Figure - available from: Shock and Vibration
This content is subject to copyright. Terms and conditions apply.
Source publication
Intelligent mechanical fault diagnosis has developed very fast in recent years due to the advancement and application of deep learning technologies. Thus, there are many deep learning network models that have been explored in fault classification and diagnosis. However, there are still limitations in research on the relationship between fault locat...
Similar publications
Intelligent fault diagnosis of rolling bearings under non-stationary and time-varying speed conditions is still a challenging task. At the same time, a reasonable explanation for an intelligent diagnosis model based on features is currently lacking. Therefore, we exploit an explainable one-dimensional convolutional neural network (1DCNN) model by c...
Citations
... Vibration analysis, popular in machine fault detection due to its diagnostic efficacy, was utilized as a predictive maintenance technique as well as a decisionmaking tool for equipment maintenance [1]. Statistically, 40% of rotating machine failures are related to bearings, indicating the significance of this component in rotating machinery [2]. Therefore, ensuring its safety and stable functioning may save us a substantial amount of money by decreasing maintenance time and expenses. ...
Deep transfer learning has evolved into a powerful method for defect identification, particularly in mechanical systems that lack sufficient training data. Nonetheless, domain divergence and absence of overlap between the source and target domains might result in negative transfer. This study examines the partial knowledge transfer, for bearing fault diagnosis, by freezing layers in varying proportions to take advantage of both freezing and fine-tuning strategies. To assess the proposed strategy, three distinct pre-trained models are used, namely ResNet-50, GoogLeNet, and SqueezeNet. Each network is trained using three different optimizers: root mean square propagation, adaptive moment estimation, and stochastic gradient descent with momentum. The suggested technique performance is evaluated in terms of fault classification accuracy, specificity, precision, and training time. The classification results obtained using the CWRU datasets show that the proposed technique reduces training time while enhancing diagnostic accuracy, hence improving bearing defect diagnosis performance.
... e numbers represent the bandwidth information that can be used by the link, and the numbers in the rectangle around the node represent the CPU and memory resource utilization that can be used [9,10]. ...
In order to make up for the deficiency of a two-dimensional user interface, enhance the visualization of network information, and comprehensively and effectively improve the efficiency of network management, this paper attempts to use virtual reality technology to design the network management system and uses the research results of related fields for reference to carry out the specific practice, providing a new idea for the design and development of the network management system. The automatic topology discovery algorithm is introduced into the network management analysis, and the immersive network management is modularized by using the method of dividing subnets. Finally, the test results show that through the functional design of the system and the hierarchical description of the structural model, the practical method is given. Finally, the operation process of the simulation system is analyzed in detail. The proposed method can effectively reduce the network load and reduce the network bandwidth and greatly improve the slow convergence of the penetration VR network topology.
... Wang et al [24] used 1D-CNN to extract features from the original vibration signal and acoustic signal under the working condition of the bearing, and fused the extracted features. Liu et al [25] used a hierarchical multi-task CNN to improve the generalization performance of the model diagnosis. ...
Safe and reliable operation of mechanical equipment depends on timely and accurate fault diagnosis. When the actual working conditions are complex and variable and the available sample data set is small, recognition accuracy of the rolling bearing fault diagnosis model is low. To solve this problem, a novel method based on Markov transition field (MTF) and multi-dimension convolutional neural network (MDCNN) is proposed in this paper. Firstly, the original vibration signals are converted into two-dimensional images containing temporal correlation by MTF. Then, a neural network model is constructed by using multi-dimension attention (MDA) and E-Relu activation function to fully extract fault feature information. Finally, the MTF images are input into the model and the data is normalized using the group normalization method. The MDCNN model is validated on two different data sets, and the results show that compared with other intelligent fault diagnosis methods, the MDCNN has higher fault diagnosis accuracy and stronger robustness under both variable working conditions and small sample data sets conditions.
... Methods [16] develop a gearbox compound fault diagnosis system based on a convolutional neural network (CNN) to greatly shorten the training time of the model. Methods [17] propose a hierarchical multitask convolutional neural network (HMCNN) to diagnose bearing faults, which reduces the propagation of classification errors among multiple tasks and improves the accuracy of fault diagnosis. ...
