Dissertation > Excellent graduate degree dissertation topics show

Slope Fault Diagnosis and ARMA model prediction research and circuit verification

Author: ZhuLi
Tutor: WangZhiGang
School: University of Electronic Science and Technology
Course: Measuring Technology and Instruments
Keywords: slope model fault dictionary method ARMA prediction particle swarm optimozation(PSO) leapfrog circuit
CLC: TH165.3
Type: Master's thesis
Year: 2010
Downloads: 72
Quote: 1
Read: Download Dissertation

Abstract

With the increasing complexity of analog circuits, the difficulty of fault diagnosis and prediction of the circuit also constantly increase. Researchers have focused on fault diagnosis and prediction algorithms and achieved many achievements which promoted the development of algorithms. Because of lacking experimental data and the actual fault diagnosis system to evaluate the algorithms’advantages and disadvantages, so the progress of fault diagnosis and prediction research on analog circuit grow slowly. In this dissertation, the method of slope model-based fault diagnosis and the ARMA prediction model have been studied, and the leapfrog circuit has been choosed to establish a fault diagnosis and prediction system to verify the simulation result.In theory, the two algorithms have been researched in detail. First, for the method of the slope-based model fault diagnosis, on the one hand, researching the theoretical foundation and concrete steps of the method, on the other hand, using it to establish the fault dictionary. The fault dictionary of the leapfrog low-pass filter circuit has been taken as an example. The simulation result show that the optimum test node has been gotten, and the number of tested nodes have been obviously reduced; Secondly, for the ARMA prediction model, the key technologies, namely, how to establish the model has been researched in details. Based on comparing between the ME(Moment Estimation, ME) and the MLE(Maximum Likelihood Estimation,MLE), a new method--PSO (Particle Swarm Optimization,PSO) has been used to get the AR parameters and MA parameters from the angle of MLE. The simulation results show that:By means of PSO, the optimized model parameters based on MLE is more suitable for establishing ARMA model.In practice, an analog circuit verification system about fault diagnosis and prediction which include the IC module used for test, collection data module and the data processing module has been built. The Protel software has been used to design schematic and circuit board, and the Pspice and VC++software have also been used to simulate circuit and build the user interface respectively, which can diagnose the fault and predict the state online. The system has two functions. On the one hand, to provide measured data for the fault diagnosis algorithm, and implement online fault diagnosis; on the other hand, to provide source data for prediction.The results show that the system works well, and basically realize the function of fault diagnosis and prediction online. Because the established system can provide the measured data for fault diagnosis and prediction, So the research has certain significants which can evaluate the merits of an algorithm on actual circuit.

Related Dissertations

CLC: > Industrial Technology > Machinery and Instrument Industry > Machinery Manufacturing Technology > Flexible manufacturing systems and flexible manufacturing cell > Fault diagnosis and maintenance
© 2012 www.DissertationTopic.Net Mobile