STRUCTURAL RELIABILITY ANALYSIS METHOD BASED ON COLLABORATION-MEAN POINT CONSTRAINED ACTIVE LEARNING SURROGATE MODEL
WEI YanXu
ZHANG ShiLong
ZHANG YanJie
Abstract：The reliability of mechanical structures is crucial for their safe operation,to address the problem of low accu⁃racy and low efficiency in reliability analysis of complex mechanical structures,a new active learning surrogate model based reli⁃ability analysis method was proposed.The spatial location characteristics of excellent fitting samples were studied and three con⁃straints,such as surface constraint,distance constraint,and domain constraint,were proposed accordingly.Correspondingly,three control functions were established to achieve the three constraints.Then,three control functions were organically collabo⁃rated,and an effective new learning function,collaboration-mean point constrained learning(CPCL)function was proposed.Combined with the augmented radial basis function(ARBF),a collaboration-mean point constrained active learning surrogate model(ARBF+CPCL)reliability analysis method was established.Finally,three cases were employed to verify the high compu⁃tational accuracy and computational efficiency of ARBF+CPCL reliability analysis method,and the application ability of ARBF+CPCL method in practical engineering cases was proved through the reliability analysis example of the turbine disk.
Keywords：Mechanical structureReliability analysisActive learningTurbine diskFitting sample
Publication Date：2024-12-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：10( 1365-1374 )