Reliability estimation for cutting tools based on logistic regression model using vibration signals

As an important part of CNC machine, the reliability of cutting tools influences the whole manufacturing effectiveness and stability of equipment. The present study proposes a novel reliability estimation approach to the cutting tools based on logistic regression model by using vibration signals. The operation condition information of the CNC machine is incorporated into reliability analysis to reflect the product time-varying characteristics. The proposed approach is superior to other degradation estimation methods in that it does not necessitate any assumption about degradation paths and probability density functions of condition parameters. The three steps of new reliability estimation approach for cutting tools are as follows. First, on-line vibration signals of cutting tools are measured during the manufacturing process. Second, wavelet packet (WP) transform is employed to decompose the original signals and correlation analysis is employed to find out the feature frequency bands which indicate tool wear. Third, correlation analysis is also used to select the salient feature parameters which are composed of feature band energy, energy entropy and time-domain features. Finally, reliability estimation is carried out based on logistic regression model. The approach has been validated on a NC lathe. Under different failure threshold, the reliability and failure time of the cutting tools are all estimated accurately. The positive results show the plausibility and effectiveness of the proposed approach, which can facilitate machine performance and reliability estimation.     

A weighted hidden Markov model approach for continuous-state tool wear monitoring and tool life prediction

Tool wear is one of the important indicators to reflect the health status of a machining system. In order to obtain tool’s wear status, tool condition monitoring (TCM) utilizes advanced sensor techniques, hoping to find out the wear status through those sensor signals. In this paper, a novel weighted hidden Markov model (HMM)-based approach is proposed for tool wear monitoring and tool life prediction, using the signals provided by TCM techniques. To describe the dynamic nature of wear evolution, a weighted HMM is first developed, which takes wear rate as the hidden state and formulates multiple HMMs in a weighted manner to include sufficient historical information. Explicit formulas to estimate the model parameters are also provided. Then, a particular probabilistic approach using the weighted HMM is proposed to estimate tool wear and predict tool’s remaining useful life during tool operation. The proposed weighted HMM-based approach is tested on a real dataset of a high-speed CNC milling machine cutters. The experimental results show that this approach is effective in estimating tool wear and predicting tool life, and it outperforms the conventional HMM approach.     

Fault diagnosis of motor drives using stator current signal analysis based on dynamic time warping

Electrical motor stator current signals have been widely used to monitor the condition of induction machines and their downstream mechanical equipment. The key technique used for current signal analysis is based on Fourier transform (FT) to extract weak fault sideband components from signals predominated with supply frequency component and its higher order harmonics. However, the FT based method has limitations such as spectral leakage and aliasing, leading to significant errors in estimating the sideband components. Therefore, this paper presents the use of dynamic time warping (DTW) to process the motor current signals for detecting and quantifying common faults in a downstream two-stage reciprocating compressor. DTW is a time domain based method and its algorithm is simple and easy to be embedded into real-time devices. In this study DTW is used to suppress the supply frequency component and highlight the sideband components based on the introduction of a reference signal which has the same frequency component as that of the supply power. Moreover, a sliding window is designed to process the raw signal using DTW frame by frame for effective calculation. Based on the proposed method, the stator current signals measured from the compressor induced with different common faults and under different loads are analysed for fault diagnosis. Results show that DTW based on residual signal analysis through the introduction of a reference signal allows the supply components to be suppressed well so that the fault related sideband components are highlighted for obtaining accurate fault detection and diagnosis results. In particular, the root mean square (RMS) values of the residual signal can indicate the differences between the healthy case and different faults under varying discharge pressures. It provides an effective and easy approach to the analysis of motor current signals for better fault diagnosis of the downstream mechanical equipment of motor drives in the time domain in comparison with conventional FT based methods.     

储罐底板漏磁检测量化技术研究

储罐是石油、石化工业中重要的设备,储罐底板腐蚀是储罐安全隐患之一。漏磁检测方法是目前储罐底板检测研究的一个重要方向。根据缺陷漏磁信号的特征,将经验模态分解方法（EMD）与小波去噪方法相结合,对漏磁信号进行去噪处理。采用BP神经网络模型对储罐底板缺陷进行量化分析研究,构建了缺陷几何参数预测BP神经网络模型,并运用有限元分析所得到的数据为BP网络训练样本,用人工模拟缺陷的漏磁信号测试BP神经网络。网络训练和测试结果符合储罐底板缺陷量化的精度要求。     

Condition monitoring with mean field independent components analysis

We discuss condition monitoring based on mean field independent components analysis of acoustic emission energy signals. Within this framework, it is possible to formulate a generative model that explains the sources, their mixing and the noise statistics of the observed signals. Using a novelty detection approach based on normal-condition examples only, we detect faulty examples with high precision. The detection is done by evaluating the likelihood that the model, trained with normal examples, generated the signals, compared to a threshold obtained with normal examples. Acoustic emission energy signals from a large diesel engine are used to demonstrate this approach. The experiment show that mean field independent components analysis detects the induced fault with higher accuracy than principal components analysis, while at the same time selecting a more compact model.     

