December 28, 2019

Several ASME B31 & EN 13480 Issues Needed to Know for Pipe Stress Engineer

By Alex Matveev

Real Bend Wall Thickness is Greater than the Matching Pipe Wall Thickness

This is a very serious issue. It leads to an underestimated pump, nozzle, and support loads and stresses calculated by ASME B31 codes!

ASME B16.9 and all ASME B31 codes don’t regulate the bend, tee and reducer wall thickness. Only the pipe wall thickness is regulated. So many people think that the elbows and other fittings have the same or almost the same wall thickness as the matching pipe. But in most cases, the real bend, tee, reducer body wall thickness is greater than matching pipe wall thickness with the same Schedule.

For elbows, the real wall thickness can be 10%-40% greater than the matching pipe. Because bends must have greater wall thickness to hold the same pressure as the connected straight pipe (see 304.2.1 3d).

Manufacturers usually produce the bends with a greater wall thickness than matching pipe, but we can get real bend wall thickness only after contacting the manufacturer or even measure it after delivery.

Click to enlarge

Piping designers usually know nothing about it. And piping stress engineers usually use the pipe wall thickness for elbows when using piping stress analysis software. Leaving the “Fitting Thk” field blank makes software thinking that elbow has the same wall thickness as a connected pipe element. This is a serious mistake!

Click to enlarge

According to the ASME B31 and other ASME B31-based codes bend flexibility factor depends on real bend wall thickness, not on matching pipe wall thickness.

The greater bend wall thickness, the greater is bend stiffness (k-factors) and greater are loads on rotating equipment, nozzles, supports and expansion stresses in piping system.

This problem quite often comes to light when Russian companies try to check the design made according to ASME B31 codes for the Russian market. While rechecking the stress analysis using PASS/START-PROF software according to GOST codes a lot of error messages appear. They say that the wall thickness of the elbows is lower than the minimum required one to hold the pressure because it is usually left blank in CAESAR II and software takes fitting thickness equal to connected pipe WT in the piping stress model. When the real elbow wall thickness entered and model recalculated, the nozzle loads and stresses become much greater than it was calculated in CAESAR II and other software! That’s because the elbow flexibility k-factors used during analysis was incorrect.

But in real practice counterparts usually can’t provide the real bend body WT. They just don’t have this information!

We received different answers to the direct question – “what will be the real wall thickness throughout the whole bend body?”

  1. Some manufacturers do not answer at all;
  2. Others say it’s a trade secret (?!);
  3. Some manufacturers replied that the edge will be 100% consistent with that ordered according to ASME (according to schedule), but the thickness of the wall at the bend may even be 40% greater!

Only after we receive the ordered fittings (bends, reducers, and tees) from the factory, only at this time we can measure and find out what are the real wall thicknesses. The stress analysis model should be changed, nozzle loads become greater. The design should be changed to add more flexibility and reduce nozzle loads and sometimes expansion stresses. And all of this should be done after the design job by our contractors was formally “finished”. Amazing!

For example, Russian standards, which are completely different from ASME B16.9 for bends, tees, and reducers, always provide the real body wall thickness for each fitting. Manufacturers follow the standards. Every piping stress engineer knows the real body wall thickness of bends and other fittings and specifies it in START-PROF while performing piping stress analysis. Also, all the bend properties can be taken from the fitting database (see screenshot below).

Click to enlarge

All RD, GOST, and SNiP stress analysis codes (power, process, oil & gas main pipelines, etc.) provide detailed wall thickness calculation procedure for all fittings including bends, tees, and reducers. On screenshots below, you can see that the calculated bend wall thickness is always greater than pipe wall thickness for the same pressure load.

The real bend body wall thickness should be used in piping stress analysis instead of matching pipe wall thickness. To solve this problem we added the special feature in PASS/START-PROF software that allows calculating the approximate “real” bend wall thickness on-the-fly according to ASME B31.3 304.2.1 and the same requirements in other ASME B31 and EN 13480 codes. Just push the button “C” near the “Wall Thickness” field and it will be calculated according to the code requirements.

Conclusion

  1. Bend, tee and reducer wall thickness should be regulated by ASME B31 codes and provided in ASME B16.9 code. Manufacturers should produce bends with body wall thickness according to the code requirements.
  2. Until the first problem is solved, the manufacturers should provide bend wall thickness in their catalogs. It will allow designers and piping stress engineers to use the real WT in the pipe stress model and to get accurate nozzle loads and expansion stresses.
  3. There should be a special remark in ASME B31 codes that explains how to calculate flexibility k-factors for the elbows if the real body wall thickness is unknown.
  4. If the elbow wall thickness is unknown, then piping stress engineers should use WT calculated by ASME B31 code equations for bend or use pipe wall thickness multiplied by 1.4 factor. This will provide more conservative design, and after the real bend, wall, thicknesses will become available (can be measured) the changes in piping design will not be as critical, as now.

Pages: 1 2 3