Mechanical parts are an important part of the machinery industry. However, the current automated production lines for mechanical parts face great challenges. Different types and sizes of parts make automatic segmentation of parts on industrial lines prone to wrong segmentation. Therefore, this paper proposes a part classification method based on a convolutional neural network. The method can utilize a large amount of data of mechanical parts and then analyze and learn from the data through convolutional neural networks to detect and classify parts more accurately. It can effectively solve the problems of high-cost and error-prone manual parts classification and greatly reduce the cost of automatic detection and classification of enterprises. We adopt this model to conduct a series of experiments on the Pacon dataset. It includes the calculation of weights and the construction of confusion matrix values. The experimental results show that the model achieves excellent performance in parts classification and prediction.
... Therefore, the incompatible modulation of the vehicle noise vibration signal pulse is due to the transmitter noise, because the noise is caused by different kinds of spurious modulation. Vehicle noise and vibration signals present different signal characteristics [10,11]. ...
Vehicle interior noise is a harmful sound formed in the vehicle interior under the influence of engine noise, transmission system noise, and body radiation noise during vehicle operation. The intensity of noise in the vehicle increases with the increase of vehicle speed and engine speed, which worsens the operating environment of the vehicle and does harm to the mood, spirit, and physiology of drivers and passengers. Because the noise is difficult to identify in the process of vehicle vibration signal recognition, aiming at the problem of vehicle vibration signal detection under the background of strong noise, combined with the characteristics of vehicle vibration signal, the variable step size LMS algorithm is applied to vehicle noise and vibration signal recognition. In this paper, based on the signal eigenvalues obtained from the vibration signal characteristics in the noise and vibration signal, the eigenvalues of the vehicle noise can be extended from the frequency domain to the complex plane through the operation of the variable step size LMS, and the separation law corresponding to the extended vibration intensity can be used at the same time. It is used to obtain data by the signal variable step size LMS algorithm, and the obtained optimal separability feature is used as the characteristic parameter of the vibration signal. Finally, the results of an example analysis show that the algorithm proposed in this paper can be used to identify different types of vehicle noise and vibration signals, has a certain anti-interference performance against noise, and can improve the recognition rate of vehicle noise and vibration signals.
... When performing multimedia optimization particle filter spectrum analysis, Δf � f/n represents the frequency resolution, where n represents the number of sampling points, and f represents the sampling frequency and meets f ≥ 2f # (f # represents the highest sampling frequency) [8][9][10]. When applying this technical method to engineering, first install an acceleration sensor at the detection position. ...
During the overhaul and maintenance of the transmission system of the modern rolling mill, a very common problem is the positioning of the motor and the reducer or speed increaser. The operation of the equipment and the service life of the equipment will be significantly affected by the accuracy of the multimodal coupling position determination. Therefore, precise positioning and rapid positioning must be achieved. This article intends to analyze how this type of equipment achieves rapid and accurate positioning and what factors will affect the location determination process.
... Based on the basic principles of low cost, scalability, and stability, in order to effectively improve the data processing ability of sensor nodes and quickly obtain multimedia data information [7][8][9]. e design of multimedia sensor network nodes is mainly divided into 4 parts. (1) e sensor module includes cameras, microphone collection devices, vibration, infrared, and other sensors, which play a role in collecting environmental media information. ...
Through in-depth analysis of the characteristics of the quality evaluation stage of prefabricated buildings, the quality evaluation of prefabricated buildings can be divided into three stages: before, during, and after construction. According to the detailed design content of the prefabricated building construction stage, we construct the prefabricated construction quality evaluation index system, use the multimedia sensor network node method to obtain the weight of each evaluation index, comprehensively evaluate the construction quality of the prefabricated building, and finally show through the case analysis results that the multimedia sensor network node method can be of high practical value in the process of prefabricated building construction quality evaluation, and it has improved the domestic prefabricated building construction quality evaluation index system and evaluation quality and has certain reference value.
