Pub Date : 2023-12-21DOI: 10.1088/1361-6501/ad1814
Tianyu Li, Zhigui Ren, Xiaoping Pang, Dingjun Chen, Shusheng Cao
The evaluation of excavator performance relies heavily on digging force, which serves as a crucial indicator. However, the accuracy of performance assessment is hindered by the absence of a suitable method to characterize the dynamic digging capacity of excavators. This study addresses this limitation by proposing an approach to establish a set of solution-limited inequalities for dynamic digging force. The approach incorporates D'Alembert's principle and composite digging, while considering the influence of inertia force. Furthermore, to mitigate the issue of bucket tooth tip trajectory shaking caused by discontinuous posture during excavation, an amount of measurement data from a 20-ton machine is utilized to construct a consistent theoretical digging trajectory. The theoretical trajectory is subjected to numerical verification to determine the dynamic digging force along the trajectory. A comparative analysis is then conducted, contrasting the obtained dynamic digging force with different theoretical digging forces and measured resistances. Additionally, the dynamic digging forces within the selected digging area of the machine are characterized, without accounting for attitude continuity. The findings demonstrate that the dynamic digging force effectively captures the excavator's performance along the trajectory, and it also provides an excellent characterization of the digging force at discrete digging spots.
{"title":"Dynamic digging force modeling and comparative analysis of backhoe hydraulic excavators","authors":"Tianyu Li, Zhigui Ren, Xiaoping Pang, Dingjun Chen, Shusheng Cao","doi":"10.1088/1361-6501/ad1814","DOIUrl":"https://doi.org/10.1088/1361-6501/ad1814","url":null,"abstract":"\u0000 The evaluation of excavator performance relies heavily on digging force, which serves as a crucial indicator. However, the accuracy of performance assessment is hindered by the absence of a suitable method to characterize the dynamic digging capacity of excavators. This study addresses this limitation by proposing an approach to establish a set of solution-limited inequalities for dynamic digging force. The approach incorporates D'Alembert's principle and composite digging, while considering the influence of inertia force. Furthermore, to mitigate the issue of bucket tooth tip trajectory shaking caused by discontinuous posture during excavation, an amount of measurement data from a 20-ton machine is utilized to construct a consistent theoretical digging trajectory. The theoretical trajectory is subjected to numerical verification to determine the dynamic digging force along the trajectory. A comparative analysis is then conducted, contrasting the obtained dynamic digging force with different theoretical digging forces and measured resistances. Additionally, the dynamic digging forces within the selected digging area of the machine are characterized, without accounting for attitude continuity. The findings demonstrate that the dynamic digging force effectively captures the excavator's performance along the trajectory, and it also provides an excellent characterization of the digging force at discrete digging spots.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"58 20","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138952140","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-21DOI: 10.1088/1361-6501/ad1816
Weili Ding, Zhipeng Zhang, Guo Xinya, Liancheng Su, Changchun Hua
In this paper, we propose a two-stage algorithm, named watershed-constrained image segmentation, for exploring complete edge-closed regions from edges. In the first stage, the input image is pre-processed and the image gradient information is obtained using a gradient operator. Anchors are then obtained from the gradient information. Finally, initial edges are obtained by intelligently connecting the anchors. In the second stage, a marker-based watershed algorithm is adopted to obtain marker points from the gradient information obtained in the first stage. A Gaussian filtered image is then used as the input image to obtain a watershed hyper-segmented edge map. Finally, complete edge-closed regions are obtained by combining the initial edges and the hyper-segmented edge map and searching for weak edges. The image segmentation results are then obtained from the edge-closed regions, demonstrating the excellent performance of our proposed algorithm on various images and videos.
