期刊：
Journal of Energy Storage,2025年116:115934 ISSN：2352-152X
通讯作者：
Yong Li
作者机构：
College of Electrical and Information Engineering, Hunan University, Changsha, 410082, China;Greater Bay Area Institute for Innovation, Hunan University, Guangzhou, 511300, China;[Yijia Cao] School of Electrical & Information Engineering, Changsha University of Science and Technology, Changsha, 410114, China;[Yixiao Wang] School of Automation, Central South University, Changsha, 410082, China;[Yinglong Zhao; Yong Li] College of Electrical and Information Engineering, Hunan University, Changsha, 410082, China<&wdkj&>Greater Bay Area Institute for Innovation, Hunan University, Guangzhou, 511300, China
通讯机构：
[Yong Li] C;College of Electrical and Information Engineering, Hunan University, Changsha, 410082, China<&wdkj&>Greater Bay Area Institute for Innovation, Hunan University, Guangzhou, 511300, China
摘要：
Lithium-ion batteries are now widely used as energy storage units in electric vehicles. Achieving high accuracy in state of charge (SOC) estimation in the battery management system (BMS) is critical for safe operation of electric vehicles. However, accurate SOC estimation remains a challenging task due to the complex dynamics of batteries and the wide range of ambient temperature. Here we propose a new method called ResNet-GRNN for accurate SOC estimation. Our approach combines a Residual network (ResNet) and a gated recurrent neural network (GRNN). Compared to traditional GRNNs, the proposed method can improve the accuracy and generalization of SOC estimation without altering the original GRNN output. The proposed method is tested on datasets collected from two lithium-ion batteries under dynamic drive cycles at different temperatures. The results show that the mean absolute errors (MAEs) of the proposed method is 80% and 56% lower than those of GRNNs and Deep-GRNNs, respectively. Particularly at low temperatures, the ResNet-GRNNs reduce MAEs by 86% and 79%. Moreover, the proposed method achieves low MAEs of 0.51% and 1.14%, respectively, under untrained varying temperatures. Finally, upon testing in a practical BMS, the proposed method achieved the highest level of accuracy while reducing memory consumption by 70%, demonstrating its superiority in practical applications.
Lithium-ion batteries are now widely used as energy storage units in electric vehicles. Achieving high accuracy in state of charge (SOC) estimation in the battery management system (BMS) is critical for safe operation of electric vehicles. However, accurate SOC estimation remains a challenging task due to the complex dynamics of batteries and the wide range of ambient temperature. Here we propose a new method called ResNet-GRNN for accurate SOC estimation. Our approach combines a Residual network (ResNet) and a gated recurrent neural network (GRNN). Compared to traditional GRNNs, the proposed method can improve the accuracy and generalization of SOC estimation without altering the original GRNN output. The proposed method is tested on datasets collected from two lithium-ion batteries under dynamic drive cycles at different temperatures. The results show that the mean absolute errors (MAEs) of the proposed method is 80% and 56% lower than those of GRNNs and Deep-GRNNs, respectively. Particularly at low temperatures, the ResNet-GRNNs reduce MAEs by 86% and 79%. Moreover, the proposed method achieves low MAEs of 0.51% and 1.14%, respectively, under untrained varying temperatures. Finally, upon testing in a practical BMS, the proposed method achieved the highest level of accuracy while reducing memory consumption by 70%, demonstrating its superiority in practical applications.
摘要：
During the battery cathode materials preparation, the temperature correlation, the external environment disturbance, and the system instability caused by control updating widely exist. All these make it difficult to accurately control the temperature of roller kiln. For this reason, an event-triggered decentralized H infinity control method based on adaptive dynamic programming is proposed. First, the temperature interconnection model is established by describing the relationship between temperatures of the atmosphere outlet and each temperature zone. Then, as for temperature interconnection and disturbance, an auxiliary subsystem is introduced and a cost function including upper bound of interconnection term, temperature state, control input, auxiliary control law and disturbance is designed. Next, the event-trigger mechanism is introduced. The event-triggering condition is designed by considering the temperature interconnection, temperature state, control input and disturbance. It is proved that the temperature event-triggered decentralized H infinity control problem can be converted to solve the Hamilton-Jacobi-Isaacs (HJI) equation problem of a set of auxiliary subsystems and the critic learning method is used to solve the HJI equation. The state of the auxiliary subsystem and the pulse dynamic system are proved to be uniformly ultimately bounded. Finally, the proposed control approach is implemented to roller kiln to prove its validity.
