检索结果-内蒙古大学图书馆

Semi-Supervised Probabilistic Learning Network for Soft Sensor Modeling With Partially Labeled Data

IEEE Transactions on Automation Science and Engineering 2025年 22卷 16309-16321页

作者： Shi, Xudong Li, Ronghuan Morales, Humberto Amicarelli, Adriana Huang, Wangya Xiong, Weili Jiangnan University School of Internet of Things Engineering Wuxi 214122 China Jiangnan University Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education) Wuxi 214122 China Instituto de Automática UNSJ-CONICET Ave. San Martín (Oeste) 1109 San Juan J5400ARL Argentina Tongji University Department of Control Science and Engineering Shanghai 201804 China Baoshan Iron & Steel Company Ltd Silicon Steel Business Unit Shanghai 201900 China

Deep probabilistic learning networks have been applied in industrial soft sensors. However, they face significant challenges in latent variable inference, deep learning backend implementation, and labeled data scarcity. The first challenge arises when covariates directly infer the latent variable, potentially leading to inaccuracies. The second stems from discrepancies between the theoretical probabilistic distribution and practical instance-based deep learning backends. The third is commonly encountered in soft sensor applications, where unlabeled data often go unused, reducing accuracy. To address these challenges, this work proposes a novel semi-supervised probabilistic learning network (SS-PLN) for soft sensor modeling with partially labeled data. The first issue is addressed by formulating an optimization problem as the model’s learning objective. This optimization is efficiently solved by investigating the input of inference network through analyzing the optimal solution’s structure. For the second issue, mean and covariance equations are used to represent probabilistic distributions, ensuring effective deep learning backends. The third issue is addressed by integrating supervised and unsupervised probabilistic learning networks to form the SS-PLN model, thereby maximizing the use of unlabeled data and enhancing the soft sensing performance. The feasibility and effectiveness of the proposed SS-PLN are validated through comparisons with recent semi-supervised learning methods, using data from two industrial processes. © 2004-2012 IEEE.

关键词： data-driven deep neuron network semi-supervised learning Soft sensor variational inference

来源：评论

学校读者我要写书评

暂无评论

Revealing the Evolution Mechanism from Spot Segregation to Banded Defect Above the Austenitization Temperature

SSRN

引用

SSRN 2023年

作者： Yao, Cheng Wang, Min Ni, Youjin Gong, Jian Gong, Tao Xing, Lidong Bao, Yanping State Key Laboratory of Advanced Metallurgy University of Science and Technology Beijing Beijing100083 China Technical Support Centre for Prevention and Control of Disastrous Accidents in Metal Smelting University of Science and Technology Beijing Beijing100083 China Qianshun Technology Center Shougang Qian'an Iron and Steel Company Hebei Qian'an064404 China

Spot segregation and banded defect are unsolved technical problems of high-quality steel. In this paper, spot segregation of different sizes and irregular morphology were found in equiaxed crystal zones of slab. The characteristics of speckle type and porosity type spot segregation were compared, where the segregation degree of large-size spot segregation was worse. After heat treatment and hot rolling processes, the concentration changes of carbon, manganese, and iron were obtained by EPMA and SEM-EDS. The results showed that the carbon distribution became more even but the change in manganese concentration was not obvious, and the maximum concentration of carbon was almost unchanged, which indicated that severe segregation cannot be alleviated by the existing processes. Circular and irregular atoms were found in the banded defect. These atoms cooperate with thermodynamic coupling caused by hot rolling and contribute to the composition homogenization and void healing of the banded defect. In addition, the connection between holding time and diffusion efficiency was studied by simulating the uphill diffusion process of carbon and manganese at a holding temperature of 1250 °C. When the diameter of the spot segregation was less than 3000 μm, the carbon could uniformly diffuse before 12000 s, and then if the spot segregation was kept warm for 36000 s, the mean concentration of manganese was still less than 1.0 wt%, which is too high to be considered homogenized. Finally, an evolution model was established to explain the formation process and evolution mechanism from spot segregation to band defect. © 2023, The Authors. All rights reserved.

关键词： Carbon

来源：评论

学校读者我要写书评

暂无评论

A Weighted Voting Ensemble Prediction Model For Adjusting Burden Distribution Matrix Of Blast Furnace

A Weighted Voting Ensemble Prediction Model For Adjusting Bu...

