Δυναμική ανάλυση και ανάπτυξη μεθόδων διάγνωσης σφαλμάτων ελεγχόμενου κινητήριου συστήματος αποτελούμενου από ασύγχρονη μηχανή δακτυλιοφόρου δρομέα και ηλεκτρονικούς μετατροπείς ισχύος

Abstract

In this thesis an investigation of the operation of an electric drive consisting of a slip ring asynchronous machine and a power electronic converter connected to the rotor circuit is carried out. This drive is referred to as “subsynchronous cascade drive” or “slip energy recovery drive” in the literature. The investigation is carried out via simulations in Matlab/Simulink of a 4 MW drive installed in a Greek cement industry and via experiments with a 5,5 kW drive in the laboratory. Furthermore, Motor Current Signature Analysis (MCSA) techniques for the detection and the accurate diagnosis of faults in asynchronous machines or/and power electronic converters are examined and novel improved approaches are proposed. The potential application of the analysis of the rotor current of the subsynchronous cascade drive for fault detection and diagnosis in the asynchronous machine or/and the power electronic converter is also investigated. In the beginning, simulations are carried out considering rigid shafts and rigid coupling between the motor and the mechanical load. The asynchronous machine has been modeled in a d-q-o reference frame referred to the rotor. Characteristic simulation results are presented and discussed. The general behaviour of the asynchronous machine electromagnetic quantities is analyzed and, furthermore, the cases that the slip has values at multiples of 1/6 are theoretically investigated and discussed. Additionally, the influence of the asynchronous machine speed, of the load torque characteristics and of the asynchronous machine parameters on the electromagnetic torque pulsations is examined. Afterwards, the equations of the mechanical part of the drive are formulated, considering a four-mass model and linearity of the mechanical coupling. A theoretical investigation of the torsional oscillations is carried out, and some general guidelines regarding the design of the mechanical part of the drive are discussed. Additionally, characteristic simulation results are presented and discussed. Then, the equations of the mechanical subsystem are reformulated taking into account the non-linearity of the coupling stiffness and characteristic simulation results are presented and discussed. The experimental investigation, using a 5,5 kW drive in the laboratory validates the conclusions drawn by the investigation via simulation, regarding the behaviour of the electromagnetic quantities of the asynchronous machine. In the experimental setup, sensors for all the electrical quantities have been used, as well as sensors for the measurement of the shaft torque (straingauge) and the lateral vibrations of the motor (accelerometer). The measurements taken by the straingauge have verified the pulsation of the shaft torque due to the presence of the rectifier in the rotor. The measurement of the asynchronous machine lateral vibrations has revealed the presence of harmonic components related to the operation of the rectifier in the rotor circuit. In the remainder of the thesis the application of Motor Current Signature Analysis techniques for fault detection and diagnosis in asynchronous machines and/or power electronic converters is examined. To this objective, a comparative investigation of different signal processing approaches is carried out for the diagnosis of two characteristic faults in asynchronous machines, namely eccentricity and/or mechanical imbalance as well as rotor fault. Derived from simulation and experimental results, the advantages and the disadvantages of each technique are presented and discussed in detail. Features and limitations of each one of the signal processing approaches are then compared in order to derive the most appropriate one, related to the operating conditions of the induction machine. Especially the phase method for the extraction of the ridges of the Continuous Wavelet Transform (CWT) and the Short-Time Fourier Transform (STFT) is examined, and it is proven that under certain conditions this method is advantageous for the detection of harmonic components related to the presence of faults. Furthermore, a novel approach has been proposed and investigated, for the detection of weak harmonic components that are close to the fundamental. The proposed approach is based on the convolution of the current signal with a complex wavelet and has been applied for the accurate diagnosis of rotor faults via analysis of the stator current. Finally, faults in the power electronic converter of the subsynchronous cascade drive are examined; specifically open circuits and short circuits in the rectifier’s diodes. The amplitude and the frequency of the electromagnetic torque pulsations which take place during these faulty conditions are investigated. A novel approach for the accurate diagnosis of this kind of faults is proposed and investigated via simulation and experiment. The proposed approach is based on monitoring the rotor current space vector angle and instantaneous frequency, in order to extract features which indicate the specific fault modes of operation.