Feature selection for ECG signal processing using improved genetic algorithm and empirical mode decomposition

This paper proposes a novel scheme of feature selection, which employs a modified genetic algorithm that uses a variable-range searching strategy and empirical mode decomposition (EMD). Combined with support vector machines (SVMs), a new pattern recognition method for electrocardiograph (ECG) is developed. First, the ECG signal is decomposed into intrinsic mode functions (IMFs) that represent signal characteristics with sample oscillatory modes. Then, the modified genetic algorithm with variable-range encoding and dynamic searching strategy is used to optimize statistical feature subsets. Next, a statistical model based on receiver operating characteristic (ROC) analysis is developed to select the dominant features. Finally, the SVM-based pattern recognition model is used to classify different ECG patterns. Comparative studies with peer-reviewed results and two other well-known feature selection methods demonstrate that the proposed method can select dominant features in processing ECG signal, and achieve better classification performance with lower feature dimensionality.     

Modelling the influence of machining dynamics on surface topography in turning

In finish machining, even with high precision machines, the influence of the dynamic behaviour of the machine tool and workpiece on the surface quality can be significant. This is especially true during interrupted cutting or the machining of ultra-hard materials. A method for modelling the influence of the machine dynamics, as part of an overall machining simulation framework, is presented here. Finally, two examples of the use of this simulation framework are presented and compared to typical machined surfaces and to simulation methodologies presented in the literature.     

Evaluation of the wavelet transform method for machined surface topography 2: fractal characteristic analysis

In this paper, a new method based on wavelet transform is proposed as a means for studying the fractal characteristics of rough surfaces. Through estimation of normal mathematical curves with known fractal dimensions, generated by the Weierstrass-Mandelbrot function, Majumdar-Bhushan function, Fractal Brownian motion (including three methods: the Midpoint FBm, the Additions FBm, the Interpolated FBm) and Interposed method (Kiesswetter curve), it is validated that the wavelet transform method can accurately calculate the fractal dimension. These fractal functions have been used to simulate some surface profiles. The results indicate that the Wavelet transform method is the most precise in its calculation of the fractal dimensions of the curves. It obtains more accurate results than seven other methods, named the Box counting method, the Yardstick method, the Co-variation method, the Structure function method, the Variation method, the Power Spectrum method and the Rescaled range analysis method. Precisely calculating the fractal dimensions of the curves is the first step in characterising machined surface topography. In addition, this paper aims to further develop the evaluation procedure for the fractal characteristics of machined surface topography.     

A dynamic surface topography model for the prediction of nano-surface generation in ultra-precision machining

Materials induced vibration has its origin in the variation of micro-cutting forces caused by the changing crystallographic orientation of the material being cut. It is a kind of self-excited vibration which is inherent in a cutting system for crystalline materials. The captioned vibration results in a local variation of surface roughness of a diamond turned surface. In this paper, a dynamic surface topography model is proposed to predict the materials induced vibration and its effect on the surface generation in ultra-precision machining. The model takes into account the effect of machining parameters, the tool geometry, the relative tool–work motion as well as the crystallographic orientation of the materials being cut. A series of cutting experiments was performed to verify the performance of the model and good correlation has been found between the experimental and simulation results.     

Pure harmonics extracting from time-varying power signal based on improved empirical mode decomposition

Harmonic decomposition makes a great impact on power system operation, especially devices with frequency converters are widely used in modern society. In this paper, a hybrid method based on improved empirical mode decomposition enhanced with masking signals is presented to extract single-frequency harmonics from disturbed power signals accurately. The parameters for building masking signals are optimized by cooperative chaotic particle swarm optimization, where the Logistic chaos and cooperative evolution are employed to improve the convergence accuracy and avoid trapping into local minima. For improving the performance further, the improved fast Fourier transform based on Nuttall window and harmonics pre-extracting procedure are introduced to enhance the decomposition accuracy and reduce the instantaneous magnitude error in extracting time-varying power signal. The synthetic and field experiments demonstrate that the proposed method reveals significant improvements in the integrality and decomposition accuracy of harmonics extracted from time-varying power signal.     

