Sammendrag
The integrity of SIS shall be maintained during its lifetime including operational and maintenance phase. Guidelines for follow-up SIF in the operating phase by SINTEF is one of the guidelines widely used, but it has not been updated for more than ten years. It is desirable to evaluate the applicability of this guideline for the existing maintenance data. The main objective of this master thesis is to use the failure notification data to analyze SIS performance during SIS follow-up activity. The starting point is classifying the failure notification data into DU failures. The simplified FMEDA is found as a feasible method. The OREDA Multi-Sample is used to calculate the aggregated failure rate for detector type and the detector model. The Bayesian method is used to calculate the failure rate for each model in a facility. The Bayesian method is required a priory failure rate as prior knowledge. It is investigated that the aggregated failure rate by OREDA Multi-Sample can be used as a priory failure rate. The master thesis concludes that the guideline is found practical and useful to be used in the existing facility. However, a few modifications can be valuable. The proposed modifications are defining a method to classify DU failure, updating the formula to calculate Bayesian failure rate, and updating the method of doubling or halving the test interval.