基于时变多变量Prony法的时变振动系统模态参数辨识

在经典的Prony法理论的基础上,提出了可以同时处理多维非平稳信号的时变多变量Prony法,并将其应用于时变多自由度振动系统的模态参数辨识。对传统的递推最小二乘算法加以改进,解决了时变多变量参数模型中时变参数矩阵估计的难题。对时变平面两杆操作臂系统进行仿真和分析,得出了较满意的结果。证明该算法在时变结构模态参数辨识方面,具有有效、准确的计算能力和较强的过程跟踪能力。     

Measurement of synchronous machine parameters using Kalman filter based fuzzy logic estimator

This paper presents a new Kalman filter/fuzzy logic approach for estimating synchronous machine parameters from short circuit tests. The technique uses on-line noisy measurements of the short circuit current for estimating direct axis reactances, and time constant synchronous machine parameters. The approach is based on expressing short circuit current as a discrete time linear dynamic system model suitable for the Kalman filter to estimate the parameters. Fuzzy rule-based logic is used to tune-up measurement noise levels by adjusting the covariance matrix. The results show a better convergence using fuzzy logic than those solely using the Kalman filter.     

基于声发射信号的外圆切入磨削去除率监测

针对精密外圆切入磨削加工的在线监测需求,提出一种采用声发射信号实现轴类零件材料去除率在线监测的方法。根据声发射信号强度与磨削力之间的联系,建立了声发射信号均方根曲线的预测模型,利用该预测模型研究了砂轮进给阶段和驻留阶段磨削系统时间常数的理论计算方法,推导了声发射信号均方根曲线与工件材料去除率的关系;编写了在线监测软件,利用声发射传感器实现了精密外圆切入磨削的材料去除率预测。实验证明,所建立的声发射信号均方根曲线模型具有良好的预测精度,基于该模型能够实现磨削系统时间常数在线评估,并实现精密轴类零件材料去除率的实时在线监测。     

Combating the effect of chassis squat in vehicle performance calculations by accelerometer measurements

The performance of a vehicle is measured as the time to complete an assigned task while accelerating. Non-intrusive performance measurements are often based on inertial navigation, i.e. measurements from an accelerometer are integrated once to obtain an estimate of the speed of the vehicle. A second integration results in a measure of the traveled distance. However, the suspension system of the vehicle introduces error in the measurements, owing to the tendency of the chassis to rise in the front and drop in the back during acceleration. An approach is derived based on the method of weighted least-squares for estimating the misalignment angle as a function of the horizontal acceleration. The proposed method includes a parametric model describing the effect of the tilt of the vehicle on the accelerometer measurements. Further, the method employs an external reference. It is shown that for a particular parametric model, the problem at hand has a closed-form solution. Certain practical applications are studied in some detail. A relative error of 0.5% in the elapsed time and 5% in the final speed is illustrated on 201-m straight track races.     

Latent degradation indicators estimation and prediction: A Monte Carlo approach

Asset health inspections can produce two types of indicators: (1) direct indicators (e.g. the thickness of a brake pad, and the crack depth on a gear) which directly relate to a failure mechanism; and (2) indirect indicators (e.g. the indicators extracted from vibration signals and oil analysis data) which can only partially reveal a failure mechanism. While direct indicators enable more precise references to asset health condition, they are often more difficult to obtain than indirect indicators. The state space model provides an efficient approach to estimating direct indicators by using indirect indicators. However, existing state space models to estimate direct indicators largely depend on assumptions such as, discrete time, discrete state, linearity, and Gaussianity. The discrete time assumption requires fixed inspection intervals. The discrete state assumption entails discretising continuous degradation indicators, which often introduces additional errors. The linear and Gaussian assumptions are not consistent with nonlinear and irreversible degradation processes in most engineering assets. This paper proposes a state space model without these assumptions. Monte Carlo-based algorithms are developed to estimate the model parameters and the remaining useful life. These algorithms are evaluated for performance using numerical simulations through MATLAB. The result shows that both the parameters and the remaining useful life are estimated accurately. Finally, the new state space model is used to process vibration and crack depth data from an accelerated test of a gearbox. During this application, the new state space model shows a better fitness result than the state space model with linear and Gaussian assumption.     