A simulated investigation on the machining instability and dynamic surface generation

In this paper, the authors propose the generic concept of machining instability based on the analysis of all kinds of machining instable behaviours and their features. The investigation covers all aspects of the machining process, including the machine tool structural response, cutting process variables, tooling geometry and workpiece material property in a full dynamic scenario. The paper presents a novel approach for coping with the sophisticated machining instability and enabling better understanding of its affect on the surface generation through a combination of the numerical method with the characteristic equations and using block diagrams/functions to represent implicit equations and non-linear factors. It therefore avoids the lengthy algebraic manipulations in deriving the outcome and the solution scheme is thus simple, robust and intuitive. Several machining case studies and their simulation results demonstrate the proposed approach is feasible for shop floor CNC machining optimisation in particular. The results also indicate the proposed approach is useful to monitor the machining instability and surface topography and to be potentially applied in adaptive control of the instability in real time.     

Evaluation of surface topography parameters for friction prediction in stamping

The continued globalisation of the automotive industry, leading to increasing demands for competitiveness and escalating legislative requirements, is the main driving force of research activities of steel sheet surfaces. Recent studies on the stamping process have been carried out among others within AUTOsurf, a project funded by the European Community, and by Wihlborg and Crafoord. The purpose of this study was to evaluate the viability of the proposed parameters for friction prediction. Seventeen different surface topographies were investigated. The sheet materials were either, hot-dip galvanised, electrogalvanised or galvannealed, and electron beam or electric discharged textures. The frictional response was measured in a bending under tension (BUT) test under mixed lubricated conditions. This BUT test simulates the conditions of the die radius in a stamping tool. The laboratory test differs from the experimental work performed in AUTOsurf which simulated the conditions of the holding-down plate. In spite of the differences in test equipments in AUTOsurf, e.g. the rotational friction tester (RTF), on a comparison the correlation of frictional response was significant. But neither of the proposed parameters could predict the frictional response with sufficient accuracy in this study. In addition, the friction model in AUTOsurf describes peak lubrication as a dragging phenomenon on sliding surfaces. The movement eased friction in inverse proportion to the average peak area. However, the trend in this study showed the opposite, movement eased friction proportionally to the average peak area. The result indicates a switch of dominant friction mechanism when the sliding velocity is increased, i.e. from a dragging phenomenon at low velocities to micro-hydrodynamic wedge effects at high velocities.     

A study on the effect of surface topography on rough friction in roller contact

The friction behaviour of gear teeth in the context of tribology can have a strong effect on housing vibration, noise and efficiency. One of the parameters that greatly influences the friction under certain running conditions is surface roughness. In this work, rough friction was studied in lubricated sliding of roller surfaces, which were manufactured to simulate the real gear surfaces. By examining 3D surface topography of two mating bodies, both surface roughness and its effect on friction behaviour can be studied. In a previous study, a rough-friction test rig has been designed, constructed and initially verified. The types of surfaces involved in this study are ground, shot-peened, phosphated and electrochemically deburred. These rollers were subjected to the same friction testing procedures. Roller surfaces were then examined, and correlation between the topography and the frictional behaviour was analysed. Friction behaviour was interpreted in terms of Stribeck curves (friction coefficient as the function of Hersey parameter (ην/p)). The results showed that electrochemically deburred and certain phosphated surfaces provide lower friction coefficient values which are competitive to fine-ground surfaces in lubricated rolling/sliding contact.     

The influence of stylus flight on change of surface topography parameters

Theoretical and experimental work on stylus flight is described. The paper describes the development of a simulation model for assessing the magnitude of surface topography distortion by stylus flight. Experiments on the surfaces support the theoretical model, which predicts stylus flight. The measurements of different surface topographies (including surfaces after grinding, turning, honing, milling) were done using Talyscan 150 measuring instrument with four traversing speeds (0.5, 1, 2 and 3 mm/s). The results of theoretical considerations and experiments were compared. The effect of stylus flight on surface topography parameters of measured surfaces basing on experimental investigation was assessed. The tendency was found that slope decreased, decrease of amplitude parameters and increase of horizontal parameters took place, but these effects were different for various surface types. The simulation procedure assured good accuracy of surface topography parameters changes.Based on theoretical investigation, the effect of stylus flight and stylus tip radius on parameters of computer-generated profiles was predicted. The choice of traversing speed to different types of surfaces was done. The parameters of biggest changes caused by error in measurement due to stylus kinematics were selected.     

