Sensorless control offers many advantages for the operation of electrical machines. Being able to operate the machine via field-oriented control without relying on the encoder greatly increases the reliability of the drive. In the low to zero speed region, saliency based control strategies are well suited. For several standard industrial motor designs, induction machine can exhibit multiple inherent spatial saliencies linked to position. For such multi-salient induction machines, the intermodulation saliency has shown good characteristics for sensorless control. Transient excitation can be used to extract said intermodulation saliency. However, it requires intervention in the current controller pulse-width-modulation sequence and greatly oversampled current. Alternatively, rotating high frequency injection can be used to extract the intermodulation saliency, which offers simpler implementation possibilities. The rotating voltage space vector can simply be superimposed onto the current controller output. Additionally the current does not need to be oversampled. However, since the resulting high frequency current is not controlled, it is distorted by the inverter nonlinearities. Compensating the unwanted saliencies and inverter-related harmonics, the intermodulation saliency will be used in this paper to extract the rotor positon of multi-salient induction motors from the high frequency current. The experimental position estimation quality is tested at different stationary loads and transient load steps. Applicability is verified on a traction machine designed for railway application.

In electric rail transportation, traction motors are responsible for propulsion and transmit torque to the bogie wheelsets through a gear drive system. Typically, each traction motor is fixed to the bogie frame using a suspension system and thus motors might withstand a large amount of shock and vibration arising from wheel to track contact. With developing fatigue, the motor-bogie suspension might fail. To prevent destructive contact between motor and track, bogies are often equipped with an additional safety apparatus: a safety bar fixed to the bogie frame that is axially aligned with a motor-fixed safety catch when suspension is healthy. This is the last protective measure that prevents the motor from falling to the track and, thus, it is of tremendous importance to reliably and quickly determine suspension faults. While the motor-bogie suspension is broken, the motor can move up or down within the safety catch. This results in small but abrupt angular movements of the rotor after torque transients take place. Thus, detection of motor-bogie suspension fault can be detected by inspection of the shaft encoder signal or speed sensor signal under load changes. In this work, a novel detection method for motor-bogie suspension fault is presented, using spatial saliency information extraction thus only requiring current transducers and no traditional shaft encoder nor additional sensors. The transient spectra of the motor saliencies are used to define a trigger that serves to detect motor suspension failure during a test. The method is experimentally validated in a test bench equipped with a 1.6MW locomotive traction motor and a suspension-fault emulation system.

Due to the fast slew rates of modern power switching devices, inverter-fed motors experience intensified insulation stress. In addition, inverter switching with fast slew rates excites several high frequencies including multiple antiresonance frequencies of the winding and supply cable, further suffocating winding insulation at localized winding spots. As a result of the wide range of excited frequencies, high frequency currents and voltages can be used to track insulation state. In particular, the voltage at the star point of inverter-fed motors can show the degree of motor ground-wall insulation aging. However, accessing star point might not be feasible in some applications. In this paper, a special switching scheme is proposed, turning the star neutral voltage visible at the terminal voltage without the need for additional equipment or disassembling. Via solely sensing the terminal voltage, ground-wall insulation will be monitored. In addition, the proposed switching scheme elicits additional motor resonance modes, which extends the variety of current and voltage analysis for insulation condition monitoring. Experimental validation using an inverter-fed 2MW medium-voltage traction induction motor is provided.

The extraction of motor spatial saliencies permits stable control at standstill and nearby speeds. One technique to extract saliencies is to excite the motor with voltage steps and process the resulting phase current slopes to form a saliency vector. Yet, the additional voltage steps produce unwanted current ripple and increase audible noise. Therefore, voltage steps have been recently integrated into the space vector PWM symmetrical switching scheme, defined at each period by two adjacent active states and two inactive states. In this paper, only the longest active state is used for saliency extraction. The novelty of this paper lies in the application of the one-active SVPWM excitation to a parallel dual motor configuration supplied by a single inverter and using a reduced current sensor configuration. This configuration consists of only three current sensors, two sensors attached to two phases of one motor, and a single sensor attached to the second motor. A novel signal processing scheme is presented in this paper to extract the saliencies of each motor individually under one-active SVPWM excitation. Experimental results demonstrate the ability of the proposed technique for dual motor saliency extraction under one-active space vector PWM excitation and using the special current sensor configuration.

The development of new technologies in the field of power electronics implies higher switching frequencies. Such frequencies are associated with higher transient voltages. Higher voltages mean stress on the insulation material. As result of this increased load, partial discharge (PD) probability can be increased, which may push the insulation material to its limits. Therefore, it is of high importance to recognize PDs. It is already known that PDs occur not only on the rising edge, but also in the steady state of a voltage pulse. However, double PDs may occur, too. Using a special test arrangement, it is possible to detect also weak PD activity, to localize the time instant and to determine changes in voltage and current. These changes are characterized e.g. by a sudden drop in pulse voltage.