引用

第40届中国控制会议

作者： Jiangxiao Wu Sen Zhang Wendong Xiao School of Automation University of Science and Technology Beijing Key Laboratory of Advanced Control of Iron and Steel Process (Ministry of Education) University of Science and Technology Beijing

Burden distribution is an important method to regulate the operation state of blast furnace. Reasonable burden distribution matrix of blast furnace is the prerequisite for improving resource utilization and ensuring stable blast furnace *** order to achieve the above target, an ensemble prediction model based on weighted voting and a method of calculating weights are proposed in this paper. In this model, the convolution neural network（CNN）-bidirectional gated recurrent unit（BiGRU）based on attention mechanism, logistic regression, k-nearest neighbor（KNN） and support vector machines（SVM） are integrated to predict whether the burden distribution matrix of blast furnace needs to be adjusted. The weight of each individual model is determined by its prediction accuracy and F1-score. CNN converts the extracted feature vectors into time series and inputs them into BiGRU. The attention mechanism is introduced to assign different weights to the hidden state of BiGRU, so as to enhance the long-term influence of important information. Finally, the production data of blast furnace are used to testify the performance of prediction models. In this paper, Multilayer Perceptron（MLP） model, RNN model, BiGRU model, BiGRU model based on attention mechanism, logistic regression model, KNN model, SVM model and the proposed ensemble model are used for comparison experiments, compared with other prediction models, the experimental results illustrated that the weighted voting ensemble prediction model can accurately determine whether the burden distribution matrix of blast furnace needs to be adjusted and have better prediction accuracy and generalization performance.

关键词： Blast furnace Burden distribution matrix Ensemble model Attention mechanism BiGRU

来源：评论

学校读者我要写书评

暂无评论

Two Feature Fusion Network Based on Efficient Deep Residual Shrinkage Spatial-Temporal Graph Convolution for Emotion Recognition from Gaits

Two Feature Fusion Network Based on Efficient Deep Residual ...

引用

第34届中国控制与决策会议

作者： Xiaoqing Liu Sen Zhang Wendong Xiao School of Automation and Electrical Engineering University of Science and Technology Beijing Key Laboratory of Advanced Control of Iron and Steel Process (Ministry of Education) Shunde Graduate School University of Science and Technology Beijing Beijing Engineering Research Center of Industrial Spectrum Imaging

Emotion Recognition plays a key role in human social *** current methods choose facial information for emotion *** facial information is easily limited by distance in large *** information can break this limitation,but the amount of publicly available data is *** paper proposes two feature fusion network based on efficient deep residual shrinkage spatial-temporal graph convolution to achieve automatic emotion recognition from gaits in large *** chebyshev polynomial approximation graph convolution is combined with efficient deep residual shrinkage network to obtain efficient deep residual shrinkage spatial-temporal graph convolution network to extract deep *** sixteen temporal affective features extracted from gaits are used as input of efficient deep residual shrinkage network to extract new affective *** the deep features is spliced with new affective features to divided emotion into happy,angry,sad and *** performed a ten-flod cross-validation experiment on the E-gait real *** highest recognition accuracy is 0.87,with an average of *** results show that the proposed model is effective in dealing with emotion recognition promblem from gait in fewer sample situation.

关键词： emotion recognition from gaits chebyshev polynomial approximation deep residual shrinkage Network efficient channel attention mechanism

来源：评论

学校读者我要写书评

暂无评论

Preparation of Biomass Hydrochar and Application Analysis of Blast Furnace Injection

SSRN

引用

SSRN 2022年

作者： Wang, Guangwei Li, Renguo Dan, Jiayun Yuan, Xiang Liu, Jiawen Xu, Kun Ning, Xiaojun Wang, Chuan State Key Laboratory of Advanced Metallurgy University of Science and Technology Beijing Beijing100083 China Hunan Valin Xiangtan Iron and Steel Co Ltd Xiangtan411101 China Swerim AB LuleåSE-971 25 Sweden Thermal and Flow Engineering Laboratory Åbo Akademi University ÅboFI-20500 Finland

Hydrothermal carbonization technology(HTC) was used to carbonize and improve biomass raw material to get hydrochar. The effects of HTC temperature and holding time on the yield, composition, structure, combustion behavior and safety of hydrochar were studied systematically. The results show that with the increase of HTC temperature and the prolongation of holding time, the yield of hydrochar is gradually reduced, the fixed carbon content of hydrochar increases, the volatile content decreases, and a large number of ash and alkali metals enter the liquid phase and are removed. The analysis of the combustion properties and structure of hydrochar can be seen that the HTC process promotes the occurrence of polymerization reactions, the specific surface area is gradually reduced, the degree of carbon ordering increases, and the combustion curve moves towards the high temperature zone and gradually approaches bituminous coal. Since biomass hydrochar has the characteristics of carbon neutrality, blast furnace injection hydrochar can reduce CO 2 emissions, and every 1 kg/tHM of biomass hydrochar can reduce CO 2 emissions 1.95 kg/tHM. © 2022, The Authors. All rights reserved.

