The cascaded solid-state transformer (SST) has garnered significant attention in recent years due to its modular advantages, offering improved scalability, efficiency, and fault-tolerant capabilities. However, the modeling of N-module cascaded SSTs and their impedance characteristics remains insufficiently explored, which could hinder the reliable integration of SSTs. To address this issue, this paper proposes a matrix-based modeling method to characterize the N-module SST and establish its corresponding impedance model. A universal module equivalent block diagram, accounting for arbitrary-order harmonic disturbances, is constructed through an initial matrix aggregation. Building upon this, a second matrix aggregation is performed to develop a system-wide equivalent block diagram, which accommodates any combination of modules, thus enabling standardized representation and modular expansion of the N-module SST. Based on the system equivalent block diagram, a transfer-matrix-based method is used to flexibly compute the SST impedance expression. Using the derived impedance model, the impact of factors such as module differences, operating conditions, and hardware parameters on port impedance is discussed. The influence of these factors on overall system stability is also discussed. Finally, the accuracy and validity of the impedance model, along with the related stability analysis, are verified through a hardware-in-the-loop (HIL) experimental setup.

The increasing integration of converters brings instability challenges to DC power electronic systems (PESs). Instability detection based on data-driven analysis provides a promising noninvasive solution for “black-box” DC PESs, enabling reliable indirectly from DC bus voltage signals without requiring internal system information. However, data-driven instability detection approaches still face two major challenges: limited adaptability to oscillation frequency variations and the scarcity of labeled instability data. To solve these challenges, a contrastive learning with deep embedded clustering framework (CoDeepCL)-based instability detection approach is proposed for DC PESs by this paper. Specifically, CoDeepCL employs temporal-noise augmented contrastive learning with a convolutional neural network with long short-term memory (CNN-LSTM) to capture oscillations’ common characteristic and learn discriminative difference between oscillation and non-oscillation features, enabling oscillation-adaptive function. In addition, intrinsic unstable-voltage feature guided contrastive learning with deep embedded clustering is integrated to achieve more accurate label-free instability detection. As a result, the proposed CoDeepCL can detect instability reliably even under varying oscillation conditions without requiring any labeled data. Experiments verified the proposed approach achieves accurately instability detection across diverse working conditions in DC PES.

Facial depression recognition is one of the current hot topics. Mainstream methods mainly focus on how to design deep models to effectively extract the difference in facial movements between depressed patients and healthy people. However, this difference changes when the stimulus source to which the subjects are exposed changes. This leads to the performance degradation in cross-stimulus situation and limits the practical application of this technology. We hold the opinion that why depressed patients show behavioral characteristics different from healthy people is that they have a specific stable pattern of responding to stimulus. Therefore, we incorporate stimuli into the modeling process for the first time and employ deep networks to learn stable representations between stimulus and response to achieve stable and effective modeling. Specifically, we propose a deep modeling framework to learn the stimulus-response pattern of the subject through the interaction relationship between the stimulus videos and the subject’s facial movements. We constructed a balanced depression dataset of 364 individuals with three different stimulus videos to verify the effectiveness of our method. The results show that our method achieves state-of-the-art and the best generalization performance in depression recognition. This stimulus-response pattern modeling provides a new perspective for recognizing depression.

In DC distribution system, load switching events and disturbances may cause significant overshoot in DC bus voltage under regular control strategies. To further enhance the dynamic performance of the system, a coordinated control strategy of real-time compensation utilizing EtherCAT communication is proposed in this paper. In this strategy, the state variation command is issued from the master controller to the AC-DC converter. Meanwhile, a compensation value is introduced to the DC-DC converter synchronously based on EtherCAT communication which owns the real-time and high-speed features. Theoretically, it can mitigate the transient process of DC bus voltage during master-regulated state variations. The proposed strategy is verified in a cascaded system of Dual Active Bridge (DAB) and a boost-inductorless single-phase bidirectional isolated AC-DC converter. In order to obtain the precise expression of the compensation, the small signal analysis upon both topology and control loop of the two converters is conducted. Finally, a 3.3kW cascaded experimental platform is built to verify the correctness and effectiveness of the proposed strategy.

