| Connecting rod bushing is the key part of diesel engine.Its main function is to avoid the direct contact between the connecting rod and the piston pin,which can affect the life of the connecting rod.The connecting rod bushings manufactured by the power spinning process not only have excellent surface quality,but also have high mechanical properties and long service life,which can meet the use of high power and high specific pressure diesel engine under harsh conditions.Since the relationship between the process parameters and the mechanical properties of the power spinning connecting rod bushing is complicated,there is no accurate and reliable mathematical relationship that can be expressed,and it is usually set by experience.In this paper,the relationship between the power spinning connecting rod bushing and its mechanical properties is studied and analyzed based on artificial neural network and the main process parameters were optimized by multiobjective optimization,which is of great value and significance.The influence rule and order of power spinning connecting rod bushing process parameters(thinning rate,heat treatment temperature,feed ratio)on its mechanical properties(hardness,yield strength,elongation,tensile strength)were gained through the method of orthogonal test and range analysis.Based on the results of orthogonal test,the nonlinear relationship between RBF and BP neural network was established.The neural network was trained with the training sample data,the predictive ability of the neural network was detected by the test sample,and the prediction accuracy of two kinds of neural network models were tested and analyzed,found that RBF neural network model has a more accurate prediction ability.RBF neural network model has the advantages of strong prediction abilityćshort modeling time and so on,which can be used to design the process parameters in practical production and can effectively improve the design efficiency and reduce the cost of the actual test.According to the influence order of process parameters on the mechanical properties gained by orthogonal test results and range analysis,The nonlinear relationship between the main process parameters(thinning rate,feed ratio)and the main mechanical properties(tensile strength and elongation)of the RBF neural network was established,then using this nonlinear relationship as the fitness function to establish the multiobjective optimization model of the main process parameters based on the MATLAB numerical simulation platform and genetic algorithm theory,and the corresponding multiobjective Pareto optimal solution and the relationship between tensile strength and tensile strength of two main mechanical propertiesis were obtained.The feasibility of these optimal solutions is verified by experimental analysis,which can effectively improve the design efficiency of the process parameters and improve the mechanical properties of the products effectively. |
