In this paper we discuss the convergence of recent advances in deep neural networks (DNNs) with design of robotic mechanisms, which entails the conceptualization of the design problem as a learning problem from the space of design specifications to a parameterization of the space of mechanisms. We identify three key inter-related problems that are at the forefront of using the versatility of DNNs in solving mechanism design problems. The first problem is that of representation of mechanisms and their design specifications, where the representation challenges arise primarily from the non-Euclidean nature of the data. The second problem is that of developing the mapping from the space of design specifications to the mechanisms where, ideally, we would like to synthesize both type and dimensions of the mechanism for a wide variety of design specifications including path synthesis, motion synthesis, constraints on pivot locations, etc. The third problem is that of designing the neural network architecture for end-to-end training and generation of multiple candidate mechanisms for a given design specification. We also present a brief overview of the state-of-the-art on each of these problems and identify questions of potential interest to the research community.
{"title":"Deep Learning-Driven Design of Robot Mechanisms","authors":"A. Purwar, N. Chakraborty","doi":"10.1115/1.4062542","DOIUrl":"https://doi.org/10.1115/1.4062542","url":null,"abstract":"\u0000 In this paper we discuss the convergence of recent advances in deep neural networks (DNNs) with design of robotic mechanisms, which entails the conceptualization of the design problem as a learning problem from the space of design specifications to a parameterization of the space of mechanisms. We identify three key inter-related problems that are at the forefront of using the versatility of DNNs in solving mechanism design problems. The first problem is that of representation of mechanisms and their design specifications, where the representation challenges arise primarily from the non-Euclidean nature of the data. The second problem is that of developing the mapping from the space of design specifications to the mechanisms where, ideally, we would like to synthesize both type and dimensions of the mechanism for a wide variety of design specifications including path synthesis, motion synthesis, constraints on pivot locations, etc. The third problem is that of designing the neural network architecture for end-to-end training and generation of multiple candidate mechanisms for a given design specification. We also present a brief overview of the state-of-the-art on each of these problems and identify questions of potential interest to the research community.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80491098","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}
In recent years, Augmented Reality (AR) has been successfully applied in various fields to assist in the execution of manual tasks. However, there is still a lack of complete set of criteria for interface design for generating real-time interactive functions and effectively improving the task efficiency through AR. In this study, subjects performed two kinds of trajectory tracking tasks in AR, the simple trajectory and complex trajectory. Their task performance under five different sensory feedbacks, namely, central vision, peripheral vision, auditory sensation, tactile sensation and no feedback, were compared. The results show that in the trajectory tracking task in complex trajectories, the feedback information should not only provide prompts of error states, but also provide suggestions for correcting the actions to the subjects. In addition, compared with visual sensation and auditory sensation, the feedback information of tactile sensation has better adaptation. Furthermore, the subjects tend to rely on the real-time feedback of tactile sensation to complete difficult tasks. It was found that in the manual trajectory tracking task, determining whether the trajectory tracking task is within the acceptable trajectory range will be affected by the postures subjects use for the tasks, and will change the task performance. Therefore, it is suggested that when designing auxiliary functions, the limitations of the postures of the task should be considered. The experimental results and findings obtained in this study can provide a reference for the auxiliary interface design of manual tasks in AR.
