Xiangjun Du

Dr. Xiangjun Du


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 May 2025 in Measurement Science and Technology
Verified
Journal Article
Closed access
XDXiangjun Du
Xiangjun Du
HZHongjie Zhang
Hongjie Zhang
LYLing Yu
Ling Yu
DYDengjie Yang
Dengjie Yang
With the increasing complexity of industrial equipment, the fault diagnosis of rolling bearings, which are important rotating mechanical components, is particularly critical. Rolling bearing fault diagnosis under small-sample conditions faces challenges of data scarcity and insufficient model validity. In this paper, a fault diagnosis method based on multi-sensor data fusion is proposed to solve the problem of rolling bearing fault diagnosis under small-sample conditions. First, the multi-sensor time–frequency image fusion module is designed to convert the multi-source 1D vibration signals into multi-source 2D fused images by short-time Fourier transform, preserving the time–frequency characteristics of the original signals. Second, the expansion residual shrinkage weighting module is constructed to extract multi-scale features from multi-source information by using multi-scale expansion convolution and improved residual shrinkage module, while the global information selection mechanism is designed to allow the network to focus on key features. Finally, a pyramid feature fusion network is constructed to minimize the information loss phenomenon, and the identification of different states is performed using Softmax classifiers. Experimental validation was performed using the Case Western Reserve University bearing dataset bearing dataset and Southeast University bearing dataset bearing dataset, and the results show that the proposed method has excellent fault diagnosis and good generalization ability under small-sample conditions.

21 March 2025 in Measurement Science and Technology
Verified
Journal Article
Closed access
XDXiangjun Du
Xiangjun Du
HZHongjie Zhang
Hongjie Zhang
LYLi Yu
Li Yu
DYDengjie Yang
Dengjie Yang
In the realm of industrial conveyor belt systems, the belt conveyor drum is a critical component whose fault-free operation is essential for seamless material transportation. Existing methods for detecting faults in these drums, including manual inspection and sensor-based techniques, have been hampered by limitations such as slow detection speed, high cost, and insufficient accuracy and reliability. This study addresses these challenges by proposing a novel and highly efficient lightweight network model, namely YOLOv8n + EMBC + SCC. The developed model is engineered to proficiently identify and classify a diverse range of drum faults, such as damage, wear, cracks, and lagging shedding, through video-based analysis. Leveraging the inherent strengths of YOLOv8n in terms of target detection accuracy and speed, we have meticulously designed the effective MB Conv module (EMBC) module. This module, underpinned by depth-wise separable convolution efficient convolution, replaces the conventional C2F module in the backbone and neck segments of the network. This strategic substitution has yielded a significant 14.5% enhancement in processing speed and a notable 0.7% improvement in detection accuracy. Complementing this innovation, the Spatial and Channel Reconstruction Convolution (SCC) efficient convolution module has been incorporated into the detection head. By supplanting the standard Conv module, it has effectively reduced the computational load while further augmenting the model’s performance. This has led to an additional 11.73% increase in speed and an extra 0.7% elevation in accuracy. To rigorously evaluate the efficacy of the proposed model, an extensive experimental campaign was conducted using a comprehensive dataset. The experimental results unequivocally demonstrate that the YOLOv8 + EMBC + SCC model attains an outstanding detection accuracy of 93.4%, surpassing the baseline (BL) YOLOv8n by 0.9%. Compared with YOLOv10, the YOLOv8 + EMBC + SCC model shows superiority. In terms of detection accuracy, the 93.4% of YOLOv8 + EMBC + SCC is higher than the assumed 92.3% of YOLOv10. Regarding speed, the frames per second (FPS) of 38.21 for YOLOv8 + EMBC + SCC is faster than the assumed 35.7 of YOLOv10. Moreover, the model exhibits an enhanced FPS value of 38.21, representing a substantial 3.56 f s−1 improvement over the BL. The heatmap analysis provides conclusive visual evidence of the model’s superior capabilities in precisely localizing and identifying faults with high accuracy. In summary, this research makes a substantial contribution to the field of conveyor belt drum fault detection. The proposed model not only combines rapid detection speed with high accuracy but also offers reliable fault judgment, thereby positioning it as an ideal solution for industrial applications. It holds great potential for enhancing operational efficiency and streamlining maintenance practices in industrial settings, ultimately ensuring the reliable and uninterrupted operation of conveyor belt systems.

