Search

To address the low accuracy of conventional bearing fault diagnosis methods that rely on empirical parameters in the scenarios with different noise degrees, this paper presents an adaptive bearing fault diagnosis method combining empirical mode decomposition (EMD) and tile coding. Firstly, the mature adaptive decomposition algorithm, namely the EMD, was employed to decompose the bearing vibration signal. Based on specific criteria, some components were selected as effective signals and further denoised based on spectral averaging to realize adaptive filtering of the bearing vibration signal. Secondly, feature extraction of the filtered signal was conducted using spectral coding based on tile coding. The coding results were input into the K-nearest neighbors (KNN) classifier as features. Finally, for validation, a test dataset with varying noise degrees was constructed to conduct a comparison between the proposed adaptive method and the classic adaptive filtering-based fault diagnosis methods. The results show that when the signal-to-noise ratio is greater than －10 dB, the diagnostic accuracy of the proposed approach exceeds 90% and outperforms that of other methods.

During the daily operation of locomotive, affected by the locomotive itself and external environmental factors, the height of the sand pipe may change, which will lead to wheel slippage failure in severe cases, resulting in safety problems. Therefore, a sand pipe height measurement algorithm based on deep learning is proposed in this paper. Firstly, the object detection of sand pipe in 2D image and the fitting segmentation of rail plane were realized through the deep-learning-based YOLOv3-tiny algorithm and DeepLabv3+ model. Secondly, for the noise point cloud in the measurement environment, this paper implements point cloud filtering through depth value filtering and clustering segmentation. For the foreign matters on the sand pipe that affect the height measurement, the edge extraction and false edge removal are designed to eliminate the influence of foreign matters on the measurement accuracy. Finally, the rail plane is fitted by RANSAC algorithm to achieve 3D high-precision measurement of sand pipe height. Using the proposed algorithm and a 10-division vernier caliper, a comparative experiment is conducted on the height measurement of sand pipes with different positions, shapes and light intensities. The result shows that using this algorithm, the height measurement error of sand pipes can be basically controlled within ± 1 mm, with an average measurement accuracy of 98.74%, indicating that the algorithm has a certain degree of robustness.

Inspection of pantograph state of EMU is one of the basic tasks of vehicle daily maintenance. There are three common inspection methods, including manual visual inspection, fixed-point image detection and on-board image detection. Due to the small number of pantograph fault samples that can be obtained, the present pantograph anomaly identification algorithms based on deep learning perform poorly. Therefore, this paper proposes an on-line recognition method based on feature analysis and independent of the number of fault samples. Visual features of the visible light images of pantograph which collected in real time through 3C device of EMU are used to identify abnormal states of pantograph. Firstly, the light intensity of the operating environment is judged by the image gray feature, and the region of interest (ROI) of the pantograph is segmented by the tracking matching algorithm. Then, the object features in the ROI of each frame such as threshold, contour thickness and tilt angle are extracted to detect abnormal states respectively, including sheep angle missing, front carbon sliding plate defect and bow head tilt, etc. Finally, the characteristic rules of pantograph anomaly determination are designed to realize the identification of abnormal state of pantograph. Test results show that the method and system proposed in this paper can detect abnormal states of the pantograph, such as missing horns, front carbon sliding blocks and bow head tilting, in both day and night environments. The average detection rate of defects is greater than 98%, which shows the method has great stability and practicability.

Aiming at the problem of insufficient detection frame size accuracy in the deep learning based wheelset tread defect detection algorithm, this paper proposes a tread defect detection algorithm based on YOLOv3-tiny and traditional image algorithm, which can realize fast positioning of defects with low CPU consumption and precise measurement of geometric parameters. First, image enhancement is performed on the small samples obtained in the industrial scene, and then the YOLOv3-tiny algorithm is used to perform migration learning on tread defects to achieve rough localization of defects. In order to solve the key problems that the detection frame is too large and too small, traditional image algorithms such as Fourier transform, band-stop filter, and threshold segmentation are used to construct a defect size measurement model, contours of roughly located defects are extracted and detection frame size is optimized, and finally location and size of defect are accurately calculated. Defect location experimental results show that the average accuracy of defect identification is 89.4% and the CPU consumption does not exceed 10% when IoU threshold is 0.5. Experimental results of defect measurement show that the algorithm can optimize 74 out of 90 inspection frames and obtain a more accurate defect size. The above experimental results show that the detection algorithm in this paper is effective in improving size accuracy of detection frame.

Periodic detection of wheel tread is necessary for train operation safety. Reference positioning of inner side benchmark is a traditional method for tread detection but there are problems of positioning error caused by field factors in wheel tread dynamic detection such as reference tilt by hunting, foreign matter and light interference. A reference positioning method for wheel tread dynamic detection is proposed in this paper. After extracting the point cloud data through structured light calibration, center-line extraction and other algorithms, reference feature points of the inner side are segmented combined with tread features, and an iterative re-weighted least squares line fitting (IRLS-LF) method is used to realize robust positioning of the inner side benchmark. Under the dynamic tilt condition, the experimental man-machine comparison deviations of flange height and thickness based on IRLS-LF positioning results are ±0.1mm and ±0.2mm respectively, and the both deviation range widths based on LSLF positioning results and fixed-parameter algorithm positioning results are about 0.8 mm. Experimental results show that this method can effectively solve the datum positioning deviation caused by field factors, and effectively ensure the measurement accuracy and robustness of wheel tread dynamic detection.

