A Crack Growth Prediction Method Based On SIR Particle Filter

The residual life management of the structure is one of the key technologies in the area of aircraftfailure prediction and health management system. This paper proposes a kind of particle filter methodand combining the structural health monitoring method of the fatigue crack propagation prediction,and it can effectively eliminate the effect of uncertainty of the fatigue crack propagation parametersand other engineering. Innovation and work of this paper are as follows:(1) As for the dispersion of the material parameters and the nonlinearity of fatigue crackpropagation, fatigue crack propagation prediction method is studied based on the sequentialimportance sampling and resampling algorithm. It can effectively eliminate the dispersion of thefatigue crack growth caused by the uncertainty of the parameters.(2) High precision simulation of the high stress gradient at the crack tip is realized by using thesingular element. Using Abaqus_Python Secondary development on the crack-tip singular elementautomatic mesh generation, the crack tip stress intensity factor range of different crack length, angle iscalculated effectively, and wrong operations during the repeated modeling is avoidable at the sametime. The two parameters Paris rules is used as the state equation of particle filter predicting thefatigue crack growth.(3) The transmission characteristics of Lamb wave in the aluminum plate is studied by using thefinite element dynamic analysis method. The method of establishing the observation equation is givenby using the damage index to represent the crack extension combining the experiment verification.(4) The hole edge crack and edge crack of aluminum plate is used to verify the effective of theabove fatigue crack extension method based on particle filter algorithm under laboratory conditions.The experimental results show that error of crack length after20000load cycles is less than2%. It caneffectively predict crack propagation, and eliminate the error accumulated by the Paris rule predictionand various of uncertainty in engineering application.