Thermal expansion on long structures like pipe rack 5

[abcmex]

Which are the typical considerations against thermal expansion of pipe racks?

I can imagine that a pipe rack (the structure, not the pipes)would be damaged if there are no design considerations like sliding supports, separation between segments ect.

Kind regards

 

[dik]

If exposed to the elements, in these environs, I generally limit the length of members to about 30' with sliding splices at cantilevered 'stubs'.

 

[Lutfi]

Make sure you provide proper pipe expansion joint (accordion like). Also, make sure that you restrain the pipe from bucking/getting out of plane by providing proper anchorage at the expansion joint.

Good luck

 

[Puthuran]

The ideal thing is to miss one bay of the structure (only beams) every 45.0m. This would allow the pipe rack to expand with in that length and no aditional precautions required on any connections.

If you haven't understood my answer, please feel free to write back

Regards
Puthuran

 

[PXC]

I would be interested to know what other people do with longitudinal frictional forces where pipes slide metal on metal. I presume that as the temperature increases, the longitudinal force between a pipe and the support beam would increase then decrease when sliding occurs, resulting in a "saw-tooth" variation with time. I have always assumed that the maximum longitudinal forces from adjacent pipes all occur simultaneouly. Is this what others do?

 

[abcmex]

thanks for your information! Is there any reference available (if possible @Internet)?

I have seen already supports like DIK mentioned.

The idea of a separation of one bay like PUTHURAN proposed in my case (small pipes and big pipe with different span req.) would require additional foundations or additional supports.

The expansion of pipes mentioned by LUTFI is a matter for other forum (see piping and fluid mechanics engineering). The use of an expansion joint is not as simple as it seems because it adds substantial forces to the pipe rack.

The question of PXC is probably one for the piping forum. A conservative approach may do the worst case approach (all at the same time). This may also be the situation in a plant start-up.
Kind regards

 

[havesealwilltravel]

Typically for a pipe rack:

Provide a longitudinal expansion joint every 150'.

For high expansion items such as steam pipes, make piping provide expansion loops and sliding bearings and design for the anchor forces. Be sure they do their calculation assuming FLEXIBLE supports.

Use 10% of the vertical load of the pipe as a longitudinal force. (Accounts for friction and thermal expansion.) I know some of you will think this is low but I've done many massive piperacks and these numbers work.

 

[Dinosaur]

Regarding the question by PCX, In bridge design a support that is truely free to expand need only be designed for the friction force developed as the structure (pipe) slides over it. If you have a quality roller, you may want to consider rolling friction rather than sliding friction, but if you are worried about the roller freezing up, then you are back to sliding friction. If the support and structure are restrained, then a stiffness analysis (or other like method) is needed to predict the force. For the pipes themselves, I recommend the "accordion like" expansion configuration where the pipe turns back and forth. This allows the pipe to flex as the temperature changes. Other types of relief mechanisms are available, but I like the accordion types. Good Luck. - Ed

 

[abcmex]

dannym,
two questions:
1. Do you mean 10% of the vertical load as distributed load in the longitudinal direction?
2. What do you mean by assuming flexible supports?
thanks!

 

[havesealwilltravel]

abcmex,

1. Yes - 10% of the vertical (dead + live) load is used as a longitudinal force to be resisted by the beams in minor axis bending locally and at the braced bay globally. This accounts for friction of the pipes due to trermal, pressure, etc. This 10% may seem low for friction, but it is common in many chemical company standards. I've designed many piperacks using it and never had a problem.

2. When calculating the thermal loads from steam pipes etc. don't let the piping stress people assume fixed ends. The supports are beams (weak axis) and they therefore have some flexibility. This flexibility may seem small but it relieves much of the force. Also, for high temperature lines have the pipers design expansion loops to relieve the loads on the steel.

Keep in mind that a pipe rack typically uses smaller members and connections. It is flexible. You don't have the stiffness (and loads) that you have in a highway bridge.

 

[PXC]

Dannym

I am pleasantly suprised that you use only 10% of the vertical load as the longitudinal force. I have previously used a much larger number and this is very significant for design.

