Klemen Drobnic

Predictive Direct Control Applied to AC Drives and Active Power Filter

This paper presents different applications of a method called direct control. The previously developed approach has been redefined into a generalized form. The method relies on the prediction of either current or flux in discrete-time intervals and, consequently, selects the inverter voltage vector that produces the fastest possible transient. Depending on the task, two possible variants have been developed, offering a compromise between ripple in the controlled variable and switching frequency. A special effort has been made to overcome problems due to various delays (processing time, acquisition, gate driver delay, etc.) in the prediction routine, thus achieving maximum performance. The approach has been upgraded for application in AC drives, which allows additional torque control. The functional versatility of the approach has been demonstrated on different applications of power electronics (active power filter, induction machine, surface-mounted permanent-magnet synchronous machine). All applications have been tested on different laboratory models and have confirmed the validity of the approach.

Detection of Broken Bars in Induction Motor Through the Analysis of Supply Voltage Modulation

This paper presents an application of a novel method for the diagnostics of electric and magnetic asymmetries of rotor cage in induction motor (IM) due to broken rotor bars. An increasing anomaly in magnetic field distribution results in degradation of steady-state and dynamic performance of an IM. This degradation can be determined through the analysis of the average duty cycle of the modulated supply voltage. Broken rotor bars would cause torque and speed ripple which is mitigated by an efficient speed-control algorithm. Consequently, specific oscillation in the duty cycle of the modulated stator voltage appears. This effect can be simply detected without additional hardware and therefore provides a correct evaluation of faulty motor performance, which is a very significant part of condition monitoring and diagnostic procedure in modern supervision systems for electrical drives.

Direct Current Control of a Synchronous Machine in Field Coordinates

This paper presents an implementation of a modified method for direct current control (DCC) in a permanent-magnet synchronous motor (SM). The control is performed using a predictive algorithm in rotor field coordinates due to its simpler implementation. Its task is to minimize the final current error at the end of the sampling interval. Depending on the desired performance, two variants are proposed, yielding either lower switching frequency or lower current ripple. Both variants of DCC, named DCC I and DCC II, have been rearranged for applications in SMs with either surface-mounted permanent magnets or interior (buried) permanent magnets. A comparison with field-oriented control using space-vector modulation and synchronized on/off modulation shows the advantages and drawbacks of the proposed method. Simulations and measurements on a laboratory model with SM having low inductances (thus expecting high ripple) and very high number of poles (high stator supply frequency) confirm the validity of the approach. Additional considerations regarding practical problems, some of which can be found in similar predictive methods, are also presented.

Direct current control of active power filter without filter current measurement

The paper presents a three-phase parallel active power filter without filter current measurement. Only supply grid voltage sensors, filter capacitor voltage sensor and supply line current sensors are used to provide filterpsila proper operation, aiming to reduce reactive power and line current distortion, caused by a non-linear load. Due to filterpsilas inverter-like topology with known parameters in filter branch, the predictive direct current control (DCC) method has been successfully applied, thus obtaining good dynamic response, reducing the commutation frequency of the active power filterpsilas switches and consequently decreasing the filter branch losses.

Pseudo-salient model of induction machine with broken rotor bars

In this paper, a dynamic model of a squirrel cage induction motor with broken rotor bars, is presented. The model is based on the assumption that the asymmetry due to broken bars is reflected through different resistances and inductances in a two-phase rotor model. Their joint effect is best expressed through rotor time constant. Mathematical equations, based on this premise, have been developed in rotor reference frame, thus obtaining a very simple yet effective model of the machine. Additionally, a method for measurement of both time constants of a faulty rotor at standstill is explained. Simulation results, based on this model, have been experimentally verified for healthy and faulty machine, showing very good agreement at different operational points. Finally, a study of the impact of various relationships between the parameters in both rotor axes on the behavior of the simulation model, is performed.

Predictive torque control of interior permanent magnets synchronous motors in stator co-ordinates

This paper proposes a possible approach to control the synchronous machines with interior (buried) permanent magnets (IPM SM). If properly controlled, this construction allows for an increase in torque through a reluctance component that adds to the main torque caused by the permanent magnets. Since the overall torque depends on the load angle and stator current magnitude, the optimal relationship between these two variables can be pre-calculated prior to the machine start. The inverse procedure allows a determination of the unique load angle and current reference, which, in turn, form a stator flux. The latter is calculated from discretized basic voltage equation in each sampling interval. This procedure is called Predictive Torque Control (PTC). An output variable (reference stator flux) is then passed to the block for Immediate Flux Control (IFC), already tested on induction machines. Its aim is to generate the actual stator flux by selecting proper inverter voltage vectors and determining their application time. The results have been verified on experimental model of a real machine.

Dynamic model of induction machine with faulty cage in rotor reference frame

This paper presents a simplified model of an induction machine with broken rotor bars. The model is founded on the fact that broken bars cause asymmetries in the normally homogeneous cages resistance and inductance. Thus, the rotor becomes pseudo-salient. This effect could be best comprised through the rotor time constants that differ in the orthogonal axes of a rotor reference frame. A method for their evaluation is also presented. Consequently, a choice of rotor co-ordinates for the model of a faulty machine to be developed in, is obvious. Experiments show good agreement with simulation results obtained from the developed model.

Estimation of parameters of induction motor with broken rotor bars

This paper describes the evaluation of parameters of a three-phase induction motor with broken rotor bars. Such fault causes electric and magnetic asymmetry in the machine. Using standard d-q model of induction motor this asymmetry can be expressed through differently modified rotor parameters in both axes. In pursuing the goal to establish accurate motor parameters, a non-standard, one-phase measuring method was employed. In this way angle dependence of rotor parameters following a sinusoidal shape was determined. Additionally to measurements an extensive finite element analysis has been performed for two purposes. Firstly, to assess the measurement results of the method and, secondly, to predict various situations which can not be easily achieved by measurements due to hardware limitations (e.g. different numbers of broken bars). Finite element analysis evidences the correctness of the machine parameters obtained from proposed measurement method.

Detection of broken bars in induction motor using voltage pattern analysis

The paper presents an application of a rotor broken bars detection method in induction motor drives with Field Oriented Control (FOC). Rotor fault distorts the magnetic symmetry and subsequently deteriorates the steady-state and dynamic performance. As fast speed control loop compensates for the unevenly distribution of the field, the inverter supply voltage is adjusted accordingly. The modulated supply voltage pattern is then analyzed and its average duty cycle is determined. When broken rotor bars are present, specific oscillations in the average duty cycle arise. The presence of these oscillations is sufficient for a successful detection. The computation is performed using simple mathematics and demands no significant computational power. The results confirm the suitability of the proposed detection method for FOC scheme.

Induction motor parameters in case of rotor electrical asymmetry

The paper presents estimation of parameters alternation in a three-phase induction motor due to broken rotor bars, which cause electric and magnetic asymmetry in the machine. Using d-q model of induction motor this asymmetry is reflected as pseudo-saliency through different resistances and inductances in both axes. An extensive study has been performed to simulate various cases of faulty cage, which cannot be achieved by measurements on actual industrial drives. Simulation results can be efficiently used for analysis of closed-loop induction motor drives, in which determination of rotor faults, due to the intrinsic speed correction performed by the controller, is much harder to achieve.