基于HPSO优化BP神经网络的刀具磨损状态识别

针对BP神经网络容易陷入局部极值导致识别精度低的问题,文章提出了一种基于混合粒子群算法(HPSO)的BP神经网络优化算法。在刀具磨损监测实验过程中,采集刀具切削的声发射(AE)信号,利用小波包分解算法对AE信号进行滤波,并进行特征提取。将频带能量特征和切削参数分别作为主特征和辅助特征,并对其对归一化处理。采用混合粒子群优化算法(HPSO)对BP神经网络预测模型进行优化,利用优化后的模型对测试样本进行模式识别,结果表明,优化后的HPSO-BP模型能够有效地降低神经网络陷入局部极值的情况,提高刀具磨损识别精度。     

On-line detection of the breakage of small diameter drills using current signature wavelet transform

This paper presents on-line tool breakage detection of small diameter drills by monitoring the AC servo motor current. The continuous wavelet transform was used to decompose the spindle AC servo motor current signal and the discrete wavelet transform was used to decompose the feed AC servo motor current signal in time–frequency domain. The tool breakage features were extracted from the decomposed signals. Experimental results show that the proposed monitoring system possessed an excellent on-line capability; in addition, it had a low sensitivity to change of the cutting conditions and high success rate for the detection of the breakage of small diameter drills.     

Drilling wear detection and classification using vibration signals and artificial neural network

In automated flexible manufacturing systems the detection of tool wear during the cutting process is one of the most important considerations. This study presents a comparison between several architectures of the multi-layer feed-forward neural network with a back propagation training algorithm for tool condition monitoring (TCM) of twist drill wear. The algorithm utilizes vibration signature analysis as the main and only source of information from the machining process. The objective of the proposed study is to produce a TCM system that will lead to a more efficient and economical drilling tool usage. Five different drill wear conditions were artificially introduced to the neural network for prediction and classification. The experimental procedure for acquiring vibration data and extracting features in both the time and frequency domains to train and test the neural network models is detailed. It was found that the frequency domain features, such as the averaged harmonic wavelet coefficients and the maximum entropy spectrum peaks, are more efficient in training the neural network than the time domain statistical moments. The results demonstrate the effectiveness and robustness of using the vibration signals in a supervised neural network for drill wear detection and classification.     

基于声发射信号模式的齿轮特征识别研究

鉴于声发射信号对齿轮早期裂纹具有独特的敏感性,对早期齿轮声发射信号的特征识别具有重要意义。介绍小波变换理论及其原理,建立齿轮疲劳试验平台,利用小波阈值降噪对不同工况下齿轮声发射信号进行预处理,获取高能量频段的信号并提取时域、频域特征参数,将其作为BP神经网络的输入,以识别不同工况下的声发射信号。实验结果表明：与去噪后的全频段信号相比,基于高能量频段信号所提取的特征参数具有更高的识别率,为早期齿轮故障信号分析和检测提供借鉴。     

基于递归小波与ART2网络的刀具状态智能监测

为实现刀具的实时状态监测,以超高斯函数为基础,构造出一类用递推公式进行小波变换的小波基,提出该系列小波基的优化方法,对其时频特性进行了分析.对刀具AE信号进行递归小波分解,提取特征并应用ART2网络识别刀具状态.结果表明,基于递归小波与ART2网络的刀具状态监测具有鲁棒性强、实时性好的特点.     

基于小波包分解和奇异谱分析方法的立铣加工颤振监测

在立铣加工过程中,颤振是加工过程失稳的一个最重要的原因。颤振将会严重影响工件表面质量和材料去除率,加剧刀具磨损和恶化工作环境。虽然大部分颤振监测系统可以监测到颤振发生,但颤振发生时已经对工件和刀具产生了严重的损伤,因此,需要提前监测到颤振特征。在颤振发生过程中,振动信号具有在时域中不断增大,在频域中能量频移的特性。考虑这两个振动信号特征,提出了一种颤振特征提取方法。提取颤振发生频带中振动信号的能量比和奇异谱熵系数作为两个颤振特征,并通过人工神经网络模型实现切削颤振的识别。文中提出的颤振监测系统包括特征提取和分类,能够精确辨识立铣加工中的稳定、过渡和颤振状态。     

基于声发射的铣刀破损信号特征的提取与优化

铣刀破损监测对实现加工自动化具有重要的意义.提出了基于小波变换的铣刀声发射破损特征提取与优化方法.首先,采用小波变换对铣刀声发射信号进行多分辨率分解,然后提取各频段子信号的能量比作为刀具破损监测的特征量.通过对正常切削、随机冲击和刀具破损这三类信号的比较分析,证明了该特征提取方法能够有效地反映刀具状态.最后,通过正交统计,分析了切削速度、进给速度和切削深度对特征量的影响,并对特征量进行优化.     

