Reliability assessment for liquefied natural gas ship's distributed control system

Abstract

Distributed Control System (DCS) is installed onboard liquefied natural gas (LNG) ships to ensure safety and efficiency during all modes of operations. The aim of this paper is to establish reliability assessment for ship’s DCS and measure the probability of failure for DCS equipment by using field failure data. DCS components failure data of four series of LNG ships was taken from the company maintenance database. Reliability of individual component was calculated by fitting the two parameter Weibull model to failure data. The shape parameter, ß and scale parameter, ? are deduced from linear regression and subsequently the reliability for each DCS component can be calculated. The reliability block modelling was developed and then the overall reliability of DCS system of a ship can be calculated by substituting each individual component reliability into the equation of total system reliability. There are five main components of DCS i.e. FTA, IO, NIM, CPU and HIS. The failure of IO and HIS made up 88% of total failure. Ship A and B having large number of failures with 36% and 46% each. The linear regression for all ships showing good fitting as the R2 value is all above 0.8 even for small sample data. IO on all ships was showing short lifetime with low ? value of about 10,000 hours while HIS was even lower with 3,000 to 8,000 hours. Other component ? value was minimum 13,000 hours. The total reliability for Ship A, B, C and D was 53%, 0.1%, 2% and 6% respectively. Ship A reliability at this point of time is showing the highest reliability compared to all other ships. Other ships’ low reliability was due to either high failure rate of one of the components or early infant failure. It can be concluded that reliability of DCS are not influenced by the number of component failures. Even though Ship A has the most number of component failures but the reliability is the highest. The DCS system total reliability depends on the failure rate, the criticality of the components whether connected serially or parallel and the date of when the last time the component was replaced.