Novel Convolutional Neural Network (NCNN) for the Diagnosis of Bearing Defects in Rotary Machinery

Kumar, Anil; Vashishtha, Govind; Gandhi, C.P.; Zhou, Yuqing; Głowacz, Adam; Xiang, Jiawei

doi:10.1109/tim.2021.3055802

Cited by 109 publications

(41 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this regard, in future work we propose to evaluate novel deep learning methods showing excellent classification results. The inclusion, after the face detection stage, of novel CNNs such as recurrent CNNs [35] or novel CNNs [36] will probably increase the percentage of emotions correctly detected.…”

Section: Discussionmentioning

confidence: 99%

Facial Emotion Recognition from an Unmanned Flying Social Robot for Home Care of Dependent People

et al. 2021

View full text Add to dashboard Cite

This work is part of an ongoing research project to develop an unmanned flying social robot to monitor dependants at home in order to detect the person’s state and bring the necessary assistance. In this sense, this paper focuses on the description of a virtual reality (VR) simulation platform for the monitoring process of an avatar in a virtual home by a rotatory-wing autonomous unmanned aerial vehicle (UAV). This platform is based on a distributed architecture composed of three modules communicated through the message queue telemetry transport (MQTT) protocol: the UAV Simulator implemented in MATLAB/Simulink, the VR Visualiser developed in Unity, and the new emotion recognition (ER) system developed in Python. Using a face detection algorithm and a convolutional neural network (CNN), the ER System is able to detect the person’s face in the image captured by the UAV’s on-board camera and classify the emotion among seven possible ones (surprise; fear; happiness; sadness; disgust; anger; or neutral expression). The experimental results demonstrate the correct integration of this new computer vision module within the VR platform, as well as the good performance of the designed CNN, with around 85% in the F1-score, a mean of the precision and recall of the model. The developed emotion detection system can be used in the future implementation of the assistance UAV that monitors dependent people in a real environment, since the methodology used is valid for images of real people.

show abstract

Section: Discussionmentioning

confidence: 99%

Facial Emotion Recognition from an Unmanned Flying Social Robot for Home Care of Dependent People

et al. 2021

View full text Add to dashboard Cite

show abstract

“…A variety of approaches has been proposed for analysing vibration measurements in related applications, e.g., monitoring of bearings [17][18][19] or bridges [20]. In many approaches, time-frequency-based methods such as empirical mode decomposition [17], wavelets [18] and correlation measures [20] were combined with machine learning and artificial intelligence (AI).…”

Section: Related Literaturementioning

confidence: 99%

“…In many approaches, time-frequency-based methods such as empirical mode decomposition [17], wavelets [18] and correlation measures [20] were combined with machine learning and artificial intelligence (AI). For example, Kumar et al [19] detected bearing defects by using a sparse cost function and convolutional neural networks and therewith addressed the challenges of a small training dataset. However, labelled datasets and precise physical models are unavailable for measurements on wind turbine blades-as occurring for many large-scale structures [21].…”

Section: Related Literaturementioning

confidence: 99%

Vibration-Based Fingerprint Algorithm for Structural Health Monitoring of Wind Turbine Blades

Loss

Bergmann

2021

Applied Sciences

View full text Add to dashboard Cite

Monitoring the structural health of wind turbine blades is essential to increase energy capture and operational safety of turbines, and therewith enhance competitiveness of wind energy. With the current trends of designing blades ever longer, detailed knowledge of the vibrational characteristics at any point along the blade is desirable. In our approach, we monitor vibrations during operation of the turbine by wirelessly measuring accelerations on the outside of the blades. We propose an algorithm to extract so-called vibration-based fingerprints from those measurements, i.e., dominant vibrations such as eigenfrequencies and narrow-band noise. These fingerprints can then be used for subsequent analysis and visualisation, e.g., for comparing fingerprints across several sensor positions and for identifying vibrations as global or local properties. In this study, data were collected by sensors on two test turbines and fingerprints were successfully extracted for vibrations with both low and high operational variability. An analysis of sensors on the same blade indicates that fingerprints deviate for positions at large radial distance or at different blade sides and, hence, an evaluation with larger datasets of sensors at different positions is promising. In addition, the results show that distributed measurements on the blades are needed to gain a detailed understanding of blade vibrations and thereby reduce loads, increase energy harvesting and improve future blade design. In doing so, our method provides a tool for analysing vibrations with relation to environmental and operational variability in a comprehensive manner.

show abstract

“…To verify the occurrence of overfitting in the model development phase, one should start from the simplest architecture then gradually increase the complexity when observing the validated result [42]. A. Kumar et al [43] used a novel fuzzy cross-entropy loss together with transfer learning to increase the accuracy of the model. The work from M. Piekarski et al [44] mainly focused on the implementation of transfer learning by utilizing various pre-trained networks where the VGG16 outperformed among all models considered in the study.…”

Section: Introductionmentioning

confidence: 99%

Detection of Myocardial Infarction Using ECG and Multi-Scale Feature Concatenate

Jian

Ger

Lai

et al. 2021

Sensors

View full text Add to dashboard Cite

Diverse computer-aided diagnosis systems based on convolutional neural networks were applied to automate the detection of myocardial infarction (MI) found in electrocardiogram (ECG) for early diagnosis and prevention. However, issues, particularly overfitting and underfitting, were not being taken into account. In other words, it is unclear whether the network structure is too simple or complex. Toward this end, the proposed models were developed by starting with the simplest structure: a multi-lead features-concatenate narrow network (N-Net) in which only two convolutional layers were included in each lead branch. Additionally, multi-scale features-concatenate networks (MSN-Net) were also implemented where larger features were being extracted through pooling the signals. The best structure was obtained via tuning both the number of filters in the convolutional layers and the number of inputting signal scales. As a result, the N-Net reached a 95.76% accuracy in the MI detection task, whereas the MSN-Net reached an accuracy of 61.82% in the MI locating task. Both networks give a higher average accuracy and a significant difference of p < 0.001 evaluated by the U test compared with the state-of-the-art. The models are also smaller in size thus are suitable to fit in wearable devices for offline monitoring. In conclusion, testing throughout the simple and complex network structure is indispensable. However, the way of dealing with the class imbalance problem and the quality of the extracted features are yet to be discussed.

show abstract

Novel Convolutional Neural Network (NCNN) for the Diagnosis of Bearing Defects in Rotary Machinery

Cited by 109 publications

References 37 publications

Facial Emotion Recognition from an Unmanned Flying Social Robot for Home Care of Dependent People

Facial Emotion Recognition from an Unmanned Flying Social Robot for Home Care of Dependent People

Vibration-Based Fingerprint Algorithm for Structural Health Monitoring of Wind Turbine Blades

Detection of Myocardial Infarction Using ECG and Multi-Scale Feature Concatenate

Contact Info

Product

Resources

About