检索结果-内蒙古大学图书馆

Simulation and Experimental Research of Thermal Power Plant Pipeline Displacement via deep learning algorithms

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2023年 72卷 1页

作者： Dai, Minglu Jian, Yanchen Liu, Qi Gao, Ben Zhao, Xuhao Zhang, Jianyong Zhou, Bin Southeast Univ Sch Energy & Environm Nanjing 210018 Peoples R China Teesside Univ Sch Comp Engn & Digital Technol Middlesbrough TS1 3BA England

The displacement of pipelines in thermal power plants is critical to ensure the safe operation of the thermal power unit. A 3-D online displacement monitoring network, the deep displacement network (DDN), has been designed to obtain real-time displacement data for the entire pipeline. The DDN utilizes a displacement network comprising a 3-D CNN-attention-gate recurrent unit (GRU)-3-D transposed CNN structure and a dataset established by CAESAR II to establish a displacement prediction model for four key piping systems. In addition, to address the limitations of traditional mechanical pointers, which can only measure 2-D displacement and cannot measure displacement online, a real-time stereo displacement measurement device (SDMD) has been developed based on the theory of stereoscopic vision and artificial neural networks (ANNs). The reliability of the DDN and SDMD has been validated through simulations, experiments, and field applications in power plants. Field application results have demonstrated that the absolute error of the SDMD is less than 2 mm compared with traditional mechanical pointers, and the relative error between the predicted value by DDN and the measured value of SDMD is less than 9%. Furthermore, the attention layer can improve efficiency in reducing the relative error, while the 3-D transposed CNN structure can minimize the maximum relative error. Finally, the DDN can solve the issue of requiring a large number of measurement points for the SDMD and significantly reduce the cost of measurement equipment.

关键词： Attention structure deep learning algorithms displacement stereo displacement measurement devices (SDMDs) thermal power plant pipeline

来源：评论

学校读者我要写书评

暂无评论

The Influence of deep learning algorithms Factors in Software Fault Prediction

引用

IEEE ACCESS 2020年 8卷 63945-63960页

作者： Al Qasem, Osama Akour, Mohammed Alenezi, Mamdouh Yarmouk Univ Informat Syst Dept Irbid 21163 Jordan Al Yamamah Univ Comp Engn Dept Riyadh 13541 Saudi Arabia Prince Sultan Univ Comp Sci Dept Riyadh 11586 Saudi Arabia

The discovery of software faults at early stages plays an important role in improving software quality;reduce the costs, time, and effort that should be spent on software development. Machine learning (ML) have been widely used in the software faults prediction (SFP), ML algorithms provide varying results in terms of predicting software fault. deep learning achieves remarkable performance in various areas such as computer vision, natural language processing, speech recognition, and other fields. In this study, two deep learning algorithms are studied, Multi-layer perceptron & x2019;s (MLPs) and Convolutional Neural Network (CNN) to address the factors that might have an influence on the performance of both algorithms. The experiment results show how modifying parameters is directly affecting the resulting improvement, these parameters are manipulated until the optimal number for each of them is reached. Moreover, the experiments show that the effect of modifying parameters had an important role in prediction performance, which reached a high rate in comparison with the traditional ML algorithm. To validate our assumptions, the experiments are conducted on four common NASA datasets. The result shows how the addressed factors might increase or decrease the fault detection rate measurement. The improvement rate was as follows up to 43.5 & x0025;for PC1, 8 & x0025;for KC1, 18 & x0025;for KC2 and 76.5 & x0025;for CM1.

关键词： Software deep learning Prediction algorithms Software algorithms Biological neural networks Predictive models Computer architecture deep learning algorithms software fault prediction classification hyper parameters

来源：评论

学校读者我要写书评

暂无评论

Adaptive investigation of the optical properties of polymer fibers from mixing noisy phase shifting microinterferograms using deep learning algorithms

引用

MICROSCOPY RESEARCH AND TECHNIQUE 2022年第2期85卷 667-684页

作者： Abo-Lila, Gehad Sokkar, Taha Seisa, Eman Omar, Emam Mansoura Univ Fac Sci Phys Dept Mansoura Egypt New Mansoura Univ Fac Sci Phys Dept New Mansoura Egypt

In this article, an adaptive denoising method is suggested to accurate investigate the optical and structural features of polymeric fibers from noisy phase shifting microinterferograms. The mixed class of noise that may produce in the phase-shifting interferometric techniques is established. To our knowledge, this is an early study considered the mixing noises that may occur in microinterferograms. The suggested method utilized the convolution neural networks to detect the noise class and then denoising, it according to its class. Four convolution neural networks (Googlenet, VGG-19, Alexnet, and Alexnet-SVM) are refined to perform the automatic classification process for the noise class in the established data set. The network with the highest validation and testing accuracy of these networks is considered to apply the suggested method on realistic noisy microinterferograms for polymeric fibers, polypropylene and antimicrobial polyethylene terephthalate)/titanium dioxide, recoded using interference microscope. Also, the suggested method is applied on noisy microinterferograms include crazing and nanocomposite material. The demodulated phase maps and the three-dimensional birefringence profiles are calculated for tested fibers according to the suggested method. The obtained results are compared with the published data for these fibers and found to be in good agreements.