关键词：
Fault location;Hybrid lines;Traveling wave method;Virtual time difference value;SOGWO
摘要：
Fault location accuracy of traveling wave method is constrained to the reliable traveling wave velocity. In this paper, a novel fault traveling waves location method is proposed to eliminate the effects of velocity uncertainty on hybrid lines’ fault location. The initial amplitude ratio coefficient (IARC) K is defined based on propagation characteristic analysis of initial fault traveling wave (IFTW) signals for hybrid lines, and the corresponding identification principle of fault line sections is derived from numerical relationships within different fault points’ K values. Then, taking into account the uncertainties of fault traveling wave (TW) velocity, a fault location method based on the virtual time difference is proposed. By comparing the difference value between the virtual time difference value and the real arrival time difference value of IFTW signals, the minimum difference value is defined as the objective function H. The fault location result is equal to the optimal fault distance value of the objective optimal model, so that the negative effects of velocity uncertainty are reduced. Besides, in order to search for the optimal value, a selected opposition-based grey wolf optimizer (SOGWO) is employed. Various fault scenarios are simulated, and the obtained results validate that the proposed approach is reliable for identifying fault sections and locating fault points in hybrid lines.
Fault location accuracy of traveling wave method is constrained to the reliable traveling wave velocity. In this paper, a novel fault traveling waves location method is proposed to eliminate the effects of velocity uncertainty on hybrid lines’ fault location. The initial amplitude ratio coefficient (IARC) K is defined based on propagation characteristic analysis of initial fault traveling wave (IFTW) signals for hybrid lines, and the corresponding identification principle of fault line sections is derived from numerical relationships within different fault points’ K values. Then, taking into account the uncertainties of fault traveling wave (TW) velocity, a fault location method based on the virtual time difference is proposed. By comparing the difference value between the virtual time difference value and the real arrival time difference value of IFTW signals, the minimum difference value is defined as the objective function H. The fault location result is equal to the optimal fault distance value of the objective optimal model, so that the negative effects of velocity uncertainty are reduced. Besides, in order to search for the optimal value, a selected opposition-based grey wolf optimizer (SOGWO) is employed. Various fault scenarios are simulated, and the obtained results validate that the proposed approach is reliable for identifying fault sections and locating fault points in hybrid lines.
期刊：
Applied Bionics and Biomechanics,2025年2025(1):6125695- ISSN：1176-2322
通讯作者：
DaoDe Zhang
作者机构：
[ShaoSheng Fan] School of Electrical Engineering , Changsha University of Science and Technology , Changsha , Hunan , China , csust.edu.cn;[Yu Yan] State Grid Hunan Ultra High Voltage Transformer Company , Transformer Intelligent Operation and Inspection Laboratory , Changsha , Hunan , China;[ZhiYong Yang; Wang Tian; HaoYang Wang; Xu Liu; DaoDe Zhang] Hubei Key Laboratory of Modern Manufacturing Quantity Engineering , School of Mechanical Engineering , Hubei University of Technology , Wuhan , Hubei , 430068 , China , hbut.edu.cn
通讯机构：
[DaoDe Zhang] H;Hubei Key Laboratory of Modern Manufacturing Quantity Engineering , School of Mechanical Engineering , Hubei University of Technology , Wuhan , Hubei , 430068 , China , hbut.edu.cn
关键词：
α-shape;Monte Carlo method;snake-like robot;volume calculation;workspace
摘要：
The method is applicable for solving the obstacle avoidance workspace of a snake-like robot working on high-voltage transmission cables, based on an improved Monte Carlo method, to address the issues of uneven distribution of scattered points, difficulty in extracting point cloud boundaries, and insufficient accuracy in traditional Monte Carlo methods. The proposed method first generates a seed workspace for the snake-like robot using traditional Monte Carlo method and then envelops the seed workspace with a cube and divides it into several smaller cubes that contain points in the workspace equally. Next, Gaussian distribution probability density function is used to extend and sample the seed workspace of the robot, generating the workspace of the snake-like robot. Finally, the α − shape algorithm is used to extract the point cloud boundaries of the snake-like robot workspace and calculate its volume, accurately determining the workspace. Simulation experiments comparing the reconstructed surface obtained from the α − shape algorithm with the point cloud of the snake-like robot workspace show high accuracy.
