Fault Diagnosis Of Single Phase Cascaded H-Bridge MLI Using K-NN Algorithm

A novel Five-Level cascaded multilevel inverter is proposed in this thesis.This thesis consists of proposing a cascade multilevel inverter fault diagnosis strategy based on the PPCA and Machine Learning k-NN Algorithm.Nowadays great progress has been made in developing the multilevel inverter in renewable energy sources and other electrical drive applications.Multilevel Inverter is very popular in industrial medium and high voltage power.The power electronics device convert Direct Current(DC)power to Alternative Current(AC)power at required output voltage&frequency level is known as Inverter.An inverter can be broadly classified into different ways,like based on level of output voltage,then single level inverter and multilevel inverter.The Multilevel inverter as compared to single level inverter has advantages like minimum Total Harmonic Distortion(THD)and can operate on several voltage levels.A multi-stage inverter is being utilized for multipurpose applications such as active power filter,Static-VAR Compensator and machine drive for sinusoidal and trapezoidal current application.As power electronics equipment reliability is very important and to ensure multilevel inverter systems stable operation,it is important to detect and locate faults as quickly as possible.It is difficult to diagnose a fault in multilevel inverter using a mathematical model because its consist of many switching devices in this context and to improve the fault diagnosis accuracy and fault rate is decrease of a Cascaded H-Bridge Multilevel Inverter(CHMLI),a fault diagnosis strategy based on the Probability Principle Component Analysis(PPCA).The output Voltage signals under various fault condition of Cascaded H-bridge Multilevel Inverter is taken as fault features by using phase shift pulse width modulation(PS-PWM)technique.The output Voltage signal under different fault condition of Cascaded Multilevel Inverter is taken as the fault characteristics signal to avoid the load variation on the fault diagnosis.The main purpose of PPCA is to reduce the samples dimension without changing the original properties of the Input data.The simulation results are given for 5-Level CHMLI different Amplitude Modulation(AM)indices and shown that this method is accurate for detection of the fault and their location.Finally,k-NN algorithm is used to identify the accurate fault location and diagnosis the fault.The proposed technique is validated by conduction the experiment using FPGA to control the CHMLI using PSPWM technique.The simulation results show that the proposed fault diagnosis method reduce the fault diagnosis time and improved the accuracy compared to other fault diagnosis methods.The experimental result is also considered and compared the different techniques like FFT-PCA-K-NN and PPCA-k-NN.