Finite Element Analysis And Research On Neural Network Controller In CFRP Drilling Radial Hole

Because carbon fiber reinforced resin matrix composites(CFRP) has high strength and light weight, such as a lot of excellent condition.Therefore, it has a very high utilization rate in the aircraft fuselage and a lot of automobile bodies, it is even used in sport and strategic weapons and other areas.If the appropriate machining parameters and tools are not used in the machining process, there may be many defects in the entrance and exit of the hole, so that the quality of the hole can be reduced.In this paper, the CFRP cylindrical shell is regarded as the research object, and the process of drilling radial hole is simulated by ABAQUS software, analyzed the relationship between the axial force and drilling parameters of CFRP drill radial hole;A BP neural network controller is designed to control the axial force of drill radial holes..Specifically made the following work:1、The structural properties and development status of CFRP are analyzed, and the factors that cause the hole defects in the process of CFRP are analyzed.Analysis and comparison that, if the axial force of the drill hole is more than the critical value,it will produce a very serious phenomenon at the exit of the hole.2、Due to the CFRP drilling radial hole drilling process may produce hole deficiency and lead to deterioration in quality of the hole, so the subject established the finite element model of CFRP cylindrical shell radial holes, and using ABAQUS software to simulate the radial holes of different drilling parameters of drilling process, get carbon fiber composite drilling radial hole of axial force with the change rule of cutting parameter.3、Due to the axial force of drilling radial hole Shitai will lead to large hole defects, thus affecting the quality and efficiency of drilling radial hole.In order to improve the CFRP drilling quality and the drilling efficiency of the radial Kong Shizhong hole. This topic using neural network knowledge, conceive of a neural network controller, the drilling radial hole of the axial force has been maintained in the below the critical value.The controller comprises a controller and an identifier,which can be trained and tested by the MATLAB software.The controller is trained to test can be used to CFRP drilling radial hole machining process, through real-timecontrolling the process of drilling parameters, to adjust the CFRP drilling radial hole of the axial force, so as to improve the radial hole quality and processing efficiency.