摘要：
Time series classification is a significant and complex issue in data mining, it is prevalent across various fields and holds substantial research value. However, enhancing the classification rate of time series data remains a formidable challenge. Traditional time series classification methods often face difficulties related to insufficient feature extraction or excessive model complexity. In this study, we propose a self-optimizing polynomial neural network with a temporal feature enhancement, which is referred to as OPNN-T. Existing classifiers based on polynomial neural networks (PNNs) struggle to achieve high-quality performances when dealing with time series data, primarily due to their inability to extract temporal information effectively. The goal of the proposed classifier is to enhance the nonlinear modeling capability for time series data, thereby improving the classification rate in practical applications. The key features of the proposed OPNN-T include the following: (1) A temporal feature module is employed to capture the dependencies in time series data, providing adaptability and flexibility in handling complex temporal patterns. (2) A polynomial neural network (PNN) is constructed using sub-datasets combined with three types of polynomial neurons, which enhances its nonlinear modeling capabilities across diverse scenarios. (3) A self-optimization mechanism is integrated into iteratively optimized sub-datasets, features, and polynomial types, resulting in significant improvements in the classification rate. The experimental results demonstrate that the proposed method achieves superior performances across multiple standard time series datasets, exhibiting higher classification accuracy and greater robustness than the existing classification models. Our research offers an effective solution for time series classification, and highlights the potential of polynomial neural networks in this field.
关键词：
Electric vehicle;Lithium-ion battery;Internal temperature;State-dependent model;Extended Kalman filter
摘要：
Accurate internal temperature estimation of lithium-ion batteries (LIBs) plays an important role in their safe and economical application. However, the traditional observer estimation method based on linear thermal models cannot accurately describe the nonlinear dynamics of the LIB, and the data-driven estimation approach based on an open loop structure makes it difficult to obtain satisfactory robustness. To bridge this research gap, we construct a state-dependent model to represent the LIB's nonlinear dynamics. The coefficients of this model are implemented by a set of radial basis function neural networks, thus the model considers the actual state of the LIB under different working conditions. Based on the identified state-dependent model offline, an online estimation method of the internal temperature is implemented using the extended Kalman filter (EKF). Furthermore, the robustness is further verified by the extended state observer-EKF. The validation results for different batteries and working conditions show that the root mean square error (RMSE) does not exceed 0.30 °C in the presence of wrong initial values, and it does not exceed 0.28 °C in the presence of bias noise.
Accurate internal temperature estimation of lithium-ion batteries (LIBs) plays an important role in their safe and economical application. However, the traditional observer estimation method based on linear thermal models cannot accurately describe the nonlinear dynamics of the LIB, and the data-driven estimation approach based on an open loop structure makes it difficult to obtain satisfactory robustness. To bridge this research gap, we construct a state-dependent model to represent the LIB's nonlinear dynamics. The coefficients of this model are implemented by a set of radial basis function neural networks, thus the model considers the actual state of the LIB under different working conditions. Based on the identified state-dependent model offline, an online estimation method of the internal temperature is implemented using the extended Kalman filter (EKF). Furthermore, the robustness is further verified by the extended state observer-EKF. The validation results for different batteries and working conditions show that the root mean square error (RMSE) does not exceed 0.30 °C in the presence of wrong initial values, and it does not exceed 0.28 °C in the presence of bias noise.
通讯机构：
[Zhao, YR ] H;Huazhong Univ Sci & Technol, State Key Lab Adv Electromagnet Engn & Technol, Wuhan 430074, Peoples R China.
关键词：
Uncertainty;Games;Programming;Indexes;Stochastic processes;Resistance heating;Load modeling;Energy hub;stochastic distributionally robust chance-constrained optimization;Wasserstein metric;Stackelberg game;scenarios and moment information
摘要：
As the increasing complexity and uncertainty of integrated energy systems, the cooperative interaction under uncertainty between the energy hub (EH) and multi-energy users has become crucial for the efficient operation of energy systems. This paper proposes a unified collaborative and interactive framework to optimize operations between the EH and multi-energy users under uncertainty. Firstly, in the horizontal dimension, the stochastic distributionally robust chance-constrained (S-DRCC) optimization method with the Wasserstein metric is proposed to hedge against the uncertainty of EH considering the multiple scenarios and moment information of uncertain variables. Then in the vertical dimension, a bilevel Stackelberg game model is developed for EH and users to facilitate the collaborative interaction for multi-agents. Furthermore, several efficient methods are leveraged to reformulate the originally intractable framework into a tractable mixed-integer linear programming (MILP) model that can be directly solved using commercial solvers. Finally, the effectiveness of the proposed model and method are demonstrated by numerical case studies and corresponding simulation results.