钻头磨损检测与剩余寿命评估

对钻头的磨损程度进行实时检测有助于对钻削加工过程实施预防性维护,提醒及时换刀。针对自动化生产中的刀具监测问题,给出一个基于主轴电流检测的钻头磨损状态分析和剩余寿命预测的应用策略。通过主轴电流传感器采样加工过程的电流信号,使用一个滑动窗口从连续采样数据中得到真实加工段数据,采用小波包分解的方法进行特征提取。基于Fisher标准筛选出最能表达磨损过程的若干特征。最后利用逻辑回归法和自回归滑动平均模型相结合的方法评估当前钻削加工的可靠性,预测钻头的剩余寿命。试验证明此方法的有效性,可为换刀决策提供依据。     

Relative vibration identification of cutter and workpiece based on improved bidimensional empirical mode decomposition

Frontiers of Mechanical Engineering - In the process of cutting, the relative vibration between the cutter and the workpiece has an important effect on the surface topography. In this study, the...     

Semi-supervised weighted kernel clustering based on gravitational search for fault diagnosis

Supervised learning method, like support vector machine (SVM), has been widely applied in diagnosing known faults, however this kind of method fails to work correctly when new or unknown fault occurs. Traditional unsupervised kernel clustering can be used for unknown fault diagnosis, but it could not make use of the historical classification information to improve diagnosis accuracy. In this paper, a semi-supervised kernel clustering model is designed to diagnose known and unknown faults. At first, a novel semi-supervised weighted kernel clustering algorithm based on gravitational search (SWKC-GS) is proposed for clustering of dataset composed of labeled and unlabeled fault samples. The clustering model of SWKC-GS is defined based on wrong classification rate of labeled samples and fuzzy clustering index on the whole dataset. Gravitational search algorithm (GSA) is used to solve the clustering model, while centers of clusters, feature weights and parameter of kernel function are selected as optimization variables. And then, new fault samples are identified and diagnosed by calculating the weighted kernel distance between them and the fault cluster centers. If the fault samples are unknown, they will be added in historical dataset and the SWKC-GS is used to partition the mixed dataset and update the clustering results for diagnosing new fault. In experiments, the proposed method has been applied in fault diagnosis for rotatory bearing, while SWKC-GS has been compared not only with traditional clustering methods, but also with SVM and neural network, for known fault diagnosis. In addition, the proposed method has also been applied in unknown fault diagnosis. The results have shown effectiveness of the proposed method in achieving expected diagnosis accuracy for both known and unknown faults of rotatory bearing.     

Simulation of surface topography with the method of movable cellular automata

Development of surface topography of two solids in a frictional contact is studied with the method of movable cellular automata. After the running-in process, the bodies are separated and the surface power spectra of both bodies are determined. The power spectra show a dependence on the wave vector, which is typical for fractal surfaces. It is shown, that roughness parameters of friction surface depend on relative velocity of sliding and external pressure.     

Dynamic interaction of friction and surface topography in brake systems

In brake systems the surface topography of the brake pad is characterized by an equilibrium of flow of hard patches, which are built up and destroyed, depending on the load history. Since these patches carry the main friction power, the friction coefficient itself depends on the load history, too. So the friction coefficient has to be described by differential equations instead of algebraic equations, connecting intimately the friction- and wear process. With these dynamic dependencies we are able to build up a new model, using the technique of Cellular Automata, which gives a dynamic and detailed insight into the brake pad's surface topography during the friction process.     

