Testing and performance of production- and safety critical valves used in gas pipelines

Abstract

Production- and safety critical valves are important safety barriers used in a pipeline network for production and transport of hydrocarbon gas. These valves, as part of a larger emergency shut down (ESD) systems, protect the facilities and plants that are part of the gas transport system against hazardous situations. They prevent accidents from occurring and escalating. The Petroleum Safety Authority (PSA) demands that these safety barriers are kept safe and reliable throughout the whole lifetime of the facilities and plants, and it is therefore important that Gassco, as the operator of the gas transport system, ensures that the valves are sufficiently maintained, monitored and tested during operation. A sufficient testing regime is a necessary part of the overall maintenance program related to barrier valves in order to verify and ensure that their performance is good and that they function when demanded. Gassco has to verify to the PSA that the overall performance of their valves is in accordance to the safety requirements set for them. This thesis analyses the test results of the production- and safety critical valves that are operated by Gassco. This has been done to evaluate their current testing regime and overall valve performance in accordance to the SIL (Safety integrity level) requirements in IEC 61508. The reported test results include measurements of closing times, internal leak rates and verification of barrier function for all valves. The methods and procedures discussed for analysis of failures reported in test results have to a great extent been based on the recommendations given in OLF 070, which is the guidelines to the IEC 61508 standard, as well as the OREDA handbook 2002 and a selection of other reports that refer to IEC 61508. Based on the literature and reported test results a procedure has been proposed that includes the steps from evaluating the risk of the reported failures and failure rate estimation to performance verification and updating of the test interval. Due to lack of information in the reported test results a number of assumptions had to be taken in the analysis, though the uncertainties related to the analysis are thoroughly discussed. Nevertheless, the results of the overall performance of the valves seem to be good. In the last part a discussion is given as to how and how often the valves should be tested in the future and what is necessary in order to maintain the overall valve performance. The testing routines and methods used and the possible effects that the supplement of partial stroke testing (PST) and valve condition monitoring may have on valve performance are discussed.