摘要：
This paper presents a method of rotor position estimation for switched reluctance motors suitable for saturation. The effects of saturation as well as voltage changes are taken into account at the same time. It is based on the inductance in the unsaturated region. When the phase inductance is equal to the threshold, it is defined as a characteristic point. Meanwhile the characteristic pulse signal is triggered. Different inductance intersection thresholds are determined when the phase current and bus voltage change. The rotor position is estimated by interval speed. Compared with the traditional inductance method, the position estimation error is smaller. Finally, the correctness and effectiveness of the proposed method are verified by simulation and experiments.
作者机构：
[Wang, Feipeng; Du, Guoqiang; Pan, Lei; He, Yushuang; Li, Jian] Chongqing Univ, Sch Elect Engn, State Key Lab Power Transmiss Equipment Technol, Chongqing, Peoples R China.;[Yang, Hongming; He, Yushuang] Changsha Univ Sci & Technol, Sch Elect & Informat Engn, Changsha, Peoples R China.;[Zhang, Xiao] Naval Univ Engn, Natl Key Lab Sci & Technol Vessel Integrated Power, Natl Key Lab Vessel Integrated Power Syst Technol, Wuhan, Peoples R China.;[Zhang, Zhicheng] Xian Jiaotong Univ XJTU, Sch Chem, Xian Key Lab Sustainable Energy Mat Chem, Xian, Peoples R China.;[Wang, Kaizheng] Kunming Univ Sci & Technol, Fac Elect Engn, Kunming, Peoples R China.
通讯机构：
[Wang, FP ] C;Chongqing Univ, Sch Elect Engn, State Key Lab Power Transmiss Equipment Technol, Chongqing, Peoples R China.
摘要：
Metallised film capacitors (MFCs) are renowned for their unique self-healing (SH) properties, which bestow them with exceptional reliability and stability in the face of intense electric fields, high voltages, and pulse power applications. Nonetheless, the exploration of SH characteristics concerning single-layer dielectric film remains insufficient for advancing MFC reliability evaluation. To establish the theoretical correlation of SH characteristics from the device to the film in the MFCs, this work developed a simulation model to analyse the SH dynamic behaviour in the MFCs. The effects of coupling capacitors, arc resistance and insulation resistance on the macroscopic characteristics (voltage drop and pulse current) are focused during the SH process in MFCs. The results indicate that SH is primarily associated with the voltage drop duration rather than the sampling current. Consequently, the SH process in MFC is characterised as an abrupt decrease in voltage to its minimum value. This refinement enhances the SH energy dissipation model of MFC. The quantified relationship between the macroscopic characteristics and microstructure evolution (polypropylene decomposition and aluminium electrode vaporisation) is established in MFCs under diverse SH energy levels. As SH energy and duration increase, the proportion of energy attributed to polypropylene decomposition increases, resulting in multi-layer ablation and adhesion within the metallised film and a pronounced deterioration in MFC electrical performance. The examination of macro–micro perspectives sheds new light on the intricate mechanisms governing the SH behaviour in MFCs, offering valuable insights for the advancement of their design, reliability evaluation, and performance optimisation in diverse electrical applications.
摘要：
Increasing multi-energy coordination in the ship necessitates advanced operation strategies to achieve greenhouse gas reduction and energy efficiency improvement in the maritime industry. However, previous research always overlooks onboard heterogeneous energy carriers and ship power distribution networks (SPDN), as well as underwater radiated noise (URN) generated by ship propellers. This will pose a huge threat to the operational safety of the multi-energy ship microgrids (MESMs) and further harm normal marine life. Hence, this paper formulates a coordinated model for a MESM with combined power, thermal, hydrogen, and freshwater flows. First, the joint energy management and voyage scheduling are modeled for the MESM, considering SPDN constraints and URN limits. Then, a copula-based two-stage operation structure with stochastic programming (SP) and rolling horizon (RH) methods is designed to tackle diverse uncertainties from onboard multi-energy loads and renewable energy. Finally, a progressive hedging (PH) algorithm is developed to support distributed calculation and accelerate the solution. Numerical case studies based on a real voyage in the Nordic countries are used to validate the effectiveness and superiority of the proposed model and method.
