Design and calculation of pipe bendings 1

[carletes]

Dear all!!

I am facing a piping system where elbows are going to be replaced by pipe bendigs carried out on site in cold condition, for small pipe sizes (NPS<2&quot.

Is there any code that states the minimum radius of bending as a function of material, pipe thickness etc or a method to calculate it? I have been reading ASME B 31.1 and I am afraid this case is not covered.

I know some rules like the minimum radius must be 2 times the diameter of the pipe, for instance, but I would like to check my particular case as the thickness of the pipe is quite small.

Thanks a lot for any help.

Regards

 

[es335]

Hi.

I think that you can find data for pressureloss and bends in almost every fluid dynamics book.
Try eg. introduction to fluid dymanics by Fox & McDonald.
Look under internal incompressible viscous flow.

Otherwise try looking in VDI or other collections.

But I think you are correct. If the bend diameter is approx 2-3 times larger than the tube diameter, a minimum flow loss should be obtained. But be aware. The loss rises when the bend diameter grows and this is proberly due to the length og the pipe begins to have an influence, hence it is not only the bend itself that is giving the loss. This might clutter the division between minor and major losses.

Hope it helps.

 

[davefitz]

I interpret your question as a fabrication question and not a pressure drop issue.

Each material has different requirements and limitations regarding the amount of cold work permitted without hot bending or without heat treatment . Also, there is a permissible amount of out-of roundness that connot be exceeded, which is related to the bend radii.

Finally, B31.1 does not provide any allowance for thinning of the bend extrados, so the bent section will need to be a heavier schedule of pipe than the straight sections in order to have the same design pressure.

 

[samuelliu]

The cold bending ranges can vary significantly with the process and degree of specialized tool used. For details, refer to PFI standard ES-24-2001 &quot;Pipe Bending Methods, Tolerances, Process and Material Requirements&quot;.

 

[1969grad]

Normal practice is to use elbows with 1.5D bend radius. Unless you intend to pig the line, why do you need a larger bend radius?

 

[TEV]

The concept that pipe/tubing needs to be increased in wall thickness to allow for thinning during bending is not universally true. It is strongly dependent on the alloy and its condition. Austenitic stainless steel tubing, for instance, is generally supplied in the annealed (soft) condition. The strain hardening due to bending more than compensates for wall thinning. We have hydrostatic tested many, many bent stainless tubing samples, .035, .049, and some heavier wall, to failure. They ALWAYS fail in the straight section, not in the bend. Austenitic stainless steels and similar alloys (nickel base CRA alloys, for instance) strain harden rapidly and have good ductility in the annealed condition. They can be readily bent without concern for wall thinning.

Samuelliu: Can you provide further information on where to find the reference you mention, PFI Standard ES-24-2001? Thanks!

 

[davefitz]

TEV:
Yes you are correct that it is not required to increase the wall thickness due to physical realities, however some codes ( ASME B31.1) do not recognize that aspect of reality and mandate that the bends shall be made thicker.

The german boiler code TRD has a explicit calcularion procedure to allow determination of the need to add more thickness- I think it is based on using toroidal pressure equations instead of cylindrical equations.

 

[NimalJayaratne]

See ASME B31.3, Clause 332, you will find some guidelines. The essence is the limitation of residual strain in the material. Acceptable residual strain is dependant on the service of the pipe. As you are using B31.1, I believe your service may be water or steam and these services do not have an impact on the degree of residual strains in the pipe material. e.g. in sour services like wet H2S the hardness related to residual strains are important. One way to remove these built in strains is to heat treat (PWHT) the pipe to resotre the microstructure.

Make sure that the final thickness on the outer radius of the bend (extrados) is adequate to withstand the design internal pressure.

 

[kyong]

Some codes require thinning due to bending be compensated by using thicker pipe. However, the term &quot;THICKER&quot; doesn't necessary mean &quot;thciker than straight region&quot;. Rather it means the pipe before bending should be thicker than REQUIRED thickness after bending. TEMA has some guidelines for bending of U-tubes. It may give conceptual help.

RCB-2.31
to = t1 * (1 +do/(4R))
to = tube wall original thickness before bending
t1 = minimum tube wall thickness
do = OD of the tube
R = mean radius diameter

further, it limits thinning to be not over 17%
of original thickness for materials which are relatively
non-work hardening and of suitable temper.

and the flattening shall not exceed 10% of tube OD.

I think actual thinning may vary depending on many factors. Method of bending would be one of the factors.

 

[Kencintube]

From a pipe and tube benders perspective, a good rule of thumb is that the radii for bending be a multiple of 3 or 5 times the nominal pipe size.

These are common bending sizes that will allow for the replacement in most cases Long Radius (LR) fittings for bends. Pressure drop is also reduced by approx 35% to 40% for each size.

 

[gstark]

I work for an aerospace pneumatics system supplier and we routinely design for and manufacture swept mandrel bends in 1.5-D and 1.0-D configurations. Many of these are in Inco alloys and comm pure titanium, with gages from .016 inches to .020 inches thick. This is a region far away from BPV code work, etc, with D/t ratios from 62 to over 400. Typical line sizes are from 1.5&quot; dia to 6.00&quot; dia, with nominal wall thickness of .016 to .032 inches. We look to codes for SIF values and other general guidance, such as local flange influence on stiffnesses, etc. I've had some interesting personal dialogue with Rodabaugh regarding this very subject.

The comments regarding strain hardening are very appropriate - not to mention the thinning that occurs can easily exceed 30%. The thinning is typically localized over an area along the outer radius. This does not automatically dictate a gage increase - stress analysis and FE results, and particularily fatigue considerations rule this world.

Whether one know it or not, any time you fly in a commercial aircraft, you are airborne with pneumatic systems that are not governed by any specific codes such as 31.1/31.3, etc. SAE ARP699 provides general design guidance, but the general system design is defined, analyzed, managed, and developed by individual companies. Basic FAR/JAR requirements are met during certification and qualification testing. Rest assured that a significant amount of evaluation and analysis goes into these systems, perhaps more so than other industries. 600 psig/1200F bleed air systems are not uncommon, so rigorous testing is conducted. After all, when you are over the mid-Atlantic, who wants an engine shutdown on a twin-jet? Gives you a new perspective to ETOPS certification (Extended-range twin-engine operations, or as other say it: Engines Turning or Passengers Swimming).

 

[zeven]

There is no design limit to the bend radius - just the capabilty of bending the pipe without flattening and reducing the wall thickness below the minimum wall thickness. Rippling of the intrados should also be avoided.
Small bore pipes (NB2&quot; and below) have usually enough extra wall thickness to avoid any thinning problems. Flatting or ovality, depends on the material, radius and method of bending. 3 x D is usually the minimum radii.
As far as the line flexibility is concerned, it will be stifffer and the stress range loads will go up.
However stresses may be reduced at bends due to the lower sif than the replaced elbow.
Stresses for other cases should go down.