For what PED category is piping stress analysis compulsory

If we consider only PED requirements (so we are not considering Temperature, lines connected to pumps, and so on), when does PED require a compulsory stress analysis?
I see many customers that, considering PED, require stress analysis only for PED III lines.
For example, for PED III lines, according yo Module B (for example) , “results of calculations made” shall be provided.
But this requirement is present in all available modules for PED I and PED II lines as well.
So looks like stress analysis is always compulsory for PED category I, II and III.
What do you think?


We have dealt with this issue in the past as well, and have been in contact with Lloyd’s Register on this. This is their point of view on things

  • For cat III, a stress analysis is required since the applicable modules requires a review of the calculation by a NoBo (= notified body). This is NoBo-indepedent, so to speak
  • For anything greater than cat art 4.3 (aka module SEP), Lloyd’s said a calculation needs to made. They refer to Annex I for this, where several sections mention something about calculations. Specifically for piping, the PED lists in Annex I, para 6 some requirements which Lloyds used as argument.

The above is not per se my point of view, but how Lloyds explained this in one of their seminars. Other Nobo’s may or may not agree on this approach. Personally I don’t agree, but in the end, PED requirements per Annex I must be met. There’s a lot of grey area, esp. for para 6 of Annex I, e.g.


Design and construction shall ensure:

(a) that the risk of overstressing from inadmissible free movement or excessive forces being produced, e.g. on flanges, connections, bellows or hoses, is adequately controlled by means such as support, constraint, anchoring, alignment and pre-tension;

How does one define ‘design’ and ‘adequately controlled’? Are basic pipe span charts sufficient, or other rules of thumb? Or is a comprehensive anaylsis (e.g. using CII or AutoPIPE) always required?

The requirements are the same regardless of module (for most things - there are exceptions, see Annex I 3.1.2, 3.1.3, 4.2© and 4.3). However, experimental methods for piping obviously don’t make any sense. Piping almost never is the same, and how would you want to do such a test?

On the other hand, we have built numerous cat I and cat II piping systems in our installations; not one time a NoBo did ask for the calculation. In the end it boils, I guess, that you, in some way, need to convince your NoBo you have fulfilled the requirements of Annex I. Keep in mind one simple way to do so is by applying a harmonized standard. For piping this is EN13480. If you take that code and follow it 100%, you should be good. However we use B31.3 for our piping systems, also for plants in the EU, as EN 13480 still contains too much bugs and other issues.

Hope this helps.

Above is a snippet.

@XL83NL, This is a very thorough and thoughtful reply. Thank you for using forums to share and learn.