摘要 : Real industrial scenarios struggle with the issues of a limited number of labeled samples and difficulty in accessing, which results in deep learning-based fault diagnosis models having poor generalization capabilities and decreased diagnostic accuracy. To address this problem, a semi-supervised prototype enhancement network (SSPENet) is proposed for rolling bearing fault diagnosis in this study. Firstly, a dual pooling attention residual network is proposed to be used in the feature extraction module. The goal is to efficiently extract the hidden features within rolling bearings, thus enabling the accurate classification of different sample categories. Subsequently, the Hungarian algorithm is utilized to design a strategic approach to update prototypes with pseudo-labels, which achieves the effect of augmenting prototypes by accurately adjusting the prototype position of each class of limited labeled samples through unlabeled samples, to improve the discriminative ability of the network model for fault classes. Finally, validation and experimental analysis are carried out on two bearing datasets, which achieve the average diagnostic accuracy of the proposed model to be above 90 % for both 1-shot and 2-shot cases, obtaining more satisfactory diagnostic results.

摘要 : Owing to the advantages of monolithic structure and little need for assembling, compliant guiding mechanisms appear to be an effective solution for decoupling multi-freedom precision motions but are still prone to geometric nonlinearities of parasitic error and stiffening effect for large strokes. This paper proposes a coiled L-shape compliant guiding mechanism featuring millimeter-scale strokes with a compact structure, constant stiffness, and minimized parasitic error. The coiled compliant guiding mechanism is formed by convolving L-shape flexure beams in a zigzag configuration with decoupled XY motions achieved. Its geometrically nonlinear parasitic error, variation in stiffness, and primary vibration are captured by using a dynamic beam constraint model (DBCM). It is theoretically, numerically, and experimentally found, by comparing with double parallel guiding mechanisms, that the kinetostatic and dynamic behaviors of the coiled L-shape compliant mechanism are nearly independent on the applied force within intermediate-deformation ranges. Such a weak geometric nonlinearity with the minimized influence of axially loaded stiffening and kinematics-arching effects is much different from the double parallel guiding mechanisms. The obtained results indicate that large strokes with constant stiffness and invariable resonance frequency can be realized, which also allows small parasitic errors.

摘要 : TC4 titanium alloy is utilized in aerospace and a wide range of other applications due to its high strength and corrosion resistance, but its poor thermal conductivity and high ductility introduce challenges in machining. In this paper, firstly, a mathematical model of workpiece surface morphology under tilted ultrasonic elliptical vibration cutting (TUEVC) was established based on the tip trajectory and material removal mechanism, and the variation of workpiece surface morphology was investigated at different tool tilt angles θ, and it was found that the tilt θ in a certain range could significantly reduce the surface residual height compared with that of ordinary ultrasonic elliptical vibration cutting (UEVC). Second, the TUEVC experiment of TC4 titanium alloy was carried out to comparatively analyze the changes in the surface morphology, surface profile, and surface roughness of the workpiece under different tilt θ, the effect of each machining parameter (cutting speed, feed, and ultrasonic amplitude) on surface roughness was explored. The experimental results indicate that as the tilt θ changes from 0° to 90° throughout the process, the workpiece surface morphology flatness decreases and then increases. When the tilt angle θ is 45°, workpiece cutting surface roughness is minimized ( Sa = 0.157), compared with the ordinary ultrasonic elliptical vibration cutting roughness is reduced by 47.6 % maximum. Both the surface morphology flatness and the surface roughness of the workpiece are at their smallest, whereas the theoretical profile curve and cutting surface profile curve are at their most consistent. Under the same machining parameters, TUEVC can reduce the surface roughness more effectively compared with UEVC, this technique reduces surface roughness by 16 %, 23 %, and 26 % at maximum for different cutting speeds, feeds, and ultrasonic amplitudes, respectively.

