Research On Closed-loop Design Method Of Cathode Tool In Electrochemical Machining Based On The BP Neural Network

With the development of modern manufacturing,the parts have gradually developed in directions of high precision,complicated surface and difficult-to-machine.Electrochemical machining(ECM),one of main non-traditional machining methods,has become one of the key technologies in modern manufacturing with its unique advantages.In ECM,precise cathode tools are important for qualified workpiece surface.However,due to the complex machining process of the ECM,it is not easy to obtain cathode tools with high-precision.Traditionally cathode tools are designed with the "open-loop" method,in which the cathode design is calculated.After massive manual corrections the cathode is used to machine workpieces.In this thesis a “closed loop” cathode design method is proposed.In the closed loop design process,the initial cathode is applied to predict the workpiece's shape and be corrected by the shape errors of anode workpiece.The closed loop design process will be iterated till the precise workpiece's shape is obtained.It is helpful to improve efficiency and accuracy of the cathode design.According to the ECM theories and the frequency domain transfer function theories of the linear control system,the relationship between the anode workpieces' s shape and the cathode tool is established.The cathode modification is realized by the Fast Fourier Transformation(FFT).According to the theories of electrolytic flow in the ECM,the distribution of the electric conductivity in the electrolytic flow field is corrected and the initial cathode tool is modified.The modification methods are compared and the “closed loop” cathode design is completed.In order to realize the prediction of the anode workpiece in the cathode closed loop design process,the artificial neural network is applied.With the help of the MATLAB software,the relationship between cathode tool and workpiece anode is established by the BP neural network model.The main contents include the structure of the BP model and the data collection and preprocessing.Finally experimental results are used to train and verify the BP model.With the BP model in the thesis the relative errors of anode workpiece prediction are less than ±7%.Taking a complicated surface of an engine blade as the research object,the anode workpiece is predicted with the BP model according to the initial cathode tool.Through the prediction results the initial cathode is modified with the FFT method.The correction coefficient in modifying cathode tool is determined by the golden section method.And in the second modification method the distribution of the electric conductivity in the electrolytic field is corrected.The comparison of two cathode correction schemes show that the cathode correction based on FFT method is more conducive to improve of cathode design accuracy and efficiency.By means of the MATLAB software and UG software platform the cathode closed loop design system is developed.The system include: the initial cathode design,the anode workpiece prediction,the cathode correction based on FFT method and the cathode correction based on variable conductivity method.The closed loop software is good for improvement of the accuracy and efficiency of the cathode design in ECM.