Reliability Analysis And Structure Optimization Design Of Soft Seal Gate Valve Based On Neural Network Model

Soft sealing gate valve is widely used in urban underground water supply system.Due to the buried underground connection water supply pipeline,the maintenance and replacement of the gate valve is very inconvenient,and the water supply is indispensable for every resident of the city.Therefore,it is very important to ensure the reliability of the soft seal gate valve.There are many underground gate valves.If the quality of the gate valve is effectively reduced on the premise of ensuring its reliability,the cost of gate valve production can be reduced.Based on the reliability theory and structural optimization design method,the reliability calculation and lightweight design of the soft seal gate valve are carried out,and the research results are verified by experiments.This paper first introduces the basic concepts of neural network model and reliability,discusses the three traditional structural reliability calculation methods of center point method,checking point method and Monte Carlo method,and expounds the basic theory of structural optimization and two multi-objective optimization methods of response surface method and neural network method.Mainstream research methods provide a theoretical basis for subsequent gate valve reliability calculation and structural optimization.Secondly,the initial size variables are determined by the structural analysis of the gate and valve body of the gate valve,and the parametric modeling of the gate and valve body is carried out.The static analysis of the gate and valve body is completed by Ansys Workbench finite element analysis software.The initial size variables were further screened by sensitivity analysis,and the original samples needed to fit the neural network model were obtained by Latin hypercube design method.Then,through a series of steps such as the construction of neural network structure,the normalization of original sample data,the determination of activation function,and the training of sample set by network learning method,the final weights and thresholds of the neural network model are obtained,and the maximum stress and quality of the gate and valve body are fitted.The reliability of gate and valve body is calculated by combining the maximum stress prediction model with Monte Carlo method.The maximum stress prediction model and the quality prediction model are further fitted,and the maximum stress and quality of the gate and valve body are optimized by the method of finding the maximum value.Finally,the stress-strain test of the optimized model is carried out,and the test results are compared with the numerical simulation results to verify the rationality and correctness of the neural network multi-objective optimization method.