{"title":"EWSeg: A Fast Segmentation Algorithm for Images based on Edge Linking and Watershed Constraints","authors":"Weili Ding, Zhipeng Zhang, Guo Xinya, Liancheng Su, Changchun Hua","doi":"10.1088/1361-6501/ad1816","DOIUrl":"https://doi.org/10.1088/1361-6501/ad1816","url":null,"abstract":"\u0000 In this paper, we propose a two-stage algorithm, named watershed-constrained image segmentation, for exploring complete edge-closed regions from edges. In the first stage, the input image is pre-processed and the image gradient information is obtained using a gradient operator. Anchors are then obtained from the gradient information. Finally, initial edges are obtained by intelligently connecting the anchors. In the second stage, a marker-based watershed algorithm is adopted to obtain marker points from the gradient information obtained in the first stage. A Gaussian filtered image is then used as the input image to obtain a watershed hyper-segmented edge map. Finally, complete edge-closed regions are obtained by combining the initial edges and the hyper-segmented edge map and searching for weak edges. The image segmentation results are then obtained from the edge-closed regions, demonstrating the excellent performance of our proposed algorithm on various images and videos.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"38 4","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138949825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-21DOI: 10.1088/1361-6501/ad1803
Enrique Covián Regales, V. Puente, Miguel Casero, Pablo Cienfuegos Suárez
Volume represents a measurand of great interest in civil engineering and construction works. The estimation of this measurand is a problem already solved by surveying engineering, but the quantification of its uncertainty has been overlooked. As a result, the inaccurate estimation of the volume can lead to significant deviations in the execution costs of earthworks. Moreover, it is not possible to comply with the internationally accepted requirements concerning the expression of measures with an indication of its uncertainty, i.e., the guidelines of the BIPM (Bureau International des Poids et Mesures). In this context, this paper presents a methodology for the quantification of uncertainty in the surveying measurement of volumes, which is generally carried out through digital terrain models processed by CAD or specific surveying software. Two methods for volume estimation are presented and the variance-covariance propagation laws are applied to each of them, leading to the computation of volume uncertainty from measures of position coordinates for which uncertainties are known. Then, the developed methods for uncertainty estimation are successfully tested in different scenarios. The conceptual and mathematical developments for the uncertainty quantification in the computation of volumes resulted in closed-form algorithms implemented in MATLAB that can be potentially incorporated into commercial surveying software.
{"title":"Uncertainties estimation in surveying measurands: application to volumes.","authors":"Enrique Covián Regales, V. Puente, Miguel Casero, Pablo Cienfuegos Suárez","doi":"10.1088/1361-6501/ad1803","DOIUrl":"https://doi.org/10.1088/1361-6501/ad1803","url":null,"abstract":"\u0000 Volume represents a measurand of great interest in civil engineering and construction works. The estimation of this measurand is a problem already solved by surveying engineering, but the quantification of its uncertainty has been overlooked. As a result, the inaccurate estimation of the volume can lead to significant deviations in the execution costs of earthworks. Moreover, it is not possible to comply with the internationally accepted requirements concerning the expression of measures with an indication of its uncertainty, i.e., the guidelines of the BIPM (Bureau International des Poids et Mesures). In this context, this paper presents a methodology for the quantification of uncertainty in the surveying measurement of volumes, which is generally carried out through digital terrain models processed by CAD or specific surveying software. Two methods for volume estimation are presented and the variance-covariance propagation laws are applied to each of them, leading to the computation of volume uncertainty from measures of position coordinates for which uncertainties are known. Then, the developed methods for uncertainty estimation are successfully tested in different scenarios. The conceptual and mathematical developments for the uncertainty quantification in the computation of volumes resulted in closed-form algorithms implemented in MATLAB that can be potentially incorporated into commercial surveying software.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"61 15","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138952316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-21DOI: 10.1088/1361-6501/ad1805
Yi Liao, Weiguo Huang, Tianxu Qiu, Juntao Ma, Ziwei Zhang
Monitoring vibration signals from a fault rotatory bearing is a commonly used technique for bearing fault diagnosis. Owing to harsh working conditions, observed signals are generally contaminated by strong background noise, which is a great challenge in extracting fault bearing signal. Sparsity-assisted signal decomposition offers an effective solution by transforming measured signals into sparse coefficients within specified domains, and reconstructing fault signals by multiplying these coefficients and overcomplete dictionaries representing the abovementioned domains. During the process, observed vibration signals tend to be decomposed, and fault components are extracted while noise is diminished. In this paper, a nonseparable and nonconvex log (NSNCL) penalty is proposed as a regularizer for sparse-decomposition model in bearing fault diagnosis. A convexity guarantee to the sparse model is presented, so globally optimal solutions can be calculated. During the process, tunable Q-factor wavelet transform with easily setting parameters, is applie din signifying multi-objective signals with a sparse manner. Numerical examples demonstrate advantages of the proposed method over other competitors.
