Experimental study of burn classification and prediction using indirect method in surface grinding of AISI 1045 steel

Grinding burn is a discoloration phenomenon according to the thickness of oxide layer on the ground surface. This study tries to establish an automatic grinding burn detection system with robust burn features that are caused by burn and not by the design parameters. To address this issue, a method based on acoustic emission sensor, accelerator, electric current transducers, and voltage transducers was proposed in an attempt to extract burn signatures. A trial-and-error experimental procedure was presented to find out burn threshold. Vitrified aluminum oxide grinding wheel and AISI 1045 steel workpiece were used in the grinding test, as they were the most commonly used wheel–workpiece combinations in conventional grinding process. With the help of fast Fourier transform and discrete wavelet transform, the spectral centroid of AE signal, the maximum value of power signal, and the RMS of the AE wavelet decomposition transform from wavelet decomposition levels d1 to d5 were extracted as burn features. The spectral centroid of AE signal was believed not to be affected by grinding parameters. A classification and prediction system based on support vector machine was established in order to identify grinding burn automatically. Results indicate that the classification system performs quite well on grinding burn classification and prediction.     

GRINDING WHEEL CONDITION MONITORING WITH HIDDEN MARKOV MODEL-BASED CLUSTERING METHODS

Hidden Markov model (HMM) is well known for sequence modeling and has been used for condition monitoring. However, HMM-based clustering methods are developed only recently. This article proposes a HMM-based clustering method for monitoring the condition of grinding wheel used in grinding operations. The proposed method first extract features from signals based on discrete wavelet decomposition using a moving window approach. It then generates a distance (dissimilarity) matrix using HMM. Based on this distance matrix several hierarchical and partitioning-based clustering algorithms are applied to obtain clustering results. The proposed methodology was tested with feature sequences extracted from acoustic emission signals. The results show that clustering accuracy is dependent upon cutting condition. Higher material removal rate seems to produce more discriminatory signals/features than lower material removal rate. The effect of window size, wavelet decomposition level, wavelet basis, clustering algorithm, and data normalization were also studied.     

GRINDING WHEEL CONDITION MONITORING WITH HIDDEN MARKOV MODEL-BASED CLUSTERING METHODS

Hidden Markov model (HMM) is well known for sequence modeling and has been used for condition monitoring. However, HMM-based clustering methods are developed only recently. This article proposes a HMM-based clustering method for monitoring the condition of grinding wheel used in grinding operations. The proposed method first extract features from signals based on discrete wavelet decomposition using a moving window approach. It then generates a distance (dissimilarity) matrix using HMM. Based on this distance matrix several hierarchical and partitioning-based clustering algorithms are applied to obtain clustering results. The proposed methodology was tested with feature sequences extracted from acoustic emission signals. The results show that clustering accuracy is dependent upon cutting condition. Higher material removal rate seems to produce more discriminatory signals/features than lower material removal rate. The effect of window size, wavelet decomposition level, wavelet basis, clustering algorithm, and data normalization were also studied.     

Grinding Burn Detection Based on Cross Wavelet and Wavelet Coherence Analysis by Acoustic Emission Signal

Grinding burn monitoring is of great importance to guarantee the surface integrity of the workpiece. Existing methods monitor overall signal variation. However, the signals produced by metal burn are always weak. Therefore, the detection rate of grinding burn still needs to be improved. The paper presents a novel grinding burn detection method basing on acoustic emission(AE) signals. It is achieved by establishing the coherence relationship of pure metal burn and grinding burn signals. Firstly, laser and grinding experiments were carried out to produce pure metal burn signals and grinding burn signals. No-burn and burn surfaces were generated and AE signals were captured separately. Then, the cross wavelet transform(XWT) and wavelet coherence(WTC) were applied to reveal the coherence relationship of the pure metal burn signal and grinding burn signal. The methods can reduce unwanted AE sources and background noise. Novel parameters based on XWT and WTC are proposed to quantify the degree of coherence and monitor the grinding burn. The grinding burn signals were recognized successfully by the proposed indexes under same grinding condition.     

