In this paper, a sequence-to-sequence deep learning architecture based on bidirectional Gated Recurrent Unit (Bi-GRU) for type recognition and time location of combined power quality disturbance is proposed. Specially, the proposed methodology can determine the type of each element in input sequence, which is different from existing sequence-to-sequence model employing encoder-decoder network. Firstly, the input sequence is normalized and batched. Secondly, deep features are extracted from input sequence by constructing Bi-GRU recurrent neural network, where multiple Bi-GRU layers are stacked together in both forward direction and backward direction. Thirdly, according to above extracted features, fully connected layer and Softmax are employed to calculate the corresponding probability indicating the category that each element in input sequence is classified to. Fourthly, Argmax or Top_K operation is further integrated to determine the type of each element in input sequence by selecting the maximal probability. Lastly, the type is recognized, and meanwhile, starting-ending times of disturbances are also located just at the moment when the type is changed. The proposed model is further validated and tested by synthetic signals and practical field signals, respectively. Experimental results demonstrate that the accuracy of type recognition is over 98% for 96 kinds of disturbances including single and combined disturbances with signal-to-noise ration being 20dB. Besides, the stating-ending times are also located with the absolute error less than 6 sampling points when sampling frequency is 256 points per cycle with noisy environment.

In this paper, a priority sorting approach based on simplified model predictive control (MPC) is proposed for modular multilevel converter (MMC). It aims at reducing the computational burden of conventional MPC method while maintaining the system performance, especially under high voltage levels. The proposed approach mainly consists of three parts, i.e., grid-side current control (GCC), circulating current control (CCC), and capacitor voltage balancing control (CVBC). The GCC and CCC are separately designed with simplified MPCs, avoiding the weight factor. Meanwhile, the redundant calculations are eliminated in GCC by considering the desired predicted output voltage of equivalent MMC model. To further minimize the optional combinations of the switching states, the CCC is constructed by utilizing the output of GCC and the arm current. Besides, a novel priority sorting approach is proposed for the CVBC to alleviate the sorting operation. The submodules are divided into three groups according to the detected capacitor voltages. Moreover, the groups are assigned with different priorities based on the arm current, and only one group needs the sorting process. Additionally, a reduced frequency approach is introduced to decrease the power loss in the steady state. The effectiveness of the proposed approach is validated by both simulation and experimental results.

A novel sliding mode observer (SMO) with optimized equal reaching law for speed estimation in sensorless induction motor drives is proposed in this paper to balance the dilemma between the requirement of fast reaching transient and the chattering reduction on sliding mode surface. By considering the reaching way to the sliding mode surface, a widely-used reaching law called equal reaching law (ERL) is introduced into SMO. However, SMO with ERL can not satisfy the rapidity of transient performance and the low chattering level simultaneously. To solve this problem, an optimized equal reaching law (OERL), which is based on the options of an exponential term that adapts to the variations of sliding mode surface and system states, is introduced into SMO. The proposed SMO with OERL is able to decrease the reaching time and suppress the chattering problem simultaneously. Simulation and experimental results both show the validity of the proposed approach.

In order to improve the efficiency of the electric locomotive, the MW-level direct-drive permanent magnet synchronous motor (DD-PMSM) is used to replace the original traction motor and gearbox in the electric locomotive. However, the DD-PMSM parameters will change with the operating conditions during the operation of the electric locomotive, and the absence of the gearbox in direct drives increases the sensitivity of the control performance to uncertainties in the drive system. In order to solve the problems, a novel robust adaptive (RA) current control of MW-level DD-PMSM for traction drives based on internal model control with disturbance observer (IMCDO) is proposed in this article. IMCDO is simply designed as the estimated uncertainty function, and simple control law is derived based on the Lyapunov function to guarantee the system stability and improve the antidisturbance performance. Then, in order to adjust the internal model disturbance observer properly, the observer gain selection criteria are given by using the stability analysis based on the discrete Lyapunov function. The proposed method in the article is verified on a 1.2-MW DD-PMSM, and the comparative experiments demonstrate the validity and effectiveness of the proposed method.

It is well known that loads of inductive power transfer systems are often energy storage devices. The major benefit of an LCC–LCC compensation topology is that without the additional compensation elements and control methods, the load-independent constant current (CC) and constant voltage (CV) outputs, required by energy storage load, can be achieved at different resonant frequencies. However, the misalignment between two coils at the transmitting and receiving sides is inevitable and can lead to the variation of the resonant frequency in CC or CV output. Therefore, a unique design approach to the LCC–LCC compensation topology is proposed in this article. With the presented method, the resonant frequency, at which load-independent output characteristics can be also implemented, is constant and unaffected by the coupling coefficient, whether in CC mode or CV mode. Furthermore, the transmission gain, which is only impacted by the coupling coefficient, can be conveniently modified by the conventional phase shift control in two modes. Finally, a 120-W prototype is designed to verify the validity of the theoretical analysis. The output current and output voltage fluctuate by only 2.4% and 1.0% in the coupling coefficient range from 0.23 to 0.3 and the large load range, according to experimental results.

The stability analysis methods of single-bus dc distribution power systems (DPSs) have been widely studied; however, the multivoltage-level dc DPSs, which have been used in data center, all-electric ship, and so on, its stability analysis methods are rarely reported. The main focus of this article is to study the stability analysis methods, especially the impedance ratio of multivoltage-level dc DPSs. From the point of view of whole system, the two-port small-signal models of multivoltage-level dc DPSs are established and the equivalent loop gains are obtained and used to judge the system stability. On the other hand, the multivoltage-level dc DPSs can be regarded as conventional single-bus dc DPSs at different interfacing locations, and the impedance-ratio type stability criteria are obtained. And it is proved that the two stability analysis methods, namely the system equivalent loop gain and impedance-ratio type stability criterion at different interfacing locations, have the same stability results, which clarifies the stability status by considering different interfacing locations. Finally, the analysis is verified with simulations and hardware-based experiments of a three-converter three-stage dc DPS.