摘要：
Process optimization is a highly successful method for achieving optimal efficiency in industrial production. The conventional optimization approach presupposes that the operational parameters should align with the optimization settings. However, it fails to consider that, influenced by the stochastic performance of the control loops, the operating parameters may deviate from the optimal operating settings. Consequently, this results in the violation of constraints in the optimization results and affects production safety. Therefore, this paper proposes an uncertainty optimization method that considers the stochastic performance of control loops to accurately determine the optimal operational performance that can be practically achieved in industrial production. Firstly, a multi-optimization variational mode decomposition strategy is developed to precisely extract the smooth random and trend terms of the control loop output data. Secondly, the random grouping smooths out the random terms and accurately characterizes the uncertainty associated with these terms. Subsequently, a moment uncertainty set with mild mean-zero net condition is then defined to construct an improved distribution robust optimization model considering the stochastic performance of control loops. Finally, the validation of the proposed optimization method in the actual hydrocracking process shows that the optimization error of the proposed method is reduced by more than 10%, and the constraint violation rate is reduced by 14%, which fully proves the effectiveness and applicability of the method.
摘要：
This study aims to investigate SF 6 decomposition gases, specifically H 2 S, SO 2 , SOF 2 , and SO 2 F 2 , as a means of diagnosing faults in GIS. Using first-principles density functional theory (DFT), simulations were conducted on Zr- and In-modified WTe 2 from several perspectives, including adsorption energy, charge transfer, adsorption distance, density of states, differential charge density, and desorption times at different temperatures. The results demonstrate that the modification of WTe 2 with Zr and In atoms is spontaneous. The pristine WTe 2 exhibits weak physisorption towards all four SF 6 decomposition gases. In contrast, the Zr-WTe 2 system shows strong chemisorption for all four decomposition gases, while the In-WTe 2 system exhibits robust chemisorption specifically for SO 2 , SOF 2 , and SO 2 F 2 , with adsorption energies of −1.169 eV, −1.371 eV, and −1.255 eV, respectively. The Zr-WTe 2 system also demonstrates prolonged desorption times for H 2 S, SO 2 , SOF 2 , and SO 2 F 2 , suggesting its potential as a scavenger for SF 6 decomposition gases. Notably, the In-WTe 2 system exhibits a rapid desorption time of only 0.892 s for SOF 2 at room temperature, indicating its potential for detecting H 2 S gas under ambient conditions.
通讯机构：
[Huang, JC; Peng, ZY ] C;Changsha Univ Sci & Technol, Sch Energy & Power Engn, Key Lab Efficient & Clean Energy Utilizat, Changsha 410111, Peoples R China.
摘要：
Despite the advancements in film fabrication techniques for emerging perovskite solar cells, achieving a high-quality film by solution processing, while maintaining considerable performance remains a significant challenge. To tackle the issue of inferior CsPbI2Br perovskite films deposited via solution-based methods, a novel thermal conduction heating approach was devised and implemented, significantly enhancing film uniformity. Crucially, aliphatic amine acetates (3A) were introduced into the precursor solution to regulate the crystallization process and therefore to mitigate defects. Systematic investigation into the impact of 3A molecules featuring varying alkyl chain lengths on defect passivation revealed that the molecular dipole moment of these additives contributed to both defect mitigation and grain size refinement. Notably, the integration of alkyl chains significantly bolstered the hydrophobic properties of the perovskite film. Consequently, an impressive efficiency of 13.50% for HTM-free carbon-based CsPbI2Br perovskite solar cells was achieved, and the device exhibited robust stability retaining 92.4% of its initial efficiency at room temperature after being stored in dry air for 5400 h. This research offers profound insights into defect passivation mechanisms and perovskite crystallization dynamics, paving the way for further advancements in the field of perovskite solar cell technology.
摘要：
To address the issues of extensive computation and laborious weight coefficient selection in model predictive torque control (MPTC) for permanent magnet synchronous motors (PMSM), an optimized model predictive flux control (MPFC) is proposed based on three-phase voltage vector duty ratio modulation. First, we established a prediction model of PMSM based on stator flux linkage, constructed a prediction error cost function of stator flux linkage without weight coefficient. Meanwhile, the duty ratio of neighboring voltage vectors in each sector and the stator flux reference value were computed using the stator flux deadbeat control concept, and the stator flux reference value is computed. Finally, we proposed an optimal stator flux predictive control method based on three-phase voltage vector duty ratio modulation. Its effectiveness and reliability were verified by simulation and experiment. This method had excellent steady-state performance in the voltage vector output range, requires fewer predictions than the classic model predictive torque control method, and greatly reduces both the stator flux and the electromagnetic torque ripple.