Real-time detection of the breakage of small diameter drills with wavelet transform

This paper presents a real-time tool breakage detection method for small diameter drills using acoustic emission (AE) and current signals. Using the transmitted properties of the AE signal, apparatus for detecting the AE signal for tool breakage monitoring was developed for a machine centre. The features of tool breakage were obtained from the AE signal using typical signal processing methods. The continuous wavelet transform (CWT) and the discrete wavelet transform (DWT) were used to decompose the spindle current signal and the feed current signal, respectively. The tool breakage features were extracted from the decomposed signals. Experimental results show that the proposed monitoring system possessed an excellent real-time capability and a high success rate for the detection of the breakage of small diameter drills using combined AE and current signals.     

Real-time fuzzy-clustering and CART rules classification of the characteristics of emitted acoustic emission during horizontal single-grit scratch tests

During the unit event of material iteraction in grinding three phenomena are involved, namely: rubbing, ploughing and cutting. Where ploughing and rubbing essentially mean the energy is being applied less efficiently in terms of material removal. Such phenomena usually occurs before or after cutting. Based on this distinction, it is important to identify the effects of these different phenomena experienced during grinding. Acoustic emission (AE) of the material grit interaction is considered the most sensitive monitoring process to investigate such miniscule material change. For this reason, two AE sensors were used to pick up energy information (one verifying the other) correlated to material measurements of the horizontal scratch groove profiles. Such material measurements would display both the material plastic deformation and material removal mechanisms. Accurate material surface profile measurements of the cut groove were made using the Fogale Photomap Profiler which enables the comparison between the corresponding AE signal scratch data. By using short-time Fourier transforms (STFT) and filtration, the salient features for identifying and classifying the phenomena were more distinct between the three different levels of single-grit (SG) phenomena. Given such close data segregation between the phenomenon data sets, fuzzy clustering/genetic algorithm (GA) classification techniques were used to classify and verify the demarcation of SG phenomena. After the cutting, ploughing and rubbing gave a high confidence in terms of classification accuracy, the results from the unit/micro-event to the multi/macro-event, both 1-μm and 0.1-mm grinding test data, were applied to the named classifier for classification. Interesting output results correlated for the classifier signifying a distinction that there is more cutting utilisation than both ploughing and rubbing as the interaction between grit and workpiece become more involved (measured depth of cut increases). With the said classifier technique it is possible to get a percentage utilisation of the grit and material interaction phenomena. In addition, optimised fuzzy clustering was verified against a classification and regression tree (CART) rule-based system giving transparent rule classification. Such findings were then realised into a Simulink model as a potential control system for a micro-grinding simulation or, for real-time industrial control purposes.     

In-process endpoint detection of weld seam removal in robotic abrasive belt grinding process

This paper proposes a novel approach for in-process endpoint detection of weld seam removal during robotic abrasive belt grinding process using discrete wavelet transform (DWT) and support vector machine (SVM). A virtual sensing system is developed consisting of a force sensor, accelerometer sensor and machine learning algorithm. This work also presents the trend of the sensor signature at each stage of weld seam evolution during its removal process. The wavelet decomposition coefficient is used to represent all possible types of transients in vibration and force signals generated during grinding over weld seam. “Daubechies-4” wavelet function was used to extract features from the sensors. An experimental investigation using three different weld profile conditions resulting from the weld seam removal process using abrasive belt grinding was identified. The SVM-based classifier was employed to predict the weld state. The results demonstrate that the developed diagnostic methodology can reliably predict endpoint at which weld seam is removed in real time during compliant abrasive belt grinding.     

Grinding Chatter Detection and Identication Based on BEMD and LSSVM

Grinding chatter is a self?induced vibration which is unfavorable to precision machining processes. This paper proposes a forecasting method for grinding state identification based on bivarition empirical mode decomposition(BEMD) and least squares support vector machine(LSSVM), which allows the monitoring of grinding chatter over time. BEMD is a promising technique in signal processing research which involves the decomposition of two?dimen?sional signals into a series of bivarition intrinsic mode functions(BIMFs). BEMD and the extraction criterion of its true BIMFs are investigated by processing a complex?value simulation chatter signal. Then the feature vectors which are employed as an amplification for the chatter premonition are discussed. Furthermore, the methodology is tested and validated by experimental data collected from a CNC guideway grinder KD4020 X16 in Hangzhou Hangji Machine Tool Co., Ltd. The results illustrate that the BEMD is a superior method in terms of processing non?stationary and nonlinear signals. Meanwhile, the peak to peak, real?time standard deviation and instantaneous energy are proven to be e ec?tive feature vectors which reflect the di erent grinding states. Finally, a LSSVM model is established for grinding status classification based on feature vectors, giving a prediction accuracy rate of 96%.     