Using a probabilistic model to estimate the reliability of locating acoustic emission sources

A probabilistic model for estimating the error in determining the coordinates of acoustic emission (AE) sources based on experimental data is considered. A spherical indentor was pressed in a test object to obtain a large number of AE signals. Processing of signal parameters has made it possible to reveal the dependence of the likelihood of determining the coordinate with a prescribed accuracy on the maximum amplitude and AE-signal rise time. The possibility for choosing the method of detection of the AE-signal time of arrival at receivers depending on the signal parameters has also been examined.     

Statistical monitoring for non-Gaussian processes based on MICA-KDR method

The focus of the current work attempts to propose a purely data-based model for generating residuals for non-Gaussian process monitoring purposes, the idea of residual generation is borrowed from the field of model-based fault detection and applied in statistical monitoring, the generated residual instead of the measured variables is thus modeled and monitored. The proposed approach first employs the modified independent component analysis (MICA) algorithm to extract independent components (ICs) from a given dataset. Secondly, through assuming but only one variable is missing at one time, the known data regression (KDR) method dealing with missing data problem is then used for estimating the corresponding ICs. The inconsistency between the actual and estimated ICs is called residual and may present much lower level of non-Gaussianity, in contrast to the actual ICs. Thirdly, a principal component analysis based statistical monitoring model can be utilized for online fault detection based on the generated residual. Finally, the superiority and efficiency of the MICA-KDR approach over its counterparts are validated by implementing comparisons on two industrial processes, the proposed MICA-KDR method is demonstrated to be a comparative alternative in monitoring non-Gaussian processes     

Estimating local characteristics of object motion in a seismic monitoring system

The estimation of the characteristics of motion of the object controllable by means of seismic monitoring system is considered. It is assumed that the motion in a local area of space is uniform and rectilinear, and the problem is therefore reduced to estimating the coordinates of the initial point and velocity vector. The input data are measured time delay differences in receiving seismic signals by various sensors of the system. Usually, these measurements are formed as a sequential flow in real systems. The obtained estimates of the trajectory also have the form of a sequence, and their accuracy increases with increasing number of primary data. The accuracy of the proposed method are analyzed by statistical simulation.     

Using ultrasonic pulse-echo B-scan signals for estimation of time of flight

Time of flight based methods are extensively used for detecting, locating and sizing faults in ultrasonic non-destructive testing and evaluation. In this paper, we investigate model-based estimation of the ultrasonic time of flight using B-scan signals. The Cramer–Rao bounds on the time of flight estimator for B-scan signals are derived and then compared to the Cramer–Rao bounds on the time of flight estimator for A-scan signals. Through this theoretical analysis, we show that the estimation based on B-scan signals significantly reduces the Cramer–Rao bound on the time of flight estimator. In addition, the resulting theoretical equation allows evaluating the improvement in the accuracy of estimating the time of flight. The theoretical lower bound is then compared to the variance of a maximum likelihood estimator which is obtained using a Monte Carlo simulation. The results show that the maximum likelihood estimator can achieve the lower bound on the variance of the time of flight estimator and hence it is an efficient estimator. This numerical result will complement the theoretical analysis by showing that the Cramer–Rao bound can be reached if a proper estimator is selected.     

基于振动相对量法的齿轮敲击振动辨识

整车转毂振动测试中测得的手动变速器箱体振动信号包含了多种部件的振动信号,传统方法无法从测试信号中直接获取齿轮敲击振动信号,故无法定量评价齿轮敲击振动水平。本文提出了运用敲击振动相对量来辨识齿轮敲击振动的方法。该方法首先对箱体振动信号进行人耳特性滤波,然后进行回归和平滑处理获得稳态振动信号。将滤波后的振动信号减去稳态振动信号,得到的振动相对量即为非承载齿轮对的瞬态敲击振动信号,最终可辨识出齿轮敲击振动的发生时刻、频率范围和水平。在实车试验中,采用该方法在3.5s的测试时间内,识别出振动相对量最大的134个齿轮敲击振动信号,其发生时刻与敲击振动信号回放得到的134个敲击噪声发生时刻完全相同,辨识结果与人的主观感受一致,即准确辨识出了齿轮敲击振动。得到的辨识结果可用于定量评价齿轮敲击振动水平,校核理论模型的正确性,研究不同参数对齿轮敲击振动水平的影响规律,找出关键影响因素并优化处理,从而改善齿轮敲击性能。     

Control design of a pneumatic cylinder with distributed model of pipelines

In this paper, the control design with distributed model of pipelines is proposed to make the cylinder side be free of pressure sensors. In this research, long connected pipelines are used. The pipeline is designed as a one dimensional distributed model. The model of pipelines is based on the discretization of four equations, as state equation of air, motion equation, continuity equation, and energy equation. The distributed model estimates the pressure losses and time delay through long connected pipelines in real time. To confirm the control method with distributed model of pipelines, a simulation model of the whole system is designed. Compared simulation and experimental results, it has been found that the model represents the real system well. In the experiments, the pressure values in the cylinder chambers estimated by the distributed model in real time played as control signals. Compared with the estimated and measured pressure values in the cylinder chambers, it is found that with this distributed model, the pressure values in the cylinder chambers is precisely estimated in real time using the measured values at the control ports of the servo valve. The experimental results demonstrate that the position accuracy is almost the same with that of using the measured pressure signals in the cylinder chambers. The cylinder side is free of pressure sensors with the proposed control method.     