通讯机构：
[Li, XP ] U;Univ Elect Sci & Technol China, Sch Math Sci, Chengdu 611731, Peoples R China.
关键词：
Phase measurement;Estimation;Termination of employment;Wrapping;Synthetic aperture radar;Resists;Redundancy;Noise;Geoscience and remote sensing;Simulation;Digital elevation model (DEM) reconstruction;phase unwrapping (PU);robust Chinese remainder theorem (CRT);synthetic aperture radar (SAR) interferometry
摘要：
Multichannel interferometry synthetic aperture radar (InSAR) systems enable the reconstruction of terrain height profiles by integrating multiple interferograms obtained from multifrequency or multibaseline configurations. In this article, we present a robust redundant remainder number system (RRNS) specifically designed for the high-precision determination of terrain height profiles. The proposed method consists of two main steps: First, erroneous remainders are clustered, and the common remainder is estimated optimally. Next, the integral portion is determined using the RRNS. Furthermore, we derive a robust estimation condition that generalizes existing results. To apply the proposed method to real data, we have extended the robust RRNS technique to accommodate practical data conditions. The main improvements are as follows: first, the phase unwrapping (PU) model for real numbers is adapted to the integral case by selecting suitable prime numbers. Second, the process of selecting the optimal reconstruction result is enhanced by incorporating redundancy congruence, which involves using two equations for each case. Simulation results demonstrate that the extended method outperforms both the least-squares (LS) method and the minimum cost flow (MCF) method.
关键词：
Fault detection and estimation;Generative adversarial networks;Fault elimination mapping;Chemical processes
摘要：
Thanks to the strong generalization and generation capabilities, Generative Adversarial Networks (GANs) have been developed for fault diagnosis schemes in recent years. Most of them are constructed as data enhancers for generating few-shot faulty samples and serve for the fault classification task. However, classifier-based fault diagnosis frameworks have difficulty in recognizing new fault classes and estimating fault magnitudes. In order to exploit the potential of GAN for fault detection and estimation, this paper proposes a Fault-estimable AutoEncoder-GAN (FAE-GAN), which uses a new framework to learn the fault elimination mapping. In FAE-GAN, faulty data are first expanded by sampling from non-normal data domains. Then, pairs of normal and faulty samples are fed into the FAE-GAN for mapping training, aiming to learn the conditional distribution from faulty to normal. In this way, the fault contained in the process data can be eliminated adaptively, thereby achieving the fault estimation purpose. Finally, a numerical example and a chemical simulation verify the proposed approach.
Thanks to the strong generalization and generation capabilities, Generative Adversarial Networks (GANs) have been developed for fault diagnosis schemes in recent years. Most of them are constructed as data enhancers for generating few-shot faulty samples and serve for the fault classification task. However, classifier-based fault diagnosis frameworks have difficulty in recognizing new fault classes and estimating fault magnitudes. In order to exploit the potential of GAN for fault detection and estimation, this paper proposes a Fault-estimable AutoEncoder-GAN (FAE-GAN), which uses a new framework to learn the fault elimination mapping. In FAE-GAN, faulty data are first expanded by sampling from non-normal data domains. Then, pairs of normal and faulty samples are fed into the FAE-GAN for mapping training, aiming to learn the conditional distribution from faulty to normal. In this way, the fault contained in the process data can be eliminated adaptively, thereby achieving the fault estimation purpose. Finally, a numerical example and a chemical simulation verify the proposed approach.
