Thermal interaction analysis in pipeline systems - A case study

Authors
Citation
A. Mazzola, Thermal interaction analysis in pipeline systems - A case study, J LOSS PREV, 12(6), 1999, pp. 495-505
Citations number
21
Categorie Soggetti
Chemical Engineering
Journal title
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
ISSN journal
09504230 → ACNP
Volume
12
Issue
6
Year of publication
1999
Pages
495 - 505
Database
ISI
SICI code
0950-4230(199911)12:6<495:TIAIPS>2.0.ZU;2-9
Abstract
The assessment of the consequences of high pressure releases of flammable g ases is a fundamental requirement for the safe design and operation of indu strial installations, plants and pipework. A scenario of interest concerns a high pressure jet-fire following the ignition of a gas jet release which results in a thermal loading to the surroundings and possibly leads to acci dent escalation. In the present paper, a case study is presented: two parallel-laid natural gas pipelines have been considered, the accidental scenarios which may poss ibly occur as a consequence of a pipeline failure have been discussed and t he thermal effects caused by the jet-fire developing from different rupture sizes have been assessed. Three scenarios have been analyzed, considering the pipelines being within a highly congested area: (i) large failure and vertical jet with detached f lame; (ii) small failure with jet fire directly impinging on the parallel p ipeline; (iii) small failure with pipeline engulfed within fire. Once the temperature gradient through the pipeline wall has been found, the stresses deriving from pressure load and steel differential expansion have been analytically calculated and compared with the yielding stresses at th e temperature achieved by the pipeline wall. In the first scenario the pipeline is able to resist without major problems ; in the second case the pipeline rupture is likely to occur; in the third scenario the pipeline resists to the applied loads bur with a low margin to yielding. It is understood that the analysis results are very much dependent on the u tilized hypotheses, therefore a sensitivity analysis was performed in order to assess the variation of the results as a function of the variation of p roblem data; this analysis identifies the large influence of the parameters on the final result. (C) 1999 Elsevier Science Ltd. All rights reserved.