Power signal separation in milling process based on wavelet transform and independent component analysis

Among many machining condition monitoring systems, a spindle motor power monitoring system is considered as one of the most popular systems for plant floor applications. However, in practice, power signals are mixed with many signal sources relevant to cutting tool, cutting conditions as well as components of a machine tool, which contaminate with each other in feature extraction processes and decrease the monitoring reliability. In this paper, modified blind sources separation (BSS) technique is used to separate those source signals in milling process. A single-channel BSS method based on wavelet transform and independent component analysis (ICA) is developed, and source signals related to a milling cutter and spindle are separated from a single-channel power signal. The experiments with different tool conditions illustrate that the separation strategy is robust and promising for cutting process monitoring.     

A brief review: acoustic emission method for tool wear monitoring during turning

Research during the past several years has established the effectiveness of acoustic emission (AE)-based sensing methodologies for machine condition analysis and process monitoring. AE has been proposed and evaluated for a variety of sensing tasks as well as for use as a technique for quantitative studies of manufacturing processes. This paper reviews briefly the research on AE sensing of tool wear condition in turning. The main contents included are:

1. The AE generation in metal cutting processes, AE signal classification, and AE signal correction.

2. AE signal processing with various methodologies, including time series analysis, FFT, wavelet transform, etc.

3. Estimation of tool wear condition, including pattern classification, GMDH methodology, fuzzy classifier, neural network, and sensor and data fusion.

A review of AE-based tool wear monitoring in turning is an important step for improving and developing new tool wear monitoring methodology.     

Real time monitoring of tool breakage in a milling operation using a digital signal processor

A monitoring system that can detect tool breakage and chipping in real time was developed using a digital signal processor (DSP) board in a face milling operation. An autoregressive (AR) model and a band energy method were used to extract the features of tool states from cutting force signals. Then, two artificial neural networks, which have a parallel processing capability, were embedded on the DSP board to discriminate different malfunction states from features obtained by each of the two methods of signal processing. In experiments, we found that feature parameters extracted by AR modeling were more accurate indicators of malfunctions in the process than those from the band energy method, although the computing speed is slower. By using the selected features, we were able to monitor malfunctions in real time.     

高速铣削加工中刀具监测系统研究

在铣削加工中,在刀具急剧磨损的初级阶段,表征刀具磨损的信号较弱,而此时工件精度已早有明显变化。小波神经网络虽能有效处理各种频段信号,但对较弱信号还是存在漏检现象。开发针对高速铣削的刀具在线监测系统,通过监测工件表面精度的变化,及时修正小波变换参数,提高了监测微弱信号的能力,有效降低了刀具监测的漏检、误报率。     

基于小波包神经网络的轴承故障识别模型

针对传统故障特征提取过程复杂、诊断方案单一且准确性差等问题,提出了基于多阈值小波包和深度置信网络(DBN)的轴承故障识别方案。本文作者采用最优小波基函数和软硬阈值结合方法对原始振动信号进行三层分解降噪处理,得到8个从低频到高频段的信号成分,对其进行组合重构作为神经网络的输入样本;通过DBN在数据处理上的特征重构优势,建立了DBNBP神经网络的轴承故障识别模型,确定模型的各类参数。经多次实验,探究不同样本输入对模型识别率的影响,并与传统的浅层神经网络识别模型做对比分析,结果表明:经训练的DBNBP轴承故障识别模型可从原始数据、小波包分解信号实现轴承故障信号的准确特征学习和分类,结合识别率和诊断时间考虑,经小波包分解信号输入具有更优的诊断效率。     

刀具磨损声谱特征的分析

刀具在切削过程中，会产生大量的声信号。文中运用统计方法对切削声信号进行了深入分析，并对可听阈内的声信号按低频带、中频带和高频带分别进行了功率谱分析。研究结果表明，切削声信号中100—8kHz频带内的特征频率分量反映了刀具磨损状态及其变化规律，这为利用声信号进行刀具实时在线监测奠定了基础。     

Sensing tool breakage in face milling with a neural network

A new approach using a neural network to process the features of the cutting force signal for the recognition of tool breakage in face milling is proposed. The cutting force signal is first compressed by averaging the cutting force signal per tooth. Then, the average cutting force signal is passed through a median filter to extract the features of the cutting force signal due to tool breakage. With the back propagation training process, the neural network memorizes the feature difference of the cutting force signal between with and without tool breakage. As a result, the neural network can be used to classify the cutting force signal with or without tool breakage. Experiments show this new approach can sense tool breakage in a wide range of face milling operations.     

