| With the development of reliability theory, Reliability analysis is becoming more and more important in engineering. How to improve the calculation accuracy of reliability has been one of the hottest issues. As we all know, linear approximation method can improve the calculation accuracy of the reliability. What’s more, it’s more accurate than other methods of reliability analysis. Especially, the combined model of linear performance function will affect the calculation accuracy of linear approximation method. Based on the minimum cut set theory, the thesis presents a method which using the minimum cut set of the failure area of linear performance function express the failure area of the original performance function. Finally, the examples verytify that the method can express the original failure area more accurately and further improve the calculation accuracy of reliability. This thesis made the following studies:1. Studying and analysing the major linear approximation methods (such as feng, mahadevan and guan), a principle problem of the linear combination model is produced when calculate the complex issues. That’s why the major linear approximation methods are difficult of further improving the calculation accuracy of reliability.2. Using the correlation computing method of system reliability analysis, the thesis presents a method of establishing the sub-linear performance functions by rotating the main linear performance function, which can avoid the arbitrariness of establishing sub-linear performance function in previous methods.3. Based on the mimimun cut set theory, the minimum cut set model of the failure area of the main and sub-linear performance function, which expresses the failure area of the original performance function is obtained, which can avoid the principle problem of the linear combination model when calculate the complex issues. Finally, the reasonable and correctness of the application of minimun cut set theory is verified by examples.4. Using the function contains three variables as an example, the paper discuss the method of improving the calculation accuracy of multidimensional reliability. The results shows that as long as the linear combination model is correct, we can get sufficient accuracy of reliability calculation by given enough linear approximation function. |