Pipe Support Friction Factors 3

[DSB123]

Just wanted to bring this old one up again.
Friction Factors for pipe supports where a pipe shoe sits on steelwork. Nearly everyone uses a friction factor of 0.3 for steel to steel contact. How can this be correct for pipe systems running along pipe racks or outside where we all know that maintenance is not what it used to be. The general friction factors for steel to steel are for static friction are around 0.72 to 0.74 and for sliding friction the values are quoted as 0.57. So I ask what is the justification of using 0.3 for pipe stress analysis ? Does the pipesupport "know" it is a pipe support therefore the maximum frictional load cannot be greater than 0.3 times the perpendicular load ? No is the answer so come on the Pipe Stress guys out there make a justification for using 0.3 in your pipe stress analysis when before movement occurs a load of 0.72 times the perpendicular load needs to be achieved and to keep the pipe moving a loading of 0.57 times the perpendicular load needs to be maintained?

 

[TGS4]

Not that I am necessarily disputing what you are saying, but do you have a reference for your values of 0.72-0.74 for static coefficient and 0.57 for dynamic coefficient?

 

[DSB123]

TGS4,
Here are a few References:-

http://www.engineershandbook.com/Tables/frictioncoefficients.htm

http://www.google.co.uk/url?sa=t&rc..._RgbAL&usg=AFQjCNGdX3f15XxhUyARaqnEyaeTh_hSXQ

http://www.carbidedepot.com/formulas-frictioncoefficient.htm

So how can the pipe stress guys legitimately use 0.3 is my question?

 

[BigInch]

Because it's been done relatively seccessfully for what I'd guess is well more than 200 years and has certainly become accepted as a nearly universal practice by millions of engineers, IMO essentially OVERRULES the internet's opinion.

I hate Windowz 8!!!!

 

[DSB123]

Good Technical conclusion there BigInch. Also do not think it's been used for 200 years!!! - bit of an over-exageration there!!! Test data is the normal way to determine engineering values not guess work!!!!

 

[BigInch]

Some things are not entirely concluded by mathematics. I can't tell you the number of times that I selected a pipe wall thickness of 0.25", just because I knew 0.100034 was not a sufficient result, even though the mathematics proved it was.

What are you saying. So nothing before it apppeared on the internet is valid?
They invented a rail running locomotive in 1712, 302 years ago. I think it did not take them 102 more years to develop the concept of steel on steel friction factors.

I hate Windowz 8!!!!

 

[BigInch]

Make that 1802. I'd hate to diverge to arguing about when a rail running locomotove was invented and the time it took to reach 0.3

I hate Windowz 8!!!!

 

[DSB123]

Let's not get too carried away here. Yes maths can say you need a 0.10000000034" pipe thickness but you use some form of logic and select a more applicable thickness. And no the internet does not make everything before invalid.

My question is still the same friction tests , whether now or a few hundred years ago (if conducted correctly), would show the friction factors applicable so why do pipe stress engineers not utilise the relevant data.

Your rail running locomotive would not have rail to wheel friction factors as low as 0.3.

Your logic of "it's been used by millions of engineers" is not really a valid answer. The simple question is since friction factors have been proven to be 0.57 (sliding) and 0.74 (static) where is the engineering justification in using 0.3 ?

 

[LittleInch]

If you look at the results posted, the high values come from clean dry steel with no oxide layer. Point me to a pipe support that is like that after about a week on a plant, never mind a year or two and then I might accept your point. I suspect it is also why many pipe supports end up being a point load on a bar rather than a flat surface which can either corrode together or could have a high FF.

As BI says, clearly this has become a universal practice based on many decades of plant design and I haven't seen many piping systems collapse due to incorrect use of the FF.

what you maybe need to do is see what the impact is between 0.3 and 0.5. The lower FF will probably end up with larger movements at the elbows due to longer lengths moving, but run the numbers and see which is the more conservative. You might be surprised....

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[DSB123]

LittleInch,
Some sanity back. Yes I agree supports become corroded and "self-welding" due to corrosive environments can occur so Friction Factors could be greater. If the supports are well maintained and "greased" then 0.3 would probably be valid. However in the real world they are not. I worked for a world renowned chemical company who insisted on friction factor for steel/steel to be taken as 0.5 as a minimum after extensive testing over years. They found that friction factors could increase to around 0.8. My original query was to question to validity of using friction factors as low as 0.3!!!

 

[LittleInch]

As I said, 0.3 is not necessarily less worse than a higher FF. Maybe some sensitivity runs need to take place to see if it actually has any impact or not.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[DSB123]

A 0.3 friction factor is less worse for the design of the support and you cannot argue against that fact!!! The pipe stress engineers normally provide the loadings to the support guys do they not?

