Monitoring and diagnostics of stresses in reciprocating machines

This work presents a study for monitoring stresses in links of reciprocating machines by means of indirect measurements. A model of such a machine is a slider-crank mechanism actuated by an asynchronous electric motor with reduction gear. For this study, an experimental simulation apparatus was constructed, composed of a reciprocating machine, a direct-current tachometer as diagnostic sensor, and the necessary instrumentation for measurement of loads acting on the slider and thus stresses on the connecting rod.

The excellent agreement between the theoretical and experimental results shows the validity of the solution for monitoring and diagnostics of reciprocating machines using a simple direct-current tachometer connected to an A/D converter and then to a PC able to execute the fast Fourier transforms of the acquired signals.

The proposed method is particularly useful when direct measurements of stresses are difficult or particularly laborious. It also allows continuous monitoring of the machine throughout its life time, with user access or denial to computerized information, as desired.     

Optimal filtering of gear signals for early damage detection based on the spectral kurtosis

In this paper, we propose a methodology for the enhancement of small transients in gear vibration signals in order to detect local tooth faults, such as pitting, at an early stage of damage. We propose to apply the optimal denoising (Wiener) filter based on the spectral kurtosis (SK). The originality is to estimate and apply this filter to the gear residual signal, as classically obtained after removing the mesh harmonics from the time synchronous average (TSA). This presents several advantages over the direct estimation from the raw vibration signal: improved signal/noise ratio, reduced interferences from other stages of the gearbox and easier detection of excited structural resonance(s) within the range of the mesh harmonic components. From the SK-based filtered residual signal, called SK-residual, we define the local power as the smoothed squared envelope, which reflects both the energy and the degree of non-stationarity of the fault-induced transients. The methodology is then applied to an industrial case and shows the possibility of detection of relatively small tooth surface pitting (less than 10%) in a two-stage helical reduction gearbox. The adjustment of the resolution for the SK estimation appears to be optimal when the length of the analysis window is approximately matched with the mesh period of the gear. The proposed approach is also compared to an inverse filtering (blind deconvolution) approach. However, the latter turns out to be more unstable and sensitive to noise and shows a lower degree of separation, quantified by the Fisher criterion, between the estimated diagnostic features in the pitted and unpitted cases. Thus, the proposed optimal filtering methodology based on the SK appears to be well adapted for the early detection of local tooth damage in gears.     

Optimal machining cycles for numerically controlled metal-cutting machines

The design of optimal automatic machining cycles for numerically controlled metal-cutting machines is considered, in theoretical terms. Dynamic programming permits the optimization of an unlimited number of control parameters for all the machining operations in numerically controlled metal-cutting machines.     

Optimal materials for the manufacture of metal-cutting machines

With the goal of improving machine-tool design, attention focuses on the possibility of manufacturing the components of metal-cutting machines from effective new nonmetallic materials that not only meet all the current requirements on such machines but also in most cases improve their performance. The best results are obtained with metal–concrete structures in which the nonmetallic binder consists of modified cement with chemically active additives. Results from materials science are compared with multiyear production tests of metal-cutting machines based on metal–concrete components, which are characterized by stable performance over time. On that basis, an effective new approach is proposed for the modernization of machine tools at the end of their effective working life, without replacing the fundamental structures. The design of the basic structures is analyzed. The production of metal–concrete components in a specialized department of the manufacturing plant is recommended.     

Practical constraints on real time Bayesian filtering for NDE applications

An experimental evaluation of Bayesian positional filtering algorithms applied to mobile robots for Non-Destructive Evaluation is presented using multiple positional sensing data – a real time, on-robot implementation of an Extended Kalman and Particle filter was used to control a robot performing representative raster scanning of a sample. Both absolute and relative positioning were employed – the absolute being an indoor acoustic GPS system that required careful calibration. The performance of the tracking algorithms are compared in terms of computational cost and the accuracy of trajectory estimates. It is demonstrated that for real time NDE scanning, the Extended Kalman Filter is a more sensible choice given the high computational overhead for the Particle filter.     

Fault diagnostics based on particle swarm optimisation and support vector machines

In the fault diagnosis based on support vector machines (SVM), irrelevant variables in the fault samples spoil the performance of the SVM classifier and reduce the recognition accuracy. On the other hand, some SVM parameters are usually selected artificially, which hampers the efficiency of the SVM algorithm in practical applications. A new method that jointly optimises the feature selection and the SVM parameters with a modified discrete particle swarm optimisation is presented in this paper. A correct ratio based on a new evaluation method is used to estimate the performance of the SVM, and serves as the target function in the optimisation problem. A hybrid vector that describes both the fault features and the SVM parameters is taken as the constraint condition. This new method can select the best fault features in a shorter time, and improves the performance of the SVM classifier, and has fewer errors and a better real-time capacity than the method based on principal component analysis (PCA) and SVM, or the method based on Genetic Algorithm (GA) and SVM, as shown in the application of fault diagnosis of the turbo pump rotor.     