Deep learning methods have been widely adopted in the field of computer-aided EEG diagnosis, one of the major topics to be investigated is the input format of EEG data. Related researches have reported the application of Fast Fourier Transform (FFT) to topology-preserving multi-spectral images generation on three frequency bands (i.e. theta, alpha and beta) jointed to preserve the spatial information. In our work, we proposed a new approach of using Empirical Mode Decomposition (EMD) instead of FFT algorithm to generate topology-preserving multi-spectral images on intrinsic mode functions (IMFs) jointed and only on the single IMF. Meanwhile, images generated on three frequency bands jointed and the single frequency band were also used for comparison. We then applied two convolutional neural network (CNN) structures to distinguish depression patients and normal subjects from the topology-preserving multi-spectral images. As a result, both convolutional networks obtain better accuracies on three IMFs jointed (about 5% better, ≈ 75% vs. ≈ 70%) than three frequency bands jointed. Analysis based on the single frequency band indicates that alpha band performs the best, with an accuracy of more than 78%, and among all classification results, the best classification accuracy obtained is 80.23% on IMF2. The results are encouraging, despite the limited size of our cohort, the use of EMD and our findings cast a new light on application of deep learning method to EEG-based depression recognition.

The adoption of bipolar multi-port converters to integrate renewable energy/energy storage and bipolar bus in bipolar dc microgrid systems is considered as a promising solution, owing to its reliability and efficiency advantages. This paper presents a bidirectional bipolar multi-port dcdc converter with low voltage stress and high gain, that can achieve single-stage power conversion between the source and bipolar dc bus. High voltage gain and low port current ripple are achieved, which is very suitable as an interfacing converter for renewable energy and energy storage. Furthermore, the low voltage stress and soft switching performance of semiconductor devices can effectively improve the efficiency of the converter. In addition, the bidirectional power flow capability and the simple control method make the proposed structure a popular solution for bipolar dc microgrids. The operation principle, steady-state performance, and design considerations of the converter are discussed in detail. Meanwhile, a ±200V/2kW simulation model is constructed to verify the validity of the theoretical analysis.

Misalignment between the magnetic coupled coils of inductive power transfer (IPT) system is inevitable, which significantly reduces the power transfer capability and system efficiency of IPT system. In this paper, an intermediate coil and dual decoupled receivers are employed to achieve high misalignment tolerance of IPT system. Besides, constant current (CC) output characteristic is obtained. Thanks to the intermediate coil, the inverter current can be limited when the receiver side moves out of the operating region (the mutual inductances are close to zero). With the proposed decoupled method, the design of the magnetic coupler is significantly simplified. In order to demonstrate the validity of the proposed method, a 3.2kW experimental prototype is designed and implemented. Experimental results show that the proposed IPT system can tolerate ±210mm x-axis and -20mm~60mm z-axis misalignment while the fluctuation of the output current is within ±5%. The maximum efficiency can reach 96.7% at full load.

The concept of smart grid has been introduced as a new vision of the conventional power grid to figure out an efficient way of integrating green and renewable energy technologies. In this way, Internet-connected smart grid, also called energy Internet, is also emerging as an innovative approach to ensure the energy from anywhere at any time. The ultimate goal of these developments is to build a sustainable society. However, integrating and coordinating a large number of growing connections can be a challenging issue for the traditional centralized grid system. Consequently, the smart grid is undergoing a transformation to the decentralized topology from its centralized form. On the other hand, blockchain has some excellent features which make it a promising application for the smart grid paradigm. In this article, we aim to provide a comprehensive survey on the application of blockchain in smart grid. As such, we identify the significant security challenges of smart grid scenarios that can be addressed by blockchain. Then, we present a number of blockchain-based recent research works presented in different literature addressing security issues in the area of smart grid. We also summarize several related practical projects, trials, and products that have emerged recently. Finally, we discuss essential research challenges and future directions of applying blockchain to smart grid security issues.

This article proposes a data-physical hybrid-driven air balancing (DPH-AB) method for the ventilation system. First, a data-driven model is proposed to establish the relationship among the airflow, path pressure drop, pressure, and damper angle in the duct system. Specifically, the airflow-path pressure drop relationship is modeled by the denoising autoencoder neural network (DAENN) while the pressure-angle relationship is obtained by the ridge regression method. Then, the physical information of the duct system is considered together with the proposed data-driven model to find the optimal angle for each damper to minimize the total fan power. With the proposed DPH-AB method, the airflow of all terminals can be accurately adjusted to the desired value while the energy consumption of the duct system can reach the lowest value. It is easy to identify which damper needs to be fully open to save the energy. The experiments verify the accuracy of the proposed DPH-AB method and its energy saving potential. The experiments also demonstrate that the DPH-AB method is robust against the noise in the actual duct system.

