Search

In this paper, the formation tracking consensus problem is investigated for multi-agent systems with a leader subject to bounded input based on the dynamic event-triggered control. A novel event-triggered control protocol is designed to achieve the desired formation consensus. First, the formation compensation input is proposed by solving the time-varying formation tracking feasible constraint equation. Second, for the leader's bounded control input, a nonlinear auxiliary function is constructed to compensate the input influence. Then, the dynamic auxiliary variables are introduced in dynamic event-triggered functions, which can adjust the triggering conditions, such that the number of unnecessary triggers can be effectively reduced and communication efficiency can be improved. Finally, a simulation example is presented to illustrate the effectiveness of the result.

One major challenge regarding series-connected IGBTs is the protection against over-current faults, besides voltage balancing among the devices. In this paper, an over-current protection method based on desaturation detection is proposed for series-connected IGBTs which are controlled by a digital signal processor (DSP) to realize voltage sharing. A hardware test platform was built, which is capable of generating current spikes of various magnitudes and durations. Different aspects of system behavior under fault are characterized to provide design guidance to ensure safety. Finally, experimental results show that the proposed series IGBT control can achieve voltage balance without slowing down the switching transients. Additionally, the desaturation circuit can detect an over-current in less than 1 μs. Furthermore, with the proposed protection method, the fault can be fully cleared depending solely on voltage balancing circuits, while maintaining voltage balance during the entire process.

Modular multilevel converters (MMC) and series-connected IGBTs are two promising solutions to high-voltage dc (HVDC) transmission. It has been shown that, when integrated with series IGBTs, a MMC can benefit from a reduced number of sub-modules (SMs), simpler control, and improved reliability. In this paper, a MMC prototype, with the capability to integrate series IGBTs to replace single IGBTs in the sub-modules, is designed and developed. The MMC control is realized with a real-time simulator for versatile software implementation. The voltage sharing among the series IGBTs is regulated locally without the intervention of the central control. Furthermore, voltage balance under bi-directional current through the series strings is achieved. Experiments were conducted where individual IGBTs in one sub-module were replaced with a series of three IGBTs each. The results verify that the MMC is functional since the local series IGBT control and the MMC central control are independent. It is also verified that both series strings can maintain voltage balance at the same time even though the string currents shift directions.

Digits/numerals in the Indian pin-code of handwritten postal documents may touch each other and hence digit string recognition is a very challenging task. In this paper, we propose a digit string recognition system for Indian postal documents written in Hindi. Unlike normal text string, in a string of digits there is no contextual information among the digits as a digit may be followed by an arbitrary digit in a string of digits. Because of this, here we propose a new architecture which is based on CNN (Convolutional Neural Network) and CTC (Connectionist Temporal Classification), without using RNN for Hindi numeral string recognition. Also to connect CNN with CTC, we transform the outputs of CNN to a two-dimension vector to meet the feeding requirement of CTC. Furthermore, we utilize dense blocks to build CNN part to extract efficient image features. Comparative studies with the state-of-the-art methods show that the proposed method outperforms the other existing methods on Hindi numeral string recognition.

Silicon carbide (SiC) device is expected to replace silicon (Si) device in many high voltage and power application due to potentially higher switching frequency and voltage capabilities. Even though its potential voltage capability, the highest voltage rating of a single commercial SiC MOSFET is 1.7 kV. An alternative method to achieve higher voltage is a device series connection. However, it causes a voltage unbalance issue during turn-off state that can harm the devices. Therefore, the voltage balancing control (VBC) must be considered in device series connection. In this paper, a VBC method with active gate driver (AGD) is proposed. A gate resistance modulation method with a signal time delay is adopted for dynamic voltage sharing. The proposed method is experimentally verified.

In this paper, we investigate a novel learning-based method for dynamic scene deblurring. Since the inference model is formulated as an encoder-decoder, the core task has turned to learning blur-invariant hidden features from the blurred images to a great degree. To achieve state-of-the-art results in terms of both deblurring accuracy and efficiency, a simple, robust and computationally efficient deep auto-encoder is developed tailored specifically for blind deblurring, which is learned in an adversarial fashion based on use of the recent Wasserstein generative adversarial networks. Thanks to the designed framework, the new model has shown comparable or superior performance both qualitatively and quantitatively to existing state-of-the-art methods. What is more important, instead of exploiting the multi-scale strategy as previous methods, our model is just single-scale capable of achieving 6 times more efficiency than the closest competitor by Nah et al. [1], which is a more complex multi-scale deep method.

