Final piping layout and support design are highly dependent of its global mechanical evaluation and should be properly defined during design phase. When necessary, several layout iterations may be needed until a satisfactory solution is found. These iterations are time consuming and all changes must be propagated to other disciplines involved in the project (i.e.: civil works, stationary equipment, rotary equipment, instrumentation, etc...). Even though during its operational life a piping system may experience several loading conditions, one of the most important is thermal expansion. The importance of this loading condition increases along with pipe lengths and temperatures, in many cases dictating the pipe route that must be followed.

Piping Stresses and Loads

Some possible failure modes that can be associated with these loading conditions based on type of loading and/or environmental condition.

Failure modes for piping and their cause

Sample Results
Consider a sample piping system consisting of four parallel air-coolers. Each air-cooler has four inlet nozzles that are connected to the process unit through a 20” diameter, ¼” thick, ASTM A672 B60 pipe header. The final section of the feed pipe consists of an extremely flexible pipe arrangement in order to reduce piping loads on the equipment nozzles. The fluid is a hydrocarbon gas at 450 °C.
The PCS model consisted of 142053 linear quadrilateral elements of type S4R, geometrically non-linear and the material model was linear elastic. The evaluation of the four load steps took less than 6 minutes on a standard desktop computer (i7-4790 @ 3.6 Ghz), running Abaqus Standard on a single processor.
Figure 15 present PCS stress results for the load step WNC+P1+T1, which consists of loadings from piping weight with no contents (WNC), pipe pressure (P1), and pipe temperature (T1). These results were compared to the ASME B31.3, calculated using Coade Caesar II 5.1, a standard pipe flexibility analysis software.
The stresses computed with Caesar II converted into cycles through Markl’s methodology were compared to cycles calculated according to ASME master S-N curve. The master S-N curve approach resulted on a fatigue life 20% higher than Markl’s. Differences of up to 120% on nozzles reaction forces were found between the two approaches. These differences may be attributed to inaccuracies of ASME B31.3 flexibility factors that are of great importance on compact regions with several pipe fittings, such as the nozzle region of this model.
The Caesar II calculation, indicates that the piping system needs to be redesigned due to overstress, while calculation through PCS and ASME VIII div. 2 part 5 clears the system for safe operation.

Sample air-cooler model

PIPE SUPPORT TYPES

Standard Beam Support
Pipe is rested on or secured to a support member usually consisting of a
standard structural shape. I-beam, wide flange beam, angle, channel etc.
The pipe may be secured to this member with a stabilizing U-bolt. (Fig. 1 & 2.)

Figure 1. Typical I-Beam Pipe Support

Figure 2. U-Bolt Stabilized Beam Supports

Saddle Clamp
Pipe is clamped between two rolled plates, one of these plates has a structural element welded to it which attaches the pipe to the support structure. (Fig. 3 & 4.)

Figure 3. Typical Half Saddle Clamp

Figure 4. Full Saddle Clamp

Welded Supports
This type of support involves welding a part to the pipe and then it is usually
free to move at the interface to the support. There are a number of variations
on this theme, this is a common approach for insulated piping systems. (Fig.
5.)

Figure 5. Typical Welded Pipe Support

Others
There are a number of other methods used, such as flange bolt supports,
various type of pipe hangers and other specialty type supports, however the
first two categories account statistically for better than 95% of support points
on a typical offshore structure.

The industry has long been aware of the problem, but has failed to appreciate the true causes; this is evidenced by some of the solutions that have been implemented to stop the problem which have actually accelerated the problem.

Rubber Pads & Liners
As previously stated, it was thought that the metal-to-metal contact was the
main problem, hence if this were eliminated the problem would also be
eliminated. The use of rubber pads of some type has been and is still thought
to solve the problem. Not true (Fig. 8.) In fact, rubber pads under pipes do a
wonderful job of reducing the life of the pipe. The crevice that was formed
without the rubber pad is mild in comparison to the new crevice, which now
has the ability to actually suck water in (by capillary action). Not only is it
better at getting water in, it is better at holding it, since air circulation and
natural evaporation is eliminated. The situation is further worsened by the
length of the crevice which allows an oxygen concentration gradient to go
from full natural concentration to anaerobic in a few centimeters.

Figure 8. Rubber Pads Accelerate Crevice Corrosion

Fiberglass Pads
Contoured pads attached to the pipe at support points (Fig. 9.) Obviously another attempt to eliminate metal to metal contact. This is better than the rubber pads but still allows a crevice to be formed at the pipe surface.

Figure 9. Fiberglass Contoured Pads Still Risk Crevice Corrosion