Research On Quantitative Reliability Model For Safety Instrumented Systems (SIS) Based On SPN

Safety Instrumented Systems (SIS), known as automatic safety protection systems, have broadly been applied in the process industry. As a matter of fact, more demands are placed on the reliability of SIS given the facts that the process industry involves complex large-scale installations, long-period operation of equipments in addition to high-parameter. To this end, many researches on the reliability of SIS have been conducted. Currently, four techniques, simplified equations, reliability block diagrams (RBD) model, fault tree (FT) model and Markov model, are frequently employed in quantitative analysis on SIS reliability. Simplified equations, RBD model and FT model are easy to use and construct, their advantage is undermined by their limits in representing dynamic behaviors of the system. Markov model, on the other hand, excels capturing many aspects of system dynamic behavior and, meanwhile, obtaining several reliability metrics of the corresponding system. However, the model is subject to model building, exponential restriction and state explosion problems.In copying with the existing difficulties in conventional reliability models for SIS, this thesis proposes a new reliability model based on SPN theory (SPN model). It introduces the building and analysis procedure of SPN model, then builds typical architectures of SIS by considering many related safety factors of SIS, including failure mode, periodic testing, Common Cause Failure (CCF), diagnostic function, repair strategy, etc, and finally analyzes these models with supplementary variables method. Several reliability indexes of system, such as PFD , PFDavg , MTTF and MTTFS , are obtained. This paper demonstrates that SPN model provides a new approach in terms of the reliability research of SIS, which possess certain practical engineering value.The main work of this research can be summarized as follows:1. The advantages and disadvantages as well as the application scope of conventional quantitative reliability model for SIS is analyzed and compared.2. The building and analysis procedure of SPN model for the reliability of SIS is proposed in detail by analyzing the logical relation of the structure of SIS and its subsystem with RBD model.3. Considering the building and analysis procedure of SPN model, corresponding SPN models for safety related factors of SIS are constructed. On the basis of these models, SPN models for typical SIS architectures are constructed. With regard to complex SIS, high-level SPN model for SIS is presented based on high-level SPN theory.4. In order to analyze Non-Markov SPN models of typical architectures of SIS, supplementary variables method is used, several important reliability metrics are derived, including PFD , PFDavg , MTTF , MTTFS , etc, and the feasibility of the SPN model and related algorithms are illuminated by comparing with the analysis results of FT model and Markov model.5. A reliability analysis prototype tool based on SPN named RelSPN@zer is developed on MATLAB platform. Many reliability metrics of system can be automatically obtained both under stationary and transient state. With a practical example, the modeling process of SPN model is illustrated, meanwhile, the analysis and calculation functions of the tool are verified.