查看更多>>摘要：The remaining useful life (RUL) prediction is of significance to the health management of bearings. Recently, deep learning has been widely investigated for bearing RUL prediction due to its great success in sequence learning. However, the improvement of the prediction accuracy of existing deep learning algorithms heavily relies on feature engineering such as handcrafted feature generation and time-frequency transformation, which increase the complexity and difficulty of the actual deployment. In this paper, a novel spatial attention-based convolutional transformer (SAConvFormer) is proposed to establish an accurate bearing RUL prediction model based on raw vibration data without prior knowledge or feature engineering. In this algorithm, firstly, a convolutional neural network enhanced by a spatial attention mechanism is proposed to squeeze the feature maps and extract the local and global features from raw bearing vibration data effectively. Then, the extracted senior features are fed into a transformer network to further explore the sequential patterns relevant to the bearing RUL. An experimental study using the XJTU-SY rolling bearings dataset revealed the merits of the proposed deep learning algorithm in terms of root-mean-square-error (RMSE) and mean-absolute-error (MAE) in comparison with other state-of-the-art algorithms.
查看更多>>摘要：To effectively monitor the operation state of in-wheel motors used in electric vehicles and ensure the safety of the whole vehicle, a diagnosis method based on hidden Markov model (HMM) and Weibull mixture model (WMM) is proposed for mechanical faults in in-wheel motors, known simply as the WMM-HMM diagnosis method. Firstly, vibration signals of the in-wheel motor are extracted for sensitive symptom parameters which are used to characterize the operation state and establish the observation sequence. Secondly, WMM is employed to expand the limited observation sequence under various operating states of in-wheel motors to obtain sufficient observation sequence as the training sample set of HMM, and HMM parameters are determined through combining supervised learning with unsupervised learning algorithm. Then the WMM-HMM diagnosis models are constructed under low and medium speed conditions respectively. Finally, the corresponding faults in-wheel motors are customized and the test bench is built to verify the proposed method. The test results show that the proposed method can accurately identify the mechanical fault state of in-wheel motors under different conditions and has good generalization and applicability in traditional methods comparison.
Ji, ShanshanWang, JinruiHan, BaokunZhang, Zongzhen...
11页
查看更多>>摘要：Machinery fault diagnosis is crucial for maintenance cost reduction and accident prevention. Vibration signal monitoring is an effective and feasible method for machinery fault diagnosis. However, extraction of the fault-related periodic impulses from weakly monitoring signals is basic but difficul . In this paper, a new weak feature extraction model using Laplacian eigenmaps and parallel sparse filtering (LE-PSF) is presented for mechanical weak fault diagnosis. Specifically, the weak vibration signal is measured from the machinery pedestal. Then, LE is used to extract principal components of the overlapped signal segments, and PSF is employed for weak feature extraction from the principal components. Finally, the extracted features are inputted to softmax regression for fault classification. A simulation study and two experimental cases are employed to testify the effect of the LE-PSF model. Experimental performances show that the LE-PSF can not only achieve accurate fault classification but also is superior to other traditional methods.
查看更多>>摘要：3D particle streak velocimetry (3D-PSV) and surface flow visualization using tufts both require the detection of curve segments, particle streaks or tufts, in images. We propose the use of deep learning based instance segmentation neural networks Mask region-based convolutional neural network (R-CNN) and Cascade Mask R-CNN, trained on fully synthetic data, to accurately identify, segment, and classify streaks and tufts. For 3D-PSV, we use the segmented masks and detected streak endpoints to volumetrically reconstruct flows even when the imaged streaks partly overlap or intersect. In addition, we use Mask R-CNN to segment images of tufts and classify the detected tufts according to their range of motion, thus automating the detection of regions of separated flow while at the same time providing accurate segmentation masks. Finally, we show a successful synthetic-to-real transfer by training only on synthetic data and successfully evaluating real data. The synthetic data generation is particularly suitable for the two presented applications, as the experimental images consist of simple geometric curves or a superposition of curves. Therefore, the proposed networks provide a general framework for instance detection, keypoint detection and classification that can be fine-tuned to the specific experimental application and imaging parameters using synthetic data.