This paper presents an optimization strategy for the passive thermosiphon beam system to maintain the indoor thermal comfort and reduce the energy consumption. The terminal unit model, indoor built model and energy models of the components in the passive thermosiphon beam (PTB) system are developed respectively. To obtain the optimal operating points, the optimization problem is formulated with respect to the system constrains and components interactions. The genetic algorithm is utilized to find the optimal fresh air supply and chilled water flow rate of the PTB system. The result indicates that the optimization strategy can reduce the energy consumption of the PTB system by 9.3%.

This paper presents a new diagnosis method for open-switch faults in power converters based on Artificial Neural Network (ANN). The ANN inputs comprise both sampled signals and control signals. Only the signals of one switching period are used in the method. The combination of control signals and output signals enables the trained ANN to represent the internal characteristics of converter behaviors, which is crucial for fault diagnosis. Compared with other data-driven methods, the ANN approach is simpler, making it easier to be applied in microcontrollers. Besides, the ANN responds quickly to the fault due to the training with instant signals. Therefore, easy operation and fast diagnosis can be both achieved. Finally, the open-switch fault diagnosis in a two-level three-phase converter is studied for method validation. In this case, an ANN is trained with 9 input elements, 7 output elements, and 10 neurons in the hidden layer. Simulation results are given to demonstrate the good performance of the ANN method.

This paper proposes an energy-saving oriented air balancing method based on the extreme learning machine. First, the relationship between the design airflow and the damper angles of all terminals is modeled by the extreme learning machine. Then, a top-down searching method for the minimal required fan voltage is proposed. The optimal damper angles and the minimal required fan voltage can be determined. Under the minimal fan voltage, the energy consumption of the air balancing system can reach the smallest value for each case. This air balancing method can accurately adjust the airflow in each terminal to the design value. The effectiveness of this method is verified by the experiments.

In industry, Boost converter is widely used in PV systems, microgrids and electric vehicles. In the components of Boost converter, the LC filter plays important roles to ensure the high-performance of Boost converter. To guarantee the economic and reliable operations of Boost converter, the efficiency of Boost converter is a significant aspect to be optimized with respect to the LC filter design. However, the optimization of the efficiency of Boost converter will always lead to the deterioration of the power density, so the trade-off relationships between efficiency and power density need to be considered in the design of Boost converter. In this paper, to consider both the efficiency and power density simultaneously in the design of LC filter, a three-stage design methodology is proposed. Following the design methodology, some design examples are given and verified experimentally.

The symmetrical CLLC-type DAB converter has become more popular as a DC transformer (DCT) in the DC microgrid for its galvanic isolation, high power efficiency and high power density. As a DCT, the CLLC-type DAB converter is expected to regulate the DC voltages on the bus and keep a robust voltage conversion gain (VCG). However, because of the practical values of the inductors and capacitors may fluctuate with varying temperature and operating power, VCG of the CLLC-type DAB converter may deviate from the designed one, negatively affecting power supplies. Therefore, this paper proposes a swarm intelligence aided parameter design approach for the CLLC-type DAB converter to ensure its robust VCG against fluctuating inductors and capacitors. In this design approach, the particle swarm optimization (PSO) algorithm is adopted to facilitate design process with high accuracy and computational speed. Finally, the effectiveness of the proposed design approach has been verified with experiments.

Testing, Adjusting and Balancing (TAB procedure) is an important issue of heating, ventilation, and air conditioning (HVAC) systems. This paper proposes an adaptive distributed consensus control-based air balancing (ADCC-AB) method for the HVAC systems, which aims to achieve air balancing via neighboring communication and parameter self-tuning. Comparing with the traditional air balancing methods, the proposed ADCC-AB method has the following advantages. 1.) This method only need to communicate with neighboring terminals, thus eliminating the necessity of a centralized control unit. 2.) It is mode-free method that requires no system topology and parameters and is therefore easy to apply. 3.) The proposed method can automatically adjust the control parameters to achieve system stability. The simulation results verified the performances of the proposed method.

