A non-probabilistic reliability assessment method for gear transmissions

Abstract

The invention discloses a non-probabilistic reliability evaluation method for a gear transmission device, and the method comprises the steps: describing uncertain factors in a reliability evaluation process of the gear transmission device through an evidence theory, and building a limit state function of the gear transmission device; to improve computational efficiency, constructing an approximatemodel of a real limit state function through a radial basis function, improving the fitting precision of the approximate model through local encryption of sample points, and meanwhile, solving an approximate reliability evaluation problem based on a first-order approximate reliability analysis method, so that the credibility and the likelihood for measuring the reliability of the gear transmission device are obtained. According to the method, the uncertainty variable in the design process of the gear transmission device can be effectively constructed, the calculation efficiency and the solving quality of solving credibility and likelihood can be essentially improved, and the method has wide engineering application value in the field of reliability evaluation of the gear transmission device.

Description

technical field [0001] The invention relates to the field of reliability evaluation of gear transmission devices, in particular to a non-probability reliability evaluation method of gear transmission devices. Background technique [0002] Gear transmission is one of the most common forms of transmission in mechanical transmission systems. The reliability of the gear transmission will directly affect the transmission efficiency of the mechanical system, and during the design process of the gear transmission, the maximum stress borne by the gear must be less than its maximum allowable stress. [0003] For the reliability evaluation and design of gear transmission, the existing methods have the following problems: [0004] 1. Most of the existing design methods for gear transmissions are studied under the deterministic condition of the system parameter model. However, in the actual design process of gear transmissions, there are errors in boundary conditions, initial condition...

AI Technical Summary

Problems solved by technology

[0004] 1. Most of the existing design methods for gear transmissions are studied under the deterministic condition of the system parameter model. However, in the actual design process of gear transmissions, there are errors in boundary conditions, initial conditions, and measurement conditions. or uncertainty, if these factors are still treated as deterministic factors, it will lead to a large deviation between the system response and the actual response

[0005] 2. For the reliability evaluation method of gear transmission, some existing methods use probability models to describe uncertain design variables, such as the patent "A Method for Reliability Evaluation of Spur Gear Transmission System" with the publication number CN105138794A, but the actual In the design process of many uncertain variables, the probability distribution density is unknown due to the lack of sufficient sample data, which makes it difficult to accurately obtain the probability distribution of uncertain design variables

[0006] 3. For the reliability evaluation methods of gear transmissions, some existing methods use non-probability convex set models to describe uncertain design variables, such as the patent with publication number CN102446239A "Gear Transmission Considering Cognitive and Random Uncertainty Multidisciplinary Reliability Analysis Method”, this patent describes the uncertain design parameters through the probability model and the non-probability convex set model based on the completeness of the design parameter uncertainty data, although the non-probability convex set describing the uncertain design parameters The model does not require an accurate probability distribution, but because the non-probability convex set model emphasizes too much extreme conditions, it will result in a conservative design

[0007] 4. For complex reliability evaluation problems such as gear transmissions, time-consuming numerical analysis models are often involved

Therefore, in order to overcome the shortcomings of low optimization efficiency, a small number of existing technologies use approximate model technology, but some patents only construct an approximate model once, and for complex engineering problems such as gear transmissions, an approximate model The establishment of the method cannot guarantee the accuracy of reliability evaluation; although the patent "a gear transmission reliability evaluation method based on the Kriging model" with the publication number CN107273609A realizes the reconstruction of the multiple approximate model of the gear transmission through the Kriging model, but The design parameters of this patent are all described by random variables, that is, the distribution type of random variables needs to be obtained through data collection and hypothesis testing, which will make it difficult to accurately obtain the probability distribution of random variables. The Monte Carlo method is used for the reliability evaluation of the structure function function, which requires the collection of a large number of sample points, resulting in a decrease in computational efficiency

Images