Multivariate statistical signal processing technique for fault detection and diagnostics

The normal fluctuation of wideband signals in process industry systems exhibits behavior that is characteristic of process dynamics, sensor dynamics, vibration of components, and product quality. A baseline statistical signature behavior can be established by a systematic processing of multivariate signals and determining the cause and effect relationship among the process variables characterizing a subsystem. Both theoretical and computational basis for processing a set of signals using the multivariate autoregression (MAR) modeling has been developed and applied to establish frequency domain statistical signatures for an aluminum rolling mill. A systematic procedure is developed to interpret the causal relationships for the detection and isolation of process anomalies and sensor maloperation. This digital signal processing technique and its implementation have clearly demonstrated the applicability of this method of characterizing and monitoring complex industrial processes.     

The spectral kurtosis: application to the vibratory surveillance and diagnostics of rotating machines

In the previous paper, the authors have demonstrated the high potential of the spectral kurtosis (SK) to detect and characterise non-stationary signals. The present paper brings together these ideas and shows how the SK can be efficiently used in the vibration-based condition monitoring of rotating machines. First, and in contrast to classical kurtosis analysis, the SK provides a robust way of detecting incipient faults even in the presence of strong masking noise. Second, the SK offers an almost unique way of designing optimal filters for filtering out the mechanical signature of faults. The first property is of practical importance for monitoring purposes, whereas the second one proves very useful in diagnostics. Another originality of the paper is the introduction of the concept of kurtogram, from which optimal band-pass filters can be deduced, for instance as a prelude to envelope analysis. All the original findings presented in the paper are illustrated using actual industrial cases.     

Application of Bayesian networks for diagnostics in the assembly process by considering small measurement data sets

The ability of variation source diagnosis in the auto body assembly process plays an essential role in the success of the manufacturing enterprises. However, it is more challenging to identify the process faults associated with the compliant sheet metal parts based on small measurement data sets. A new Bayesian networks (BN) modeling approach under the condition of small data sets is proposed. The main causal links are identified based on mapping of the variation sensitivity matrix. The interaction effects are detected according to the conditional mutual information tests. After the network structure is determined, the Bayesian approach is used to obtain the conditional probability tables by incorporating prior probability distributions. The evaluation of diagnostic performance concerning evidence number and log-odds noise levels is also presented. A real bracket assembly case was used to illustrate the whole procedures for fixture fault diagnosis. The examined test cases demonstrate the proposed BN approach is practical and effective, even when incomplete evidences are observed and a medium-level noise is present.     

A Bayesian machine learning method for sensor selection and fusion with application to on-board fault diagnostics

In applications like feature-level sensor fusion, the problem of selecting an optimal number of sensors can lead to reduced maintenance costs and the creation of compact online databases for future use. This problem of sensor selection can be reduced to the problem of selecting an optimal set of groups of features during model selection. This is a more complex problem than the problem of feature selection, which has been recognized as a key aspect of statistical model identification. This work proposes a new algorithm based on the use of a Bayesian framework for the purpose of selecting groups of features during regression and classification. The hierarchical Bayesian formulation introduces grouping for the parameters of a generalized linear model and the model hyper-parameters are estimated using an empirical Bayes procedure. A novel aspect of the algorithm is its ability to simultaneously perform feature selection within groups to reduce over-fitting of the data. Further, the parameters obtained from this algorithm can be used to obtain a rank order among the selected sensors. The performance of the algorithm is first tested on a synthetic regression example. Finally, it is applied to the problem of fault detection in diesel engines (30,000 data records from 43 sensors, 8 classes) and used to compare the misclassification rates with a varying number of sensors.     

Optimal parallel machines scheduling with machine availability and eligibility constraints

In reality, the machine might become unavailable due to machine breakdowns or various inevitable reasons, and machine might have different capability to processing job. Motivated by this, we consider the problem of scheduling n non-preemptive and independent jobs on m identical machines incorporating machine availability and eligibility constraints while minimizing the maximum lateness. Each machine is capable of processing at specific availability intervals. We develop a branch and bound algorithm applying several immediate selection rules for solving this scheduling problem.     

Bayesian model for detection and classification of meningioma nuclei in microscopic images

Image segmentation aims to determine structures of interest inside a digital picture in biomedical sciences. State‐of‐the art automatic methods however still fail to provide the segmentation quality achievable by humans who employ expert knowledge and use software to mark target structures on an image. Manual segmentation is time‐consuming, tedious and suffers from interoperator variability, thus not serving the requirements of daily use well. Therefore, the approach presented here abandons the goal of full‐fledged segmentation and settles for the localization of circular objects in photographs (10 training images and 20 testing images with several hundreds of nuclei each). A fully trainable softcore interaction point process model was hence fit to the most likely locations of nuclei of meningioma cells. The Broad Bioimage Benchmark Collection/SIMCEP data set of virtual cells served as controls. A ‘colour deconvolution’ algorithm was integrated to determine (based on anti‐Ki67 immunohistochemistry) which real cells might have the potential to proliferate. In addition, a density parameter of the underlying Bayesian model was estimated. Immunohistochemistry results were ‘simulated'for the virtual cells. The system yielded true positive (TP) rates in the detection and classification of real nuclei and their virtual counterparts. These hits outnumbered those obtained from the public domain image processing software ImageJ by 10%. The method introduced here can be trained to function not only in medicine and morphology‐based systems biology but in other application domains as well. The algorithm lends itself to an automated approach that constitutes a valuable tool which is easy to use and generates acceptable results quickly.     

