Modeling valves in piping stress analysis

[ZippyDDoodah]

Every pipe stress software program I've seen treats valve weights at piping centerline as if the valve is a pipe, and valve vendor technical sheets aren't much help either, since they provide a single weight only for the entire valve, as if it's credible/reasonable to ignore the centroid and moments of intertia about the centroid of the valve. Any quick look at valves with long heavy bonnets with handwheel make it clear that moments of intertia would have a significant impact, particularly under lateral loads, yet I've never seen this issue discussed on a forum.

Based on experience, have piping engineers decided that the centroid/moments of intertia of different valve sections can be credibly ignored in a pipe stress analysis? I realize that you can model a rigid element or valve element to sort of account for this, but without knowing the centroid and moments of inertia of a particular valve, even that approach would be very crude. I'm a structural engineer who's still a newb in piping analysis and design, but this issue struck me as one in which piping engineers seem to routinely model/analyze valves incorrectly. Has it really become accepted practice to lump the entire valve weight at the piping centerline?

 

[SNORGY]

I think this would be an interesting post in the COADE forum, and I am sure that someone there has given it some thought.

I would venture to suggest that if you are already trying to establish true centroids and apply what would then presumably amount to torsional moments about the pipe axis in combination with the valve weight, you are probably doing more than what is normally done, and certainly more than what I normally do.

Regards,

SNORGY.

 

[DSB123]

ZippyDDoodah,
I concur with your thoughts. Rotational inertias are important for valves especially when subject to seismic loading. The software I am very familiar with needs to have the rotational inertias input and there is an entry for a value to be input. If you do not enter a value then the program flags up a warning. What you also need to consider for gate valves is that there are two values for the valve whether it's open or closed.

 

[ZippyDDoodah]

Thanks SNORGY and DSB for your replies. If we agree that rotational inertias are important when modeling valves, particularly for seismic loading, where/how do you obtain the centroid and moments of inertia for different valves? I haven't seen any vendor data that provides this information.. have you? I would love to hear any ideas or rules of thumb that would improve upon the current default of lumping valve weights at centerline.

DSB, this is probably a rookie question, but from a pipe stress analysis standpoint, why would it make a difference if the gate valve is open or closed? Are you referring to the effect of fluid contents weight?

 

[DSB123]

ZippyDDoodah,
Yes I have received valves from a valve vendor. You just need to specify the requirement for the information in the order documentation. (i.e. he does not get the money until he provides the values for the valves).

For gate valves the C of G is different hence the rotational inertia is different if the valve is open or closed. For ball and globe valve there is no difference but for gates or parrallel slide valves there is. The "gate" position is different for open and closed position.

 

[Hurricanes]

ZippyDDoodah,

Maybe I am not reading your post properly, but AutoPipe allows you to specify valve weights offset from the pipe centreline. I do this all the time.

When using 600mm angle valves with large actuated handles sitting about 800mm from the pipe centreline, it does make a difference.

 

[SNORGY]

As crude as it sounds, when I need stuff like that (centroids of different parts of a valve, etc.) I usually just give it "...the old college try..." and take a flyer at it. Not academically accurate but better than nothing...

We do most of our work in places where we don't have earthquakes, so seismic loading isn't even considered in the design, for the most part.

Regards,

SNORGY.

 

[SNORGY]

I was actually not going to post that before editing it...I am not sure what I pressed but I am sure it wasn't [ENTER].

Anyway...

In most of the stuff I deal with (12 NPS and down in non-seismic zones) there is enough torsional rigidity (would stiffness be a better word choice here?) that the eccentric loads imparted by valves whose centroids are not coincident with pipe centreline are negligible. We do see some hefty actuators on occasion and might design for those loads.

As for other cases of relevance, if we have (for example) a big Barton-Woods P/L isolation valve or other valves with significant length and mass relative to pipe diameter, we usually try to orient them vertically or in such a manner as to have the centre of gravity closely above pipe axis.

But, like I said...most of our concerns are in non-seismic zones with small diameter, fairly stiff pipe.

Regards,

SNORGY.

 

[ZippyDDoodah]

Hurricanes, I appreciate your suggestion, but I'm familiar with the Autopipe weight offset feature. That feature is specific to heavy valve operators/actuators offset from centerline which is a separate issue relatively easy to model and deal with. My concern is somewhat different, involving valve bodies like gate valves which typically have a centroid significantly offset from the piping for seismic loading. I am looking for advice as to how to reasonably model this in order to realistically account for moments of intertia about the centroid. Unless valve vendors supply the data we need, I'm leaning towards SNORGY's advice to take an 'eductated guess' at centroid and moments of inertia as a better approach than what most piping engineers appear to be doing now, which is assuming that the valve weight is lumped at the piping centerline. My next step however, is to try to muscle valve vendors into supplying centroid data on valves per DSB's approach. I'll try and post whether this approach yields any success.