Kwanghee Nam

A fast dynamic DC-link power-balancing scheme for a PWM converter-inverter system

The authors propose a new power converter control scheme for a converter-inverter system. The strategy is to fully utilize the inverter dynamics in controlling the converter dynamics. The authors obtain the power dynamics for both converter and inverter systems, and control the converter power so that it matches the required inverter power exactly. Then, in the ideal case, no power flows through the DC-link capacitors and, thus, the DC-link voltage does not fluctuate even though a very small amount of the DC-link capacitance is used. In forcing the converter power to match the inverter power, the authors utilize the master-slave control concept. They control the DC-link voltage level indirectly through the stored capacitor energy in order to exploit the advantage of the linear dynamic behavior of the capacitor energy. This helps them to circumvent a complex control method in regulating the DC-link voltage. Through simulation and experimental results, the superiority of the proposed converter control scheme is demonstrated.

PI type decoupling control scheme for high speed operation of induction motors

In the high-speed operation of a vector controlled induction motor, decoupling current control is a very important problem. However, the conventional decoupling scheme does not perform well when there are mismatches in the motor parameters. The authors propose a PI type decoupling current controller. A great advantage of this PI type decoupling controller is its robustness to motor parameter errors. Hence, the proposed scheme appears suitable for practical applications where the exact parameters are unknown. Through simulation results, the authors show that the proposed controller achieves the desired performances.

A load torque compensation scheme under the speed measurement delay

The average speed detection method involves a measurement delay, which can cause a serious instability problem to the unknown load torque observer. The instability can be cured by inserting an artificial delay into the torque-filtering path of the observer. Also, by utilizing the concept of the phase lead compensator, we propose a method of designing an arbitrary high-order low-pass filter (Q filter). Through the results of simulation and experiments, we show that our proposed method yields more robust and improved results than the conventional load torque observer.

A new parallel hybrid filter configuration minimizing active filter size

Conventional parallel hybrid active filters suffer from the problem of heavy current ratings of devices used in the inverter. In general, this problem has been solved by adjusting the turn ratio of a matching transformer. However, adjusting the transformer to a high turn ratio may not be possible for high voltage power systems due to their requirements for high voltage insulation. In this paper, a new configuration is proposed for a parallel hybrid active filter. In the proposed hybrid active filter, the active filter is connected to the passive filter inductor in parallel through a matching transformer with the aim of reducing the current rating of the inverter. As a result, the fundamental component of the current flowing through the passive filter is divided by the parallel paths of two inductors located at the inverter side and passive filter side, respectively. Hence, the current rating of the inverter can be reduced. Additionally, a harmonic elimination method is suggested for the proposed active filter. Through computer simulations and experiment, we have verified the effectiveness of the proposed topology.

Diagnosis of rotating machines by utilizing a backpropagation neural net

The authors utilize a backpropagation neural net for the diagnosis of rotating machines. The abnormal vibrations due to imbalances, axis misalignments, and bolt-loosening have different spectra. Similar to a pattern recognition technique, the spectra of abnormal vibrations is used in obtaining characteristic feature vectors. For an experiment, a vibration test bench was constructed in such a way that artificial faults could be realized easily. The feature vectors of abnormalities obtained from the test bench were used for training the neural net. The performance of the trained neural net was tested in recognizing the causes of vibrations.<<ETX>>

A simple DC-rail soft switched voltage source inverter

A simple DC-rail soft switched resonant circuit is proposed for a three-phase voltage source inverter. The proposed resonant circuit consists of a DC-rail switch with an antiparallel diode and two auxiliary switch. The DC-rail switch is turned on and off under zero voltage condition, and the auxiliary switches are turned on and off under zero current condition. Since the resonant inductor is disconnected at zero current instant, the proposed circuit requires no resonant inductor freewheeling diodes. Furthermore, the freewheeling diodes in the bridge are softly turned off. Also, the proposed circuit has another advantage that the required power ratings of the resonant inductor and capacitor are small. With the proposed scheme, it is possible to make zero DC-rail voltage with resonance in a PWM period whenever a switching transition is required. The rule of selecting the switching time instant is very simple and less dependent on the system. Furthermore, the proposed scheme can be applied with space vector modulation PWM for motor drives without any modifications. The authors analyze the proposed circuits, and demonstrate its performance through simulations, as well as through experiments.

Sensorless vector control in the presence of voltage and current measurement errors by dead-time [induction motors]

A sensorless vector control scheme for induction motors is proposed in the presence of voltage and current measurement error caused by dead-time. To overcome the pure integrator problem, the authors modified the stator flux observer through the injection of the current observation error so that poles lie in the left half complex plane. Further, upper bounds of the observation error and the speed estimation error are derived in the presence of voltage and current measurement errors. Experimental results show that the proposed scheme yields fairly good dynamic and steady-state performances irrespective of applied load torque at low speed.

A real-time parameter identification scheme for the sensorless control of induction motors using a reduced order model

An observer-based systematic parameter identification scheme for an induction motor is proposed by utilizing the Lyapunov stability criteria. Assuming that all the motor parameters except the stator resistance are not known, a parameter update law is derived based on a reduced (second) order dynamic model. The model reduction is achieved by substituting the estimated value of the rotor flux into the stator current equation. The rotor flux is calculated from the stator current and the stator flux, and the stator flux is estimated by integrating voltage terms numerically. As practical examples, two identification problems are considered. Both simulation and experiment results show that the estimated values converge to the real ones with the proposed identification algorithm. This algorithm can be applied to the real sensorless vector control of induction motors.

A robust adaptive sensorless field-oriented control using a modified stator flux observer

A sensorless vector control scheme is proposed based on a reduced order dynamic model of an induction motor and an adaptive parameter identification technique. A very special feature of this work is the fact that the observer dynamics and the adaptive algorithm are solved concurrently via the Runge-Kutta 4th order method in the hardware environments. This approach improves the numerical stability of the proposed control scheme over whole speed range. Another characteristic feature is that it avoids the use of the problematic pure integrator in the stator flux observer. To overcome the pure integrator problem, the authors modified the stator flux observer so that the poles lie in the left half complex plane. Experimental results show that the proposed sensorless control scheme yields fairly good dynamic and steady-state performances irrespective of applied load torque.