频带能量特征法在声发射刀具磨损监测系统中的应用

基于对声发射(AE)信号特点的分析和小波包分解理论对不平稳信号特征提取的优势,提出一种利用AE信号的能量变化来监测刀具磨损状态的方法。该方法利用db8小波基对AE信号进行5层小波包分解,将分解后各频带上的能量值作为特征参数,并组成特征向量。分别提取在新刀和刀具磨损状态下的特征向量,根据其变化即可判别刀具磨损的程度。试验结果验证了该方法在刀具磨损判析中的可用性。     

基于小波多分辨率分析和RBF神经网络的刀具磨损状态识别

刀具磨损监测对于提高加工过程的精度和自动化程度具有重要意义。本文提出一种基于RBF函数神经网络的刀具磨损状态监测模式。该系统利用声发射传感器对切削过程进行监测,采用多分辨率小波分解技术从声发射信号中提取反映刀具磨损的特征向量,并输入RBF神经网络,实现了刀具磨损的自动识别。     

采用小波神经网络的刀具故障诊断

为了有效的进行刀具状态监测，采用小波神经网络的松散型结合对刀具进行故障诊断。通过小波变换提取刀具磨损声发射（AE）信号的特征．即对AE信号进行小波分解，提取了5个频段的均方根值作为神经网络的输入，来识别刀具磨损状态。试验表明，均方根值完全可以作为刀具磨损过程中产生AE信号的特征向量。仿真结果表明，基于小波神经网络的刀具故障诊断对刀具磨损状态的识别效率高．该方法是有效的。     

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。     

Intelligent monitoring and prediction of tool wear in CNC turning by utilizing wavelet transform

In order to realize an intelligent CNC machine, this research proposed the in-process tool wear monitoring system regardless of the chip formation in CNC turning by utilizing the wavelet transform. The in-process prediction model of tool wear is developed during the CNC turning process. The relations of the cutting speed, the feed rate, the depth of cut, the decomposed cutting forces, and the tool wear are investigated. The Daubechies wavelet transform is used to differentiate the tool wear signals from the noise and broken chip signals. The decomposed cutting force ratio is utilized to eliminate the effects of cutting conditions by taking ratio of the average variances of the decomposed feed force to that of decomposed main force on the fifth level of wavelet transform. The tool wear prediction model consists of the decomposed cutting force ratio, the cutting speed, the depth of cut, and the feed rate, which is developed based on the exponential function. The new cutting tests are performed to ensure the reliability of the tool wear prediction model. The experimental results showed that as the cutting speed, the feed rate, and the depth of cut increase, the main cutting force also increases which affects in the escalating amount of tool wear. It has been proved that the proposed system can be used to separate the chip formation signals and predict the tool wear by utilizing wavelet transform even though the cutting conditions are changed.     

基于小波包分解的意识脑电特征提取

针对2种不同意识任务(想象左手运动和想象右手运动)的脑-机接口(brain-computer interface,BCI)设计,提出了基于小波包分解的特征提取方法。首先深入研究了小波包变换,结合事件相关去同步化(event-related desynchronization,ERD)/事件相关同步化(event-related synchronization,ERS)现象,提出以小波包分解系数来考虑特征,然后对C3、C4导联脑电信号进行小波包分解系数方差和相对能量2种特征的提取,最后采用最简线性分类器进行分类。结果表明,2种特征对应的最大分类正确率均达到了85%,对应时间分别为4.34 s和4.39 s。因此,在保证分类正确率的前提下,所提方法更加简单和有效,为大脑意识任务分类提供了新思路。     

基于光纤光栅和支持向量机的声发射定位系统

利用光纤光栅传感器和边缘滤波原理构建传感系统,结合小波分解与重构和支持向量机算法,对铝合金板声发射定位进行了研究。根据划分区域进行声发射实验,探索声发射源所在区域与信号特征之间的关系。在对声发射信号进行小波分解的基础上,使用近似系数和细节系数进行重构,并对重构后的各信号计算其振荡能量作为信号特征,进行声发射区域识别。以重构信号的振荡能量作为输入、声发射区域位置类别作为输出构建支持向量机多分类模型,实现了声发射区域定位识别。实验结果表明,在400mm×400mm×2mm的铝合金板上对36个测试样本进行了多次声发射区域定位识别,在180次模拟实验中实现了176次声发射区域准确定位,正确率达到97.78%,声发射区域识别精度为30mm×30mm。该研究结果为机械结构的声发射区域定位检测提供了有效方法。     

磨削砂轮状态在线监测方法研究

提出了用声发射（AE）信号在线监测砂轮状态的方法.利用该方法可以监测工件材料、加工要求和磨削参数经常变化环境下的砂轮钝化程度和破碎情况;并采用神经网络建立了传感器信号与砂轮状态之间的非线性关系.     