摘要 : Developing highly reliable infrared radiation materials with broadband high emissivity at high temperatures, low thermal conductivity, and excellent thermal stability is highly desirable for aerospace thermal protection applications. However, it remains a huge challenge to take into account infrared radiation heat dissipation and blocking heat transfer through low thermal conduction simultaneously. In this work, we reported a broadband high emissivity Ca 2+ -doped YbCrO 3 ceramic with emissivity above 0.89 at room temperature across the entire range of wavelength (1–14 μm). This doping strategy leads to the introduction of impurity energy levels into the band gap of YbCrO 3, which increases the possibility of light absorption to promote electron transition, improving the emissivity of the near-infrared band (1–3 μm). Simultaneously, un-equivalent doping induces electron exchange between chromate ceramic ions, which complicates the electronic structure (producing lattice distortion and extra multi-mode vibrations) and reduces the band gap width, thus boosting the emissivity in the mid-infrared band (3–14 μm). More importantly, (Yb 0.8 Ca 0.2 )CrO 3 presents a high emissivity (0.76) at an elevated temperature of 1200 °C, together with low thermal conductivity (2.5 W m −1 K −1 at 1000 °C) due to strong phonon scattering. Moreover, the doping-dominating phase stabilization effect contributes to impressive thermal stability (stable at 1300 °C for 50 h) and a high coefficient of thermal expansion (9.0–9.5 × 10 −6 K −1 ), which makes it suitable for long-term high-temperature thermal protection application. All these merits render the development of thermally stable high-temperature infrared radiation ceramic materials core competitive.

摘要 : Remaining useful life (RUL) prediction of rolling bearing is one of the important measures to ensure the reliable operation of mechanical equipment. Most of the existing methods are domain adaptation (DA) based RUL prediction on the same machine with different conditions, but few on cross-machine. DA can cope with the data distribution discrepancy (domain shift) under different machines or other conditions, but the potential negative transfer will affect the effect of DA and prediction performance. Therefore, an enhanced residual convolutional domain adaptation network (ERCDAN) is designed for cross-machine rolling bearing RUL prediction. Firstly, the enhanced residual convolutional module (ERCM) is designed for degradation feature extraction from limited data, and with the convolutional block attention module (CBAM) to enhance the extracted features. Secondly, the DA module with a collaborative full connection structure and attenuation multi-kernel maximum mean discrepancy is designed for mitigating negative transfer to effective domain-invariant feature extraction. Finally, the experimental analysis of cross-machine rolling bearing RUL prediction is conducted on the PHM2012, XJTU-SY, and EBFL datasets. The results show that the proposed method can not only effectively achieve cross-machine RUL prediction, but also has good cross-bearing prediction performance with different conditions on the same machine, reflecting good generalization performance.

摘要 : A three-parameter Weibull small sample aggregation method is proposed to address the problem of small sample fatigue data in practical engineering. First, based on the empirical assumption that the shape parameter is a constant and data equivalent conversion and aggregation, the small sample fatigue life data are converted into "equivalent large samples." Second, through the linear relationship between logarithmic scale parameter and stress, an iterative procedure is established to estimate the scale parameters of small sample data. At the same time, the moment and rank estimation methods are used to solve the location parameters quickly. Finally, P–S–N curves with different survival rates are given, and a 7-series aluminum alloy is selected to verify the correctness and reliability of the method proposed in this study. The results show that the predicted life obtained by the method still has high accuracy under a high survival rate, which can provide a theoretical basis for solving practical engineering problems.

摘要 : This investigation focuses on the feasibility of the reaction synthesis of Ti–C–N phase using plasma spraying powder mixture of Ti and graphite under nitrogen ion flame flow. To regularize the microstructure and properties of the sprayed Ti–C–N coating, the microstructure evolution of the Ti-graphite-N 2 system was analyzed based on water quenching tests. The reaction pathway of the Ti-graphite-N 2 system has been investigated by thermodynamic calculations and thermal analysis, revealing the mechanisms of the reaction during plasma spraying. The results show that the as-deposited coatings exhibit a typical layered structure consisting mainly of TiC x N y and TiO 2 . The quenched powders exhibit a melting process from irregular particles to smooth spherical particles, forming complete droplets with increasing spray distance. Based on phase evolution of the quenched powders and the thermal analysis of Ti powders or Ti/graphite powders under N 2 /Ar atmosphere, it found that the Ti powder can reacted with graphite or N 2 by solid-solid or solid-gas reaction to form TiC x and TiN y, and TiN y can be formed preferentially as compared with TiC x phase. As the spraying distance increases, the Ti and graphite reactions are enhanced, and a pronounced TiC x phase can be detected once a large liquid Ti phase is formed. As the spray distance increases, formed TiN y and TiC x phases aggregate into droplets and rapidly transform into TiC x N y phase by solid-solution reactions. In addition, titanium oxide coexists in lamellae of Ti–C–N coatings due to the absorption and dissociation of O2 from the atmosphere at the droplet surface.

