Pipe stress modelling techniques – ask the expert

Q&A from our pipe stress modelling webinar

By Chris Harper, PEng, Principal Consultant

Watch the recording

The answers below were given in the context of the 2017 webinar and do not address all aspects of the issues discussed. Questions and answers from the 2022 session will be added shortly.
For more comprehensive information or application support, we strongly encourage you to contact our experts directly.

1 How to model piping stress

a) Miscellaneous

How would you consider radial expansion or contraction of a pipe clamp?

Radial expansion of a pipe due to temperature change is not thought to affect a hold-down clamp significantly, as the clamp is expected to come to be at the same temperature as the pipe to which it is in contact and thus experience the same radial growth.

For vertical supports, how do you account for both preload and pipe weight in the system (using Caesar II)?

A set of free body diagrams, one of the upper clamp and one of the lower wear pad, would show that they both experience equal and opposite clamp preload reactions, but only the lower wear pad would experience load due to the weight of the pipe. The weight load is accounted for using a coefficient of friction on the vertical restraint, while the clamp preload is accounted for using the bilinear clamp technique.

Do you consider tolerances like 2mm as gaps?

We typically see gaps with shoe-type supports with guides or limits stops, which was not the topic of the webinar. For hold-down clamps, we typically assume no gaps, as hold-downs are intended to be in contact with the pipe at all times.

For the example in case 5, what are the results compared to? Were there any load tests done? What does a ratio of 1.0 mean?

Case 1 included a model of the structure beneath the pipe clamp. What case 5 shows is that hand calculations of the stiffness of the beam result in a very good estimation of the total stiffness of the support (using the formula shown on slide 29). The ratio of 1.0 means the stresses, deflections, and loads are all identical to case 1, where the structure was actually modeled. This is good news for piping stress engineers who may not want to add a bunch of structure to their models.

How do you differentiate between off-skid and on-skid piping analyses? On on-skid piping, short distances make it is easier to manage thermal stresses, but the management of stiffness is much harder on off-skid piping. When piping is on pipe racks, the thermal expansions are much larger, and clamps are not practical sometimes, especially around elbows where appreciable thermal expansion is expected.
How do we manage such situations? Should we provide more lateral or longitudinal slots on the clamps?

The techniques we showed in the webinar cannot replace all the regular pipe stress mitigation strategies that are currently used (e.g., thermal loops). Instead, these techniques are meant to be used in addition to the regular pipe stress mitigation strategies. For piping in vibratory service, where hold-down clamps are required to manage the vibratory load, it may be that an additional loop is required to lessen the expansion at elbows so that hold-downs can be used, or the support structure itself might be designed to allow for the thermal growth. Using your standard solution set along with a combination CL-1-T, CL-1-ST (with slots), and CL-1-T-ST (with slots) clamps can most often provide acceptable solutions for pipe stress and vibration.

Is it OK to consider total Mu when the normal forces are different as shown in your example?

The total Mu should only be used when there are two contact surfaces, like in the case of a hold-down type clamp. Further, this total Mu should only be used when calculating the friction breakaway force (eg, the "Fy" in the bilinear stiffness restraint in Caesar II). The Mu used in Caesar II for the vertical restraint should be the Mu of the contact surface between the pipe and the structure beneath.

Why do we need to model the first three supports accurately, and why not at the location where we think failure is most probable?

The recommendation for the first three supports to be accurate is out of the DNV-RP-D101 document, out of concern that loads on machinery nozzles can increase the risk of failure. We, however, would recommend that supports are modeled accurately wherever possible, and particularly in any location that you are worried about problems, not just near machinery nozzles.

What do you think of installing expansion joints on the suction and discharge of rotating equipment to prevent transmitting vibration from the equipment to the suction/discharge piping system?

We see this approach attempted quite often, sometimes with poor results. For mechanical vibration directly from the rotating equipment, it may help a bit, but we have typically seen that either the expansion joint fails in fatigue, or the forces transmit through the joint anyway. For pulsation-induced vibration, the pulsations can travel through expansion joints and cause vibration elsewhere.

Can the clamping force (Mu) be higher than the dead weight support?

This depends on the pipe diameter: for small piping, the clamping force tends to be higher than the dead weight of the pipe. For larger pipe diameters, the clamping force will tend to be lower than the dead weight.

b) Modeling pulsation forces in a pipe stress analysis

Can pulsation loads be considered in a pipe stress analysis?

In our experience, static or quasi-static loads can be included in most pipe stress analysis software, but true dynamic loads (such as pulsation forces) cannot be included. Doing so would be a very difficult process, as dynamic loads not only have a magnitude, a frequency at which they occur, but also a phase relationship.