{"title":"Assistive Sensory Feedback for Trajectory Tracking in Augmented Reality","authors":"I-Jan Wang, Lifen Yeh, Chih-Hsing Chu, Yan-Ting Huang","doi":"10.1115/1.4062543","DOIUrl":"https://doi.org/10.1115/1.4062543","url":null,"abstract":"\u0000 In recent years, Augmented Reality (AR) has been successfully applied in various fields to assist in the execution of manual tasks. However, there is still a lack of complete set of criteria for interface design for generating real-time interactive functions and effectively improving the task efficiency through AR. In this study, subjects performed two kinds of trajectory tracking tasks in AR, the simple trajectory and complex trajectory. Their task performance under five different sensory feedbacks, namely, central vision, peripheral vision, auditory sensation, tactile sensation and no feedback, were compared. The results show that in the trajectory tracking task in complex trajectories, the feedback information should not only provide prompts of error states, but also provide suggestions for correcting the actions to the subjects. In addition, compared with visual sensation and auditory sensation, the feedback information of tactile sensation has better adaptation. Furthermore, the subjects tend to rely on the real-time feedback of tactile sensation to complete difficult tasks. It was found that in the manual trajectory tracking task, determining whether the trajectory tracking task is within the acceptable trajectory range will be affected by the postures subjects use for the tasks, and will change the task performance. Therefore, it is suggested that when designing auxiliary functions, the limitations of the postures of the task should be considered. The experimental results and findings obtained in this study can provide a reference for the auxiliary interface design of manual tasks in AR.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47766946","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 introduction of the idea of “carbon neutrality” gives the development of low carbon and decarbonization a defined path. Climate change is a significant worldwide concern. To offer a theoretical foundation for the implementation of carbon reduction, this research first analyzes the idea of carbon footprinting, accounting techniques, and supporting technologies. The next section examines carbon emission reduction technologies in terms of lowering emissions and raising carbon sequestration. Digital intelligence technologies like the Internet of Things, big data, and artificial intelligence will be crucial throughout the process of reducing carbon emissions. The implementation pathways for increasing carbon sequestration primarily include ecological and technological carbon sequestration. Nevertheless, proving carbon neutrality requires measuring and monitoring greenhouse gas emissions from several industries, which makes it a challenging undertaking. Intending to increase the effectiveness of carbon footprint measurement, this study created a web-based program for computing and analyzing the whole life-cycle carbon footprint of items. The practical applications and difficulties of digital technologies, such as blockchain, the Internet of Things, and artificial intelligence in achieving a transition to carbon neutrality are also reviewed, and additional encouraging research ideas and recommendations are made to support the development of carbon neutrality.
{"title":"Carbon Neutrality: A Review","authors":"Bin He, Xin Yuan, Shusheng Qian, Bi Li","doi":"10.1115/1.4062545","DOIUrl":"https://doi.org/10.1115/1.4062545","url":null,"abstract":"\u0000 The introduction of the idea of “carbon neutrality” gives the development of low carbon and decarbonization a defined path. Climate change is a significant worldwide concern. To offer a theoretical foundation for the implementation of carbon reduction, this research first analyzes the idea of carbon footprinting, accounting techniques, and supporting technologies. The next section examines carbon emission reduction technologies in terms of lowering emissions and raising carbon sequestration. Digital intelligence technologies like the Internet of Things, big data, and artificial intelligence will be crucial throughout the process of reducing carbon emissions. The implementation pathways for increasing carbon sequestration primarily include ecological and technological carbon sequestration. Nevertheless, proving carbon neutrality requires measuring and monitoring greenhouse gas emissions from several industries, which makes it a challenging undertaking. Intending to increase the effectiveness of carbon footprint measurement, this study created a web-based program for computing and analyzing the whole life-cycle carbon footprint of items. The practical applications and difficulties of digital technologies, such as blockchain, the Internet of Things, and artificial intelligence in achieving a transition to carbon neutrality are also reviewed, and additional encouraging research ideas and recommendations are made to support the development of carbon neutrality.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78989807","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}
Yi Sun, Liping Xie, Chihua Lu, Zhien Liu, Wan Chen, Xiaolong Li
Acoustic sensitivity analysis is an essential technique to determine the direction of structural-acoustic optimization by evaluating the gradient of the objective functions with respect to the design variables. However, acoustic sensitivity analysis with respect to acoustic impedance, which is an important parameter representing the interior absorbent material in automotive acoustics, is lacking in the study. Moreover, acoustic sensitivity analysis implemented with conventional numerical methods is time and effort-consuming in automotive acoustics, due to the large-scale mesh generation. In this work, the impedance sensitivity analysis for automotive acoustics based on the discontinuous isogeometric boundary element method is presented. The regularized boundary integral equation with impedance boundary conditions is established, then the sensitivity is derived by differentiating the boundary integral equation. The efficiency of the proposed method is improved by employing the parallel technique and generalized minimal residual solver. A long duct example with an analytical solution validates the accuracy of the proposed method, and an automotive passenger compartment subjecting to impedance boundary conditions illustrates that the computing time of the proposed method is one order of magnitude less than the conventional method. This work presents an easily implementable and efficient tool to investigate acoustic sensitivity with respect to impedance, showing great potential in the application of automotive acoustics.
