A reliability model for optimization of test schemes for fire and gas detectors

Safety system detectors must be tested periodically to check that they would respond to an actual demand. A detailed reliability quantification model is here given for fire and gas detectors. The main objective of the model is to provide a tool for a realistic quantification of the safety of a detector, using combinations of periodic tests (both manual and automatic). The model can be combined with a life cycle cost (LCC) model to identify a combination of detector design and test scheme for optimization of safety and acquisition/maintenance costs. The reliability modelling is based on a detailed failure classification and accounts for all types of safety-critical failures actually observed in operation. Thus, failures that are caused by design errors (and later triggered by specific inputs) are handled by the model. Also, failures introduced during the testing are modelled. The modelling includes an approach for assessing the total coverage of any combination of tests, utilizing expert judgements. A quantitative example comparing two different designs/test schemes is also presented.