Studies On Some Relative Theories And Experiments For Fault Diagnosis Of Magnetic Bearings

A magnetic bearing is a complex electromechanical system.The technologies of a magnetic bearing involve multisubjects such as mechanics, electronics, computer and control technology, sensor and material science and so on. So far the main interests about magnetic bearings have been focused on the design of the structure and the control system in our country and abroad, and few report is about fault diagnosis of magnetic bearings. On the basis of summarising the development of magnetic bearings and their fault diagnosis, some relative fundamental theories, key technologies and problems are deeply studied, and some creative progresses are got as follows:Data fusion of the multisensors of magnetic bearings is researched. The mathematical model of the data of the multisensors of magnetic bearings is established. The variance analysis method and the range analysis method are proposed to be used to test the mean of the data, and whether the mean of the data has changed can be known from the result. It is proposed to adopt the maximum variance method in the variance test of the data, and whether the variance of the data are equal can be known from the result. Data fusion way is researched of the two sensors on a same degree of freedom. It is pointed out how to test a single sensor and get the data. It is also analyzed how to select the significance level in the statistic test. The proposed data fusion methods are useful to the state monitoring and fault diagnosis of magnetic bearings.Multivalued-logic-based fault diagnosis of the sensors used in magnetic bearings is researched. Multivalued logic algebra based on the sequential variables is presented, and the theory is applied in fault diagnosis of the sensors of magnetic bearings. A few theorems are presented and proved.The states of the sensors can be deduced according to the states of the magnetic bearings, and the deduction process is analyzed in detail, which has been realized by program.Fault diagnosis of the control system of magnetic bearings is researched. The mathematical model of the control system of magnetic bearings is established. The characteristics of the control system are analyzed by simulation. It issimulated to clarify the influence to the control system from two typical fault sources: circuit delay and random noise. The results show that the delay can lead to high frequency vibration, and noise has no very obvious influence to the control system. So the delay should be as small as possible in the system design. The influence of the delay can also be cut down by adjusting the PID parameters of the control system. Fault diagnosis of data acquisition and process system is also researched, and some useful test and measurement technologies are presented. Fault diagnosis of the control circuit is researched, and some relative diagnosis ways are given for the test of a few fault phenomena of the combinational logic circuit and the sequential circuit. It is also analyzed about how the control system searches the axis of the rotor.The task scheduling of the control system of magnetic bearings is studied considering reliability and fault diagnosis. A scheduling algorithm is presented, which unites the characteristics of the RM(Rate Monotony) and the EDF(Earliest Deadline First) algorithms. The algorithm can schedule the tasks efficiently by changing the periods of the tasks, and improve the reliability. Fault diagnosis can proceed on line. The principle of interrupt-mode fault alarm is also discussed in the task scheduling.Experimental tests are carried out for fault diagnosis of magnetically suspended rotors. Many experimental studies reveal that, to create an effective fault diagnosis system for the rotors suspended magnetically, experimental tests should be carried out in multiaspects including vibration, noise, temperature, model frequency and load-carrying property. Fault diagnosis and fault positioning can be completed according to the experimental results.Fault characteristics of passive magnetic bearings are also researched. The mathematical model of passive magnetic bearings is established. The amplitude and the frequency characteristics are studied. Simulation is given, and the vibration and fault characteristics are analyzed.At last an outlook of fault diagnosis of magnetic bearings is given.