Based On Wavelet Transform And Neural Network-voltage Inverter Fault Diagnosis System

With the rapid development of power electronics and microelectronics, frequency conversion velocity modulation as the most main velocity modulation get great development with each passing day. Frequency converter is used extensiverly in many fields for its advantages of excellent control performance and prominent effect of power saving. As the other equipments, frequency converter often goes wrong for the antinomy. The fault of electronic components is the main reason for result the system invalidations. Statistics have indicated that, 80% of the control system invalidations are due to the fault of the electronic components. So the research about the fault of main circuit is very valuable in the engineering.Because of the limit of condition, we can't get the fault information on the condition of fault which can be used as the base of fault detection and diagnosis. At first, the simulation model of varied-frequency system was built. And on this base, aim at the popular voltage source frequency converter in the market, this paper build its simulation model and simulate some kinds of frequency converter main circuit faults that can be met usually, such as branch open circuit fault which is caused by SCR open circuit, fuse fusing, pulse loss etc. The key part of frequency converter is inverter module. So we can extract the three-phase current on the condition of the fault which can be used as fault diagnosis.For the character of nonlinear, inverter can't use accurate mathematical model for fault diagnosis, a novel algorithm based on wavelet analysis and neural network was presented to detect and diagnose in inverter. Nonlinear MATLAB simulation model of three-phase sine-PWM inverter was setup. Three-phase current acted as fault information, drawing out the fault characteristics by wavelet analysis as the inputs of neural network to posit the inverter bridge with switch fault. Lastly, a simple judge strategy for locating the fault switches based on the symmetry of three-phase fault was proposed. The simulation results show that the algorithm can be used to diagnose the PWM inverter, and the algorithm is faster and more accurate than other algorithm in the convergence rate and fault diagnosis.In the realizing aspect, TMS320LF2812 is used as core control chip and the fault diagnosis system for controlled voltage source frequency converter on wavelet analysis is developed. The fault diagnosis hardware includes for five parts: signal adjustment and communication interface, A/D, DSP, power manage. Software includes wave show, harmonic's analysis, wavelet analysis, fault diagnosis and identify algorithm, and so on.