{"title":"Impedance Sensitivity Analysis Based on Discontinuous Isogeometric Boundary Element Method in Automotive Acoustics","authors":"Yi Sun, Liping Xie, Chihua Lu, Zhien Liu, Wan Chen, Xiaolong Li","doi":"10.1115/1.4062544","DOIUrl":"https://doi.org/10.1115/1.4062544","url":null,"abstract":"\u0000 Acoustic sensitivity analysis is an essential technique to determine the direction of structural-acoustic optimization by evaluating the gradient of the objective functions with respect to the design variables. However, acoustic sensitivity analysis with respect to acoustic impedance, which is an important parameter representing the interior absorbent material in automotive acoustics, is lacking in the study. Moreover, acoustic sensitivity analysis implemented with conventional numerical methods is time and effort-consuming in automotive acoustics, due to the large-scale mesh generation. In this work, the impedance sensitivity analysis for automotive acoustics based on the discontinuous isogeometric boundary element method is presented. The regularized boundary integral equation with impedance boundary conditions is established, then the sensitivity is derived by differentiating the boundary integral equation. The efficiency of the proposed method is improved by employing the parallel technique and generalized minimal residual solver. A long duct example with an analytical solution validates the accuracy of the proposed method, and an automotive passenger compartment subjecting to impedance boundary conditions illustrates that the computing time of the proposed method is one order of magnitude less than the conventional method. This work presents an easily implementable and efficient tool to investigate acoustic sensitivity with respect to impedance, showing great potential in the application of automotive acoustics.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89086508","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}
Conceptual design is the design phase that deploys product functions and structures based on user requirements and ultimately generates conceptual design solutions. The increasing diversification of products has led to the promotion of customized design that involves deep user participation. As a result, there has been a growing focus on user-centric conceptual design. In this regard, the relationship among users, designers, and design solutions has been subtly changed, which has brought challenges to the traditional designer-oriented design model. To address the complex understanding and decision-making problem caused by deeper user participation, emerging new user-centric product conceptual design models need to be discussed. In the new design model, addressing the changing or growing requirements of users through the design of solutions and leveraging multidisciplinary knowledge to guide the conceptual design process are the critical areas of focus. To further describe this design model, this paper examines the user-centric interconnection among users, designers, design solutions, and multi-domain knowledge. In order to optimize design solutions, the solution resolution process and knowledge mapping based on design deviations are considered effective approaches. In addition, the paper also presents the types of design deviations and the multi-domain knowledge support techniques.
