Ship Steel Plate Of High Frequency Induction Thermal Stress Forming Experiment And Prediction Model Of Neural Network

The hull steel forming is an important link in the shipbuilding industry, the worldshipyards mainly rely on the forming used the oxyacetylene flame fire and water. Thefactors of the fire and water heating forming are complex, so is difficult to adapt to theshipbuilding of automation and high efficiency and facing the aging of skilled workers.These factors severely limit the development of shipbuilding, So a new heat source isneeded urgent to instead of the oxyacetylene flame.Laser and high-frequency induction heating are new heat plate bending, becauseusually ship plate, forming the greater curvature, so the laser heating is limited for shipforming. High-frequency induction heating is almost integrated the advantages of flameheating and laser heating, this is the best alternative source of the ship plate thermalstress forming.The Q345marine steel plate as the research object to explore the characteristics ofhigh-frequency induction heating thermal stress forming and the significance of theshipbuilding industry, different thermal stress forming according to different types ofheat sources are summarized,and each type of domestic and foreign research status.According to the characteristics of high frequency induction heating and the basicprinciple, analyzes the temperature field of high frequency induction heating, and byusing ANSYS finite element software to simulate temperature distribution during theheating process. According to the calculate, the parameters of induction heating aredetermined, and the high frequency induction heating test equipment and measuringequipment are made, then single-pass and multi-pass of heat thermal stress formingexperiments have done. The deformation effect is obtained by experiments of themarine Q345plate steel under different induction heating process parameters. Finally,on the basis of the experimental data, the multiple linear regression and BP neuralnetwork prediction model are established for predict the bending angle under differentinduction heating process parameters and different thickness basis.The heating plate temperature field distribution is analyzed through the principle ofhigh-frequency induction heating, and is consistent with the finite element simulation.Homemade and buy high frequency test equipment can meet the basic formingexperiment and measurement of high frequency induction heating. Througha-single-channel induction heating experiment, we use the Marine Q345plate at20 KHz, heating power to20kw,25kw,30kw and35kw, the heating time of4s,8s,12sand16s, bending deformation Angle of0~2.1°, radius of curvature is81m~1.5mm,then we get a conclusion that bending–Angle will increase with the drop of thethickness, biger power and heating time last, the radius of curvature will decrease withthe enlargement of bending Angle, but when reduced to a certain value,it remains thesame; According to the analysis of the results,we proposed induction heatingtemperature and steel plate gravity effect on the deformation; Select heating time is12s，heating Marine Q345plate6times under different heating power. Except the platealmost no change when the heating power is20kw, after heating medium thicknesssix times under other heating power the planking of bending Angle has increased3~5times. According to the results of the single channel heating experiment, using BPneural network theory, establish a Marine Q345plate bending Angle and thickness,heating power and heating time prediction model, this model has a certain accuracy, cansatisfy the engineering application.