Research On Intelligent Method Of Hull Curved Plate Forming

This paper chooses the feature subset with the greatest correlation through the scientific feature subset selection method,predicts the deformation with low error through the neural network algorithm,and designs an intelligent decision-making system,embeds the prediction mechanism into the system,provides decision support for the processing workers,and designs a concise human-computer interaction field.It is convenient for processing workers to record processing data.For the screening of processing factors of hydrothermal bending plate,this paper adopts the methods of literature review and intelligent screening of support vector machine.Firstly,through reading the literature,we summarize the influencing factors that predecessors thought had great influence on deformation,and consider the influencing factors as a whole,and select the feature subset which had the greatest impact on the final deformation by using support vector machine.For the study of parameters of hydrothermal bending plate,the heating process of hydrothermal bending plate was simulated by ANSYS finite element analysis software.98 groups of processing cases were obtained,and the relationship between processing parameters and deformation was studied qualitatively.Then the relationship between acetylene flow rate,heating line length,heating speed,heating width,plate thickness and deformation is determined by neural network algorithm,so as to achieve low error prediction.Finally,under the condition that the influencing factors and reasoning mechanism have been determined,an intelligent decision-making software for hydrothermal bending plate is designed,which makes the system have both the functions of decision-making guidance and case accumulation for hydrothermal bending plate processing.