6 August 2025 in Engineering Research Express
Verified
Journal Article
Closed access
LYLing Yu
Ling Yu
XDXiangjun Du
Xiangjun Du
DYDengjie Yang
Dengjie Yang
YLYi Liu
Yi Liu
Detecting bearing faults is of crucial importance for ensuring the reliability and performance of modern industrial equipment. In this research, a method integrating the Gramian Angular Difference Field (GADF), an improved Swin Transformer, and a Convolutional Neural Network with Global Attention Module (CNN-GAM) is proposed. The proposed approach utilizes GADF to convert one-dimensional vibration signals into two-dimensional images, which not only preserves the time-varying characteristics of the signals but also enhances the feature representation. The improved Swin Transformer efficiently extracts multi-scale and multi-level features, reducing computational complexity and broadening the receptive field of the model. The CNN-GAM module then optimizes the feature representation, with the convolutional neural network capturing local features and the global attention module assigning precise attention weights to accentuate crucial information, thereby significantly enhancing the accuracy of the model. This method has been validated using the bearing dataset from Case Western Reserve University (CWRU). After comprehensive data cleaning and processing, the model exhibited remarkable performance. In the first round of training, the accuracy reached 88%, and within 25 iterations, the validation accuracy was stably maximized at 99%, while the training loss was as low as 0.01. Compared with existing models, this approach demonstrates greater robustness, requires fewer training rounds, and maintains excellent stability. Even under simulated complex working conditions, it can accurately classify a wider range of fault types with minimal misclassifications. It achieves high precision and high reliability, and thus holds significant theoretical value as well as promising potential for industrial applications.

1 July 2024 in Research Square
Verified
Preprint
Open access
PDF file
XDXiangjun Du
Xiangjun Du
LYLi Yu
Li Yu
JWJun Wang
Jun Wang
DYDengjie Yang
Dengjie Yang
YZYao Zheng
Yao Zheng
YWYimin Wang
Yimin Wang
This paper presents a comprehensive study on enhancing the accuracy, real-time performance, and reliability of fault detection in conveyor belt drums. Leveraging insights from two distinct approaches, a novel lightweight network model, YOLOv8n + EMBC + SCC, is proposed. The model integrates the strengths of YOLOv8n in target detection accuracy and speed with innovative modules designed for improved performance. Firstly, the EMBC module, based on DSC high-efficiency convolution, replaces the traditional C2F module in the backbone and neck segments, resulting in a notable 14.5% increase in speed and a 0.7% enhancement in accuracy. Secondly, the SCC efficient convolution module replaces the Conv module in the detection head, further optimizing computational load and model performance, leading to an additional 11.73% increase in speed and a 0.7% improvement in accuracy. Experimental results demonstrate the efficacy of the proposed model, achieving a detection accuracy of 93.4%, surpassing YOLOv8n by 0.9%. Moreover, the model exhibits an improved Frames Per Second (FPS) value of 38.21, representing a 3.56 f/s advancement over YOLOv8n. Heatmap analysis validates the model's superiority in terms of high detection accuracy, precise fault identification, and clear fault localization. This research contributes to the development of a fast, precise, and reliable fault detection system suitable for conveyor belt drum applications, with implications for improving operational efficiency and maintenance practices in industrial settings.

4 August 2025 in Engineering Research Express
Verified
Journal Article
Closed access
XDXiangjun Du
Xiangjun Du
LYLing Yu
Ling Yu
DYDengjie Yang
Dengjie Yang
A belt conveyor roller defect detection method based on genetic algorithm feature selection and stacking learning was proposed. Roller defect data from 1620 samples of three types of roller defects was first collected. To improve the detection efficiency of roller defects, a feature selection method based on a genetic algorithm for selecting roller defect features was proposed. The model classification efficiency of roller features after feature selection by the genetic algorithm was found to be higher. The feature subset with the highest fitness had a feature number of 35. A stacking learning model for detecting roller defects was also constructed. The stacking method enhanced the comprehensive performance of the base learner model for detecting roller defects. The stacking learning model finally achieved a classification accuracy of 94.86% on the roller defect dataset. The classification performance of the stacking learning model was better than that of its base learning model. The results showed that the stacking learning model was preferred for detecting roller defects. The proposed conveyor belt roller defect detection method, while achieving high accuracy, may be limited by potential overfitting and sensitivity to environmental conditions. The next steps in refining the conveyor belt roller defect detection method will focus on dataset expansion, algorithm parameter tuning, and environmental robustness testing.