查看更多>>摘要：A portable and light-weight aerosol homogenizer has been designed and validated experimentally. The design relies on large-scale primary standards for particle number and mass concentration previously developed for metrology applications, but the dimensions have been scaled down to produce a versatile and user-friendly apparatus for everyday applications in aerosol sciences. The homogenizer is a 0.8 m long cylinder made of steel with an inner diameter of 50 mm, equipped with three inlets for primary aerosols and up to four outlets for sampling homogenized aerosol mixtures. Mixing is achieved by three turbulent air jets. The aerosol spatial homogeneity in the sampling zone was within +/- 1% and +/- 4% for 2 and 5 mu m polystyrene (PS) particles, respectively. The possibility to supply and control independently aerosol flows with pressure-sensitive generators and the short equilibration time (<1 min) have also been demonstrated. The homogenizer allows for mixing various aerosol components, such as soot, inorganic species and mineral dust particles, to generate ambient-like aerosols in the laboratory or industrially manufactured particles such as PS spheres as model aerosols. We believe that it could have applications in applied aerosol research, health-related studies, and instrument calibration.
查看更多>>摘要：Three-dimensional (3D) visualization has provoked interest in electrical capacitance tomography (ECT) due to recent progress in the development of capacitance measurement circuits and data-acquisition systems. This paper proposes image reconstruction using spatio-temporal sampling in the Z-axis for 3D ECT to improve spatial image resolution. In a typical application of ECT, multi-phase flow imaging is performed through two-dimensional (2D) data acquisition and 2D reconstruction of image frames over time. In the presented method, a volume of interest (VOI) is reconstructed from several time samples obtained using a 3D sensor. The procedure of Jacobian matrix extension for the spatio-temporal data is described. The experiments were carried out using the EVT4 system and a 32-electrode sensor. The reconstruction results of a VOI from a single 3D measurement and spatio-temporal sampling are compared. A phantom for 3D imaging to assess spatial image resolution using a modulation transfer function is proposed.
查看更多>>摘要：Rotating machinery is widely used in industrial production facilities, and once a failure occurs, it can be catastrophic. Alerting to potential defects in time to prevent further equipment degradation is a challenging task. In this paper, a novel two-stage fault warning framework is proposed for early fault warning of rotating machinery. Specifically, a new method based on intra-class and inter-class neighborhood information graph embedding orthogonal discriminant projection is firstly adopted in this framework to extract the global distribution feature information and local geometric structure information of the data so that the homogeneous distance is compressed and the heterogeneous distance is distanced. Secondly, the minimum quantization error between the sample to be measured and the optimal winning neuron weight vector is calculated by self-organizing map to characterize the health state change, and combined with the Beta distribution self-learning technique to establish the fault warning threshold to circumvent the defects brought by the traditional fixation and it. Finally, the effectiveness of the proposed method is verified in the bearing and planetary gearbox test cases, and exciting conclusions are obtained under different working conditions in the gearbox case.
查看更多>>摘要：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 and E-rectified linear units (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.
查看更多>>摘要：Micro-force measurement with high resolution, accuracy, and reliability is of interest in a broad range of applications including gravitational-wave detection, intelligent healthcare, bionic robotics, and micromanipulation. Herein, the researchand development in recent years of micro-force sensors based on various principles is reviewed thoroughly, presenting their characteristics and applications, as well as summarizing their advantages and limitations. The most indispensable component of force sensors, elastic sensitive elements, is underlined. Next, four kinds of not widely used but promising sensors are also introduced briefly. Finally, the traceable reference forces are analyzed, concluding with a future perspective into the corresponding challenges and opportunities of micro-force sensors for future research. This review aims at providing references for developing micro-force sensors and improving their performance.
查看更多>>摘要：System-level remaining useful life (RUL) estimation is difficult due to multiple degrading components, external disturbances, and variable operational conditions. A similarity-based approach does not rely on health assessment and is more suitable for system-level RUL estimation. However, for practical applications, how to capture effective degradation features from raw data, how to fuse multiple nonlinear sensor data, and how to handle multiple source uncertainties need to be considered. To solve the above challenges, this study focuses on RUL estimation for systems under variable operational conditions. A similarity-based probabilistic RUL estimation strategy is proposed and verified using the NASA aeroengine dataset. First, measurement uncertainty can be addressed. Proper degradation features are extracted by three defined indicators. Subsequently, multiple nonlinear sensor data fusion and unsupervised synthesized health index construction can be realized using the proposed deep autoencoder-based polynomial regression approach. Finally, this strategy can handle the modeling and prediction uncertainties, including providing probabilistic RUL estimation results by well-trained residual-based similarity models. The verification results indicate the effectiveness and feasibility of the proposed strategy.
NSTL