How to consider pulsation forces in a pipe stress analysis

A pipe stress analysis typically doesn't include pulsation forces, and pipe stress analysis software are typically not capable of performing a pulsation/acoustical analysis. There are other commercially available software packages that are used for an API 618 DA2 pulsation and mechanical study. (We at Wood have our own in-house software package that has proven its accuracy over the last 40 years. If you are in need of pulsation analysis, please contact us).

Do you consider pulsation forces when calculating brake friction forces in a reciprocating compressor system?

Typically, the pulsation forces at a restraint will be much lower than the friction force, and we design for it not to slip. If pulsation force breaks friction force, you generally have big problems with pulsation design!

c) Pipe stress modeling software, CAESAR II

Can you correctly predict MNFs for B31.1 using Caesar II, and can you also get same frequency using modal analysis in Caesar II?

Caesar can predict MNFs, but the correctness of the predictions depends on the accuracy of the support stiffnesses.

Is Caesar or AutoPIPE suitable if your small diameter pipe is supported by a larger header pipe that may itself experience deflection? If not, can you recommend a software that can model this?

We recommend that you include structure common to multiple pipe runs (such as racks) wherever possible into your PSA, as a pipe with large deflection may drag other pipes with it - and cause overstress. Both Caesar and AutoPIPE can include structure in this fashion. If you are looking for a small-bore piping vibration analysis, we often perform this type of FEA analysis, generally using ANSYS. We are experts at doing this fast and cheaply. If you need support, please contact us.

Any guideline for modeling bilinear supports in AutoPIPE?

At this time no, but you should request bilinear functionality for applications other than buried pipe from AutoPIPE to be able to do this analysis.

How are vibration absorbers and damper braces modeled in Caesar II?

Our vibration absorbers and DamperX braces require more advanced modeling techniques than those available with Caesar II software. We use FEA software such as ANSYS to apply these products effectively.

d) Stiffness

How far would you go with modeling flexibility into the model? For example, would you model earth stiffness for pedestal supports?

Often this is a matter of engineering judgment - knowing that the component with the lowest stiffness will dominate the overall support stiffness, you can generally identify which component is the determining factor, whether it be the clamp, the structure, or the soil. We typically do not model earth stiffness of the supports, but where it is thought to be significantly low such that it will affect the overall stiffness, it might be worth doing. This calculation can be done with other software packages such Dyna.

How do you account for interaction between soil, structure and clamp in obtaining dynamic stiffness and damping of pipe supports?

We typically do not model earth stiffness of the supports, but that can be done with other software such as Dyna.

How do you account for dynamic stiffness being a function of frequency?

We have two solutions for this:

1. We will include the structure into our model, so that the dynamics of the structure are considered, or

2. We assume (and confirm) that the natural frequency of the structure is 2x the natural frequency of the vibrating pipe, which means at the vibrating frequency, the structure dynamic stiffness is approximately equal to its static stiffness. Static stiffness is simply the dynamic stiffness at frequency of 0 Hz.

(We do our detailed vibration modeling using FEA software and not Caesar II.)

How would you match API recommended stiffnesses with clamp stiffnesses?

If you know the minimum support stiffness (Ksupport) and you know the clamp stiffness (Kclamp), then you can calculate the minimum structure stiffness (Kstructure) by rearranging the formula on slide 29. It is: (1/Kstructure)=1/[(1/Ksupport)-(1/Kclamp)]. Also, given we know the clamp stiffness for a typical hold-down (see table values in our design guide), we can back-calculate the required structural stiffness. Knowing this stiffness allows you to inform the structural engineer of your requirements.

What is the stiffness of a clamped shoe or a welded shoe?

We do not have detailed data on the stiffness for a shoe at this time, and stiffness would vary depending on the design. If the shoe has gussets, then the stiffness should be close to the stiffness for a hold-down type clamp. If the shoe is tall or does not have gussets, then the lateral stiffness will be lower.

2 How to model pipe supports

a) Miscellaneous

How do you consider pipe support spacing in a pipe rack?

Pipe racks typically provide very little support stiffness. This will result in pipe MNFs that are low compared to high stiffness supports. You can compensate for low support stiffness by using tighter support spacing than with ground supports, or by stiffening the structure to raise the stiffness. For situations where your rack does not provide the minimum required stiffness for a vibratory node, Caesar II can predict pipe MNFs so that you easily check if your spacing is adequate, as can most other FEA programs you have access to.

Our pipe flexibility engineers recommend supports with more gaps, which is in conflict with vibration requirements. How do you solve this problem?

