| As a new kind of Electrical Discharge Machining (EDM) technology developed in 90's of 20th century, electrical Discharge Milling (ED-Milling) applies simple electrode which moves along certain tracks to different discharge place between electrode and workpiece for required shape. It not only can machine the metal workpiece materials with high hardness and high intensity and toughness, but also it can execute multi-axis NC machining as well as NC milling machining. Because of its outstanding ability in the machining of the middle- or small-size workpieces with 3D surface, ED-Milling is highly attached importance by researchers in the areas of Electrical Machining in the world.However, traditional control techniques could not satisfy demands in the ED-Milling process, due to the complexity and randomness of the discharging process. Application studies of intelligent control in the ED-Milling process are developed with the help of the flourishing theory and technology of intelligent control. The dissertation includes the following six aspects: (a) the research of the servo feed system of ED-Milling based on fuzzy control technology, (b) the research of the tool wear prediction; (c) the research of the compensation method of tool inthe ED-Milling process, (d) the research of ED-Milling CAM system, (e) the research of developing techniques of real-time system in Windows2000, (f) the development of Economic ED-Milling experiment device. The results of main research work in the dissertation are as follows:(1) On the basis of that servo feed motion of ED-Milling is carefully analyzed, a servo control system is designed based on fuzzy control technology, and then a fuzzy control algorithm with parameter adaptation is presented. The self-adaptive fuzzy control system is composed of two parts: fuzzy controller ( FC ) and parameter adjuster. The input parameters of the fuzzy controller include the error and error change of short circuit-discharge arc rate, the error and error change of no discharge rate, and the output is the feed frequency of motor. Furthermore, a parameter adjuster with two inputs: error and error change of discharge rate, is developed to adjust proportionality factors of the FC. The presented fuzzy control method with parameter adaptation has higher machining rate and better adaptability and stability compared with traditional adaptive control used in EDM.(2 ) In according with the non-linear character in the process of ED-Milling, a tool wear prediction model is established based on artificial neural network for the first time so far. The model can be used for the prediction of machining rate and tool relative loss, and thus the amount of practical tool wear can be figured out in the process, which will lay the foundation for tool compensation on-line. Due to the faults in the neural network BP algorithm, an improved genetic algorithm is proposed to optimize network weights and structure, which could adaptively adjusts probability of mutation and quantity of mutation, in order to enhance evolutionary speed and network prediction precision.(3 ) According to electrode wear prediction model, the formulae used for tool compensation are derived, and corresponding algorithms for machining two typical trajectory, straight line and arc obtained by discharging from the side surface of tool are presented. A compensation method of tool for eliminating the slope of part are investigated under condition of uniform wear machining. At last, the motion path of ball electrode is analysed in 3D surface machining.(4) A ED-Milling CAM system is discussed, in which two technologies: optimization of machining parameters and data exchange between CAD/CAM systems are studied. A genetic algorithm is applied to optimization of machining parameters on the basis of the above neural network prediction model. Two data exchange technologies are studied according to machining methods in ED-Milling. For a method, discharging from the side surface of tool, interface program for plane polygon is designed based on DXF f...
|
PDF Full Text Request