Choice of Demand Mode for Subsea Safety Systems
MetadataShow full item record
In the real industry project where new technology and systems have blossomed and been codified, many problems and challenges regarding the application of international safety systems standard IEC 61508 in terms of low demand and high demand have been encountered. For instance, the automatic train protection system (ATP) is argued to be a both low-demand system and a continuous mode system. A low-demand blow-out preventer (BOP) system will operate in the high-demand mode to withstand the well pressure for hours and weeks when it is activated to full closure. In the real case, both reliability assessments for low and high demand modes could be requested due to the vague and ambiguous statement of concepts and definitions in the IEC standard. The current existing research placed the main emphasis on low demand systems, but with few addresses on high/continuous demand systems. This master's thesis which is written in cooperation with Aker Solutions aims at discussing those problems confronted in IEC 61508 and summarizing the existing academic work, mathematical models as well as relevant industry guidelines. By assuming the system will operate in both low demand and high/continuous demand mode, the Subsea High Integrity Pressure Protection System as a case example is used to illustrate the problems in the case study. Both the probability of failure on demand (PFD) and the probability of dangerous failure per hour (PFH) are calculated as per the PDS method in the low and high/continuous demand mode reliability assessment. The results from the case study in terms of SIL are found to be inconsistent. The problem is discussed and traced back to the general quantitative SIL allocation method in IEC 61508-5. The thesis therefore attempts to develop a general consistent SIL range by carrying out SIL calibrations with Matlab. The new correction factor is calculated and a new proposed SIL table is proposed. The decisions in the following IEC 61508 SIL table are further argued and both pros and cons of different solutions are compared.
Full text not available