| Two-axis synchronization control in CNC machine tool, that is to say, a machine axis is driven by two servomotor axes with coincident position. In two synchronized axes driven by rotary servomotors through ball screw transmission, because of the flexibility of ball screw in high-speed and the strongly coupled dynamics, it is difficult to control this complex two-degree-freedom control system. The purpose of this dissertation is to investigate and solve some of the problems that are confronted in two-axis synchronization control, under supporting by national science and technology funds. The work in this dissertation is as follows:It takes the representative prototype machine with compounded axes and two synchronized ball screw transmission as an example. A finite element model for it has been developed base on mixing finite element modeling with lumped parameter modeling. And the modal analysis and model identification show the effectiveness of the model. Then, base on the modal analysis conclusion, a lumped parameter, coupled dynamics model for two-axis synchronization control system has been developed by simplifying current loop in servo devices.Base on theory analysis, three kinds of synchronization control strategy, such as the torque compensation method, the instruction precompensation method, and the speed loop variable gain method,were proposed. At the same time, a PID neural network control method which doesn't depend on model precision was adopted. And its assessment criterion has been modified to fit on on-line control.The character of several common speed planning instructions was investigated from frequency domain and time domain. And the synchronization error was simulated base on the influence of the different speed planning and the position loop bandwidth. As a result, the following conclusions were put forward. To achieve the ability of lg average acceleration, position loop bandwidth of synchronized axes should satisfy with 40Hz.Trapezoid acceleration speed planning is desirable for synchronization control, because it will cause a smaller following error, synchronization error, and demand a less interpolation time than the others. On the analysis of the S-shaped acceleration and deceleration curve model which consists of seven sections, it puts forward an interpolation algorithm which can calculate the number of interpolation cycle of every section. Taking fully into account the influence of filters in the servo devices, it builds a discrete-time model with 40 Hz position loop control bandwidth for next simulation. At the same time, the torsional vibration of ball screw has been compensated. The effectiveness of discrete model has been verified through experiments.Four kinds of synchronization control method proposed were investigated in simulation. Specially, the instruction precompensation method and the speed loop variable gain method were tested on the prototype machine which was controlled by the field bus NCUC-Bus. The results of simulation and experiments are shown as follows:â‘ The torque compensation method has best synchronization ability, but it is difficult to use in the practical.â‘¡The instruction precompensation method is not affected by CNC system structure. It is easy to be applied to the practical, but excessive compensation can cause bumpy motion. The simulation shows that synchronization error decrease is prominent. The experiments shows that synchronization error can reduce about 40%.â‘¢The speed loop variable gain method has a good comprehensive performance, but it can be used only on field bus CNC. The simulation shows that synchronization error decrease is obvious. The experiments shows that synchronization error can reduce about 50%.â‘£Although the PID neural network control method has the ability of Self-study and Self-adaptation, it can be used in the practical need to modify the structure of the servo devices and cooperate between the CNC system and the servo devices. The simulation shows that synchronization error can reduce 40%.
|
PDF Full Text Request