This paper presents a method to extend the DC bus utilization on an induction motor (IM) by using a combination of Space-Vector Modulated Direct Torque Control (DTC-SVM) and conventional DTC. DTC-SVM along to stator flux orientation has several advantages over the more standardized indirect rotor field-oriented control (IRFOC). One of the most important advantages of DTC-SVM is its total independency on motor parameters, which are difficult to determine and dependent on external factors such as temperature. On the other hand, both DTC-SVM and IRFOC share a disadvantage, which is limited DC bus utilization during the linear region. However, the fact that torque and flux are directly controlled for DTC-SVM, allows a straightforward transition to the conventional DTC. Resulting in instantaneous transient response, motor parameters independency, and six-step operation capability. The scheme proposed in this paper exploits the advantages of both control methods. It allows to have low torque ripple and low current harmonic distortion (THD) during the linear region, as well as the fastest torque response up to the six-step operation region, always keeping the independency on rotor parameters. DTC-SVM is applied while the modulation index (MI) is within the linear modulation region limit. As soon as this limit is exceeded, the control algorithm will switch to conventional DTC. The paper describes a way to provide smooth transition between the two control schemes. Simulation and experimental results are presented to verify the feasibility of the proposed method.
This paper proposes a charge current controller based on a concept of partial-rated converter for capacitor discharge type pulse voltage generators. The required rated voltage of the partial converter can be 10% ∼ 20% of the system voltage rating of the pulse voltage generator. This paper presents an example of control method for this charge current controller and its pre-charge operation for capacitors. A laboratory-scale prototype was fabricated and experimental verification with the prototype was conducted. An estimation of the volume increase in capacitors is presented as a case study. The estimation indicates that the volume increase in capacitors by adding the proposed current controller is 3% for the case.
To decrease voltage stress of power devices and enhance LED average current accuracy, the dual-input LED driver with an adaptive off-time (AOT) control is presented in this paper. Thanks to the dual-input architecture, its low-voltage power switches can be on-chip to significantly reduce power loss with the same chip area. The AOT control with a current sensing signal automatically regulates the off-time duration based on the LED number and input voltages. Thus, precise average inductor current can be achieved with slight extra power. Experimental results demonstrate the theoretical analysis of the dual-input LED driver.
This paper proposes a mixed modulation combining unipolar and bipolar modulations with the off-time discrete control for the discontinuous current mode (DCM) grid-tied inverters to improve the efficiency. The proposed method has the advantages of the two conventional modulation. The off-time discrete control, which discretely adjusts the oscillation cycles during the zero-current period of the DCM, enables zero-voltage switching (ZVS). This paper presents a strategy to mix both modulations for achieving ZVS and the peak current reduction. The performance of the proposed method was validated by experiments with a 400 W class prototype.
This paper discusses a power loss, which occurs in a load-resonant inverter operating at megahertz for wireless power transfer (WPT) systems. A calorimetric method is applied to measure the power losses in the mega-hertz operation accurately. The loss is measured using a single chamber with temperature control of the internal chamber using a Peltier device. The loss can be measured within 10% error when the loss is less than 23 W with a prototype chamber in experiment. A drive circuit loss, conduction loss of GaN devices, and PCB pattern loss are measured using the measurement site. Moreover, the total loss of the inverter with 6.78-MHz operation is measured under the zero voltage switching (ZVS) condition. The losses on the inverter are separated by a combination of each measurement result. As a result, the loss, which is related to the mega-hertz operation under the ZVS condition, is 9.9 W when the output power is 417 W at 6.78-MHz operation.
In the application of electric vehicle battery systems, an auxiliary power module with wide range is required to convert the high-voltage battery pack to the low-voltage auxiliary battery. A two-stage converter with wide-range is implemented with silicon carbide power devices in this paper. The first stage is a boost converter and the second stage is a full-bridge resonant converter. According to the output voltage range of the system, the corresponding output voltage range through pulse width modulation control is designed for the first-stage converter. In this way, the second-stage converter can be operated through near series resonant frequency control, the magnetizing inductance can be designed to be large to reduce circulating current, and with the synchronous rectification technology of the secondary side, the overall system efficiency can be improved. In this paper, a prototype with input voltage (VH) 220-450 V, output voltage (VL) 9-16 V, and rated power 2.5 kW is built with a digital signal processor TMS320F28335. As the experimental results shown, the maximum efficiency of the system is 97.74% with VH = 450 V and VL = 9 V at 20% load condition.
This paper presents analysis of unidirectional isolated Y-Y connection three-phase DC-DC converter. The waveforms of the proposed converter is similar to conventional Y-Y connection three-phase dual-active-bridge (DAB) DC-DC converter. Therefore, by deriving the phase current and power flow of each DC-DC converter, the waveforms, power factor and efficiency of the proposed converter will be compared to the DAB to clarify the effectiveness of the proposed converter. Finally, the characteristics of the converter is verified by simulation and experiment.