{"title":"Sparsity-assisted signal decomposition via nonseparable and nonconvex penalty for bearing fault diagnosis","authors":"Yi Liao, Weiguo Huang, Tianxu Qiu, Juntao Ma, Ziwei Zhang","doi":"10.1088/1361-6501/ad1805","DOIUrl":"https://doi.org/10.1088/1361-6501/ad1805","url":null,"abstract":"\u0000 Monitoring vibration signals from a fault rotatory bearing is a commonly used technique for bearing fault diagnosis. Owing to harsh working conditions, observed signals are generally contaminated by strong background noise, which is a great challenge in extracting fault bearing signal. Sparsity-assisted signal decomposition offers an effective solution by transforming measured signals into sparse coefficients within specified domains, and reconstructing fault signals by multiplying these coefficients and overcomplete dictionaries representing the abovementioned domains. During the process, observed vibration signals tend to be decomposed, and fault components are extracted while noise is diminished. In this paper, a nonseparable and nonconvex log (NSNCL) penalty is proposed as a regularizer for sparse-decomposition model in bearing fault diagnosis. A convexity guarantee to the sparse model is presented, so globally optimal solutions can be calculated. During the process, tunable Q-factor wavelet transform with easily setting parameters, is applie din signifying multi-objective signals with a sparse manner. Numerical examples demonstrate advantages of the proposed method over other competitors.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"44 21","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138952513","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-21DOI: 10.1088/1361-6501/ad1815
Xiangyu Liu, Xuehui Gan, An Ping
As an intermediate material for carbon fiber composites, surface flaws inevitably occur during carbon fiber prepreg preparation, which will seriously affect the quality of carbon fiber composite products. The current approaches for identifying flaws on carbon fiber prepreg have the drawbacks of being labor-intensive and inefficient. This research puts forward a novel model for identifying surface flaws on carbon fiber prepregs using an improved single-shot multibox detector (SSD), called CFP-SSD model. A machine vision-based platform for surface flaws identification on carbon fiber prepreg is created. Additionally, the modified-Resnet50 backbone employed in the proposed CFP-SSD model can enhance the effectiveness of network feature extraction. Then, the multi-scale fusion remote context feature extraction module is designed to efficiently fuse the information from the shallow and deep layers. The findings of performance comparison experiments and ablation experiments indicate that the proposed CFP-SSD model achieves 86.63% mean average precision (mAP) and a detection speed of 47 frames per second (FPS), which is sufficient for real-time automatic identification of carbon fiber prepreg surface flaws.