 

[LittleInch]

I worry about the pipe not the support so much. This is where you could get two different worst cases...

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[DSB123]

if the pipe support is not designed for the correct frictional load then yes I would be worried about the pipe not being supported when the support fails!!!

 

[BigInch]

Most of the time 0.3 is fine. If you know that it will be higher, use the higher.

I hate Windowz 8!!!!

 

[TGS4]

I will object to your statement

A 0.3 friction factor is less worse for the design of the support and you cannot argue against that fact!!!

This might not always be the case.

My approach to friction is that where it helps you, it is best to assume that it isn't there. And where it hinders you or indeed makes things worse, then it is best to assume a high value. In this approach, it is best to bracket the range of possible solutions to determine whether or not the friction is helpful or hindering. (I can provide good examples of both, BTW). In various designs, that not only include piping, I have designed assuming a friction coefficient of 0.0 all the way up to 2.0.

One of the interesting aspects of friction testing is that the coefficient is often a function of the contact pressure. This is published for low-friction materials, but such testing is often not performed for "regular" materials.

Another aspect to be considered for design of process piping is that these pipes vibrate, shake, rattle and roll. A short time after the initial heat-up or cool-down, it is quite likely that the frictional forces have vanished. I understand that the frictional forces can impart substantial lateral loads on the supporting structural steel, both at each individual support, as well at at special supports such as guides and directional anchors. However, after the vibration "relaxes" the friction forces, the loads generally decay to almost nothing.

I suspect that even if the 0.3 coefficient of friction for steel-on-steel is low based on testing, from a practical perspective it has demonstrated itself to be adequate for design purposes. It could very well be that we are relying on the Code design margins and eroding the ultimate factor of safety (note that anyone who has read anything that I have published know that I don't use that phrase lightly, and I use it in a very specific sense). But, until I can be pointed to an actual failure in process piping where the root cause has been an underestimation of the coefficient of friction, I will stick to experience.

 

[DSB123]

TGS4,
Can you explain in engineering terms how you beleive a vertical load times a low friction coefficient to give a lateral load on a pipe support is worse than the same vertical loading times a higher friction coefficient to give a larger lateral load on a pipe support to substantiate your objection to what I stated. If you have a pipe sat on a support how can you justify using a friction factor of 0.0 unless you are designing the support for a specific deflection of the pipe (in which you do not consider friction loading as it is deflection governed)

 

[SNORGY]

I think TGS4 has the correct approach. If you have an unlimited run version of whatever software you are using, you can easily run a "low friction" and a "high friction" case, and pick the case that governs. If they both "pass" ( "pass" meaning more than just Code Compliance Check - meaning acceptable loads plus acceptable displacements plus acceptable stresses and all the rest of it) then the friction coefficient really doesn't matter much to anyone other than the structural guy(s) - who may be one and the same as the stress guy(s). I also think that you will find a lot of cases where friction serves to reduce displacements at points of interest, which will have the effect of lowering expansion case / bending stresses so that, often, assuming zero friction yields the most conservative design.

In short, there's more to stress engineering than running the software. As DSB123 suggests, the stress analyst needs to put intelligent thought into the data input and be prepared to back up the assumptions made; whereas as TGS4 suggests, it's good practice to look at ranges or envelopes and run a few scenarios in order to arrive at a design that is not uber-sensitive to the assumed friction factor.

 

[DSB123]

SNORGY,
Agree with your synopsis. That is what I was trying to get at but many pipe stress engineers just bang in the 0.3 as a routine not looking at whether or what effect any variation higher or lower friction factors have on the results. And generally they never look at the systems without friction!!! A standard 0.3 is applied to all supports. As you say a sensitivity analysis approach is required to cover all bases. But I presume you agree that an individual support must be designed for the maximum frictional loading not the minimum as TGS4 was intimating by criticising my statement?

 

[LittleInch]

DSB123,

I think what was being talked about was the effect on the pipe, after all you are in the pipelines, piping and fluid mechanics forum not the pipe support forum... We tend to be more concerned about the pipe and assume the structural guys add quite a bit onto the loads given to them. Many supports also have multiple pipes, not all of which are going to be providing loads at the same time in the same direction and in my experience are designed and built quite big and strong.

Apart from individual pipe supports badly supported on sand, I've not heard of supports bending or collapsing because of extra sideways load, but there are many examples of pipes being damaged due to expansion and this is where a low friction factor can provide the largest deflection, movement and stress.

I agree that there should be some sensitivities run to determine what impact different FFs have.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way