Pipe line failure statistics

[ssk]

Thanks SAFETYUSER and ARMOFI. Sorry for the delay in thanking you for useful inputs. Still I am little bit confused. Why one should consider rupture scenarios ( either vessel or pipe line) when there is no past history of ruptures in that particular type of Industry. Failure frequencies given in many data bases are based on the data published in 1970 and mainly based on steam generators or chlorine vessels or Ammonia vessels and that too on older technologies. Suppose I have got a vessel which handles Hydrogen Sulfide. Now, the data given in the failure data bases can not be applied for my vessel as services are different. But I have got 100 years of data on Hydrogen sulfide vessels and pipe lines, which says that there is not even a single rupture of either pipeline or vessel, only pinhole/----- failure were reported. Now for my QRA, whether should I consider the rupture scenario? Going by World Bank guidelines, IAEA guidelines, AIChE publication (QRA with data tables), HSE guidelines, one should give preference to plant specific data. NOW, please let me know whether should I consider rupture scenarios? If it is to be considered, which did not happen for 100 years of experience, then plant can not run as H2S, you know, highly toxic.

[ARMOFI]

Thanks SAFETYUSER and ARMOFI. Sorry for the delay in thanking you for useful inputs. Still I am little bit confused. Why one should consider rupture scenarios ( either vessel or pipe line) when there is no past history of ruptures in that particular type of Industry. Failure frequencies given in many data bases are based on the data published in 1970 and mainly based on steam generators or chlorine vessels or Ammonia vessels and that too on older technologies. Suppose I have got a vessel which handles Hydrogen Sulfide. Now, the data given in the failure data bases can not be applied for my vessel as services are different. But I have got 100 years of data on Hydrogen sulfide vessels and pipe lines, which says that there is not even a single rupture of either pipeline or vessel, only pinhole/----- failure were reported. Now for my QRA, whether should I consider the rupture scenario? Going by World Bank guidelines, IAEA guidelines, AIChE publication (QRA with data tables), HSE guidelines, one should give preference to plant specific data. NOW, please let me know whether should I consider rupture scenarios? If it is to be considered, which did not happen for 100 years of experience, then plant can not run as H2S, you know, highly toxic.

As far as my knowledge goes, consider the pipeline/vessel failure frequency irrespective of its' content. Content properties are analyzed in consequence analysis such as flammability, Radiation, toxic effect, over pressure effects. etc. QRA considers all potential release models. Subsequently it's consequence, frequency, severity ,risk rating predicted.

[safetyuser]

Now, the data given in the failure data bases can not be applied for my vessel as services are different. But I have got 100 years of data on Hydrogen sulfide vessels and pipe lines, which says that there is not even a single rupture of either pipeline or vessel, only pinhole/----- failure were reported. Now for my QRA, whether should I consider the rupture scenario?

One could reply that you must consider pipeline rupture because so is requested by the regulation in force and it is the" good engineering practice" or the state of the art in the QRA field.

Another one (included me) could reply that my failure database do consider at least 1 rupture of pipelines and that is in principle a theoretical chance. Theoretical, of course, but not sufficient to be neglegted on "a priori" basis.

If it is to be considered, which did not happen for 100 years of experience ...

Also Piper Alpha destruction accident did not occur before of that time, and I guess some people made the same consideration: "It never happened for 100 years of experience". Think about Deepwater Horizon ...... again in that case the specific problem "never happened for 100 years of experience" but ... shit it happened!

Please bear in mind that we're talking of events with a failure period in the order of thousands of years (1E-3 for acceptance criteria): neither me nor (possibly) the majority of this forum members will survive so long to witness such a catastrophic failure event........but it is still a credible event to account for.

NOW, please let me know whether should I consider rupture scenarios? If it is to be considered, which did not happen for 100 years of experience, then plant can not run as H2S, you know, highly toxic....

Again, you are here misusing the concept of "credible events" with "tolerable or acceptable risk" .....