基于智能算法的全断面掘进机刀盘优化设计

分析了全断面掘进机刀盘的破岩机理,提出了如何提高刀盘破岩性能的方案.构建了带复杂约束的多变量非线性目标优化数学模型,采用智能算法开发了一种针对盾构刀盘上刀具布置规律的优化程序,以EPB(Earth Pressure Balance)6.28型全断面掘进机刀盘为例进行了优化设计.通过对两种刀盘刀具磨损性能的比较分析,数据结果表明与原刀盘相比新设计的刀盘具有更加良好的实用性.     

基于非线性扩散滤波的边缘检测和图像测量

提出了一种基于自适应非线性扩散方程滤波的边缘检测方法, 以提高图像测量的精度。首先对原始图像实施一种非线性扩散处理,即沿着等照度线(边缘)的梯度方向实施反向扩散以锐化边缘, 而相反地沿切线方向实施正向扩散以去除噪声和锯齿伪像; 然后应用经典的微分算子来检测边缘。实验结果表明,相对于经典的边缘检测算子, 本算法得到了尖锐而平滑的单像素宽的图像边缘,较好地定位了边缘, 相对误差为0.03。当图像边缘模糊和存在附加噪声时, 测量结果将会受到很大影响。本方法较好地定位了边缘像素, 对于微小尺寸测量显示出它的优越性.     

Computational and experimental methods in the systems of current diagnostics of thermal stresses in elements of machines and structures

Application of the inverse-problem methods of experimental mechanics for diagnosing thermal states of elements and structures is considered. An inverse problem of nonstationary thermal conductivity and an iteration method for its solution on the basis of an alternating procedure are studied.     

Fault detection and classification in automated assembly machines using machine vision

The geometry of rotary aircraft engine components is usually defined by thin mechanical elements and complex surfaces that are only achievable by 5-axis machining due to either limited access or the design itself. Such thin-walled characteristics make these components susceptible to vibrations while machining and usually require careful manipulation of the toolpath parameters to minimize cutting forces and vibration. Moreover, the tool suppliers’ approach leans towards the feature-build design of cutter geometry to increase the productivity and quality of a machined surface. Some examples of those recent improvements for rotary aircraft engine components are barrel-shaped tools that attempt to increase the contact radius on the tool-part interface to minimize step-over while conserving the scallop height to meet roughness tolerances. This research aims to fill a gap in the current literature by proposing a stability model for barrel-shaped tools. Stability contour maps make use of a mechanistic dynamic force model for barrel-shaped tools. The force model is also capable of including tool runout and orientation angles, tilt and lead, as named in most CAM software. By simulating dynamic forces on the time domain, a contour map is presented to address unstable vibrations. Since forced vibrations and surface location error (SLE) may also appear when milling aircraft parts, SLE and surface roughness are also determined. Finally, given the complexity and number of parameters, validation of the stability maps is performed through experimental chatter tests using a thin wall component.     

Generalized singular value decomposition in multidimensional condition monitoring of machines—A proposal of comparative diagnostics

With the modern metrology, we can measure almost all variables in the phenomenon field of working machine, and many of measuring quantities can be symptoms of machine condition. On this basis, we can form the symptom observation matrix (SOM) for condition monitoring. From the other side we know that contemporary complex machines may have many modes of failure, so-called faults, which form the fault space. Even if we apply some modern tool like singular value decomposition (SVD) for the fault extraction purpose, this multidimensional problem is not a simple one. Therefore, the question remains if one can learn considering similar problem when having SOM of similar machine observed just before. In this way, we can consider the application of generalized singular value decomposition (GSVD) to the machine condition monitoring problems, and uncover some new possibilities.     

模糊图像边缘精确定位的滤波算法

图像的边缘是图像最基本也是最重要的特征之一。实际应用中,光学系统离焦和图像传感器噪声源引起图像模糊,导致边缘自动定位算法无法精确确定边缘点。从边缘检测理论出发,根据噪声及图像模糊模型,提出了基于空间域和频域的滤波方法对模糊边缘图像去噪,从而消除模糊图像对于边缘定位的影响。实验表明,各种滤波方法对于边缘点的自动提取都具有一定的效果,离焦模糊和噪声可以得到有效抑制,尤其是频域低通滤波器能够精确地提取出模糊边缘位置。