Abstract. To more accurately calculate the dynamic reliability for the center gear of power assisting cycle with various failure modes, the multi-extremum response surface method is adopted. The first, the gear's torque, material density, elastic modulus and poisson's ratio are sampled in small batches as the random input samples, the dynamic extremum of deformation, stress and strain are obtained as output response by finite element model. Then, the multiextremum response surface equation is established. Finally, multitudinous sample points are obtained by using Monte Carlo method and linkage sampling to multi-extremum response surface equation, which can used to calculate the reliability of the deformation, stress and strain of the gear in the comprehensive failure mode. The results show that the comprehensive reliability degree of gear is 0.9949 when the allowable deformation, stress and strain are 0.47 mm, 540 Mpa and 0.003 , respectively.
IntroductionThe gears support complex loads during the operation of the power assisting cycle operation. Once the gears fail, the energy stored in the spring can't be released and the whole system does not run. So, the structural reliability analysis of gear is very necessary. In recent years, a structural analysis method considering the tooth contact of internal gear system is introduced through the comparison with the simplified gear system model and applied to the structural analysis of a 2-stage differential-type gearbox for wind turbine by Cho et al [1]. The key meshing states of the gear pair for the contact fatigue and the bending fatigue was investigated based on the cumulative fatigue criterion and the stress-life equation by Deng et al [2]. Although the problem of gear contact was studied in different degrees, they were based on deterministic analysis method, and it is not discussed further from the angle of probability. Currently, scientists have presented a lot of research on the gear reliability analysis method, response surface method is the most widely used [3][4][5][6]. The reliability analysis method based on response surface and the Markov chain Monte Carlo (MCMC) were proposed to improve the precision of the response function, and the dependability sensitivity analysis was conducted to quantify the effects of assembly errors, machining errors and stochastic external loads on gear transmission reliability by Tong et al [7], then the reliability was calculated using the founded mathematical model by Monte Carlo method (MCM). However, these methods only consider the reliability of single failure mode analysis without considering the failure correlation of dynamic reliability problems. In order to solve the problem of multiple failure mode correlation, the multi-extremum response surface method was introduced in the reliability analysis of gear. The gear's torque, material density,