{"title":"Harnessing Multi-Domain Knowledge for User-Centric Product Conceptual Design","authors":"Xin-biao Guo, Zechuan Huang, Ying Liu, Wu Zhao, Zeyuan Yu","doi":"10.1115/1.4062456","DOIUrl":"https://doi.org/10.1115/1.4062456","url":null,"abstract":"\u0000 Conceptual design is the design phase that deploys product functions and structures based on user requirements and ultimately generates conceptual design solutions. The increasing diversification of products has led to the promotion of customized design that involves deep user participation. As a result, there has been a growing focus on user-centric conceptual design. In this regard, the relationship among users, designers, and design solutions has been subtly changed, which has brought challenges to the traditional designer-oriented design model. To address the complex understanding and decision-making problem caused by deeper user participation, emerging new user-centric product conceptual design models need to be discussed. In the new design model, addressing the changing or growing requirements of users through the design of solutions and leveraging multidisciplinary knowledge to guide the conceptual design process are the critical areas of focus. To further describe this design model, this paper examines the user-centric interconnection among users, designers, design solutions, and multi-domain knowledge. In order to optimize design solutions, the solution resolution process and knowledge mapping based on design deviations are considered effective approaches. In addition, the paper also presents the types of design deviations and the multi-domain knowledge support techniques.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72786581","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}
In rotating machines, roller bearings are important and prone to frequent faults. Hence, accurate classification of bearing faults is significant in maintenance of machines. Towards this, a framework using the combination of signal processing, machine learning and deep learning algorithms has been proposed in contrast to traditional approaches. The benefits of each algorithm have been reaped in the proposed framework to overcome challenges met in fault identification. In this, Ensemble Empirical Mode Decomposition is applied on bearing vibration signals to reduce non-stationarity and noise. The 12 Intrinsic Mode Function (IMF) signals of 24k length obtained for 3 bearing conditions at 4 speeds constituted feature space of dimension [36*8*24000]. IMFs that have highest correlation coefficient with raw vibration signals are selected as features [3*8*24000] and intelligent algorithms are applied. Application of Principal Component Analysis on selected IMF feature space resulted in extraction of significant features retaining temporal characteristics along 2 major components [3*2*24000]. Considering the temporal dependence of faults in signals, a Stacked Long Short Term Memory (LSTM) deep network is chosen and trained with extracted features to improve fault classification. The performance of this developed framework has been evaluated for different metrics of stacked LSTM model. The proposed framework also satisfactorily surpassed the performance of stacked LSTM model trained with raw data, capable of auto-feature learning. The comparative results inclusive of models in relevant literature illustrate efficacy of developed combinational framework in handling dynamic vibration data for precise classification of bearing faults.
{"title":"Combinational Framework for Classification of Bearing Faults in Rotating Machines","authors":"Sujit Kumar, D. Ganga","doi":"10.1115/1.4062453","DOIUrl":"https://doi.org/10.1115/1.4062453","url":null,"abstract":"\u0000 In rotating machines, roller bearings are important and prone to frequent faults. Hence, accurate classification of bearing faults is significant in maintenance of machines. Towards this, a framework using the combination of signal processing, machine learning and deep learning algorithms has been proposed in contrast to traditional approaches. The benefits of each algorithm have been reaped in the proposed framework to overcome challenges met in fault identification. In this, Ensemble Empirical Mode Decomposition is applied on bearing vibration signals to reduce non-stationarity and noise. The 12 Intrinsic Mode Function (IMF) signals of 24k length obtained for 3 bearing conditions at 4 speeds constituted feature space of dimension [36*8*24000]. IMFs that have highest correlation coefficient with raw vibration signals are selected as features [3*8*24000] and intelligent algorithms are applied. Application of Principal Component Analysis on selected IMF feature space resulted in extraction of significant features retaining temporal characteristics along 2 major components [3*2*24000]. Considering the temporal dependence of faults in signals, a Stacked Long Short Term Memory (LSTM) deep network is chosen and trained with extracted features to improve fault classification. The performance of this developed framework has been evaluated for different metrics of stacked LSTM model. The proposed framework also satisfactorily surpassed the performance of stacked LSTM model trained with raw data, capable of auto-feature learning. The comparative results inclusive of models in relevant literature illustrate efficacy of developed combinational framework in handling dynamic vibration data for precise classification of bearing faults.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81018729","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}
As technology advances, we are surrounded by more complex products that can be challenging to use and troubleshoot. We often turn to online resources and the help of others to learn how to use a product's features or fix malfunctions. This is a common issue in both everyday life and industry. The key to be able to use a product or fix malfunctions is having access to accurate information and instructions and to gain the necessary skills to perform the tasks correctly. This paper offers an overview of how Artificial Intelligence, Digital Twins, and the Metaverse - currently popular technologies - can enhance the process of acquiring knowledge, know-how, and skills, with a focus on industrial maintenance. However, the concepts discussed may also be applicable to the maintenance of consumer products.