{"title":"Automatic flaw detection of carbon fiber prepreg using a CFP-SSD model during preparation","authors":"Xiangyu Liu, Xuehui Gan, An Ping","doi":"10.1088/1361-6501/ad1815","DOIUrl":"https://doi.org/10.1088/1361-6501/ad1815","url":null,"abstract":"\u0000 As an intermediate material for carbon fiber composites, surface flaws inevitably occur during carbon fiber prepreg preparation, which will seriously affect the quality of carbon fiber composite products. The current approaches for identifying flaws on carbon fiber prepreg have the drawbacks of being labor-intensive and inefficient. This research puts forward a novel model for identifying surface flaws on carbon fiber prepregs using an improved single-shot multibox detector (SSD), called CFP-SSD model. A machine vision-based platform for surface flaws identification on carbon fiber prepreg is created. Additionally, the modified-Resnet50 backbone employed in the proposed CFP-SSD model can enhance the effectiveness of network feature extraction. Then, the multi-scale fusion remote context feature extraction module is designed to efficiently fuse the information from the shallow and deep layers. The findings of performance comparison experiments and ablation experiments indicate that the proposed CFP-SSD model achieves 86.63% mean average precision (mAP) and a detection speed of 47 frames per second (FPS), which is sufficient for real-time automatic identification of carbon fiber prepreg surface flaws.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"58 9","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138951724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The measurement of critical dimensions in the field of integrated circuits has moved from 7nm to 5nm. The existing chromium atomic lithography grating has a pitch period of 4700 l/mm and uniformity of picometre, and the interferometric signal period based on the above grating is as small as 106.4 nm, which brings new problems and challenges to the accurate processing of the signal. This paper investigates the error characteristics of ultra-high precision grating interferometric signals, establishes a Heydemann correction mathematical model for high inscribed line density grating interferometric signals, corrects the grating interferometer signals based on the random sample consensus (RANSAC), and verifies the effectiveness of the algorithm through simulation. By comparing the repeatability and linearity of the original algorithm and the self-traceable grating interferometric displacement measurement data processed by RANSAC, the conclusion that the standard deviation of the self-traceable grating interferometer repeat measurement after RANSAC is 1.60 nm in a 10,000 nm travel is obtained, and the purpose of improving the stability and uniformity of the signal solution with the algorithm of this paper is achieved, which is important for the study of laser interferometer and grating interferometer The results show that the stability and uniformity of the signal solution can be improved by the algorithm of this paper, which is of great significance for the study of the displacement solution of laser and grating interferometers.
{"title":"Study of interferometric signal correction methods in ultra-precision displacement measurement","authors":"Zhangning Xie, Tao Jin, Lihua Lei, Zichao Lin, Yulin Yao, Dongbai Xue, Xiong Dun, Xiao Deng, Xinbin Cheng","doi":"10.1088/1361-6501/ad179b","DOIUrl":"https://doi.org/10.1088/1361-6501/ad179b","url":null,"abstract":"\u0000 The measurement of critical dimensions in the field of integrated circuits has moved from 7nm to 5nm. The existing chromium atomic lithography grating has a pitch period of 4700 l/mm and uniformity of picometre, and the interferometric signal period based on the above grating is as small as 106.4 nm, which brings new problems and challenges to the accurate processing of the signal. This paper investigates the error characteristics of ultra-high precision grating interferometric signals, establishes a Heydemann correction mathematical model for high inscribed line density grating interferometric signals, corrects the grating interferometer signals based on the random sample consensus (RANSAC), and verifies the effectiveness of the algorithm through simulation. By comparing the repeatability and linearity of the original algorithm and the self-traceable grating interferometric displacement measurement data processed by RANSAC, the conclusion that the standard deviation of the self-traceable grating interferometer repeat measurement after RANSAC is 1.60 nm in a 10,000 nm travel is obtained, and the purpose of improving the stability and uniformity of the signal solution with the algorithm of this paper is achieved, which is important for the study of laser interferometer and grating interferometer The results show that the stability and uniformity of the signal solution can be improved by the algorithm of this paper, which is of great significance for the study of the displacement solution of laser and grating interferometers.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"77 12","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138956763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-20DOI: 10.1088/1361-6501/ad179e
Bingnan Hou, Yanchun Wang
WiFi-based indoor positioning technology has gradually become a hot research topic in the field of indoor positioning, but the development of this technology has been facing the challenge of susceptibility to environmental interference. Therefore, in this paper, the kernel function method (KFM) with stronger interference resistance is used for positioning, and the adaptive σ algorithm is proposed for the time-consuming and laborious problem of manual parameter tuning, which incorporates the ideas of cross-validation and iteration. In addition, too many wireless access points (APs) mean higher computational cost and longer positioning time, so it is necessary to choose reasonable APs for positioning. In this paper, we use the random forest (RF) algorithm to assess the importance of APs and filter out a small number of APs with high importance. Considering the obvious differences in the WiFi signals received on different floors, a system framework for positioning by floors based on WiFi fingerprints is proposed. In the offline phase, the fingerprint library is first divided according to floors, and then perform separately AP selection and parameter tuning for each sub-fingerprint library. In the online phase, support vector machine (SVM) is used to discriminate the floors first, and then KFM is used for planar positioning. Experiments are conducted on the public dataset, and the results show that the proposed algorithm has higher positioning accuracy, more robustness, and less time-consuming compared to several common algorithms.