关键词： deep learning algorithms denoising method nanocomposite material noise phase-shifting interferometry (PSI) technique

来源：评论

学校读者我要写书评

暂无评论

Comparative analysis of Air Quality Index prediction using deep learning algorithms

引用

SPATIAL INFORMATION RESEARCH 2024年第1期32卷 63-72页

作者： Mishra, Ankita Gupta, Yogesh Banasthali Vidyapith Comp Sci & Engn Radha Kishnpura Rajasthan India

This paper comprehensively reviews and compares methodologies used to monitor air quality and their impact on human health. With urbanization and industrialization increasing in emerging nations, air pollution levels have become a significant threat to human well-being. The study highlights the importance of reducing exposure to air pollution for the improvement of public health. The paper focuses on the comparative analysis of measuring the Air Quality Index (AQI) using deep learning algorithms like Long Short-Term Memory (LSTM) and classical machine learning models such as Autoregressive Integrated Moving Average (ARIMA), Decision Tree, K-Nearest Neighbour, Extreme Gradient Boosting, Gradient Boosting, Adaptive Boosting, Huber Regressor, and Dummy Regressor for AQI prediction. The performance of these models is evaluated using daily and hourly time series data from 2014 to 2018, with the Root Mean Squared Error (RMSE) used as the performance indicator. The results demonstrate that LSTM outperforms ARIMA, particularly with hourly data. For daily data, ARIMA achieved an RMSE of 97.88, whereas LSTM obtained an RMSE of 143.07. On the other hand, for hourly data, ARIMA yielded an RMSE of 69.65, while LSTM achieved a lower RMSE of 44.6539. These findings highlight the potential of deep learning algorithms, specifically LSTM, in accurately forecasting air quality.

关键词： Air Quality Index (AQI) deep learning algorithms Classical machine learning algorithms LSTM ARIMA And Urbanization

来源：评论

学校读者我要写书评

暂无评论

Multivariate systemic risk measures and computation by deep learning algorithms

引用

QUANTITATIVE FINANCE 2021年第10期23卷 1431-1444页

作者： Doldi, A. Feng, Y. Fouque, J. -P. Frittelli, M. Univ Milan Dipartimento Matemat Milan Italy Univ Calif Santa Barbara Dept Stat & Appl Probabil Santa Barbara CA 93106 USA

In this work, we propose deep learning-based algorithms for the computation of systemic shortfall risk measures defined via multivariate utility functions. We discuss the key related theoretical aspects, with a particular focus on the fairness properties of primal optima and associated risk allocations. The algorithms we provide allow for learning primal optimizers, optima for the dual representation and corresponding fair risk allocations. We test our algorithms by comparison to a benchmark model, based on a paired exponential utility function, for which we can provide explicit formulas. We also show evidence of convergence in a case in which explicit formulas are not available.

关键词： Systemic risk measures Multivariate utility functions Primal and dual problems deep learning algorithms >

来源：评论

学校读者我要写书评

暂无评论

Soil parameter inversion modeling using deep learning algorithms and its application to settlement prediction: a comparative study

引用

ACTA GEOTECHNICA 2023年第10期18卷 5597-5618页

作者： Hu, An-Feng Xie, Sen-Lin Li, Tang Xiao, Zhi-Rong Chen, Yuan Chen, Yi-Yang Zhejiang Univ Res Ctr Coastal & Urban Geotech Engn Hangzhou 310058 Peoples R China Zhejiang Univ Architectural Design & Res Inst Zhejiang Univ Co Ltd Hangzhou 310058 Peoples R China Zhejiang Univ Ctr Balance Architecture Hangzhou 310058 Peoples R China Zhejiang Univ Sci & Technol Sch Civil Engn & Architecture Liuhe Rd 86 Hangzhou 310023 Peoples R China

