Gyo-Bum Chung

Low Frequency Pulsed Resonant Converter for Energy Harvesting

A pulsed-resonant ac-dc converter and an integrated circuit (IC) controller have been designed, fabricated, and tested for harvesting energy from low-voltage (1.2V), low-power (1-100muW) energy transducers with output frequency in the 10-Hz-1-kHz range. Simulations using foundry models suggest that the silicon loss could be as low as 0.6muW, and the efficiency could reach 70%. With the IC experimentally packaged, the measured efficiency is between 50% and 70%, depending on the size and the loss in the resonant inductor

SOC estimation of LiPB batteries using Extended Kalman Filter based on high accuracy electrical model

This paper proposes a SOC (State-of-charge) estimator of batteries using the Extended Kalman Filter (EKF). EKF can work properly only with accurate model. Therefore, this paper includes high accuracy electrical battery model for EKF state. The modeling is focused on high-capacity LiPB batteries. The battery model is extracted from single cell of LiPB 40Ah, 3.7V. The dynamic behavior of single cell battery is modeled using a bulk capacitor, two series RC networks and a series resistance. The voltage in bulk capacitor represents OCV (Open Circuit Voltage) related to SOC value. EKF is employed to estimate the OCV, and then OCV is converted to SOC value. EKF is tested with experiments data obtained by MACCOR battery tester. EKF is improved with single varying parameter of bulk capacitor which follows the SOC value. The test of estimator algorithm is done for full region of SOC. The test results show the error of estimation can reach max 5%SOC at some points.

Shape Optimization of a Large-Scale BLDC Motor Using an Adaptive RSM Utilizing Design Sensitivity Analysis

This paper proposes a new algorithm for the shape optimization of a large-scale brushless dc (BLDC) motor to reduce the cogging torque. In the algorithm, an adaptive response surface method (RSM) using the multiquadric radial basis function is employed to interpolate the objective function in design parameter space, and incorporated with (1+1) evolution strategy to find an optimal point. In the adaptive RSM, an adaptive sampling point insertion method is developed utilizing the design sensitivities computed by using the finite-element method to get a reasonable response surface with a relatively small number of sampling points. The developed algorithm is applied to the shape optimization of a 5-MW BLDC motor, and the cogging torque was reduced to 17% of the initial one

Robust digital controller design for three-phase inverter using CRA method

This paper proposes a CRA based robust digital controller for three-phase PWM VSI inverter. To regulate the inverter output voltage with constant voltage and constant frequency, the usual inverter controller consists of double looped PI controllers. Those are an outer voltage controller and an inner current controller, then the order of system characteristic polynomial is high and so the gain tuning of the controllers is difficult. Considering the limited switching frequency of the devices and sampling frequency of the digital controller, the gain tuning is practically dependent on the engineering experiences with the try and error method. In this paper, the error-space approach is used to get the system model including the controller with low order, and the characteristic ratio assignment (CRA) method is proposed for the design of robust controller which has the advantage to design the optimal gain to meet the referenced response and overshoot within the limit range. The validity of the proposed controller is verified through the PSiM simulation and experimental results.

Design and Fabrication of Electromagnetic Micromirror with Bulk Silicon Mirror Plate and Aluminum Spring.

We have designed and fabricated an electromagnetic micromirror with a flat mirror plate capable of large angular deflection. Single-crystal silicon is used as a mirror plate to obtain an optically flat reflective surface. By a combination of surface and bulk micromachining processes, the single-crystal silicon mirror plate is connected to the substrate via an aluminum spring, which occupies a relatively small area without compromising the compliance. The proposed fabrication process can be applied to the fabrication of devices in which a flat moving part is supported by a compliant spring structure. Nickel electroplated onto the mirror plate enables a mirror rotation angle of up to 84.8° by applying an external magnetic field vertical to the substrate. This designed structure can be used individually or as an array in microphotonic applications.

Improvement of electrical modeling of NiMH battery for application of Microgrid System

Simulation can be a powerful tool to evaluate the batteries for the utilization in system level. It can be a low-cost and simple evaluation method instead of the high-cost and complicate prototyping. However, to guarantee the simulation results of batteries is similar to real condition, it is necessary to get a high-accuracy electrical battery model. In this paper, battery modeling is focused on high capacity of NiMH battery for the application of Microgrid System. The tested battery is prismatic single cell module, 1.2V, 100Ah. Experiment results are shown: The SOC-OCV relation has very strong hysteresis characteristic and its relation in charging and discharging is obviously different. The terminal voltage response has a different order of exponential function between loaded and rest conditions. This paper proposes the different RC networks for loaded and rest condition. The model is developed in PSiM and is verified with the experimental data.

Power disturbance classifier using wavelet-based neural network

This paper presents a wavelet-based neural network technology for the detection and classification of the various types of power quality disturbances. For the detection and classification of the PQ transient signals, it should be done at the same time and the automatic methodology is recommended. In this paper, the hardware and software of the power quality data acquisition system (PQDAS) is proposed. In this system, the auto-classifying system combines the properties of the wavelet transform and the advantages of neural networks. Especially, the additional feature extraction to improve the recognition rate is considered. The configuration of the hardware of PQDAS and some case studies are also described.

Active damping of LCL filter without capacitor voltage sensors for three phase PWM inverter

In this paper, new sensorless active damping method of the grid-connected LCL filter for a three-phase PWM inverter is proposed. The resonance component of capacitor voltage of LCL filter is observed by a simple MRAS (Model Reference Adaptive System) observer without additional sensors or circuits. The active damping is implemented by feed-forward compensation of the observed capacitor voltage for current controller. The validity of the proposed method is verified by the simulation and experiment.

Development of artificial-intelligent power quality diagnosis equipment for single-phase power system

A DSP process-based equipment to diagnose the power quality of a single-phase power system is developed. The artificial-intelligent equipment diagnoses the transient, the voltage sag, the voltage swell and the THD among the power quality index of a power system. The 256 data sampled in a period of the single-phase voltage of the power system are used for the real-time calculation of RMS value, Fast Fourier Transform (FFT) and Discrete Wavelet Transform (DWT). The type of mother function for DWT is DB4. The data are measured at 154kV or 22.9kV substations for a year. The feature vectors extracted from the data are used to train the neural network for the artificial-intelligent diagnosis of power quality. The type of the activation function in the neural network is sigmoidal. After learning with the feature vectors, the back-propagation algorithm simulated in PSIM program and C++ code generates the weights and biases of the neural network, which are used in the DSP320C6713-based Artificial-Intelligent Power Quality Diagnosis Equipment (AIPQDE). The developed equipment detects satisfactorily the PQ problems in a real situation simulated in the laboratory.

Analysis of LLC Resonant Converter with Current- Doubler Rectification Circuit

In this paper, LLC series resonant converter with current-doubler as a rectification circuit for secondary side is proposed. The current-doubler circuit is generally used for the high voltage input and low voltage output application to get high efficiency characteristics with reduced transformer turn ratio. However, an inductive circuit is not used in a secondary side for the LLC series resonant converter generally. When there is inductive circuit like current-double on the secondary side, the resonant characteristics are changed. In this paper, LLC converter with the current-doubler using coupled inductor is proposed. We verify that the secondary side current-doubler using coupled inductor is not affect to the resonant characteristic of primary LLC by the mathematical analysis. Finally, the analysis results are proved by the simulation and experiments.