Parametric Modeling Error Reducer And Bearing Reliability Assessment Methods

CAD has brought great change to the workers engaged in the engineering design work since born in the 1960s. The use of CAD software, designers can effectively manage large amounts of information, the use of three-dimensional CAD modeling capabilities, and the subsequent analysis and optimization capabilities combine to become an indispensable tool in engineering design and analysis.The current parametric study of reducer products targeted mainly for parts, focuses on the process of establishing a standard parts library, are an ideal model and lack of interactivity. To overcome the shortcomings of traditional parametric design exists for three cylindrical gear reducer, committed to the application of the subject to focus Pro/Toolkit for Pro/ENGINEER secondary development, the introduction of a variety of error model parameters, to reflect the real situation of the gear unit parametric models, more realistically reflect the actualMeanwhile, the application of finite element analysis software ANSYS dynamics can get more accurate results, but the workload and time cost is relatively too large. In contrast, the artificial neural network method is highly nonlinear computing power and a strong self-learning and fault-tolerant capabilities for data input, through the training of the neural network to achieve the desired accuracy can be used to predict the output, efficiency has been greatly improved. In this paper, the finite element method, the neural network, Monte Carlo method and the Box-Behnken method combined for structural reliability analysis. Rapid response based on neural network models and works on the structure of the neural network to establish the reliability and sensitivity parameters assessment.