For non-vibratory service, gaps are an effective way to allow thermal growth to occur without stress developing in the piping. However, for vibratory service, supports with gaps will generally not provide the dynamic stiffness required to resist vibratory loads. This is the conflict of requirements between pipe stress design and vibration design. This webinar is trying to help bridge that gap in showing how hold-down clamps can be used to meet both requirements effectively.

What type of support do you prefer to use between a compressor package and pipe rack? Would you use a clamp support or hold-down support?

In vibrating service, we prefer the use of hold-down type clamps like shown in the webinar, even on pipe racks. With the techniques we have shared, an appropriate pipe stress solution is most often attainable while maintaining the superior vibration performance of our clamps.

When using a clamp support, should we use it with tolerance or without? If I don't have sufficient information about the support inside the packages, should I use a clamp support outside of the package?

We typically do not model gaps with our clamps.
For recip compressor packages, even with good pulsation control, there are often small residual forces in the off-skid piping. We recommend clamps in off-skid piping around reciprocating (pulsating) equipment as well.

With generic U-bolts, do the suggested clamp forces still apply, or are these only for the Wood clamps?

The suggested clamp forces are simply a way to understand and predict the breakaway friction load in order to accurately model it. Using those loads on a generic U-bolt would allow you to predict the friction force and use a bilinear element. However, U-bolts tend to bite into the pipe that they are installed on and do not allow the pipe to slip through at a predictable friction force.  As such, we do not recommend the methods in the webinar for U-bolts.

Why isn’t the weight, pressure, temperature and external forces included to calculate the breakaway force for a hold-down, but they are used for a simple rest support?

The weight, pressure, temperature, etc are used to calculate the breakaway force, but they are done automatically by Caesar II, exactly how it is done for a resting support. That is why a Mu is entered on the restraint in the vertical direction. The bilinear restraint's only purpose is to model the effects of the clamp on the restraint accurately.

b) Anti-vibration clamps and damping

What is the main difference between standard anti-vibration clamps and damping clamps?

Standard anti-vibration clamps are good at creating vibratory nodes that, when spaced appropriately, keep the piping MNFs away from high-excitation frequencies. However, this does not guarantee that the pipe will stay free of excessive vibration. DamperX™ clamps improve on standard anti-vibration clamps by lowering the vibration response of the pipe by adding damping.

In terms of the Caesar II model, the stiffness of the DamperX clamp is 10% of a regular CL-1 type clamp.

In terms of damping, the DamperX clamp has large damping compared to a steel clamp and is therefore a good option for vibrating service.

Can Wood’s clamps be used on small-bore piping (2in)?

Yes, we make DamperX clamps as small as 2in (5cm).

How much displacement is expected with a bi-linear clamp?

In the axial direction, if the force on the clamp is less than the friction force, then the expected deflection will be similar to other clamps (ie, small). If the force is higher than the friction force, then the pipe will slip through the clamp. The amount that the pipe can slip is thus not limited by the clamp – the limiting factor at the point the pipe slips will be the stress in the rest of the piping, or the allowable restraint load, etc. We see values <1.0", but it does depend on many factors, and more than 1.0" could still be acceptable.

Should we check the lateral movement when using DamperX clamps?
What if the damper is fully crushed and the stiffness reaches the max?

We publish our clamp allowable loads in the design guide provided, and recommend comparing the restraint load against these allowables in order to determine if the clamp can withstand the predicted load. We do not publish the deflection that the clamp will experience, as this number is dependent on both the clamp and the structure together, and will thus change from one structure to another.

The allowable loads we publish account for the damping material and, if observed, will prevent the damping material from being crushed. Additionally, we offer DCL-1-HT-T-ST DamperX clamps which allow for lateral motion for this reason. If you want a copy of our design guide for piping stress engineers, please contact us.

Is the mechanical natural frequency (MNF) of piping system affected by anti-vibration or bilinear clamps?

Yes, the natural frequency of a pipe span is affected by the clamp stiffness. This means that DamperX clamps, given that they are less stiff, could result in a lower MNF than a standard clamp. This effect is expected to only manifest itself for large pipe ODs (>20"), depending on the MNF guideline being applied, the number of supports on the line, etc. As for the bilinear approach, Ceasar II ignores bilinear effects for its modal calculation. This is appropriate, as anti-vibration clamps should be able to withstand vibratory loads without the pipe sliding through the clamp. Additionally, the natural frequency calculations should be more accurate when using an appropriate stiffness in Caesar II with this approach.

What is the clamp preload for thin wall pipe?

Our calculations do not account for local points of high stress in the pipe shell due to clamp loads on thin wall pipe. Those should be ensured by others.

How do you model anti-vibration clamps in a modal analysis?