NON-LINEAR SYSTEMS REPRESENTATION USING ARMAX MODELS WITH TIME-DEPENDENT COEFFICIENTS

An approach that models a non-linear process operating over a large dynamic range is developed and validated. This approach is based on stochastic Time-varying AutoRegressive Moving Average with eXogeneous inputs (TARMAX) models. The TARMAX model coefficients are explicit functions of time and vary in a deterministically organised fashion. A novel model parameter estimation method fully based on linear operations is presented. The estimation approach is characterised by a low computational complexity and requires no initial guess of the parameter values. The ability of the approach to model non-linear processes is validated by addressing problems dealing with improving the estimation of mass air flow going into an automobile engine. First, a TARMAX model is used to capture the dynamics of the engine process relating air flow provided by a laboratory grade sensor and three input signals available in the engine electronic controller. The TARMAX model is used to simulate the complex relationship between the output and the three input signals. Second, TARMAX models are used to anticipate the future response of a hot-wire-based mass air flow sensor (MAF) in order to obtain an accurate estimate of the cylinders air charge. The estimated TARMAX models prove to have good simulation and prediction capabilities. All models are estimated using actual production vehicle data.     

[设备健康管理]深度置信网络在齿轮故障诊断中的应用

针对齿轮传动系统中齿轮等零部件易出现故障或失效等问题,提出了一种基于深度学习理论的齿轮传动系统故障诊断方法。首先利用深度置信网络强大的特征自提取能力,对齿轮传动系统的振动信号进行特征提取,然后通过DBNs的复杂映射表征能力对故障信号进行故障判别。诊断实例表明,若不对齿轮振动的原始时域信号进行特征提取,直接利用DBNs对其进行诊断时,故障识别正确率只能达到 60%左右;如果对时域信号进行简单的傅里叶变换后,再利用 DBNs 对处理后的振动信号频谱进行诊断分析,正确率能达到 99.7%,从而证明了所提故障诊断方法的简易性和有效性。     

Model order determination and noise removal for modal parameter estimation

Given a noisy impulsive response function (IRF) that has been contributed by an unknown number of modes, this article proposes a different approach from the traditional methods for estimating modal parameters from this noisy IRF. The major difference lies in the way of handling noise and choosing the computational model order. Whereas the traditional approach accommodates noise by purposely increasing the computational model order, the proposed approach uses the actual system order as the computational model order and rejects noise prior to performing the modal parameter estimation. The proposed approach includes three steps: (1) model order (or number of modes) determination from the measured IRF—by finding the rank of a Hankel matrix constructed from the measured IRF, (2) noise removal from the measured IRF to obtain a filtered IRF—by implementing Cadzow's algorithm for the structured low rank approximation (SLRA) on the Hankel matrix, and (3) modal parameters estimation from the filtered IRF—by using the complex exponential method (Prony's method). Numerical studies include both synthesized and experimental data. While measured IRFs with mild and strong noise levels are simulated for a 5 degree-of-freedom mass-spring-dashpot system, the modal parameter estimations based on the filtered IRFs are very good for both noise levels. While experimental data are measured from two accelerometers mounted at a cantilever beam, the modal parameters estimated from the filtered IRFs of the two accelerometers are in excellent agreement.     

Data processing in subspace identification and modal parameter identification of an arch bridge

A data-processing method concerning subspace identification is presented to improve the identification of modal parameters from measured response data only. The identification procedure of this method consists of two phases, first estimating frequencies and damping ratios and then extracting mode shapes. Elements of Hankel matrices are specially rearranged to enhance the identifiability of weak characteristics and the robustness to noise contamination. Furthermore, an alternative stabilisation diagram in combination with component energy index is adopted to effectively separate spurious and physical modes. On the basis of identified frequencies, mode shapes are extracted from the signals obtained by filtering measured data with a series of band-pass filters. The proposed method was tested with a concrete-filled steel tubular arch bridge, which was subjected to ambient excitation. Gabor representation was also employed to process measured signals before conducting parameter identification. Identified results show that the proposed method can give a reliable separation of spurious and physical modes as well as accurate estimates of weak modes only from response signals.