Modeling Of Cutting Forces And Optimization Of Parameters For Interrupted Turning Of Hardened Steel

Hardened steel has the characteristics of high strength,high hardness and high wear resistance,and is widely used in bearings,automobiles,molds and other industrial fields.In actual machining,intermittent turning of hardened steel is often encountered,such as external turning of spline shafts and machining of complex shapes of molds.When intermittent turning hardened steel,the tool is frequently subjected to transient impact,and in order to avoid too much impact on the tool and the workpiece,more conservative cutting parameters are often used in the machining,which makes the energy consumption too high and the machining cost too high.Therefore,this paper establishes a transient impact force and quasi-static cutting force prediction model for intermittent turning of hardened steel,and optimizes the cutting parameters,mainly including the following:Firstly,the theoretical analysis of the sources of cutting force generation and geometric characteristics of cutting layer in interrupted turning hardened steel machining,the calculation formula of cutting layer area is derived,the model of quasi-static cutting force is established,the main factors affecting quasi-static cutting force and transient impact force are analyzed,and the foundation for cutting force testing and modeling is laid.Secondly,the cutting force test of intermittent turning hardened steel was conducted based on the response surface BBD method with workpiece hardness,cutting speed,feed and back draft as the influencing factors,and on this basis,the response surface prediction model was established,and the accuracy and validity of the model were verified by ANOVA and residual analysis.Then,a cutting force modeling method combining neural network and Kfold cross-validation based on response surface BBD method is proposed,through which a neural network prediction model with better prediction performance than response surface model is established.In establishing the neural network prediction model,the PSO-BP prediction model and LSO-BP prediction model were established by improving the BP neural network using particle swarm algorithm and lion swarm algorithm respectively to address its shortcomings.By comparing the evaluation indexes,the results show that the prediction performance of both the LSO-BP model and the PSO-BP model is better than that of the BP neural network model,among which the LSO-BP model has the best prediction performance.Finally,the optimization of multi-objective cutting parameters for interrupted turning of hardened steel was completed.With surface roughness,transient impact force and cutting specific energy as the objective functions and cutting speed,feed and back draft as the decision variables,a multiobjective cutting parameter optimization model for intermittent turning of hardened steel was established by considering various constraints.The model is solved based on the multi-objective marine predator algorithm,and the solution results are scored using the entropy-weight-TOPS IS comprehensive evaluation method to guide the selection of optimal parameters.Comparing the values of each objective under the optimized cutting parameters and the empirical cutting parameters,the results show that the optimized cutting parameters can achieve better results in machining.