Liquid desiccant dehumidification system (LDDS) has emerged as an energy-efficient approach for air dehumidification. In this paper, a simple model for the liquid desiccant cooling and dehumidification air conditioning (LDCDAC) system is proposed. The model is built by using artificial neural network (ANN) to describe the cooling, dehumidification and regeneration performance of the LDCDAC system. The system outlet parameters, such as chilled water temperature, air temperature and humidity, can be calculated directly from the inlet parameters. A multilayer neural network is adopted, and the ANN model is trained by the experimental data collected under different operating conditions. The model predictions of the heat and mass transfer rates are compared with the experimental values. The results indicate that the model predicting errors are within ±8%. The proposed model can be used in control and optimization applications of the LDCDAC system.

A fault detection model for air handling units (AHU) is proposed in this study based on machine learning methods. The hyperparaters of the model is tuned by the grid search method. The training accuracy, test accuracy, recall, precision, and F1 score are evaluated for this model. The result demonstrates that the model is robust enough to detect the fault for AHU with test accuracy and F1 score of 99.58% and 99.51%, respectively.

Dual active bridge (DAB) converter has become a promising solution to integrate batteries and renewable energies into DC microgrids (MG), which can reduce the number of power conversions. To improve the stability and robustness of DC microgrid, a particle swarm optimization based active disturbance rejection control (PSO-ADRC) approach is proposed and validated in a single-phase DAB topology. PSO method is employed as an automatic tune mechanism to update the parameters of the ADRC controller in real-time. The simulations show that the control method is robust against parameter variations and external disturbance.

Physical model of ventilation system requiring exact knowledge of all component parameters is usual unavailable in practice. Based on this issue, the objective of this study is to develop an air balancing method to predict the damper position provided the desired airflow rate without requiring a physical model. In the study reported here, the proposed air balancing method consists of a multilayer perceptron model and a damper control method. The multilayer perceptron model is constructed and trained to simulate the non-linear relationship between pressure drop and airflow rate at the damper, and the damper position control method is used to relate pressure drop to operating position of the damper. Experimental tests are carried out to validate the performance of the proposed method. The results show that the proposed method is powerful to balance the ventilation system.

Depression, influencing millions of people, has become a major disease in the past decade. However, the assessment methods of diagnosing depression almost exclusively rely on patient-reported or clinical judgments of symptom severity, which are associated with subjective biases and intensive labor. Some bio-signals such as EEG and eye movements are used for automatic detection but their accuracies are not accurate enough for the real application, further improvements are needed. This research proposes a content based ensemble method (CBEM) to promote the depression detection accuracy, generating data subsets by the content of the experiment, then using the majority vote of subsets to determine the subjects' label. The validation of the method is testified by two different experiments which included free viewing eye tracking and task-state EEG and these two experiments have 36, 40 subjects respectively. In these two experiments CBEM gains accuracies of 82.5% and 92.73% respectively. The results show that CBEM outperform traditional classification methods. Our findings provide an effective solution for promoting the accuracy of depression identification, and give an objective and quantitative evaluation of depression, which in the future could be used for the auxiliary diagnosis of depression.

Recurrent neural networks (RNNs) have found a great variety of application areas. As a special type of RNNs, zeroing neural network (ZNN), or termed Zhang neural network, has been reported to have powerful abilities to address various nonstationary problems. To overcome drawbacks and improve the performance of existing ZNN models, several modified ZNN models are proposed in this paper, which allow nonconvex activation functions and possess accelerated finite-time convergence property. Theoretical analyses suggest that the developed ZNN models are equipped with the global convergence property and the convergence-accelerated models are verified by the estimated upper bounds of convergence time. Finally, comparative and illustrative simulation results, including a verification on a mobile manipulator, are presented to illustrate the effectiveness and superiority of proposed ZNN models to existing models for solving nonstationary Sylvester equations.

Most existing frequency-invariant (FI) beamformer algorithms design array beampattern response across the entire frequency bands. In this paper, the design of array gain response is extended to both selective frequency and spatial region. The algorithm consists of an objective function that has a two-dimensional constraint. One dimension constraint is on frequency range; this is to ensure a selective frequency invariant region is formed. The second dimension constraint is on spatial direction; this is to maintain the array response of the beamformer constant for a small amount of angle centered at the desired direction. Having such constant gain response over a selective spatial region makes the array beamformer less sensitive to the exact position of the source. Advanced optimization method such as second order cone programming (SOCP) is used to solve this complex optimization problem with high efficiency and accuracy. Simulation results are compared with other existing algorithms. It demonstrates that the proposed method is able to achieve a constant gain over the specified sector of angle and at the same time having a lower mean square error on the FI performance over the specified frequency region.