{"title":"Exploring the Intersection of Metaverse, Digital Twins, and Artificial Intelligence in Training and Maintenance","authors":"M. Bordegoni, F. Ferrise","doi":"10.1115/1.4062455","DOIUrl":"https://doi.org/10.1115/1.4062455","url":null,"abstract":"As technology advances, we are surrounded by more complex products that can be challenging to use and troubleshoot. We often turn to online resources and the help of others to learn how to use a product's features or fix malfunctions. This is a common issue in both everyday life and industry. The key to be able to use a product or fix malfunctions is having access to accurate information and instructions and to gain the necessary skills to perform the tasks correctly. This paper offers an overview of how Artificial Intelligence, Digital Twins, and the Metaverse - currently popular technologies - can enhance the process of acquiring knowledge, know-how, and skills, with a focus on industrial maintenance. However, the concepts discussed may also be applicable to the maintenance of consumer products.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82780318","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}
Abhijeet S. Raina, Ronak R. Mohanty, Abhirath Bhuvanesh, Divya Prabha J, Manohar Swaminathan, Vinayak R. Krishnamurthy
We present an experimental investigation of spatial audio feedback using smartphones to support direction localization in pointing tasks for people with visual impairments (PVIs). We do this using a mobile game based on a bow-and-arrow metaphor. Our game provides a combination of spatial and non-spatial (sound beacon) audio to help the user locate the direction of the target. Our experiments with sighted, sighted-blindfolded, and visually impaired users shows that (a) the efficacy of spatial audio is relatively higher for PVIs than for blindfolded sighted users during the initial reaction time for direction localization, (b) the general behavior between PVIs and blind-folded individuals is statistically similar, and (c) the lack of spatial audio significantly reduces the localization performance even in sighted blind-folded users. Based on our findings, we discuss the system and interaction design implications for making future mobile-based spatial interactions accessible to PVIs.
{"title":"Pointing Tasks Using Spatial Audio on Smartphones for People With Vision Impairments","authors":"Abhijeet S. Raina, Ronak R. Mohanty, Abhirath Bhuvanesh, Divya Prabha J, Manohar Swaminathan, Vinayak R. Krishnamurthy","doi":"10.1115/1.4062426","DOIUrl":"https://doi.org/10.1115/1.4062426","url":null,"abstract":"\u0000 We present an experimental investigation of spatial audio feedback using smartphones to support direction localization in pointing tasks for people with visual impairments (PVIs). We do this using a mobile game based on a bow-and-arrow metaphor. Our game provides a combination of spatial and non-spatial (sound beacon) audio to help the user locate the direction of the target. Our experiments with sighted, sighted-blindfolded, and visually impaired users shows that (a) the efficacy of spatial audio is relatively higher for PVIs than for blindfolded sighted users during the initial reaction time for direction localization, (b) the general behavior between PVIs and blind-folded individuals is statistically similar, and (c) the lack of spatial audio significantly reduces the localization performance even in sighted blind-folded users. Based on our findings, we discuss the system and interaction design implications for making future mobile-based spatial interactions accessible to PVIs.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77841950","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}
Nagendra Singh Ranawat, Jatin Prakash, Ankur Miglani, P. K. Kankar
Rags, dusts, foreign particles etc. are primary cause of blockage in centrifugal pump and deteriorates the performance. This study elaborates an experimental and data-driven methodology to identify suction, discharge and simultaneous occurrence of both blockages. The discharge pressure signals are acquired and denoised using CEEMD. The fuzzy recurrence plots obtained from denoised signals are attempted to classify using three pre-trained models: Xception, GoogleNet and Inception. None of these models are trained on such images, thus, features are extracted from different pooling layers which include shallow features too. The features extracted from different layers are fed to four shallow learning classifiers: Quadratic SVM, Weighted KNN, Narrow Neural network, and subspace discriminant classifier. The study finds that subspace discriminant achieves highest accuracy of 97.8% when trained using features from second pooling of Xception model. Furthermore, this proposed methodology is implemented at other blockage condition of the pump. The subspace discriminant analysis outperforms the other selected shallow classifier with an accuracy of 93% for the features extracted from the first pooling layer of the Xception model. Therefore, this study demonstrates an efficient method to identify pump blockage using pre-trained and shallow classifiers.