This study shows the application of the deep learning (DL) algorithm in the inversion of the crucial constitutive model parameters directly referring to the surface settlement monitoring data. Three parameter inversion models (BiLSTM, CNN1D, and ConvLSTM) based on DL algorithm are used for a comparative study. The datasets for training and testing are synthetic data that we calculated by importing the independently designed Python script into the finite element (FE) software ABAQUS. The Mohr-Coulomb (MC) model and the Modified Cam-Clay (MCC) model are used in the numerical model. The generalization abilities of the inversion models are tested using the monitoring settlement of the new airport in Xiamen, China. The main results show that: (i) the three-parameter inversion models based on DL can accurately identify the soil parameters that have a significant influence on the settlement calculation, and (ii) in terms of training time and prediction accuracy, the inversion model based on CNN1D has a better comprehensive performance than ConvLSTM and BiLSTM. The good consistency between the settlement calculated by the parameters retrieved from the DL model and the monitoring values shows that it has great potential to use DL as a meta-model for the calculation of the foundation settlement and the prediction of future development.

关键词： ABAQUS deep learning algorithms Parameters inversion Python Secondary development Settlement prediction

来源：评论

学校读者我要写书评

暂无评论

Application of Machine learning and deep learning algorithms for Landslide Susceptibility Assessment in Landslide Prone Himalayan Region

引用

EARTH SYSTEMS AND ENVIRONMENT 2024年 1-19页

作者： Bhattacharya, Subhasis Ali, Tarig Chakravortti, Sudip Pal, Tapas Majee, Barun Kumar Mondal, Ayan Pande, Chaitanya B. Bilal, Muhammad Rahman, Muhammad Tauhidur Chakrabortty, Rabin Sidho Kanho Birsha Univ Dept Econ Purulia WB India Amer Univ Sharjah Dept Civil Engn POB 26666 Sharjah U Arab Emirates Raiganj Univ Dept Geog Raiganj 733134 West Bengal India Univ Texas Dallas Sch Econ Polit & Policy Sci Geospatial Informat Sci Program Richardson TX 75023 USA Asian Inst Technol Sch Environm Resources & Dev Khlong Nueng 12120 Phatum Thani Thailand Univ Tenaga Nas Inst Energy Infrastruct Kajang 43000 Malaysia Al Ayen Univ Sci Res Ctr New Era & Dev Civil Engn Res Grp Thi Qar 64001 Iraq Govt Gen Degree Coll Dept Zool Mohanpur 721436 India King Fahd Univ Petr & Minerals KFUPM Architecture & City Desing Dept ACD Dhahran 31261 Saudi Arabia King Fahd Univ Petr & Minerals KFUPM Interdisciplinary Res Ctr Aviat & Space Explorat I Dhahran 31261 Saudi Arabia Sidho Kanho Birsha Univ Dept Sanskrit Purulia 723104 India

Landslides are an unpredictable natural disaster, but steps can be taken to reduce their impact. The Landslide Susceptibility Index plays a critical role in minimizing the risk of living in landslide-prone areas. Effective planning and management of these locations are essential. In recent years, statistical methods and, increasingly, machine learning-based approaches have gained popularity for landslide susceptibility modeling. This study employs various machine learning and deep learning algorithms, specifically Random Forest (RF), Artificial Neural Network (ANN), and deep learning Neural Network (DLNN), to estimate landslide susceptibility in Chamoli district, Uttarakhand, India-a region that witnessed over a thousand landslides in 2023. We carefully selected relevant metrics based on existing research and conducted a multicollinearity analysis on each parameter to ensure the model's accuracy. We randomly split the data into training and validation sets in a 70/30 ratio. Among the models used, the DLNN outperformed others, superiorly predicting landslide susceptibility. These findings are valuable for local government efforts in disaster prevention and mitigation, particularly in the Chamoli District of Uttarakhand, where Geographical Information System (GIS)-based susceptibility mapping plays a critical role in identifying vulnerable areas. Overall, this model evaluation framework can be used as a guide to select the most suitable modelling strategy for assessing landslide susceptibility. This type of outcome is valuable to the decision-maker to implement a more optimal strategy for reducing the probability of landslides and its associated damages.

关键词： Chamoli District deep learning algorithms GIS-based Susceptibility Mapping Landslide Susceptibility Machine learning Natural Disaster Uttarakhand

来源：评论

学校读者我要写书评

暂无评论

The impact of deep learning algorithms with MRI Systematic review 11

The impact of Deep learning algorithms with MRI Systematic r...