关键词： Biomass

来源：评论

学校读者我要写书评

暂无评论

Revealing the Effect of Travelling-Wave Magnetic Field on Dendrite Growth of High-Strength steel Slab: Industrial Trials and Numerical Simulation

SSRN

引用

SSRN 2022年

作者： Yao, Cheng Wang, Min Ni, Youjin Wang, Dazhi Zhang, Haibo Xing, Lidong Gong, Jian Bao, Yanping State Key Laboratory of Advanced Metallurgy University of Science and Technology Beijing Beijing100083 China Technical Support Center for Prevention and Control of Disastrous Accidents in Metal Smelting University of Science and Technology Beijing Beijing100083 China Silicon Steel Department Shougang Qian'an Iron and Steel Company Hebei Qian'An 064404 China

The dendrite growth behaviours of high-strength steel during slab continuous casting with travelling-wave magnetic field was studied in this paper. The morphology of solidification structure and composition distribution were analysed in detail. The results showed that the columnar crystals could deflect and break when travelling-wave magnetic field worked with low current intensity. With the increase of current intensity, the secondary dendrite arm spacing and solute permeability decreased, and the columnar crystal transformed to equiaxed crystal. Electromagnetic force caused by travelling-wave magnetic field changed the temperature gradient and velocity magnitude, and promoted the breaking and fusing of dendrites. The order of dendrite compactness and composition uniformity from good to bad was columnar-to-equiaxed transition (high current intensity), columnar crystal zone (low current intensity), columnar-to-equiaxed transition (low current intensity) and equiaxed crystal zone (high current intensity). Additionally, combined with the verified numerical simulation results, as well as the boundary layer theory of solidification front and the dendrite breaking-fusing model, the dendrite deflection mechanism and growth process were revealed. © 2022, The Authors. All rights reserved.

关键词： Numerical models

来源：评论

学校读者我要写书评

暂无评论

Promotion Effect of Foam Formation on the Degradation of Polyvinyl Alcohol by Ozone Microbubble

SSRN

引用

SSRN 2023年

作者： Zhang, Liang Li, Pan Wang, Xuelin He, Xin Li, Enchao Lan, Qingquan Liu, Yanan Yin, Daqiang State Key Laboratory of Pollution Control and Resources Reuse College of Environmental Science and Engineering Tongji University Shanghai200092 China Shandong Jianzhu University China Research Institute of Baoshan Iron and Steel Co China Nanjing Tianqi Advanced Oxidation Co. Ltd Nanjing210044 China School of Environmental Science and Engineering Donghua University 2999 North Renmin Road Shanghai201620 China

Polyvinyl alcohol (PVA) widely applied in various industries was difficult to degrade. Ozone (O3) combined with microbubble improving residence time and oxidation efficiency of O3, which provides a possibility of efficient degradation of PVA. In this study, PVA degradation by ozone microbubble (MB/O3) and oxygen microbubble (MB/O2) was investigated. The phenomenon that bubbles evolved into the not easily broken foam was observed for the first time. The PVA removal by MB/O3 and MB/O2 followed a first-order kinetic model, whose removal rate reached 98.58% and 48.68% in the first 20 minutes, respectively. The physicochemical reaction of PVA degradation by MB/O3 and the physical removal reaction by MB/O2 were demonstrated. The microscale changes of bubbles and foam were captured, revealing the observation of self-pressurizing dissolution of MB/O3 and the formation of larger foam by the remaining unbroken microbubbles, which eventually broke. A four-stage theory was proposed to explain PVA degradation by MB/O3, where physical separation, foam film degradation, and bubble coalescence within the foam play crucial roles. This process was demonstrated under different pH and salt concentration conditions, with a 33.2% increase in the COD removal rate when the pH was decreased from 8.69 to 3.30. Mechanism analysis further confirmed the four-stage theory, demonstrating that during the degradation of PVA, MB/O3 underwent a random cleavage process, resulting in small molecule residues containing aldehyde, carboxyl, and ester groups. The trial of MB/O3 was considered as a simple and efficient treatment method for PVA degradation compared with other PVA degradation processes. © 2023, The Authors. All rights reserved.