摘要：
Automatic ultrasound image segmentation improves the efficiency of clinical diagnosis and decreases the workload of doctors. Many ultrasound image segmentation methods only focus on capturing local details and global dependencies, whereas ignoring large-scale context information. However, it is essential to extract large-scale context features for large targets in images. To enhance the capability of feature extraction of the model for targets with various sizes and improve segmentation performance, we propose an effective multilevel feature extraction network (SLG-Net) which can extract features from local small details, large-scale context to global dependencies. The SLG-Net is parallel dual-encoder architecture which consists of a CNN encoder and a transformer encoder. Specifically, the CNN encoder improves the representation and interaction of fine feature and large-scale context feature for targets of different sizes by large-small kernel attention (LSKA) modules. The LSKA module firstly extracts features by parallel small kernel module and large-scale feature selection (LSFS) module. The extracted features from above modules are added for further information interaction through a following multi-scale feature interaction module. To fully leverage the feature extraction capability of large kernel convolutions and decrease the number of parameters, we design the large kernel decomposition module (LKDM) to extract large-scale context features in LSFS module. The transformer encoder is used to capture global features for compensating the limitations of CNN encoder. To merge multilevel features, a multi-scale feature fusion module is introduced after the dual-encoder. In addition, at the skip connection, a multi-scale attention module is integrated to retain significant shallow features for subsequent fusion of deep and shallow features. Experiments on three public ultrasound datasets indicate that the proposed network accomplishes the prominent performance for ultrasound image segmentation. It shows the potential of our study to promote intelligence in clinical medicine.
期刊：
IET CONTROL THEORY AND APPLICATIONS,2025年19(1):e12785- ISSN：1751-8644
通讯作者：
Chen, N
作者机构：
[Chen, Ning; Chen, N; Chen, Jiayao; Gui, Weihua; Luo, Biao; Luo, Zeng] Cent South Univ, Sch Automat, Changsha 410083, Peoples R China.;[Chen, Ning; Chen, N] Cent South Univ, Natl Engn Res Ctr Adv Energy Storage Mat, Changsha 410083, Peoples R China.;[Li, Binyan] Changsha Univ Sci & Technol, Sch Elect & Informat Engn, Changsha, Peoples R China.
通讯机构：
[Chen, N ] C;Cent South Univ, Sch Automat, Changsha 410083, Peoples R China.;Cent South Univ, Natl Engn Res Ctr Adv Energy Storage Mat, Changsha 410083, Peoples R China.
关键词：
adaptive control;boilers;chemical industry;control non-linearities;distributed control;H∞$H_{\infty }$ control
摘要：
The roller kiln temperature model is constructed and heat transfer delay is identified. The effect of time delay on the system is estimated by adding L‐K function into the cost function. A distributed H∞$H_{\infty }$ control method is designed to achieve approximated optimal control of roller kiln temperature. The control accuracy of the model is verified based on roller kiln temperature data. Abstract The roller kiln with multi‐temperature zones used for cathode material sintering is an interconnected system with time delay in energy transfer and precise control of the temperature for the preparation of cathode materials for lithium‐ion batteries. However, the interconnection between the temperature zones, the time delay of the temperature state, and the disturbance of the external environment make it difficult to control the sintering process. For this reason, this paper develops a distributed H∞$H_{\infty }$ control of the temperature of roller kiln based on adaptive dynamic programming (ADP). Firstly, the heat transfer mechanism of sintering process is discussed; the law of energy conservation provides the physical basis for the sintering process on which the autocorrelation function method identifies the time delay. Then, the cost function is constructed by combining the Lyapunov–Krasovskii function containing the time delay, the temperature state, the heating power of the silicon carbon rod and the perturbation. Subsequently, Hamilton–Jacobi–Isaac equation with the optimal cost function and the optimal distributed control strategy are approximated by neural network of ADP. Finally, the stability of the closed‐loop system is proved by Lyapunov functional analysis, and the effectiveness of the proposed method is verified by the simulation results of roller kiln temperature control.
摘要：
Multi-energy rural microgrids (MERMs) hold both economic potential and multi-energy coordination ability, emerging as a promising energy management paradigm in rural areas. In this paper, an energy scheduling method is investigated for a MERM with renewable energy and biomass resources, aiming to satisfy the rural electrical, thermal, natural gas, and irrigation demands economically. Mathematically, biomass flows are formulated by adopting a differential dynamics model of anaerobic biomass fermentation. The irrigation system is accurately formulated by fully taking into account meteorological information such as ambient temperature and precipitation. To handle the uncertainties in precipitation, reservoir inflows, renewable power generation as well as electrical and thermal load demands, a two-stage stochastic optimization method is employed, and the proposed model is then reformulated into a stochastic mixed integer quadratic programming (SMIQP) problem. To mitigate the computational burden arising from integer variables and enhance the solution efficiency, a scenario decomposition algorithm, progressive hedging (PH), is used to decompose the SMIQP into scenario-wise subproblems, which are then solved in parallel. Finally, the simulation results demonstrate the effectiveness of the proposed MERM scheduling method and the efficiency of the PH algorithm.