The investigation presented in this paper aimed to analyze the temperature dependences of induction machine parameters. More specifically, the focus was placed on the temperature behavior of magnetic properties. The machine under test is a 100kW traction motor used in standard railway applications. Knowledge of parameters temperature dependences is necessary to ensure precise operation of the induction machine over a wide temperature range. Therefore, measurements for different stator and rotor temperatures of the machine were conducted, which revealed that the permeability of the stator and rotor side electrical steel sheet show a clear temperature dependence. Moreover, a reduction of flux density due to increasing temperature was determined. The purpose of these results is to improve the accuracy of stator or rotor equation-based flux models.

The basic spatial modulation effects of induction machines caused by slotting and saturation, have been analyzed for exploitation in speed sensorless control. With increasing load level, additional spatial harmonics arise, acting as disturbance and deteriorating sensorless control performance. In view of the fact, that induction machines for application in traction systems (e.g. public or cargo transportation) are operated at a high level of excitation, higher harmonic numbers especially of the spatial saturation become dominant and need to be compensated in order to establish speed sensorless operation. It can be observed, that especially the first order harmonic of spatial saturation shows a significant rise in magnitude when load current is increased. The presented paper thus focus on the investigation of high order spatial modulations and their deterministic properties with the goal to establish an efficient compensation of their disturbing effects. Based on the identified deterministic properties a compensation method is proposed and tested through measurements on a 15kW induction machine.

In this paper a method to consider saturation behavior of induction machines during the finite element (FE) analysis of high frequency excitations is developed. This saturation behavior manifests itself in an additional saturation and intermodulation harmonic, when evaluating the current response related to a high frequency voltage excitation. In order to correctly reproduce these harmonics within FE-analysis, the fundamental wave as well as high frequency magnetic properties of the lamination material must be represented in correct proportion. To achieve this, relative permeability is adjusted to create asymmetries in the high frequency magnetic properties. An adaptation method for the transient permeability based on the fundamental wave flux level was developed. Comparison of measurement and FE simulation shows good agreement of all main high frequency modulations present in induction machines.

Condition monitoring of inverter-fed drives is gaining a more important role. Inverter operation enables maximum utilization of a drive. However, some parasitic phenomena may occur, e.g. bearing currents, temperature increase due to additional losses and stress caused by fast voltage rise times. With focus in this work on stator winding insulation an online monitoring method is presented to detect changes in the insulation strength. The insulation health state is assessed at a thermal aged machine by voltage step excitation of the inverter. A change of the electrical strength is evaluated with the transient current response measured for each phase with the built-in sensors of the inverter. The proposed method is able to detect insulation deterioration before an actual short circuit occurs. The results are compared with the insulation condition parameters resulting from additional off-line low-voltage and high-voltage measurements.

This paper investigates the feasibility of sensorless position estimation from PWM-induced transient excitation sequences, present during inverter-driven operation of induction machines (IM) with standard space vector PWM (SVPWM). The applied method exploits existing machine saliencies and their correlation to the current slopes measured in response to a voltage pulse at the machine terminals. Determination of this quite linear current response and its position dependent slope modulation allows for extraction of the flux/rotor position information. However, an initial transient oscillation process delays the first possible measurement instant, what imposes minimum pulse duration and limits the possible operational range of SVPWM-integrated transient excitation for sensorless methods. Previous publications showed, that by application of current signal high frequency sampling it is possible to estimate the current derivative before these oscillations decay, thus further reducing the minimum pulse duration requirement. The paper at hand contributes an analysis of different signal processing algorithms and their capabilities when used for this purpose. The investigations include: mean value evaluation (MVE), Theil-Sen regression (TSR), least-squares regression (LSR) and maximum-likelihood estimation (MLE). In a first step an offline comparison is carried out with previously sampled measurement data. The second part features an online evaluation to verify the possibilities for SVPWM-integration.

Stator insulation defects in the machine winding of traction drives can be analyzed by evaluating the current response resulting to inverter induced voltage injection. Due to effects occurring during the aging of the insulation, e.g. change of parasitic electrical components, deviations of the high frequency components of the current response can be monitored. The insulation degradation of a stator winding system is a long developing process and usually takes place over a long period. With the proposed method stator insulation degradation can be detected already at an incipient stage by comparing the high frequency transient current response of the drive system from the first initial startup with continuous measurements during operation. No additional current transducers are necessary than the transducers already applied for the control of the machine. Investigations are conducted with a 1.4 MW induction machine designed for a traction application having form-wound coils equipped with fiber-insulated winding taps and a silicon carbide voltage source inverter enabling high dv/dt rates of up to 25kV/μs. Additionally, in order to investigate insulation degradation of the stator winding system, results of measurements of a form wound coil specimen are presented.