This paper proposes an estimation method of switching-waveforms in power devices based on only characteristics of drain-source capacitance, gate-source capacitance, and gate-drain capacitance and output characteristics of $i_{\text{d}}-v_{\text{d}\text{s}}$. Since the proposed method needs the $i_{\text{d}}-v_{\text{d}\text{s}}$ characteristics in wide-current and -voltage range, this paper also proposes a measurement platform having capability of the $i_{\text{d}}-v_{\text{d}\text{s}}$ characteristic measurement under both high-voltage and large-current condition. The proposed method makes it possible to reproduce gate–source voltage variation due to the short-channel effect around the Miller plateau. It is confirmed that the estimated switching waveforms well agree with the ones measured in the experiment.
The previously proposed non-isolated high-buck converter [1–7], although with a small number of components, simple control and zero voltage switching, has no isolation function. This paper proposes an ultra-low-gain buck converter with a novel capacitive isolation architecture. In addition to the traditional architecture, the converter has the following two advantages: (1) using only a single non-isolated coupled inductor and adding a capacitor. (2) The secondary-side synchronous rectification adopts independent driving to make the controller simple. Finally, a prototype will be completed and the proposed concept will be valid.
Different from conventional digital low dropout (DLDO)regulators that adaptively respond to load changes, the proposed DLDO regulator communicates with the microprocessor to achieve fast transient response at low switching frequencies. The proposed data detection (DD) technique collects data and sends it to the cache to record the operation of the entire circuit at different drive currents. Error Quantization Control (EQC) sets adequate compensation to immediately regulate the output voltage. Therefore, compensation corresponding to different load currents can eliminate unnecessary recovery time, and limit the number of switchable power MOSFETs to one to reduce the limit cycle oscillation (LCO). The test chip fabricated in CMOS 40nm process proves the settling time reduces from 1.3μs to 0.52μs. At the same time, more than 62.3% reduction in power and 60% reduction in recovery time can be achieved.
With the popularity of electric vehicles, there are more and more battery charging devices for electric vehicles, and the full bridge converter is suitable for medium and high power occasions. Therefore, the full bridge converter has become the research goal. Moreover, gallium nitride transistors based on the third generation wide-band gap semiconductor materials have more advantages than MOSFETs. Therefore, the application of gallium nitride transistor instead of MOSFET in electric energy converter has become an important research topic. In this paper, the full bridge architecture is adopted, and the phase-shift modulation is used to make the converter have zero voltage switching, and the gallium nitride transistor is used as the power switch to improve the efficiency of the converter. The proposed converter has the following characteristics: (1) It can be applied to electric vehicle charger and has zero voltage switching to improve the efficiency of the converter (2) Gallium nitride transistor is used for switching to improve the overall efficiency of the converter (3) The efficiency of the converter under light load can be significantly improved by using gallium nitride transistor. In this paper, two sets of phase shift full-bridge converters using gallium nitride transistor and MOSFET are completed and compared. The input voltage range is 300 V $\sim$ 350V, the output voltage is 96 V, the rated output power is 480 W, and the maximum efficiency is 96.15%. At light load, the efficiency of gallium nitride circuit is 6.84% higher than the circuit using MOSFET.
The use of light-emitting diodes (LEDs) in street lighting applications has been greatly welcomed with the current trends of energy saving, environmental protection, carbon reduction, and sustainable development. This paper presents a novel AC-DC LED integrated streetlight driver that combines an interleaved buck converter with a coupled inductor and a half-bridge series resonant converter with a full-bridge rectifier into a single-stage power conversion topology with power factor correction (PFC) and soft switching capabilities. The PFC is achieved by designing the coupling inductor in the interleaved buck converter sub-circuit in discontinuous conduction mode. In addition, the resonant tank in the half-bridge series resonant converter sub-circuit is designed to be similar to an inductive load, thus giving the power switch a zero-voltage switching (ZVS) function, decreasing switching losses and increasing the overall efficiency of the proposed circuit. A prototype circuit of the proposed LED integrated streetlight driver with a power rating of 165 W (235 V/0.7 A) and 110 V input utility voltage has been developed and tested. According to the measurement results, a power factor greater than 0.98, a total harmonic distortion coefficient of the input current less than 3%, and an efficiency greater than 89% were obtained in the AC-DC LED integrated streetlight driver. Therefore, the experimental results are satisfactory and demonstrate the functionality of the proposed AC-DC LED integrated streetlight driver.
