Taesic Kim

Power pulsation decoupling for a two-stage single-phase photovoltaic inverter with film capacitor

This paper presents a power pulsation decoupling strategy for a two-stage single-phase photovoltaic (PV) inverter with film capacitor, which has small capacitance and large voltage ripple. Such large voltage ripple at DC bus is propagated to the PV array and decreases the maximum power point tracking (MPPT) efficiency. To maintain the MPPT efficiency, a new small-signal model which considers the DC bus voltage ripple is derived. Based on the model, a new boost current controller is developed and it consists of two parts: a PI controller to control DC component and a repetitive controller to suppress the harmonic component caused by the DC bus voltage ripple. Simulation results show that the proposed controllers have drastically reduced the double frequency voltage of the PV array, which improves the MPPT efficiency of the PV array. Experimental result further verified the effectiveness of the controller, the measured MPPT efficiency was increased from 97.06% to 99.9%.

Internet of Things-enabled multiagent system for residential DC microgrids

This paper proposes an Internet of Things (IoT)-enabled multiagent system (MAS) for residential DC microgrids (RDCMG). The proposed MAS consisting of smart home agents (SHAs) aims to cooperate each other to alleviate the peak load of the RDCMG and to minimize the electricity costs for smart homes. These are achieved by agent utility functions and the best operating time algorithm (BOT) in the MAS. Moreover, IoT-based efficient and cost-effective agent communication method is proposed, which applies message queuing telemetry transport (MQTT) publish/subscribe protocol via MQTT brokers. The proposed IoT-enabled MAS and smart home models are implemented in five Raspberry pi 3 boards and validated by experimental studies for a RDCMG with five smart homes.

A new hybrid filter-based online condition monitoring for lithium-ion batteries

This paper proposes a novel online condition monitoring algorithm estimating battery states and model parameters. The proposed method includes: 1) an electrical circuit battery model incorporating the hysteresis effect, 2) an extended Kalman Filter-based online parameter identification algorithm for the electrical battery model, and 3) a smooth variable structure filter (SVSF)-based state estimation algorithm for state of charge (SOC) estimation. The proposed method enables an accurate and robust condition monitoring for lithium-ion batteries. Since the proposed hybrid filter further reduces the complexity compared to existing dual extended Kalman filter (DEKF), it is much more suitable for the real-time embedded battery management system (BMS) application. Simulation studies validate the effectiveness of the proposed strategy.

Fast recursive least square-based active and reactive power estimator for single-phase power electronic converters

This paper proposes a novel estimator for active and reactive power control of single-phase power electronic applications. The proposed estimator includes: 1) an instantaneous grid power regression model; 2) a fast upper-triangular and diagonal recursive least square (FUDRLS) algorithm for parameter identification of the regression model; and 3) a new self-tuning variable forgetting factor included in the FUDRLS to optimize its performance. The proposed estimator is capable of quick and accurate active/reactive power estimation. Due to its low complexity, numerical stability, the proposed method is suitable for the real-time embedded system to estimate active and reactive powers for control applications of the grid-connected converters. Simulation results are provided to validate the proposed estimation method of single-phase power electronic converters.

Low-voltage ride-through performance of sensorless vector controlled doubly-fed induction generators

This paper investigates the low-voltage ride-through (LVRT) performance of a proposed sensorless vector control method for the doubly-fed induction generators (DFIGs) under power grid voltage sags. The sliding-mode observer (SMO) is used to estimate the rotor position and speed information for the vector control of the DFIG. The performance of the SMO is compared with the model reference adaptive system (MRAS)-based sensorless method over a wide operating speed range. The effectiveness of the proposed sensorless vector control scheme under power grid voltage sags is demonstrated by simulation results for a 2-MW DFIG-based wind energy conversion system (WECS) in MATLAB/Simulink and experimental results for a 200-W DFIG-based WECS simulator.