Newsfeeds
Measurement Science and Technology - latest papers
Latest articles for Measurement Science and Technology
-
A generative zero-shot learning method for compound fault intelligent diagnosis of rolling bearings
In practical industrial applications, the scarcity of compound fault samples poses significant challenges in obtaining sufficient training data, leading to class imbalance and negatively impacting the diagnostic performance of models. To tackle this issue, we propose a generative zero-shot learning method for rolling bearing compound fault intelligent diagnosis, attempting to diagnose unknown target faults using only known fault samples. Specifically, a novel fault attribute description method is designed, which combines fault semantic information with manually defined fault description information, thereby constructing fault category auxiliary information (FCAI) to learn the correlation between single faults and compound faults. Furthermore, the continuous wavelet transform is used to data preprocess the raw vibration signals, and a wide kernel convolutional neural network is constructed to extract deep fault feature information from the samples. Finally, an adversarial training strategy is adopted to learn the mapping relationship between the fault feature information and the FCAI, and compound faults are diagnosed using a distance metric method based on the similarity relationship between fault features. Through experimental validation on the laboratory bearing dataset and the Huazhong University of Science and Technology bearing dataset, the effectiveness and superiority of the proposed method in scenarios lacking compound fault samples.
-
Aeroengine intelligent gas path simulation and diagnosis based on feature fusion and meta learning
With the rapid development of artificial intelligence technology, data-driven methods show great potential in the simulation and fault diagnosis of complex industrial systems. However, the application of traditional intelligent methods in aeroengine health management still faces many challenges, including the limited number of initial fault samples and the insufficient ability to capture gas path performance degradation. To solve these problems, this paper proposed a new dual cross-modal feature fusion module and a new model-agnostic meta-learning module to construct an intelligent simulation and diagnosis framework for engine gas path. This framework can effectively integrate physical model and sensor measurement data, and realize high-precision simulation under gas path performance degradation and new fault diagnosis under small sample condition. Through the example verification and comparative analysis, the results show that under the condition of performance degradation, the simulation error of gas path parameters is only 0.074%; in the case of limited initial fault samples, the fault diagnosis accuracy is as high as 99.4%, which is significantly better than the existing intelligent methods. The research shows that the proposed intelligent simulation and diagnosis framework can provide strong support for the aeroengine intelligent health management system.
-
Crack-YOLOv8: multi-class crack detection method for track slabs based on improved YOLOv8
Cracks in high-speed rail track slabs are important indicators of aging, and the issue of ineffective crack repairs remains a significant challenge in crack maintenance. However, research on this issue is limited. This paper proposes a novel multi-class crack detection method based on the YOLOv8 framework, addressing the problem of ineffective crack repairs. To enhance the model’s feature extraction capability, we introduce the dynamic snake convolution (DSC) module into the YOLOv8 backbone network and reconstruct the Bottleneck structure in C2f, establishing the C2f-DSC module. This modification replaces some of the C2f modules, enabling better multi-scale feature extraction, particularly for difficult-to-detect ineffective cracks. Additionally, we incorporate the CA attention mechanism in the neck network to improve the model’s ability to focus on critical features and effectively transmit subtle crack details throughout the network. We also replace the CIoU loss function with SIoU, reducing excessive penalization of geometric factors, thereby enhancing the model’s generalization ability. Finally, we validate the proposed method through a comprehensive evaluation of network structure, crack data, classification methods, and environmental conditions. Experimental results show that the proposed Crack-YOLOv8 model significantly improves detection accuracy, reducing both false positives and false negatives. Specifically, the average precision and recall are enhanced by 4.9% and 4.3%, respectively, demonstrating the effectiveness of our approach in accurately detecting ineffective crack repairs in rail track slabs.
-
Quantitative diagnosis of tooth root crack fault based on GMM and improved DANN: a transfer from simulation domain to experimental domain
Tooth root crack faults of gears have long been affecting the performance and service life of rotating machiners. The quantitative diagnosis of tooth root crack is crucial for effective equipment condition monitoring and management. However, affected by the high costs and limited availability of sample data, the problem of insufficient sample size and challenges to the quantitative diagnosis of tooth root crack fault always exist. A method is proposed to facilitate transfer diagnosis from the simulation domain (source domain) to the experimental domain (target domain), which is driven by the dynamic simulation model combining the Gaussian mixture model (GMM) and an improved domain adversarial neural network (DANN). A gear dynamics simulation model is first established based on the fundamental parameters of tooth root crack fault of gears to generate sufficient simulated data to compensate for the limited real data. Then the feature distribution of the simulated data is optimized with a small amount of real data based on GMM. The key indicators, such as amplitude and impact point characteristics, are aligned with the real data distribution. It reduces dependence on experimental data to a certain extent. Furthermore, the DANN improved by a Beta distribution DANN is employed in the feature spaces of two domains introducing interpolation processing. The distribution differences between the two domains are further reduced to enhance the cross-domain generalization ability of the model. The effectiveness of the proposed method is validated under 15 transfer tasks. The results demonstrate that the method achieves accurate cross-device and cross-condition quantitative diagnosis of tooth root crack from simulated data to experimental data, using fewer experimental data. The advantage of simulated data in addressing the limited sample sizes problem is confirmed, with accuracy improving by 8% compared to other methods.
-
Partially domain-adaptive rolling bearing fault diagnosis based on joint weighting and metric optimization
Partial-domain adaptive techniques are widely applied in cross-operational bearing fault diagnosis to address inconsistencies between source and target domain fault classes effectively. However, existing studies face challenges in feature alignment, including insufficient alignment of shared class fault features between the source and target domains, interference from outlier class samples in the source domain, and low-confidence pseudo-labels in the target domain. These issues hinder efficient alignment of shared class features, ultimately reducing fault diagnosis performance. Therefore, this paper proposes a joint weighting metric domain adaptation model. To improve the alignment of shared class fault features, a joint metric combining correlation alignment and local maximum mean discrepancy is developed. This metric complements adversarial training, reduces distributional discrepancy between domains, and optimizes feature alignment. To address interference from outlier class samples in the source domain, class-level weights are employed to effectively mitigate the negative transfer effects of these samples. Additionally, sample-level weights are introduced to reduce the negative transfer effects of low-confidence pseudo-labels and boundary samples, enhancing the accuracy and robustness of shared class feature alignment. The proposed method is validated through experiments on both public and self-constructed bearing datasets. Experimental results demonstrate that the proposed method achieves higher diagnostic accuracy than existing partially domain-adaptive methods in cross-operational diagnostic tasks involving similar and different equipment.