Sensorless field oriented control (FOC) strategies have found their place in industry applications during the last few years. Stimulated by this development scientific research continues to improve their performance, which is directly linked to the estimation accuracy of flux and rotor position without a mechanical sensor by means of high frequency or transient excitation. The influence of induction machine (IM) design parameters on the rotor saliency, originating from interactions of rotor and stator slotting, is investigated with a focus on transient electrical machine behavior. As this saliency is often used to estimate rotor position in sensorless control schemes, effects attenuating or sustaining it have a straight impact on the control's quality. The present study features finite element (FE) simulations that are compared to measurements for a number of test machines, to validate the results. In a second step it aims at generating a more comprehensive view on the dependencies between saliency and machine design by conducting parameter variations by means of FE simulations, with a focus on stator short pitching and rotor skewing.

In this work a method to detect insulation degradation of an adjustable speed drive (ASD) based on analyses of the transient current reaction caused by inverter switching is proposed. The drive system consists of an inverter fed three phase induction machine with 1.4MW. The standard current sensors are used to analyze the transient signal ringing to extract an insulation state indicator of the winding system and for the control of the machine as well. Using the inverter as a source of excitation, it is possible to perform an insulation test by evaluation of the resulting transient current sensor signals. The influences of the fundamental wave portion and transient component of the current signal on the insulation state indicator are analyzed. An effective separation of these components is investigated with an algorithm applicable for real time evaluation of the insulation state.

The fields of application of adjustable speed drives fed by voltage source inverters is constantly increasing. For better exploitation of the system's capabilities all drive components are more and more operated near and even above their rated values. This leads to more strains for the drive. However, at the same time the demands for reliability are also increasing. To reach this requirement different strategies like fault tolerant design, fault detection and condition monitoring can be implemented. However, to implement such strategies a deep understanding of the effects that lead to problems or faults in the drive system is necessary. The fast switching of modern voltage source inverters leads to transient overvoltage stressing the machine's insulation system. A short literature study presenting the characteristic of and the parameters influencing the non-linear voltage distribution in the stator winding of inverter-fed machines will be presented. Furthermore the findings will be compared with experimental results on an induction machine with tapped windings.

Adjustable Speed Drives (ASD) are the favorable choice in a fast increasing scope of applications. Although generally highly reliable, continuous condition monitoring of electrical drive systems becomes more and more important, as the adoption of such systems in safety critical applications, like x-by-wire and in the so called more electric aircraft, increases. To fulfill the demand of high efficiency and good return on investment the different drive components are operated at or near their rated values. This fact and additional stress added due to high dynamic operation, fast switching of modern power electronics like IGBT lead to increased risk of drive breakdown. In many applications this is completely unacceptable and has to be compensated by preventive maintenance, fault tolerant operation or continuous condition monitoring and detection of faults in an early stage. A variable speed drive basically consists of three main components - the inverter, the cabling and the machine itself. These components - including their parasitic elements - form a complex impedance system. The proposed method evaluates the transient current reaction on inverter switching. A change in this reaction is correlated to a change in the mentioned complex impedance system, e.g. resulting from degradation of the insulation system. Application of high-frequency current sampling and further signal processing allow a very accurate monitoring of system parameters and even small changes in magnitude and location can be detected. Experimental results proof the applicability of the proposed condition monitoring technique.

The winding system of high voltage machines is usually composed of pre-formed coils. To facilitate the winding fitting process stator slots are usually wide opened. These wide opened slots are known to cause disturbances of the magnetic field distribution. Thus losses are increased and machine's efficiency is reduced. A common way to counteract this drawback is given by placing magnetic slot wedges in the slots. During operation the wedges are exposed to high magnetic and mechanical forces. As a consequence wedges can get loose and finally fall out into the air-gap. State-of-the-art missing slot wedge detection techniques deal with the drawback that the machine must be disassembled, what is usually very time consuming. In this paper a method is investigated which provides the possibility of detecting missing magnetic slot wedges based only on measurement of electrical quantities and without machine disassembling. The method is based on exploitation of machine reaction on transient voltage excitation. The resulting current response contains information on machine's magnetic state. This information is composed of several machine asymmetries including the fault (missing wedge) induced asymmetry. A specific signal processing chain provides a distinct separation of all asymmetry components and delivers a high sensitive fault indicator. Measurements for several fault cases are presented and discussed. A sensitivity analysis shows the high accuracy of the method and the ability to detect even partially missing slot wedges.