圆弧形砂轮精密修整及其声发射在线监测技术

针对球面、非球面及自由曲面超精密磨削加工用圆弧形金刚石砂轮难以精密修整的问题,提出基于旋转绿碳化硅（GC）磨棒的端部在位精密修整方法及修整过程的声发射在线监测技术。基于圆弧形金刚石砂轮的结构特性,制订圆弧形金刚石砂轮的在位精密修整与修整过程的声发射在线监测技术方案。依据修整与在线监测方案,对D64圆弧形金刚石砂轮进行修整实验及其声发射信号采集,修整后跳动误差小于10μm,比修整前减小30μm左右,砂轮精度显著提高。利用声发射信号均方根值获取砂轮修整结束的特征预警阈值,实现了旋转GC磨棒端部在位精密修整过程的在线监测以及修整结束时间的准确判断,可以有效提高球面非球面磨削加工过程的效率。     

Analysis of acoustic emission in precision and high-efficiency grinding technology

Current demands of machining hard and brittle materials at very small tolerances have predicated the need for precision and high-efficiency grinding. In situ monitoring systems based on acoustic emission (AE) provide a new way to control the surface damage and integrality of the components. However, a high degree of confidence and reliability in characterizing the manufacturing process is required for AE to be utilized as a monitoring tool. The authors established AE based online monitoring system and studied technique parameters versus the waveforms of AE under different working conditions. The results show that there are obvious mapping relations between the technique parameters of grinding and the effective values of the AE signals. Grinding along different directions would result in different strength of AE signal. Comparing with grinding along first longitude, fewer AE signal is released when grinding along latitude and better surface quality is generated. Similar variation tendency is observed no matter between AE root mean square (RMS) and linear speed or between surface roughness and linear speed which justify some kind of correlation may exist between AE RMS and surface roughness. The distance between the AE transducer and the AE source should be less than 80 mm while monitoring the process of grinding composite ceramics.     

On-line measurement of the size distribution of particles in a gas–solid two-phase flow through acoustic sensing and advanced signal analysis

Acoustic emission (AE) technology is a promising approach to non-intrusively measure the size distribution of particles in a pneumatic suspension. This paper presents an experimental study of the AE sensing technology coupled with signal processing algorithms for on-line particle sizing. The frequency characteristics of the AE signals under different experimental conditions are studied and compared. Initially, the characteristics of the background noise and AE signals are compared in the frequency domain for different air velocities and particle feeding rates. Through short-term energy analysis the working features of the suction unit and the vibration feeder are revealed. To find the effective characteristic frequency band of the AE signals, a multiple scanning and accumulation method assisted with a Savitzky–Golay smoothing filter is used to denoise the power spectra of the signals. Wavelet analysis is also deployed to denoise the signals. The denoising performance of different wavelet parameters (wavelet function, decomposition level and thresholding) is compared in terms of signal-to-noise ratio and signal smoothness. Finally, particle size is predicted through a neural network with energy fraction extracted through wavelet analysis. Experimental results demonstrate that the relative error of the particle sizing system is no greater than 23%.     

Acoustic emission method for tool condition monitoring based on wavelet analysis

It is believed that the acoustic emission (AE) signals contain potentially valuable information for tool wear and breakage monitoring and detection. However, AE stress waves produced in the cutting zone are distorted by the transmission path and the measurement systems and it is difficult to obtain an effective result by these raw acoustic emission data. In this article, a technique based on AE signal wavelet analysis is proposed for tool condition monitoring. The local characterize of frequency band, which contains the main energy of AE signals, is depicted by the wavelet multi-resolution analysis, and the singularity of the signal is represented by wavelet resolution coefficient norm. The feasibility for tool condition monitoring is demonstrated by the various cutting conditions in turning experiments.