Using Caesar II model analysis, we recommend the same methodology. For more accurate vibration models we use FEA software such as ANSYS.

Does the clamping force have any effect on the clamp stiffness?
Will it reduce the overall reduction in stiffness for damping clamps and will it reduce the damping?

As the clamp bolts are tightened, the pipe is forced to sit deeper and deeper into the damping material on the wear pad underneath the clamp. This contact between the pipe and the damping material is crucial for effective clamp operation so that resonant vibration energy in the pipe can be directed toward and damped out by the damping material. If the bolts are not adequately tightened and the clamping force is very small, the pipe would not engage the damping material effectively, and both the stiffness and damping properties of the clamp would be lower than intended. We publish recommended torques for our clamps such that the stiffnesses and damping are adequate for vibratory service.

How would your analysis change with use of a spring support (e.g. Belleville washers)? What is the benefit of adding the springs to a hold down?

Generally, Belleville washers are used on joints subject to deformations (eg, gasket creep), and help prevent loosening of nuts. Belleville washers effectively reduce the stiffness of the bolts. Additionally, Belleville washers can help control the preload on a clamp if used as a visual indicator during installation.

Using a Belleville washer should not affect the analysis as long as the clamp pre-load is accurately calculated and modeled.

Since anti-vibration clamps are designed to fail first on the bolt, what might cause cracking on weldment area on a vibrating piping system?

The component that we intend to fail first is the bolts. If the bolts do not fail as intended, then the next component intended to fail is the clamp, which has been observed in some cases. The reason that clamps fail first instead of the bolts can vary: residual stresses from clamp forming, from welding, inadequate weld penetration, among many other reasons. This leaves the clamp as the weak link in the intended failure order of bolt, clamp, structure, pipe.

Are your anti-vibration clamps or DamperX clamps suitable for subsea use, i.e., fully submerged in seawater for 25 years?

We have not tested our DamperX technology for subsea applications yet, but are interested in trying. Please contact us if you are interested in using Wood VDN anti-vibration products in subsea testing.

Is DamperX available worldwide?

Yes, DamperX is available worldwide (shipped out of Canada).

Is there a catalog or list of Wood’s pipe support types and liners?

Yes, please find our brochures on our product pages:

• DamperX anti-vibration clamps and support
• Standard anti-vibration clamps

c) Liners and insulation

Can DamperX clamps be used for insulated piping?

Currently, insulation can be applied around DamperX clamps with the right considerations to avoid insulation crush. However, we are working on a vibration pipe shoe design that can better accommodate insulation.

When would you use PTFE clamps (since they are more expensive)?

PTFE-lined clamps can be thought of as being similar to rest-and-guide supports: they are stiff enough to resist vibratory loads, but allow growth through the clamp so as not to overstress the pipe. We use unlined clamps for the bulk of the supports in vibratory service and only use PTFE-lined clamps at specific locations to direct growth away from equipment.

What is the recommended maximum temperature for Wood’s PTFE-lined clamps?

The PTFE used in Wood VDN anti-vibration products is rated to 204° C (400° F).

What is the minimum ambient temperature for DamperX?

Our DamperX HD material is rated to -17° C (1° F). The material can get brittle at temperatures below -20C, but typically the heat of the pipe and fluid will warm the material into the acceptable range. We are interested in collecting more data on how our DamperX material performs in colder conditions. Please contact us for trial applications.

What is the service life of the DamperX material?

We expect 15 years out of the material. Currently, we have had our products in service in industry for three years with no known problems with durability. We recommend inspecting the material every five years.

What maintenance is required for DamperX clamps? Do the liners need to be replaced after a certain time?

DamperX clamps do not require any maintenance, however, we recommend inspecting the material every five years.

Do your clamp liners retain moisture?

The DamperX materials are comparable in moisture absorption to Fabreeka, Neoprene, or other common clamp liners.

Do your clamp liners affect friction?

Yes, they do affect the friction. Our design guide provided to webinar attendees contains a table of friction coefficients for each of our lining materials. Please contact us for a copy of our design guide for piping stress engineers.

3 Contact an expert

Have more questions or would like to talk to an expert about your specific situation?

Don't hesitate to book a call or contact our dedicated pipe stress and vibration specialists.

Related Pages

Webinar: Shake, rattle and grow I (2022 update)  •   Webinar: Shake, rattle and grow – part II  •   Pipe Stress Analysis  •   DamperX™ Technology  •   Pipe Clamps, incl. DamperX™ and ThermaGlide™  •   Training courses and webinars  •  

Free webinar

Understand the top three vibration threats in your piping system and how to tame them.
Register now