Would you still use 10% if there was only 1 large pipe and the rest were small?

Do you know if the value of 10% is based on site measurements or simply on the observation that pipe racks have not fallen down?

If I was to use this lower value in the future, it would be good if you (or someone else) was able to offer a reference such as a research paper or book or design code that can be quoted.

 

[beaver12]

abcmex,

The key thing to remember when designing pipe rack is that you’re providing vertical support to the pipe. Depending on the height of the pipe rack, the pipe rack (usually bents) are quite flexible in the direction of the length of the pipe. You’ll find that the force required to deflect the bents to the same magnitude as the expansion in the pipe is quite small. The beam in the pipe support bent need to be designed for the imposed vertical load and the columns will be designed for the vertical reactions from the beam(s) in the deflected shape. i.e. if the pipe grows 1” at the support location, your bent columns need to be analyze with the top of the column being 1” over the bottom. This will introduce a bending moment in your column so you will have to design the columns as a beam-column.

In long pipe runs where the growth is large, i.e. 3”, you will need to calculate the force required to move the bent 3”. If this force is larger than the lateral load (frictional force between the pipe and beam) then design the beam and bent to accommodate the frictional force. If the force required to move the bent 3” is smaller than the lateral load, then design the columns for the 3” offset.

Hope this help….write back if you require clarification.

B.

 

[PXC]

Beaver12

would you clarify what you mean by bents: do you mean free standing supports ie 2 columns and cross beams, that are not connected to other bents except by the pipes? The pipe racks I have been involved with have all had longitudinal struts/beams ie in the direction of the pipes.

 

[edba]

To everybody,

Pipe racks have transverse bents and looks like a portal frames with several level of beams to support the pipes. And bents are normally tied with longitudinal struts/beams (several level also) between column to column. Columns are supported by isolated footing (piled or unpiled) or combined footing along transverse direction.

Normally the bent spacing are spaced 20’ to 25’ only, unlike in the bridge that the pier/column spacing (longitudinally) are ranging from 50’ or more.

We try to limit the length of pipe rack to 150’ so that the longitudinal drift/deflection would not be more than 1 inch. We also provide a least one brace bay along this length of the pipe rack to ensure stability and the friction/anchor thermal load are transmitted to the foundation.

If the length for example of the pipe rack is longer than 150’, we divide this in two segments either equal or not equal in length as long as the length of each rack is not longer than 150’. We connect them with longitudinal struts with slotted holes connection on one end to prevent transfer of forces and loads run into specific braced bay.

We use 10% of pipe vertical load (pipe empty + operating), uniform loading (applied longitudinally) for a series of pipe 8 inch diameter or less. For bigger pipes, we add concentrated load 20 to 30% of vertical loads in excess of 8 inches pipe full of water. For large friction/anchor thermal loads (say for 60 inch diameter pipe), we provide horizontal bracing, however, it should be verified by pipe stress engineer if the loads really exist.

Many thanks!

 

[beaver12]

PXC,

Yes. Bents are frames with the two columns and a beam (or beams depending on number of tiers req'd). Typically, we design these bents as a moment frame in the direction perpendicular to the length of the pipe. We do join these frames together with beams and bracings. This is done to provide additional longitudinal bracing (when piping req'd). The reason we don't do this all along the length of the pipe run is that once the pipe grow the "x" amount, it does not grow any further therefore no additional lateral resitance (in the direction of lenght of pipe) is req'd.

 

[SweetNlow]

Hold on- there is no one answer! The spacing of the supports is related to the pipe diameter. The filled weight and the material of the pipe (such as steel) control if the pipe can span between the supports. I have lost the charts on diameter vs. pipe span, but it is a matter of logic and structural analysis.

I have been taught to use .2 or 20% of the gravity load as a horizontal pull or push on a seat connection. That is the coeficient of friction between steel to steel in case your bearing materials such as teflon fails. PCA has the coeficients in their manual.

By accordion type do you mean 'Z' shaped? For water pipes changing direction in a building you need an inertia block, mass of concrete, to take the push when the water starts to flow in a big pipe. A battered pile or 'A' shaped structure can handle it.