This paper considers the unweighted double dominating set problem seeking a minimum cardinality node subset of a graph such that the closed neighborhood of every node must contain at least two members of the double dominating set. This problem is motivated from power system measurement system planning (e.g. PMU placement) and cyber-physical security analysis (e.g. protection assignment). While generically the problem can be posed as an integer linear program and solved using methods such as branch-and-bound (e.g. via Gurobi), in this paper we investigate an alternative solution methodology based on branch decomposition and dynamic programming. To facilitate the computation problem reduction procedures are proposed. It is demonstrated, through benchmarks in MATPOWER 7 (including the 25,000-bus ACTIVSg25k example), that the proposed method is on par with the state-of-the-art integer programming solver Gurobi in computation speed while possessing the advantage of explicit bound in complexity.
This paper presents the use of grasshopper optimization algorithm based selective harmonic minimization pulse-width modulation technique for optimizing the switching angles of five-phase nine-level multilevel inverter. The proposed algorithm is able to provide optimum switching angles in a wide range of modulation index. Besides, the switching angles obtained by the proposed algorithm can be used to synthesize an output voltage waveform with desired fundamental component and minimized undesired low order harmonics. MATLAB simulation results show the proposed algorithm can provide a wide modulation index range of optimum solutions. Furthermore, both MATLAB simulation and PSIM experimental results obtained from five-phase nine-level cascaded H-bridge multilevel inverter show good agreement, confirming the validity of switching angles calculated using the proposed algorithm.
In the development process of power electronics system, a wide range of know-how is required for special-purpose circuits such as main circuits, control circuits and gate drive circuits. The Universal Smart Power Module (USPM) has been proposed to solve these problems by converting the above circuits into one module to reduce the design cost and development time. Since the USPM is a combination of several circuits that modularize various functions, its controller requires the ability to perform high-speed digital calculation and high-speed sampling. In this paper, verifications were carried out for a single-phase PWM inverter using a 50MHz sampling ADC with the USPM controller. The performance of USPM controller was discussed.
Multilevel technology has the advantages of reducing the withstand voltage and voltage stress of a switch and improving the quality of the output current/voltage waveform. This paper takes the T-type three-level converter as the research object and analyzes its topology and means of transforming current to establish a mathematical model. The model reference adaptive control (MRAC) strategy is used to control the output waveform of the system. Based on this, a midpoint potential balancing algorithm based on SVPWM is introduced. Finally, a 5-kW prototype is tested to verify the proposed design procedure.
Switching power supplies equipped with power factor correction (PFC) control schemes are an essential part of modern power conversion architectures, as they provide high power factor and low total harmonic distortion. Boost converters are the most commonly used topology in PFC circuits. In order to design the feedback controller for a PFC circuit, simulation is a useful tool to investigate the cir...
This paper discusses the design and analysis of a modified test bench for high-voltage direct-current (HVDC) circuit breakers which consists of a power converter based on multiple cascaded H-bridge cells, a small-sized inductor, and an auxiliary capacitor bank. It is capable of producing output voltages up to several hundred kilovolts depending on the auxiliary capacitor bank while allowing for high current controllability via the power converter which is driven by phase-shifted pulse-width-modulated (PSPWM) signals. Consequently, the test bench can generate a wide range of controllable current waveforms and high current gradients, within hardware limitations, to simulate a wide range of fault conditions. The flexibility of the proposed design is complemented by higher reliability and longer lifetime. The concept of the test bench was verified by experiments using a down-scaled model based on nine cascaded H-bridge cells with an equivalent switching frequency of 92.5 kHz.
The introduction of distributed power generation systems is expanded worldwide, many grid-connected inverters are connected to the grid. Since the grid impedance affects the stability and control performance of the grid-connected inverter, it is important to accurately estimate the grid impedance. The grid impedance has resonant characteristics because the capacitor component is connected to the grid in addition to the wiring LR component for power factor improvement. In this paper, the impedance measurement method with variable sampling frequency for grid impedance with resonant characteristics is proposed, and the measurement accuracy of the grid impedance with resonant characteristics is verified.
In recent years, solar photovoltaic (PV) generation becomes one of the most relevant energies. However, the intermittent characteristics of solar generation create significant problems to power system operations. To overcome this problem, many solar power forecasting techniques have been developed, and different forecasting horizons require different methodologies. For a short-term prediction, forecasting horizons generally require numerical weather prediction models (NWP) that provide an important estimation of weather variables such as solar irradiance, temperature, wind speed, rainfall, air pressure, etc. This research proposes a machine learning model based on Kernel Principal Component Analysis (PCA) - XGBoost to improve the accuracy of one-hour-ahead solar power forecasts. The model considered the deterministic Weather Research and Forecasting (WRFD) provided by Taiwan Central Weather Bureau (CWB). Furthermore, a XGBoost model was built on an ensemble of decision trees, providing important information and appropriate results in the forecasting process.
