wave (to the end of the run) whereas the Fluid Transient utility onlymodels the first rise
and fall in each leg.
Anyway, with the two approaches agreeing I feel more comfortable with my results.
Bye.
John Phelan
https://communities.bentley.com/products/pipe_stress_analysis/f/autopipe-forum/65534/fluid-transient-water-hammer-time-history-example 6/7
10/27/2017 Fluid Transient & Water Hammer (Time History) Example - AutoPIPE Forum - AutoPIPE - Bentley Communities
(Although for dierent leg lengths and ramp times there may not always be two
trapezoid function in the first leg.)
Being my first foray into dynamic analysis, I can't comment whether this amounts
to the same thing or not. I'd be interested if anyone has any insights though.
John
https://communities.bentley.com/products/pipe_stress_analysis/f/autopipe-forum/65534/fluid-transient-water-hammer-time-history-example 7/7