Vibration-resistant interference microscope with assistant focusing for on-machine measurement of surface topography

The interference microscope is a powerful tool for surface topography measurement, but its high sensitivity to vibration hinders its application to on-machine use. To measure surface roughness on a machine for the ultra-precision machining, a vibration-resistant interference microscope (VRIM) with an assistant focusing function is developed. The basic principle of VRIM is an error-compensated phase-shifting interferometry. An iterative algorithm is presented to calculate the surface phase with the phase shift amounts as unknown variables, where the phase shift amounts are calculated and compensated with least-squares method. A narrow bandwidth illumination is employed to alleviate coherence envelop influence, and a simplified intensity model is established to decouple the variables. Assisting the microscope to find fringe quickly, the focusing is realized by introducing an off-axis thin beam to generate two spots, of which their relative position relates to the defocus. The focusing method is directional and determinant, and has a large range up to 0.3 mm. In the vibration disturbances of 0.2 μm and 0.4 μm amplitudes over 0 Hz to 20 Hz frequency region, the roughness accuracy and repeatability of measuring an ultra-precision machined surface are both up to the sub-nanometer level. The developed instrument is applied to a single-point diamond turning machine and achieves a sub-nanometer accuracy and repeatability.     

二维经验模态分解在工程表面形貌误差评定中的应用

针对目前表面微观形貌面形误差分离方法中存在边界畸变及自适应差等缺点,提出了将具有自适应时频分辨能力的二维经验模态分解算法(bidimensional empirical mode decomposition,BEMD)应用于三维工程表面面型误差分离中,同时用Riesz变换构造单演信号,计算信号整体的频率特征,完成对二维经验模态分解算法的终止准则的改进,使其能严格按照ISO4287所规定的截止波长分离三维表面各频段形貌误差。仿真结果表明,本文新方法相比于国标ISO中高斯滤波以及常用小波滤波,在分离三维工程粗糙表面各面型误差时,所得分离图形效果远优于传统方法所得,且各频段误差对应的三维评定参数误差均小于5%。最后对光学镀膜元件实例进行分析,结果表明该算法能够很好地分离各形貌误差的的空间信息,所得参数评定基准面相对传统方法不存在边界畸变等问题,因此该方法在实际工程表面评定应用中具有可行性。     

Modeling and prediction for 3D surface topography in finish turning with conventional and wiper inserts

Wiper insert have the characteristics of achieving an excellent surface finish and improving the productivity in turning processes. Wiper insert can provide twice feed rate while maintaining the comparable surface roughness compared to that provided by the conventional insert. In the present study, surface topographies in finish turning with conventional and wiper inserts are investigated. The key element of this work is that the cutting edge path equation in the cutting tool coordinate system is transformed into the machine tool and workpiece coordinate system by the use of spatial coordinate transformation. Following that a surface topography simulation algorithm based on the cutting edge path equation and cutting parameters is put forward. The output of this work is that both the simulated surface topography and surface roughness profile are good agreement with the experimental results. Both the simulated and the actual machined surface results show that better surface topography is obtained in finish turning with the wiper insert than that with conventional insert. Burnishing effect of the wiper insert leads to half decrease of the Ra and Rz. The actual surface profiles are no longer regular wave shapes due to ploughing effect and side flow existing in the cutting zone. In addition, a surface roughness map has also been developed to optimize the selection of wiper radius and feed rate to satisfy the requirement of surface finishing with higher productivity. From the viewpoint of cutting tool design, the wiper radius with five times larger than tool nose radius can fully come into its role. This provides a novel insight into the design of wiper insert over conventional techniques. Above all, the proposed model gives a better prediction of surface roughness in finish turning process compared to the previous empirical and regression roughness models. The prediction of surface roughness in finish turning with wiper insert is also realized.     

An analysis of the three-dimensional surface topography of textured cold-rolled steel sheets

The surface topography of cold-rolled steel sheet affects its press-forming behaviour and also influences the appearance of the end product after painting. In this study, the three-dimensional surface topography of several steel sheets processed by different technologies has been measured. Using statistical and signal processing techniques, two-dimensional power spectral density function, two-dimensional autocorrelation function and two three-dimensional characteristic parameters of the surface topography of each steel sheet are obtained separately. Through analyzing these parameters and charts of the functions, an investigation on three-dimensional surface topography of these steel sheets is performed.