{"title":"Positioning by Floors Based on WiFi Fingerprint","authors":"Bingnan Hou, Yanchun Wang","doi":"10.1088/1361-6501/ad179e","DOIUrl":"https://doi.org/10.1088/1361-6501/ad179e","url":null,"abstract":"WiFi-based indoor positioning technology has gradually become a hot research topic in the field of indoor positioning, but the development of this technology has been facing the challenge of susceptibility to environmental interference. Therefore, in this paper, the kernel function method (KFM) with stronger interference resistance is used for positioning, and the adaptive σ algorithm is proposed for the time-consuming and laborious problem of manual parameter tuning, which incorporates the ideas of cross-validation and iteration. In addition, too many wireless access points (APs) mean higher computational cost and longer positioning time, so it is necessary to choose reasonable APs for positioning. In this paper, we use the random forest (RF) algorithm to assess the importance of APs and filter out a small number of APs with high importance. Considering the obvious differences in the WiFi signals received on different floors, a system framework for positioning by floors based on WiFi fingerprints is proposed. In the offline phase, the fingerprint library is first divided according to floors, and then perform separately AP selection and parameter tuning for each sub-fingerprint library. In the online phase, support vector machine (SVM) is used to discriminate the floors first, and then KFM is used for planar positioning. Experiments are conducted on the public dataset, and the results show that the proposed algorithm has higher positioning accuracy, more robustness, and less time-consuming compared to several common algorithms.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"259 1","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139170111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-20DOI: 10.1088/1361-6501/ad17a0
Xiaoxia Wang, Xiaoxuan Zhang
Domain adaptation (DA) approaches have received significant attention in industrial cross-domain fault diagnosis. However, the scarcity of sufficient labeled fault data, particularly under varying loading conditions and harsh operational environments, can give rise to distinct label spaces between two domains, thereby impeding the application of DA-based diagnosis methods. In this paper, we propose a novel dual-weight domain adversarial network (DWDAN) for diagnosing partial domain faults of feedwater heater system in a large-scale power unit, where the target label space is a subset of the source domain. Firstly, domain adversarial network with an instance-based feature learning strategy is constructed to capture domain-invariant and class-discriminative features hidden in raw process data, thereby enhancing feature extraction and generalization abilities of fault diagnosis. Furthermore, a dual-stage reweighted induction module is designed to quantify the contribution of samples from both class-level and sample-level for selective adaptation. This module can automatically eliminate outlier fault categories in the source domain and facilitates alignment of feature distributions for shared fault categories. Comprehensive experiments conducted on the feedwater heater system of a 600-MW coal-fired generating unit demonstrate the outstanding performance of DWDAN.
