Digital Control System Of Active Magnetic Bearing Based On DSP Platform

Magnetic bearing is a novel high-performance bearings which use magnetic force to make the rotor suspended in space providing a non-contact support between the rotor and stator. As the magnetic bearing between the rotor and stator without mechanical contact, it has the following advantages:without friction. The longer spindle life, and easy to maintain; lower vibration noise; high-speed rotation; low energy consumption; no lubrication, no oil pollution, can be used in the environments of vacuum, clean, heat, cold, specialty gases, and even the body, and other special environments. It is particularly suitable for high-speed, vacuum, clean, low temperature and other special circumstances. In addition to the above advantages, in active magnetic bearings, the electromagnetic force can also be adjusted by the controller, which means that adjustable bearing stiffness and damping, thus the implementation of active control of the rotor, the unbalance compensation, temperature compensation, etc., and it can help to improve the dynamics of the rotor performance. You can also apply a variety of advanced control algorithms, and realize the system monitoring and control. Because the magnetic bearings have the above advantages, so the industrial countries strongly focus on the research and development of the magnetic bearing technology. Currently, the application of magnetic bearings are mainly concentrated in rotating machinery, turbine equipment, the heart pumps and rotor flywheel and other fields.For the completion of the subjects, This paper mainly do the following work:Rotating machinery are moving into high speed, high precision and high flexibility and many other directions, and because the magnetic bearing (AMB) has many advantages:high speed, no wear, no lubrication, high reliability and dynamic characteristics of adjustable and other advantages, the past two decades in rotating machinery applied research it is taken seriously. Excellent control system can make a powerful magnetic bearing;on the other hand, the controller design is a challenging task. This paper belonging to the National Natural Science Foundation project "Intelligent inline monitoring system of AMB grinding spindle", mainly studies a robust digital control system which can be applied to the CNC grinding machine spindle and compensate for the position changes caused by the temperature rise.In order to make the digital control system to meet the grinding spindle control requirements, this platform based on DSP digital control systems and monitoring platform is designed. Among them, the digital control system is mainly composed of the displacement sensors, AD converter, DSP digital controller, FPGA digital power amplifiers; the spindle condition monitoring system mainly monitors the displacement of the rotor, the electromagnet coil current and temperature of the spindle.According to the Hoo theory and the unstable structure of robust control theory, combined with the analysis of the magnetic bearing spindle systems mathematical models and uncertainty, this paper designs a magnetic bearings robust controller. Among them, the adaptive neuro-fuzzy inference system is used for an intelligent identification of magnetic bearings non-parametric uncertainties, then the corresponding Hoo controller is designed, as opposed to only consider the parameter uncertainty of the controller. This method has better robust performance in practical applications achieved the good result.Based on studying the composition of the magnetic bearing digital control delay and on the impact on performance of the control system, a new delay compensation algorithm is proposed. The algorithm, by predicting the next sample time delay system output, eliminates the impact of the control system. Prediction algorithm is from the discrete model of magnetic bearings, and the algorithm factors are corrected by the neural network. Experimental results show that the algorithm can compensate the delay of the digital controller very well, to achieve a stable digital control magnetic bearing suspension and high-speed operation.Based on the detection of temperature rise in system, the model of the temperature rise associated with the spindle position and orientation is established; identifies the correspondence relationship of five displacement control input settings and the spindle position and orientation. According to temperature samples of the system, the5-way displacement control input settings are corrected, to achieve the inline adjustment of grinding wheel pose and complete the inline compensation of the system temperature expansion. Compensation algorithm is realized by FPGA. Experimental results show that the algorithm works well to compensate for temperature expansion, to ensure the stability and accuracy of the magnetic bearing spindle.MK2110-type grinding machine is altered to build a magnetic bearing spindle control system experimental platform. Implementation of a magnetic bearing spindle grinder stable suspension of the five degrees of freedom is successful; the rotation experiment speed is up to500Hz. At360Hz the grinding experiments are preceded. The roughness and roundness of the resulted workpieces can basically meet the processing requirements, close to the level of industrial applications.In above work, the active control is implemented by the digital controller to realize advanced control algorithms and achieve high system robustness, and the on-line compensation is employed to offset the adverse effects which is caused by temperature and other factors on the system, which reflects the advantages of magnetic bearings and is also the important and difficult parts of the research. Rotor dynamics analysis, system identification, automatic control, sensor, power electronics technology, advanced knowledge of these subjects are needed. First, the magnetic bearing is a strong non-linear and essentially unstable control object, and in grinding process, the spindle is required to have both high precision and high rigidity, the appropriate controller need to be designed carefully. As the uncertainties of system parameter and dynamic uncertainty in the model, the use of PID control or the control strategy dependent on the deterministic model can not get the ideal control effect. Thus it is necessary to design a robust controller with good performance getting used to the system model uncertainty. In the H∞control method, the choice of weighting function is a difficult problem to be solved, the choice of weighting function is to rely on the experience and repeated test. In general, it depends on the requirements and indicators of the control design objectives. By the use of the intelligent identification method for the choice of non-parametric uncertainty weighting function, to meet the design requirements, the system has good control performance. Second, the magnetic bearing system, temperature effects will affect the accuracy of the static magnetic bearing systems, deteriorate the characteristics of axial bearing, threat the system reliability. In order to solve the temperature problem, we study the temperature effect on the rotor position and posture, and use neural networks to establish a critical temperature point temperature and the rotor position offset mapping. How to use the hardware neural network to make on-line real-time compensation, is a difficult point of this paper.Mainly do the following innovations:the use of smart identification method in the magnetic bearing spindle systems to describe the non-parametric uncertainty weighting function, and for the parameter uncertainty and non-parametric uncertainty, design the Hoo controller to achieve a high robust performance; analyzes the effect of control system time delay in digitally controlled magnetic bearing system, and on this basis, presents a digital controller delay compensation algorithm. The algorithm effectively removes the impact of digital control time delay to achieve a magnetic bearing system stability; for the magnetic bearing spindle temperature problems, in the detection system based on temperature rise, establishes the model of temperature rise and the position and orientation of the spindle; identifies the correspondence of five displacement control input settings and the spindle position and orientation. Experimental results show that the algorithm works well to compensate for the temperature expansion, the magnetic bearing spindle to ensure the stability and accuracy. Based on the innovative research work, the control system in practical applications obtains good results.This paper accumulated system design and actual operating experience, has important theoretical and engineering significance.