引用

11th International Conference on Information and Communication Systems (ICICS)

作者： Al-Aiad, Ahmad Mustafa, Omara Jordan Univ Sci & Technol Dept Comp Informat Syst Irbid Jordan

ISBN: (纸本)9781728162270

The current trend in the medical diagnosis world is aiming towards utilizing Artificial Intelligence (AI) in the medical imaging field to highlight what is supposed to be abnormal signs and abnormal structures in the medical image. In our study we construct a systematic review to gather the deep learning algorithms, that have a direct impact with MRI in one single article, to analyze, synthesize and focus on their advantages, and limitations, to ensure that the future researchers will try to get over these limitations, and to take the benefits from the previous. We used Ramey and Rao's that separate the research into two main phases, In the first phase, we are planning to review the published literature on the deep learning algorithms and their impact with MRI, to identify the main algorithms to identify the main limitations and gaps in the current literature. In the second phase conducting the review, this includes how we collect the data, what is the search keywords, the databases where used, and what is inclusion and exclusion criteria. Then we extract the main themes in our researches and build taxonomy connect the whole articles with their themes. And in conclusion we can conclude the deep learning algorithms and their impact with MRI can facilitate the image diagnosis and can also make an improvement on the images whether in resolution or get thinner slice thickness from the thicker one, furthermore, they can enhance the decision-making process that has a direct impact with patient care.

关键词： MRI Magnetic Resonance Imaging deep learning algorithms Medical imaging

来源：评论

学校读者我要写书评

暂无评论

Research on Carbon-Electricity Feature Extraction Technology for Enterprises based on Blockchain and deep learning algorithms 24

Research on Carbon-Electricity Feature Extraction Technology...

引用

International Conference on Power Electronics and Artificial Intelligence (PEAI)

作者： Li, Chunzhe Zhai, Lingjiang Wang, Keqiang Zhang, Zhansheng Zhang, Chengkang Li, Xue Baicheng Power Supply Co State Grid Jilin Elect Power Co Ltd Baicheng 137000 Jilin Peoples R China

ISBN: (纸本)9798400716638

Accurate collection and data analysis of enterprise carbon emissions are fundamental for precise estimation of carbon emissions. To address the challenges related to the security of collecting diverse types of enterprise carbon emission data and the effectiveness of extracting heterogeneous data features, this paper proposes a method for extracting carbon-electricity feature information for enterprises based on blockchain and deep learning algorithms. Firstly, by leveraging the decentralized, open, autonomous, and tamper-proof properties of blockchain technology, a comprehensive architecture for collecting diverse types of enterprise carbon emission data is constructed. The blockchain network is utilized to establish an information exchange and data storage center for multiple energy sources such as coal, electricity, gas, heat, and oil. Secondly, a multi-modal fusion neural network model is developed to extract the spatiotemporal distribution features of electricity consumption and carbon emissions within the same enterprise. This model is designed to handle heterogeneous data and capture multidimensional characteristics. Finally, simulation analysis demonstrates that blockchain technology facilitates the exchange of enterprise energy information and enables the extraction of numerical features, categorical features, temporal features, and spatial features related to carbon-electricity. This validation confirms the universality of the proposed model.

关键词： carbon emissions Carbon-electricity blockchain deep learning algorithms heterogeneous data features

来源：评论

学校读者我要写书评

暂无评论

Modeling of laser aided direct metal deposition of stainless steel using supervised deep learning algorithms

Modeling of laser aided direct metal deposition of stainless...

引用

International Conference on Advances in Materials and Mechanical Engineering. (ICAMME)

作者： Singh, Abhijit Nath, Ananya Roy, Shibendu Shekhar Lohar, Aditya Kumar Natl Inst Technol Durgapur Durgapur W Bengal India CSIR Cent Mech Engn Res Inst Durgapur Durgapur W Bengal India

Laser Direct Metal Deposition (LDMD) is a very promising additive manufacturing methodology which provides metal cladding layer over the substrate with good corrosion resistance, desired surface finish and very accurate geometry. In this work, an attempt has been made to develop a forward and reverse model of laser aided direct metal deposition utilizing deep learning algorithms. Forward model aims at predicting the deposited bead height and width from the known set of LDMD process parameters such as laser power, scanning velocity, and powder flow rate. Reverse modelling shows the effectiveness of each input parameter over the chosen response factors. Therefore, in the reverse modelling, the optimized input parameters values have been tried to predict to get the desired deposition characteristics during experimentation. Three different neural networks based deep learning architectures, namely, Multi-Layer Feed Forward Neural Network (MLFFNN), Recurrent Neural Networks (RNN) and Radial Basis Function Neural Network (RBFNN), have been put forward to develop the said predictive models. The supervised learning algorithms are used to train the neural networks. The performances of MLFFNN, RNN and RBFNN models are compared among themselves. The results show that all the models are capable of making better predictions and the models can be effectively used in shop floor in optimal selection of most influential parameters for the desired outputs. Copyright (C) 2022 Elsevier Ltd. All rights reserved.

关键词： Direct metal deposition Forward modeling Reverse modeling Additive Manufacturing Supervised learning deep learning algorithms

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：