The environment of operation and maintenance of locomotives and rolling stocks is complex. Factors such as dust and reflection make point cloud missing, noise and false points, which seriously affect the accuracy of locomotive robot brake shoe image detection. To this end, this paper presents an size measurement algorithm of locomotive brake shoe for locomotive robot based on 3D image. Firstly, it realizes automatic transformation between the template region of interest (ROI) and measurement ROI according to the camera perspective transformation and point cloud registration algorithm, which makes up for the repeated positioning error of the locomotive robot to a certain extent. For the problem of point cloud missing and noise caused by reflection, RANSAC algorithm is used to fit the reference ROI in plane to make up for the loss of precision. The algorithm of automatic filtering of ordered point cloud is used to filter the point cloud between brake disc and brake shoe effectively. For the difficulty of edge discrimination by the false point cloud caused by lock brake and specular reflection of brake disc surface, this paper designs a filtering operator to preliminarily separate the edge of brake shoe and extract the ROI on the side of the brake shoe. Then, according to the imaging characteristics of the brake shoe, the gradient-based edge extraction and tracking algorithm is used to realize the accurate extraction of point cloud boundary of brake shoe and the accurate measurement of brake release gap. Finally, the accurate dimension measurement is realized by using the statistical method of sorting strategy. By the algorithm presented in this paper, accurate measurement can be realized under the condition that slack gap is less than 2mm. The method is used to measure the brake shoe in different shape, position and lighting complex environment, and the Vernier Caliper is used to measure it. The results show that the measuring accuracy of brake shoe wear size is within ±1mm range, and the algorithm has certain robustness.

Single cell sequencing (SCS) is a technique for precise analysis of cells at an individual level, which provides a new perspective on the identification of individual cell types, functions, states and interactions with other cells, and greatly deepens the understanding of cellular heterogeneity and genetic and molecular functional diversity. With the continuous optimization of various technical aspects, SCS technology has been widely used in many research areas, and the study of the hematopoietic system is an example of its application. SCS has not only assisted in the construction of relevant hematopoietic cell profiles under physiological and pathophysiological conditions in basic research, but also explored possible pathways of hematopoietic cell development and differentiation in clinical practice. Moreover, in the process of clinical diagnosis and treatment, it can effectively promote the deciphering of mechanisms of disease occurrence, progression and prognosis, and facilitate the exploration of potential factors of phenomena such as significant differences in individual outcomes and ease of relapse. Therefore, this paper reviews the technical overview and development trend of SCS and its application in the hematopoietic system, with a view to providing references for subsequent positive and reasonable application of SCS.

Rapid and high-sensitive detection of mycotoxins is believed to be of vital importance in assuring food safety. In this study, we developed a novel fluorescence immunochromatographic sensor (ICS) for the mycotoxin of zearalenone (ZEN) in cereals. This was done by using a black phosphorus-Au nanocomposite (BP-Au) as the 2D quenching platform. Herein, gold nanoparticles (AuNPs) were directly reduced on the surface of BP nanosheets (BPNSs) to form BP-Au nanocomposites, showing higher fluorescence (quantum dots, λEm = 525 nm) quenching efficiency compared to the BPNSs and AuNPs. The fluorescence quenching efficiency of the prepared BP-Au nanocomposite reached 73.8%, which was 1.73-fold and 1.44-fold higher than AuNPs and BPNSs, respectively. The density functional theory was also successfully used to explore the formation mechanism of the BP-Au nanocomposite. By introducing the quantum dots/BP-Au signal/quencher pair, a high-sensitive fluorescence quenching ICS (B-FICS) was developed for the detection and discrimination of ZEN with the limit of detection of 0.1 μg/l in pure working buffer. This was 2.5-fold more sensitive than AuNPs-based FICS (A-FICS). The B-FICS was successfully applied in real cereals detection with the sample limit of detection of 2 μg/kg. The successful construction of B-FICS offers a novel method for a rapid and high-sensitive detection of ZEN in cereals. It also provides a new practical application of 2D BPNSs in food safety sensing.

Slider is one of the main parts of a rocket projectile,which determines the stability and safety of a rocket launch. A slider structure design that satisfies the suspended-type launch was introduced in this research. Based on the theoretical analysis of offtrack speed and time,the slider structure was designed. By using numerical simulations,the strength of front slider under two different situations,ontrack and offtrack,was checked. The results show that the strength of slider meets the design requirement. The designed slider was applied to a rocket projectile flying test and the design feasibility was verified.

As the national policy encourage to reduce the fuel consumption and oil prices rising gradually,to reduce fuel consumption of the vehicle is widely concerned.The fuel consumption is closely related to the transmission efficiency,and the rear axle plays an important role in auto transmission system and different lubricant has a great influence on the transmission efficiency of rear axle,the viscosity of the lubricant,the losses of gear meshing and churning losses are analyzed,a lot of experiments are conducted to investigate the influence of viscosity of lubricant on transmission efficiency of rear axle.It is found that relatively low viscosity lubricant contributes to improving the transmission efficiency of rear axle and makes an important significance on reducing vehicle fuel consumption.

With the target of reducing the friction torque of tapered roller bearing,the globally optimal solution of bearing inner structural parameters is obtained by using the genetic algorithm.The friction torque,gearbox transmission efficiency,and vehicle fuel consumption and sliding distance are tested before and after the bearings are optimized.The results show that,the bearing friction torque is reduced,the gearbox transmission efficiency is increased,and the vehicle fuel consumption is slightly decreased while the vehicle sliding distance is highly improved.