{"title":"Fuzzy Recurrence Plots for Shallow Learning-Based Blockage Detection in a Centrifugal Pump Using Pre-Trained Image Recognition Models","authors":"Nagendra Singh Ranawat, Jatin Prakash, Ankur Miglani, P. K. Kankar","doi":"10.1115/1.4062425","DOIUrl":"https://doi.org/10.1115/1.4062425","url":null,"abstract":"\u0000 Rags, dusts, foreign particles etc. are primary cause of blockage in centrifugal pump and deteriorates the performance. This study elaborates an experimental and data-driven methodology to identify suction, discharge and simultaneous occurrence of both blockages. The discharge pressure signals are acquired and denoised using CEEMD. The fuzzy recurrence plots obtained from denoised signals are attempted to classify using three pre-trained models: Xception, GoogleNet and Inception. None of these models are trained on such images, thus, features are extracted from different pooling layers which include shallow features too. The features extracted from different layers are fed to four shallow learning classifiers: Quadratic SVM, Weighted KNN, Narrow Neural network, and subspace discriminant classifier. The study finds that subspace discriminant achieves highest accuracy of 97.8% when trained using features from second pooling of Xception model. Furthermore, this proposed methodology is implemented at other blockage condition of the pump. The subspace discriminant analysis outperforms the other selected shallow classifier with an accuracy of 93% for the features extracted from the first pooling layer of the Xception model. Therefore, this study demonstrates an efficient method to identify pump blockage using pre-trained and shallow classifiers.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85884496","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}
Product sustainability is a pressing global issue that requires urgent improvement, and low-carbon design is a crucial approach towards achieving sustainable product development. Digital twin technology, which connects the physical and virtual worlds, has emerged as an effective tool for supporting product design and development. However, obtaining accurate product parameters remains a challenge, and traditional low-carbon product design primarily focuses on design parameters. To address these issues, this paper proposes a method for data collection throughout the product lifecycle, leveraging the Internet of Things. The paper envisions the automatic collection of product lifecycle data to enhance the accuracy of product design. Moreover, traditional low-carbon design often has a limited scope that primarily considers product structure and lifecycle stage for optimization. In contrast, combining digital twin technology with low-carbon design can effectively improve product sustainability. Therefore, this paper proposes a three-layer architecture model of product sustainability digital twin, comprising data layer, mapping layer, and application layer. This model sets the carbon footprint as the iterative optimization goal and facilitates the closed-loop sustainable design of the product. The paper envisions sustainable product design based on digital twins that can address cascading problems and achieve closed-loop sustainable design.
{"title":"Digital Twin-Driven Product Sustainable Design for Low Carbon Footprint","authors":"Bin He, Hangyu Mao","doi":"10.1115/1.4062427","DOIUrl":"https://doi.org/10.1115/1.4062427","url":null,"abstract":"\u0000 Product sustainability is a pressing global issue that requires urgent improvement, and low-carbon design is a crucial approach towards achieving sustainable product development. Digital twin technology, which connects the physical and virtual worlds, has emerged as an effective tool for supporting product design and development. However, obtaining accurate product parameters remains a challenge, and traditional low-carbon product design primarily focuses on design parameters. To address these issues, this paper proposes a method for data collection throughout the product lifecycle, leveraging the Internet of Things. The paper envisions the automatic collection of product lifecycle data to enhance the accuracy of product design. Moreover, traditional low-carbon design often has a limited scope that primarily considers product structure and lifecycle stage for optimization. In contrast, combining digital twin technology with low-carbon design can effectively improve product sustainability. Therefore, this paper proposes a three-layer architecture model of product sustainability digital twin, comprising data layer, mapping layer, and application layer. This model sets the carbon footprint as the iterative optimization goal and facilitates the closed-loop sustainable design of the product. The paper envisions sustainable product design based on digital twins that can address cascading problems and achieve closed-loop sustainable design.","PeriodicalId":54856,"journal":{"name":"Journal of Computing and Information Science in Engineering","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2023-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84520600","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}