在工业跨领域故障诊断中，领域适应（DA）方法受到了极大关注。然而，由于缺乏足够的标注故障数据，特别是在不同的负载条件和恶劣的运行环境下，两个域之间会产生不同的标注空间，从而阻碍了基于 DA 的诊断方法的应用。本文提出了一种新型双权重域对抗网络（DWDAN），用于诊断大型机组给水加热器系统的部分域故障，其中目标标签空间是源域的子集。首先，构建了基于实例特征学习策略的域对抗网络，以捕获隐藏在原始过程数据中的域不变特征和类区分特征，从而增强故障诊断的特征提取和泛化能力。此外，还设计了一个双级加权归纳模块，以量化来自类级和样本级的样本贡献，从而进行选择性适应。该模块可自动消除源域中的异常故障类别，并促进共享故障类别的特征分布对齐。在 600-MW 燃煤发电机组给水加热器系统上进行的综合实验证明了 DWDAN 的出色性能。
{"title":"A dual-weight domain adversarial network for partial domain fault diagnosis of feedwater heater system","authors":"Xiaoxia Wang, Xiaoxuan Zhang","doi":"10.1088/1361-6501/ad17a0","DOIUrl":"https://doi.org/10.1088/1361-6501/ad17a0","url":null,"abstract":"\u0000 Domain adaptation (DA) approaches have received significant attention in industrial cross-domain fault diagnosis. However, the scarcity of sufficient labeled fault data, particularly under varying loading conditions and harsh operational environments, can give rise to distinct label spaces between two domains, thereby impeding the application of DA-based diagnosis methods. In this paper, we propose a novel dual-weight domain adversarial network (DWDAN) for diagnosing partial domain faults of feedwater heater system in a large-scale power unit, where the target label space is a subset of the source domain. Firstly, domain adversarial network with an instance-based feature learning strategy is constructed to capture domain-invariant and class-discriminative features hidden in raw process data, thereby enhancing feature extraction and generalization abilities of fault diagnosis. Furthermore, a dual-stage reweighted induction module is designed to quantify the contribution of samples from both class-level and sample-level for selective adaptation. This module can automatically eliminate outlier fault categories in the source domain and facilitates alignment of feature distributions for shared fault categories. Comprehensive experiments conducted on the feedwater heater system of a 600-MW coal-fired generating unit demonstrate the outstanding performance of DWDAN.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"29 3","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138955200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The existing deep learning models can achieve a high level of fault diagnosis accuracy in the case of a large number of samples. However, in actual production, data is often limited due to the difficulty of data collection and labeling. For small sample fault diagnosis, a fault diagnosis method called Diffusion Model-Overlapping-Patch Vision Transformer (DM-OVT) is proposed in this paper. The method adds Coordinate Attention (CA) to the diffusion model, so that it can consider both channel information and spatial information. In the patch embedding part of Vision Transformer (ViT), features are first extracted using convolutional layers, and then overlapping patch divisions are used to improve the correlation between each patch. To be specific, DM-OVT first uses short-time Fourier transform (STFT) to convert the one-dimensional signals into the time-frequency maps. And then inputs them into the diffusion model (DM) to generate different classes of fault data according to labels. Finally, Overlapping-Patch Vision Transformer (OVT) is used to classify the expanded data. The effectiveness of the proposed method was tested on data sets from laboratory multistage centrifugal fans and Case Western Reserve University, and the highest accuracy was achieved in the comparison experiments.
{"title":"Diffusion Model and Vision Transformer for Intelligent Fault Diagnosis under Small Samples","authors":"Jian Cen, Weiwei Si, Xi Liu, Bichuang Zhao, Chenhua Xu, Shan Liu, Yanli Xin","doi":"10.1088/1361-6501/ad179c","DOIUrl":"https://doi.org/10.1088/1361-6501/ad179c","url":null,"abstract":"\u0000 The existing deep learning models can achieve a high level of fault diagnosis accuracy in the case of a large number of samples. However, in actual production, data is often limited due to the difficulty of data collection and labeling. For small sample fault diagnosis, a fault diagnosis method called Diffusion Model-Overlapping-Patch Vision Transformer (DM-OVT) is proposed in this paper. The method adds Coordinate Attention (CA) to the diffusion model, so that it can consider both channel information and spatial information. In the patch embedding part of Vision Transformer (ViT), features are first extracted using convolutional layers, and then overlapping patch divisions are used to improve the correlation between each patch. To be specific, DM-OVT first uses short-time Fourier transform (STFT) to convert the one-dimensional signals into the time-frequency maps. And then inputs them into the diffusion model (DM) to generate different classes of fault data according to labels. Finally, Overlapping-Patch Vision Transformer (OVT) is used to classify the expanded data. The effectiveness of the proposed method was tested on data sets from laboratory multistage centrifugal fans and Case Western Reserve University, and the highest accuracy was achieved in the comparison experiments.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"115 16","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138953834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-20DOI: 10.1088/1361-6501/ad179f