关键词： Ozone

来源：评论

学校读者我要写书评

暂无评论

Two Feature Fusion Network Based on Efficient Deep Residual Shrinkage Spatial-Temporal Graph Convolution for Emotion Recognition from Gaits

Two Feature Fusion Network Based on Efficient Deep Residual ...

引用

Chinese control and Decision Conference, CCDC

作者： Xiaoqing Liu Sen Zhang Wendong Xiao School of Automation and Electrical Engineering University of Science and Technology Beijing Beijing China Key Laboratory of Advanced Control of Iron and Steel Process (Ministry of Education) Beijing China Shunde Graduate School University of Science and Technology Beijing Foshan China Beijing Engineering Research Center of Industrial Spectrum Imaging Beijing China

ISBN: (纸本)9781665478977

Emotion Recognition plays a key role in human social activities. Most current methods choose facial information for emotion recognition. But facial information is easily limited by distance in large scenes. Gait information can break this limitation, but the amount of publicly available data is small. This paper proposes two feature fusion network based on efficient deep residual shrinkage spatial-temporal graph convolution to achieve automatic emotion recognition from gaits in large scenes. Firstly chebyshev polynomial approximation graph convolution is combined with efficient deep residual shrinkage network to obtain efficient deep residual shrinkage spatial-temporal graph convolution network to extract deep features. Then sixteen temporal affective features extracted from gaits are used as input of efficient deep residual shrinkage network to extract new affective features. Finally the deep features is spliced with new affective features to divided emotion into happy,angry,sad and neutral. We performed a ten-flod cross-validation experiment on the E-gait real datasets. The highest recognition accuracy is 0.87, with an average of 0.85. Experimental results show that the proposed model is effective in dealing with emotion recognition promblem from gait in fewer sample situation.

关键词： Emotion recognition Convolution Chebyshev approximation Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Design and Development of Big Data System for Blast Furnace ironmaking

Design and Development of Big Data System for Blast Furnace ...

引用

Industrial Cyber-Physical Systems (ICPS)

作者： Qifu Chen Jinhua She Min W Zhuofu Zhang Zhenhong Liao School of Future Technology China University of Geosciences Wuhan China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan China Engineering Research Center of Intelligent Technology for Geo-Exploration. Ministry of Education Wuhan China Hunan Valin Lianyuan Iron and Steel Co. Ltd. Loudi China School of Engineering Tokyo University of Technology Tokyo Japan School of Automation. China University of Geosciences Wuhan China

The process of blast furnace ironmaking is com-posed of several sub-processes. The process of blast furnace smelting is complicated and the data of each sub-process are scattered and not concentrated. The traditional working mode of subjective judgment based on manual experience is not conducive to the stable production of the blast furnace. Therefore, this paper develops a big data system for the production process of blast furnace, which is of great significance to realize the digital production and intelligent production of blast furnace. According to the system architecture of blast furnace ironmaking, this paper mainly designs and develops the data acquisition layer, data storage layer, data analysis layer and data application layer of the big data system.

关键词： Software design Data acquisition Smelting Memory Systems architecture Production User interfaces

来源：评论

学校读者我要写书评

暂无评论

Effective Throughput Optimization of SAG Milling Process Based on BPNN and Genetic Algorithm

Effective Throughput Optimization of SAG Milling Process Bas...

引用

Industrial Cyber-Physical Systems (ICPS)

作者： Zhenhong Liao Ce Xu Wen Chen Qifu Chen Feng Wang Jinhua She School of Automation China University of Geosciences Wuhan China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan China Department of Mineral Resources Development and Utilization Changsha Research Institute of Mining and Metallurgy CO. LTD. Chagnsha China Hunan Valin Lianyuan Iron And Steel Co. Ltd. Loudi China School of Engineering Tokyo University of Technology Hachioji Tokyo Japan

Grinding is an energy-consuming process in mineral processing industry. Improving grinding processing capacity per unit power consumption is an effective means to reduce grinding production cost. In this paper, a new index for evaluating the effective processing throughput of SAG milling is proposed. The production process model is established by BP neural network (BPNN). Through combining the process mechanism and production constraints, the genetic algorithm is adopted to optimize the operating parameters of the SAG milling process to maximize the effective throughput, thus improving the grinding efficiency. The experimental results showed that through optimization of effective throughput function proposed in this paper, the SAG mill processing capacity has been increased by 4% and the operating power drawn reduced by 1.12%. It has important guiding significance for the actual production process.

关键词： Industries Power demand Neural networks Milling Throughput Minerals Materials processing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：