Optimization of power system restoration path is a combinatorial optimization problem, which is one of the sub-problems of the whole power system restoration. Because the optimization of restoration path problem is a classical NP-hard problem and time consuming, furthermore the corresponding solution algorithm for optimal path will be performed frequently during the entire optimization process. It is very important to speed-up the computational runtime of optimization procedure. In existing research, heuristic algorithm has been widely used to solve this problem due to its strong optimization capability, and has achieved a good performance. But the efficiency of existing heuristic algorithm for large scale system should be further enhanced. Therefore, an efficient heuristic algorithm based on minimum cost maximum flow modelling has been proposed in this paper. The problem of optimizing power system restoration path is formulated as a minimum cost maximum flow problem. The system topology is constructed as a single-source and single-sink network, maximizing the network flow is used to ensure the connectivity of restoration path, then the best set of energizing transmission line can be found by minimizing the cost of flow from source s to sink t. The IEEE standard test systems are used to examine the applicability of proposed method. Simulation results demonstrate that the proposed method is more efficient than traditional method.

Series-connected IGBTs, when properly controlled, operate similarly to a single device with a much higher voltage capacity. Integrating series IGBTs into a Modular Multilevel Converter (MMC) can reduce its complexity without compromising the voltage capacity. This paper presents the circuit design on the sub-modular level of a MMC in which all the switching devices are series-connected IGBTs. The voltage sharing among the series IGBTs are regulated in a self-balancing manner. Therefore, no central series IGBT controller is needed, which greatly reduces the sensing and communication complexities, increasing the flexibility and expandability. Hardware experiment results demonstrate that the series IGBTs are able to self-regulate the voltage sharing in a fast and accurate manner and the system can operate similarly to a sub-module in a MMC. Current

Series-connected IGBTs, with their ability to withstand high voltages, are necessary to fully achieve High-Voltage DC (HVDC) Transmission and Flexible AC Transmission (FACT). In this paper, a hardware implementation method of a self-balancing control of series IGBTs is developed and tested. The proposed method does not rely on either resistive balancing components because of high power loss or costly high-end digital units. Instead, it regulates the voltage sharing among IGBTs via gate pulse delays introduced at modulation stage. Experiments with 4 series modules indicate that the proposed hardware implementation method is able to enhance both the steady-state and transient voltage sharing. The auxiliary balancing circuits consume low energy loss per cycle.

Modern dc power systems are highly penetrated with power electronics circuits, which presents challenges to fault detection and protection. This paper studies the effect of de arc fault to dc power systems with constant power loads (CPLs). Experiments with various conditions were conducted and analyzed. The impact of current level to the formation of intermittent and sustained arcing is presented. Frequency spectrum and discrete wavelet decomposition is used to analyze the current and provide critical insights into the arc characteristics. The effect of controller response speed, as well as fault current coupling among two parallel CPLs will also be presented.

With the increasingly prevalent application of dc systems, protection of such systems has become an important research focus. Given that many dc systems are augmented with buck converter interfaces, this paper proposes a novel method of series fault detection. Instead of exploiting the signatures of an arc, the proposed method detects a fault by sensing the intermittent discharges at the beginning of the fault which are caused by the high input capacitance of the interface. In this paper, the characteristics of such phenomena are investigated and a detection algorithm utilizing spark signatures, as well as crosscorrelation, is proposed. Experimental validation is designed and conducted on a scaled down dc microgrid with a constant power load (CPL) and the detection is achieved with Arduino Uno and Udoo X86. The results show that the proposed method detects sparks in a timely and accurate manner.

Pulsed power generation is required by various types of industrial applications. With the current development of semiconductor devices, solid-state pulse generators have been growing. To fully integrate such devices, series connection is preferred for high-voltage applications. In this paper, a review of series IGBT control methods for solid-state switching devices are provided. These methods are categorized into three groups based on the locations of the auxiliary circuits: load-side, gate-side and both sides combined. Then, they are analyzed and compared in terms of output voltage rise time, voltage balancing, loss, and complexity, specifically for pulse generation. Finally, it is shown that each group excels only at certain aspects and that further development of series-connected devices for pulse generation is necessary.

Modular Multilevel Converter (MMC) has been an attractive option for power transmission and has been studied and used for HVDC applications. However, it still has limitations when the number of Sub-Modules per arm is large. To avoid the limitations and exploit the advantages, a method is proposed in this paper which integrates an MMC with series-connected IGBTs in each Sub-Module (SM). The proposed MMC is able to withstand a higher voltage with devices of lower voltage ratings. The series-connected IGBTs are self-balancing. Simulink and real-time simulations both show that, with the proposed method, series IGBTs are able to share a much higher MMC sub-module voltage and recover from voltage imbalances in a timely manner. The proposed series IGBT balancing method causes considerably less loss compared to RCD snubber circuit-based balancing methods.