INTELLIGENT SIMULATION METHOD FOR ANALYZING RELIABILITY BASED ON NO-STEADY MODEL

OF MATERIAL FATIGUE DAMAGE

Li Jia  Ping An  Wang Dejun

(Northeastern University)

Lu Fengwu

(Shenyang Toshiba Elevator Co. , Ltd.)

Abstract: The fatigue damage of material is impacted by many factors and regarded as an unstable process. The relationships between its momentary fatigue damage and material properties, loading stress behave so complicate and no-linear that it is very difficult to depict the process precisely applying definitive mathematical functions. Theoretical hypothesis of uniform energy dissipation and linear cumulative damage are avoided, artificial neural network technology is adopted to describe it, the fatigue damage process of material can be depicted really. Dynamic failure criterion is set up, which is considered the random of material proper and loading comprehensively, it can be used to correctly reflect the individual characters and relationship between them. Discrete simulation theory is also adopt to build an intelligent simulation system for analyzing fatigue life reliability of material. The test result has proved that the analysis accuracy is very high for normalizing 35 steel under random loading.
Key words: Fatigue damage  Artificial neural network  Reliability analysis  Intelligent simulation
CLC No: TH114  TP18
国家自然科学基金资助项目. Received 20000607, received in revised form 20000712