The Research On Fault Diagnosis Technology Of Satellite Gyro Components Based On Data-driven

Today, with the rapid development of modern science and technology, agreat progress has been made in space technology, and the status of spacecraft ismore and more important. In order to guarantee the safe operation of spacecraftand try to avoid or reduce the occurrence of fault, it has a important practicalsignificance to study the fault diagnosis technology.The main contents of this paper include:Firstly, it used the real gyro in the lab and the block diagram of satelliteattitude control system that has been set up by my brother to get the operationdata of angular velocity, then it has a study on a fault diagnosis method in thetime domain: the EMD method based on the concept of energy entropy weight. Itused EMD method to decompose the gyro signal, then got the IMFs and the REScomponent, and the first IMF is seen as the characteristic component, and usesthe concept of energy entropy weight to diagnose on fault.Then it introduced a fault diagnosis method combining frequency domainand time domain: the method based on HHT. It based on EMD, and used K-S testdistribution to test the IMFs, then combined the spectrum and marginal spectrumfeatures to realize the fault diagnosis.Finally, it studied a fault diagnosis method of multi variable statistics, that’sthe method of principal component analysis, it has a fault detection on theangular velocity signal of4gyros on the axis in system with the statistic ofsquared prediction error, and then calculate the contribution value on statisticsfor each angular velocity signal, so it can determine the source of the fault, thenfault diagnosis completely.It has a feasibility simulation on the platform of matlab for each faultdiagnosis method firstly, then has a engineering application on the fault diagnosismodule of the project in lab and displays the result on the PC interface, so it hasa certain significance for further improvement and application of these methods.