Application Of Bayes Method In The Determination Of SIL Assessment

Safety instrumented system which are composed of sensors, logic controllers and actuators is able to exercise one or more of the safety instrumented system functions. It is widely used in petroleum, chemical, metallurgy and other high-risk industries. It is an important measure to ensure safety in production. Therefore, to ensure that SIS can properly exercise their intended safety functions is significant to prevent accidents, reduce accident impact.Safety Integrity Level is a measure of the safety performance of SIS. The appropriate SIL not only can guarantee the safety of production and maintenance of personal and property safety, but also can save the production cost of enterprises.SIL quantitative classification requires complete and reliable data equipment failure.At present, our country to SIS equipment failure data collection and analysis work is lack of experience. Therefore, China does not have the molding, available database.In the quantitative SIL validation process, we can only reference overseas database,without fully tapping and using the small amount of equipment failure rate in the process of enterprise operation and maintenance. For the above, this dissertation studies the following:1. This dissertation intensively analyzes the sources,types and advantages and disadvantages of reliability data in the verification process of SIL. In the SIL verification process, forming foreign databases available include: OREDA(offshore reliability data handbook), American Chemical Society process equipment reliability database PERD(process equipment reliability data) and EXIDA Reliability Handbook and other safety equipment. This kind of data is accumulated through a long period of statistics, but it fails to collect all the information about the failure,such as the total running time, failure cause, etc. And the equipment failure rate is constant, the actual failure rate is likely to be variable.2. This dissertation introduces the whole process of the SIS assessment, andmainly studies the SIL verification process. The main verification methods are reliability block diagram, fault tree, Markov model. The first two methods in the calculation of SIL process, in order to improve the application of the simplified model and calculation, use the approximate value. But the results will increase because of the increase of lambda and function test cycle. The calculation results of the Markov model will not produce this error, it can achieve good accuracy.3. For the issue of versatility lack specificity of foreign database and utilization of enterprise samples, the dissertation introduces the Bayes theory. Based on the foreign information database, Then update the foreign database through integrating the collected enterprise sample information. This is the Bayes estimation of failure rate.4. In the petrochemical enterprises of wax oil hydrogenation unit, for example,select one recycle hydrogen desulfurization tower-level protection safety instrumented system. Firstly, we analyzed the risk, and determine the high risk of bias and consequences, in order to determine whether it needs SIS. Then we determine SIL by risk map. Finally, using Bayesian method, verify the SIL with modified equipment failure rate and compare the results of each verification method.