An Optimization of Master S-N curve Fitting Method Based on Improved Neighborhood Rough Set

In this paper, a novel master S-N curve fitting method based on improved neighborhood rough set is proposed. The neighborhood relationship is used to granulate the fatigue information system of aluminum alloy welded joints, and the neighborhood rough set model is constructed from the information perspective. Based on conditional entropy, a neighborhood rough set attribute reduction algorithm (NRSBCE) is proposed. By means of this algorithm, comprehensive mathematical analysis model for fatigue life influencing factors is established and optimization of the S-N curve fitting method is introduced. In this work, the fatigue decision system of the aluminum welded joints is constructed firstly. The key influencing factors of the welded joints are obtained by means of NRSBCE algorithm subsequently. After that, the concept of fatigue characteristic domain is introduced according to the reduction result of the NRSBCE algorithm. Then, the master S-N curve fitting is carried out according to the divided fatigue characteristic domain. Correspondingly, a set of S-N curve clusters are fitted. Finally, fatigue analysis system for welded joints based on the fatigue characteristic domain is designed and developed. In the system, the equations of master S-N curve based on the mesh-insensitive structural stress method and the master S-N curve cluster based on the fatigue characteristic domain could be both obtained by uploading the practical fatigue test data of the aluminum alloy welded joints. This work helps to further reduce the dispersion degree of the fatigue sample data of the aluminum alloy welded joints thus provide reference for fatigue design.