Fault diagnosis of railway freight car wheelset based on deep belief network and cuckoo search algorithm

Li, Honghui; Hong-kun, Wang; Xie, Ziwen; He, Mengqi

doi:10.1177/09544097211029155

Cited by 5 publications

(5 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If the state vector v of the visible layer is known, the neuron activation possibility of the concealed layer is shown in equation 3. 4.…”

Section: Construction Of Deep Belief Network Model the Research Optim...mentioning

confidence: 99%

“…In addition, the duplication of recommendation content and the inability to effectively process high-dimensional and nonlinear data will lead to inefficient resource recommendation. On the basis of the DBN model, the research conducted an in-depth discussion on the RBM and BP of each component of the DBN [3,4]. In the process of in-depth modeling, the feature vector based on "learner-course resources" and the relevant course scores provide a training example for students.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Application Analysis of English Personalized Learning Based on Large-scale Open Network Courses

Yang

2024

SCPE

View full text Add to dashboard Cite

In the context of Big data, large-scale open online courses increase learning paths for learners, but in the face of countless high-quality curriculum resources, it is easy for derivative learners to face the dilemma of rich curriculum resources but difficult to choose resources, which leads to information maze for learners. How to help learners quickly and accurately find their own learning resources in the explosive growth of MOOC resources is an urgent problem in the field of education Big data. However, the traditional Collaborative filtering recommendation technology does not perform well when dealing with sparse data and cold start. The recommendation content is repeated and can not effectively deal with high-dimensional and nonlinear data of online learning users, resulting in low efficiency of resource recommendation. Therefore, the study adopts a deep belief network (DBN) to construct a personalized resource recommendation model. The model combines the learner behavior characteristics with the curriculum resource content attribute characteristics to form the learner feature vector. The parameters of the model are adjusted according to the characteristics of learners. Through experiments, the proposed model has shown good performance. The experiment explored the effects of training set size, learner characteristics, and GPU on model performance. The experimental results show that when the training set proportion is 100%, the RMSE, Accuracy, Recall, and F1 values of the model are 0.76, 0.946, 0.957, and 0.951, respectively. When the model is trained using a training set containing learner features, the RMSE, Accuracy, Recall, and F1 values of the model are 0.75, 0.962, 0.908, and 0.958, respectively. After using GPU to accelerate the model, the running time of the model decreased from 360 minutes to 90 minutes. The results indicate that the model cannot effectively mine data information when the degree of correlation between sample information is low. The richer the relationships between samples, the better the performance of the model. Simultaneously learning hunger feature vectors and learner behavior feature vectors for training can significantly improve the recommendation accuracy of the model. The main contribution of this study is to propose a recommendation method based on DBN classification to replace traditional similarity calculation methods, using DBN's efficient feature abstraction and feature extraction capabilities to fully explore learners' interest and preference for course resources. In addition, in view of the common problems of cold start and data sparsity in traditional Collaborative filtering recommendation methods, the research deeply mines the characteristics of learners' Demographics and curriculum resources' content attributes, and constructs a learner interest model based on DBN combined with learners' behavior characteristics, which effectively solves the problems of cold start and data sparsity, as well as the inaccurate expression of learners' interest preferences for curriculum resources.

show abstract

“…If the state vector v of the visible layer is known, the neuron activation possibility of the concealed layer is shown in equation 3. 4.…”

Section: Construction Of Deep Belief Network Model the Research Optim...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Application Analysis of English Personalized Learning Based on Large-scale Open Network Courses

Yang

2024

SCPE

View full text Add to dashboard Cite

show abstract

“…Therefore, factors such as rotating reserve's capacity and response speed will directly affect system frequency's dynamic response. Formula ( 8) is used to calculate generator's rotating reserve capacity [26].…”

Section: ( ) ( )mentioning

confidence: 99%

Frequency Response Mode Prediction of Power System After Large Disturbances Based on Deep Belief Neural Network

Wang,

Liu

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The power system frequency is an important indicator that reflects the power system's operating status. Through real-time detection or prediction, it can effectively ensure stable power system operation. To provide a theoretical basis for the steady-state recovery of thepower system after major disturbances, a deep belief neural network model for predicting the power system's frequency response mode after major disturbances has been proposed. This model combines deep belief networks with deep neural networks. The feature extraction and learning abilities of deep belief networks were utilized to complete model training and learning. Deep neural networks were utilized to complete data classification and prediction. The load sudden variable is larger, the system frequency fluctuation is greater, the inertia time constant increases less, and the system frequency fluctuation is greater. In simulation testing with a training sample size of 2100 and a testing sample size of 900, the deep belief neural network took 3320 seconds, while the deep neural network took 11523 seconds. The prediction results' absolute error amplitude of this deep belief neural network is 0.025Hz. It meets the practical needs of frequency response mode prediction in power systems after major disturbances. The study analyzed the impact of load sudden changes and inertia time constants on system frequency, and successfully designed a frequency response mode prediction model for power systems after major disturbances.

show abstract

“…Fault diagnosis technologies can be used to detect early faults by utilizing physics theories or data analysis tools. To avoid the accident, fault diagnosis has widely applied to critical components in railway vehicles, such as bogies and pantographs [2]. Much research have achieved good performance in the field of railway fault diagnosis [3,4].…”

Section: Introductionmentioning

confidence: 99%

Intelligent fault diagnosis of railway pantograph using a novel graph construction methodology

Ma,

Duan,

Chen

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Railway pantographs provide power for railway vehicles by conducting electrical energy from overhead catenary. The failure of the pantograph tends to damage the contact quality between the pantograph and the catenary, reducing the transmission efficiency of electric energy. Hence, fault diagnosis of pantograph plays a significant role in expanding the service life of railway vehicles. In this work, a novel graph construction method is proposed for the fault diagnosis of pantographs combined with a graph Neural Network (GNN). In the graph construction method, 1D load signals collected from the test pantograph are firstly transformed into multiple 2D images with the same size in both time and frequency domains using Gramian Angular Field (GAF), Markov Transition Field (MTF) and Recurrence Plot (RP). Secondly, Pixel values in images are regarded as features in vertexes of graphs, and graphs can be constructed by connecting eight neighbor vertexes. Finally, the GNN model is trained by constructed graphs for obtaining the fault diagnosis model of pantographs. Laboratory experiments are implemented to show the advantages of the proposed method by comparing it with other conventional methods. There are two main innovations of this approach: 1. Handling the data with different lengths. 2. Performing perfectly in the situation of small training data size.

show abstract

Fault diagnosis of railway freight car wheelset based on deep belief network and cuckoo search algorithm

Cited by 5 publications

References 5 publications

Application Analysis of English Personalized Learning Based on Large-scale Open Network Courses

Application Analysis of English Personalized Learning Based on Large-scale Open Network Courses

Frequency Response Mode Prediction of Power System After Large Disturbances Based on Deep Belief Neural Network

Intelligent fault diagnosis of railway pantograph using a novel graph construction methodology

Contact Info

Product

Resources

About