摘要 : To improve the reliability of the dynamic system including physical and control design, the reliability-based control co-design (RB-CCD) problem has been studied to account for the uncertainty stemming from the random physical design. However, when encountering RB-CCD in the sophisticated system in which the dynamic model simulation is time-consuming or the state equation is expressed implicitly, the available RB-CCD methods will consume significant computational effort to perform numerous system simulations for the reliability analysis and deterministic optimization. Therefore, this work proposes a Dendrite Net-based decoupled framework for RB-CCD to alleviate the computational burden. Specifically, the Dendrite (DD) model constructed by the suggested training scheme integrated with an adaptive sampling strategy is used to approximate the state equation in the dynamic system. After that, the sequential optimization and reliability assessment method decouples RB-CCD into the control co-design (CCD) problem and time-dependent reliability assessment problem, which are solved sequentially based on the cheap estimations of DD model, rather than the expensive simulations of the original system. Furthermore, two numerical examples and an engineering example of 3-DOF robot system are applied to demonstrate the feasibility and efficiency of the proposed framework.

摘要 : Compliant amplification mechanisms based on the triangular principle have attracted considerable applications in precision and other engineering fields due to their compactness and efficient amplification capacity. However, a fast engineering design for geometric nonlinearity of large strokes is still much challenging. We report herein a series of analytical equations of nonlinear amplification ratio, input and output stiffness for three commonly-used triangular-amplified compliant mechanisms, namely the rhombus, diamond and bridge types. The pronounced geometric nonlinearities of axially-loaded stiffening and kinematics-arching effects are explicitly formulated. The proposed nonlinear formulas can be directly degenerated as the linear models by vanishing the nonlinear terms, which enables a comprehensive analysis of linear and nonlinear kinetostatics, and hence offers a straightforward way for the fast performance evaluation and size synthesis/optimization. This relieves an engineer's experience and knowledge for an ab initio modeling process. The prediction error is discussed and some insights into the linear and nonlinear characteristics of the three commonly-used triangular-amplified compliant mechanisms are outlined that confirms the advantage of analytical equations bringing to a fast engineering analysis and design.

摘要 : Recently, flexible pressure sensors rooting in piezoresistive mechanism have gained immense popularity due to their excellent application potential in the fields of software robotics, electronic skins and human health monitoring, etc. However, the trade-off between sensitivity and pressure range still remains challenging for the wide application of piezoresistive sensors . Herein, a novel flexible pressure sensor constructing from electrospun ionic liquids/Multi-walled carbon nanotubes/poly(vinylidene fluoride) (ILs/MWCNTs/PVDF) ternary composite that permits high-sensitivity detection over broad pressure range is implemented. Benefitting from the filamentous porous network coupled with urchin-shaped beads, the newly developed active ternary composite showed excellent compressibility and high sensitivity to pressure, endowing the resultant piezoresistive sensor with broad-pressure detection ranging from 1.0 Pa to 307 kPa with a maximum sensitivity of 0.075 kPa −1 . Simultaneously, the proposed sensor present fast response/recovery time of 120/80 ms and high operation stability over 14000 cyclic pressings. Also, the feasibility of sensor in effectively monitoring various physiological signals, recognizing keyboards and expressing spatial pressure distribution is demonstrated, implying their great potential as viable high-performance pressure sensor.