Research On CNC Machine Tool Energy Consumption Prediction And Energy Efficiency Based On SVM

Machine tool is the typical mechanical manufacturing process equipment and themain working machine. It produces a large amount of energy in the process oftransferring roughcast into parts. With the increasingly serious problem of energyshortage, energy consumption of machine tools is becoming a hot spot of internationalconcern. Based on the composition analysis of energy consumption of CNC machinetool and the quantitative analysis of energy consumption, this paper tries to proposesome energy-saving methods which takes cutting amount as input and energyconsumption as output to build the energy model. The specific processes aresummarized as follows:Firstly, the author makes a classification of the energy agency of CNC machinetools and analyzes its function. On this basis, the machine model is divided intoauxiliary system energy consumption model, processing power system energyconsumption model and energy consumption predication model. We concentrated onthe neural network theory module and SVM theory module in energy consumptionpredication model and made a comparison between them.Second, the author proposed the energy consumption prediction method based onSVM. When machine and tool has been established, the choice of cutting amount hasbecome the key influential factor of energy consumption. Analyzed the effect ofcutting parameters, the processed workpiece, the cutting tool and the machine tools theamount of energy on CNC machine tools, choosing model input as cutting parametersand the machine tool consumption as output. After analyzing the nature of SVM,taking XK713CNC milling machine for the object, based on characteristics of themachine and the actual circumstances, we chose the kernel function reasonably andadopting K-fold cross-validation method to choose the penalty factor and nuclearparameters, we established energy consumption predication model. Through thisexample, the validity of this energy prediction method has been proved. Third, the author proposed cutting parameters optimization method based ongenetic algorithm. Taking milling as an example, we took the lowest energyconsumption of the machine and low cost as the objective optimization function. Weused linear acceleration method to deal with the multi-objective function model,adopting regression analysis to get the milling force parameter, determining theconstraints according to the technical specifications of the machine. We used geneticalgorithm toolbox to optimize solution of the parameters of the model, and put theminto the predication model for verification. Experimental results show that the use ofoptimized cutting parameters can meet cutting requirements and achieve lower cuttingenergy consumption and reducing costs.