Yongyan Cao, Wei Yang, Hao Li, Hao Zhang, Minzan Li
In the practical application of farmland, the soil organic matter prediction model es-tablished by the traditional near-infrared spectroscopy is affected by factors such as soil texture, which leads to a serious decline in the accuracy of the model. To im-prove the robustness and prediction accuracy of the model, a prediction model based on near-infrared spectroscopy and image fusion is proposed. A 1D CNN organic matter prediction model (based on near-infrared spectroscopy) was established using eight characteristic wavelengths of extracted soil organic matter (932 nm, 999 nm, 1083 nm, 1191 nm, 1316 nm, 1356 nm, 1583 nm, and 1626 nm) as spectral infor-mation. A 2D CNN organic matter prediction model was established using soil RGB images as information. Based on the idea of model weight fusion, 1D CNN and 2D CNN models are fused. When using small convolutional kernels(3-layer convolu-tional kernel size: 3 * 3, 1 * 1, 1 * 1)and 1D-CNN: 2D-CNN = 6:4, the model has the highest prediction accuracy(R2=0.872). The optimal fusion model was embedded into the inspection system. The final laboratory and field testing results are as fol-lows: under laboratory conditions, the detection accuracy R2 of the 1D CNN predic-tion model, 2D CNN prediction model, and fusion model are 0.838, 0.781, and 0.869, respectively. The RMSE is 3.005, 3.546, and 2.678, respectively. The above experi-mental data indicates that the R2 of the fused model is more accurate compared to the model established with a single information. In the field test, the R2 detection accuracy of 1D CNN prediction model, 2D CNN prediction model and fusion model is 0.809, 0.731 and 0.835, respectively. The root mean square errors are 3.466, 3.828 and 2.973, respectively. The results show that this testing system can meet the needs of soil nutrient testing in farmland and provide guidance for precision agriculture management.
{"title":"Development of a vehicle–mounted soil organic matter detection system based on near–infrared spectroscopy and image information fusion","authors":"Yongyan Cao, Wei Yang, Hao Li, Hao Zhang, Minzan Li","doi":"10.1088/1361-6501/ad179f","DOIUrl":"https://doi.org/10.1088/1361-6501/ad179f","url":null,"abstract":"\u0000 In the practical application of farmland, the soil organic matter prediction model es-tablished by the traditional near-infrared spectroscopy is affected by factors such as soil texture, which leads to a serious decline in the accuracy of the model. To im-prove the robustness and prediction accuracy of the model, a prediction model based on near-infrared spectroscopy and image fusion is proposed. A 1D CNN organic matter prediction model (based on near-infrared spectroscopy) was established using eight characteristic wavelengths of extracted soil organic matter (932 nm, 999 nm, 1083 nm, 1191 nm, 1316 nm, 1356 nm, 1583 nm, and 1626 nm) as spectral infor-mation. A 2D CNN organic matter prediction model was established using soil RGB images as information. Based on the idea of model weight fusion, 1D CNN and 2D CNN models are fused. When using small convolutional kernels(3-layer convolu-tional kernel size: 3 * 3, 1 * 1, 1 * 1)and 1D-CNN: 2D-CNN = 6:4, the model has the highest prediction accuracy(R2=0.872). The optimal fusion model was embedded into the inspection system. The final laboratory and field testing results are as fol-lows: under laboratory conditions, the detection accuracy R2 of the 1D CNN predic-tion model, 2D CNN prediction model, and fusion model are 0.838, 0.781, and 0.869, respectively. The RMSE is 3.005, 3.546, and 2.678, respectively. The above experi-mental data indicates that the R2 of the fused model is more accurate compared to the model established with a single information. In the field test, the R2 detection accuracy of 1D CNN prediction model, 2D CNN prediction model and fusion model is 0.809, 0.731 and 0.835, respectively. The root mean square errors are 3.466, 3.828 and 2.973, respectively. The results show that this testing system can meet the needs of soil nutrient testing in farmland and provide guidance for precision agriculture management.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":"68 7","pages":""},"PeriodicalIF":2.4,"publicationDate":"2023-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138954740","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：