Browsing by Author "Chen, Rui"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Analysis and Design of a DCM SEPIC PFC with Adjustable Output Voltage
    Chen, Rui (Virginia Tech, 2015-03-31)
    Power Factor Correction rectifiers are widely adopted as the first stage in most grid-tied power conversion systems. Among all PFC converts for single phase system, Boost PFC is the most popular one due to simplicity of structure and high performance. Although the efficiency of Boost PFC keeps increasing with the evolution of semiconductor technology, the intrinsic feature of high output voltage may result cumbersome system structure with multiple power conversion stages and even diminished system efficiency. This disadvantage is aggravated especially in systems where resonant converters are selected as second stage. Especially for domestic induction cooker application, step-down PFC with wide range output regulation capability would be a reasonable solution, Conventional induction cooker is composed by input filter, diode-bridge rectifier, and full bridge or half bridge series resonant circuit (SRC). High frequency magnetic field is induced through the switching action to heat the pan. The power level is usually controlled through pulse frequency modulation (PFM). In such configuration, first, a bulky input differential filter is required to filter out the high frequency operating current in SRC. Second, as the output power decreases, the operating point of SRC is moved away from the optimum point, which would result large amount circulating energy. Third, when the pan is made of well conducting and non-ferromagnetic material such as aluminum, due to the heating resistance become much smaller and peak output voltage of the switching bridge equals to the peak voltage of the grid, operating the SRC at the series resonant frequency can result excessive current flowing through the switch and the heating coil. Thus for pan with smaller heating resistance, even at maximum power, the operating frequency is pushed further away from the series resonant point, which also results efficiency loss. To address these potential issues, a PFC circuit features continuous conducting input current, high power factor, step-down capability and wide range output regulation would be preferred. The Analysis and design work is present in this article for a non-isolated hard switching DCM SEPIC PFC. Due to DCM operation of SPEIC converter, wide adjustable step-down output voltage, continuous conduction of input current and elimination of reverse recovery loss can be achieved at same time. The thesis begins with circuit operation analysis for both DC-DC and PFC operation. Based on averaged switching model, small signal model and corresponding transfer functions are derived. Especially, the impact from small intermediate capacitor on steady state value are discussed. With the concept of ripple steering, theoretic analysis is applied to SEPIC converter with two coupled inductors. The results indicate if the coupling coefficient is well designed, the equivalent input inductance can be multiple times larger than the self-inductance. Because of this, while maintaining input current ripple same, the two inductors of SEPIC can be implemented with two smaller coupled inductors. Thus both the total volume of inductors and the total number of windings can be reduced, and the power density and efficiency can be improved. Based on magnetic reluctance model, a corresponding winding scheme to control the coupling coefficient between two coupled inductors is analyzed. Also the impact of coupled inductors on the small signal transfer function is discussed. For the voltage follower control scheme of DCM PFC, single loop controller and notch filter design are discussed. With properly designed notch filter or the PR controller in another word, the closed loop bandwidth can be increased; simple PI controller is sufficient to achieve high power factor; THD of the input current can be greatly reduced. Finally, to validate the analysis and design procedure, a 1 kW prototype is built. With 120 Vrms AC input, 60V to 100V output, experimental results demonstrate unity power factor, wide output voltage regulation can be achieved within a single stage, and the 1 kW efficiency is around 93%.
  • Thumbnail Image
    Application of training data affects success in broad-scale local climate zone mapping
    Xu, Chunxue; Hystad, Perry; Chen, Rui; Van Den Hoek, Jamon; Hutchinson, Rebecca A.; Hankey, Steven C.; Kennedy, Robert (2021-12-01)
    Satellite imagery has been widely used to map urbanization processes. To address the urgent need for urban landscape mapping that goes beyond urban footprint analysis, the local climate zone (LCZ) scheme has been increasingly used to reveal the urban forms and functions important to urban heat islands and micro-climates across the globe. As with most supervised classification strategies, proper application of training data is critical for the success of LCZ classification models. However, the collection and application of LCZ training areas brings with it two challenges that may affect mapping success. First, because digitizing training areas is a timeconsuming task, there is a broad effort in the LCZ mapping community to create a crowdsourced data collection among different experts. However, this strategy likely leads to inconsistencies in labels that could weaken models. Second, the LCZ labeling process typically involves the delineation of large zones from which multiple training samples are drawn, but those samples are likely spatially autocorrelated and lead to overly optimistic estimates of model accuracy. Although both effects - inconsistent labeling and spatial autocorrelation - are theoretically possible, it is unknown whether they substantially affect accuracy. We investigated both issues, specifically asking: (i) how do the discrepancies of LCZ labeling by different experts impact broad-scale LCZ mapping? (ii) to what extent does spatial correlation affect model prediction power? We used two classifiers (Random Forests and ResNets) to map eight metropolitan areas in the US into LCZs, comparing training areas drawn by different or consistent interpreters, and data splitting strategy using rules that allow or reduce spatial autocorrelation. We found large discrepancies among results built from crowdsourced training areas digitized by different experts; improving the consistency of labels can lead to substantial improvements in LCZ classification accuracy. Second, we found that spatial autocorrelation can boost the apparent accuracy of the classifier by 16% to 21%, leading to erroneous interpretation of mapping results. The two effects interplay as well: spatial auto correlation in the raw data can lead to an underestimation of the model's predictive error when modeling with crowdsourced training areas of high inconsistency. Due to the uncertainty in the labeling process and spatial autocorrelation in derived training data, broad-scale LCZ mapping results should be interpreted with caution.