... erefore, the use of fuzzy upper and lower level functions to evaluate the feasibility of the plan requires further research [9,10]. Based on the difference of architectural design evaluation and the nonlinear relationship of corresponding elements, the theory and method of system analysis are proposed, a comprehensive evaluation model is established, and the architectural design plan is comprehensively evaluated. ...
... (2) Network training: use 20 samples to train the network, and the process is shown in Figure 3. Table 4) e multimedia joint evaluation model was used to rank 30 theater architectural designs. 23, 6,14,1,5,17,2,27,19,21,29,13,7,26,4,25,9,18,3,12,20,16,22,28,24,15,11,8,10, and 30. As shown in Table 4, the results of multimedia joint evaluation have small errors and achieve the expected results. ...
This article proposes improvements to the placement, organizational structure, and adaptability of the evaluation model and proposes a comprehensive evaluation model for the multiattribute scheme of the particle swarm optimization algorithm. The structure table of the multimedia building design plan evaluation index, the plan evaluation table, and the standardized model of the plan evaluation value of the common method was proposed. It also proposes the import of “with tutor” type data of particle swarm optimization algorithm as well as the self-organizing neural network evaluation model and calculation method of “no tutor” type resource import. According to the complexity of architectural design evaluation and the nonlinear relationship between related influencing factors, we have studied the principles and methods of applying system analysis, designed comprehensive evaluation models, and comprehensively evaluated architectural design schemes.
... In order to obtain the inspection result within the limited inspection time, finally the inspection statistics are compared with the reduced threshold C to obtain the final judgment result. e threshold C is analyzed for discussion [9]. ...
According to the basic principle of piecewise linear classifier and its application in the field of infrared chemical remote sensing monitoring, the characteristics of unilateral piecewise linear classifier applied to the infrared spectrum identification of chemical agents are studied. With the characteristic of separate transmission, the characteristic recovery with the total observed deviation is used for the model. The relaxation factors are used to replace the constrained conditions that cannot be optimized into constrained separate line segment calculation conditions. Experiments show that the result of signal recovery is better than traditional Wiener filtering and Richardson–Lucy methods.
Les installations industrielles deviennent de plus en plus complexes. La surveillance de leur état est essentielle pour la sécurité du système, les économies de coûts et l'accroissement de la productivité. Ceci nécessite d'avoir recours à des techniques de surveillance sophistiquées et hautement efficaces. Dans ce contexte, Le premier objectif de ce travail est d'appréhender expérimentalement le phénomène des vibrations au sein des machines tournantes, ainsi que la concrétisation de certains défauts courants tels que les déséquilibres et les défauts de roulement. Cependant, l'aspect le plus fondamental de ce travail réside dans le développement d'une nouvelle approche de diagnostic très efficace pour la détection précoce et précise des défauts de roulements, basée sur l'apprentissage par transfert. Pour atteindre cet objectif, premièrement, un banc d'essai a d'abord été conçu et construit, capable de simuler plusieurs défaillances mécaniques. D'autre part, nous avons développé un système d'acquisition de
données à la fois simple et économique pour notre projet en utilisant un microcontrôleur Arduino UNO et un accéléromètre (ADXL-345). Deuxièmement, notre étude examine le transfert partiel de connaissances, pour le diagnostic de défauts de roulements, en gelant des couches selon des proportions variables pour tirer parti à la fois des stratégies de figement et d'affinement. Pour évaluer la stratégie proposée, trois modèles pré-entraînés distincts sont utilisés, à savoir ResNet-50, GoogLeNet et SqueezeNet. Chaque réseau est entraîné en
utilisant trois optimiseurs différents : SGD, Adam et RMSprop. Les performances de la technique suggérée sont évaluées en termes de taux de classification des défauts, de spécificité, de précision et de temps d'entraînement. Les résultats de classification obtenus à l'aide des ensembles de données CWRU montrent que la technique proposée réduit le temps d'entraînement tout en améliorant l'exactitude du diagnostic, améliorant ainsi les performances du diagnostic des défauts de roulements