Monitoring of flank wear of coated tools in high speed machining with a neural network ART2

A monitoring system for classifying the levels of the tool flank wear of coated tools into some categories has been developed using an unsupervised and self-organizing artificial neural network, ART2. The input pattern used for the ART2 was an array of normalized mean wavelet coefficients of the feed force, which was affected by not only the flank wear but also the severe crater wear observed in high speed machining. The outputs of ART2 were classified into four or five categories of wear levels: the incipient stage, one or two intermediate stages, final stage and hazardous stage. For two apparently different series of input data obtained under the same cutting conditions, which are often experienced in the experiment, the ART2 neural network showed very similar classification of tool wear levels from the beginning to the end of cutting. Further study proved that this monitoring system detected the excessive wear in the hazardous stage for different cutting speeds 5–7 m/s and different feed rates 0.10–0.20 mm/rev.     

Machine ensemble approach for simultaneous detection of transient and gradual abnormalities in end milling using multisensor fusion

In a fully automated manufacturing environment, instant detection of the cutting tool condition is essential for the improved productivity and cost effectiveness. This paper studies a tool condition monitoring system (TCM) via machine learning (ML) and machine ensemble (ME) approach to investigate the effectiveness of multisensor fusion technique when machining 4340 steel with multilayer coated and multiflute carbide end mill cutter. In this study, 135 different features are extracted from multiple sensor signals of force, vibration, acoustic emission and spindle power in the time and frequency domain by using data acquisition and signal processing module. Then, a correlation-based feature selection technique (CFS) evaluates the significance of these features along with machining parameters collected from machining experiments. Next, an optimal feature subset is computed for various assorted combinations of sensors. Finally, machine ensemble methods based on majority voting and stacked generalization are studied for the selected features to classify not only flank wear but also breakage and chipping. It has been found in this paper that the stacked generalization ensemble can ensure the highest accuracy in tool condition monitoring. In addition, it has been shown that the support vector machine (SVM) outperforms other ML algorithms in most cases tested.     

基于小波包能量和层次熵的D-S证据理论的轴承故障诊断技术

针对齿轮箱轴承故障识别率低、故障信号不平稳的问题,提出层次熵与小波包能量多源数据融合轴承故障诊断方法。采用小波包对轴承正常、内圈、外圈、滚动体故障等4种振动信号进行三层小波包分解并重构,计算各频段样本熵（即层次熵）和小波包能量作为故障特征向量集;应用归一化方法对2种特征向量处理后分别建立BP神经网络模型实现轴承不同故障模式的诊断;最后应用D-S证据理论,通过小波包能量和层次熵以及两者融合信息的故障诊断结果比较,表明基于神经网络和D-S证据理论相结合方法用于复杂机械的故障诊断是可行和有效的。     

End milling tool breakage detection using lifting scheme and Mahalanobis distance

In this paper, a novel method based on lifting scheme and Mahalanobis distance (MD) is proposed for detection of tool breakage via acoustic emission (AE) signals generated in end milling process. The method consists of three stages. First, by investigating the specialty of AE signals, a biorthogonal wavelet with impact property is constructed using lifting scheme, and wavelet transform is carried out to separate AE components from the original signals. Second, Hilbert transform is adopted to demodulate signal envelope on wavelet coefficients and salient features indicating the tool state (i.e., normal conditions, slight breakage, and serious breakage) are extracted. Finally, tool conditions are identified directly through the recognition of these features by means of MD. Practical application results on a CNC vertical milling machine tool show that the proposed method is accurate for feature extraction and efficient for condition monitoring of cutting tools in end milling process.     

A wavelet-based methodology for grinding wheel condition monitoring

Grinding wheel surface condition changes as more material is removed. This paper presents a wavelet-based methodology for grinding wheel condition monitoring based on acoustic emission (AE) signals. Grinding experiments in creep feed mode were conducted to grind alumina specimens with a resinoid-bonded diamond wheel using two different conditions. During the experiments, AE signals were collected when the wheel was ‘sharp’ and when the wheel was ‘dull’. Discriminant features were then extracted from each raw AE signal segment using the discrete wavelet decomposition procedure. An adaptive genetic clustering algorithm was finally applied to the extracted features in order to distinguish different states of grinding wheel condition. The test results indicate that the proposed methodology can achieve 97% clustering accuracy for the high material removal rate condition, 86.7% for the low material removal rate condition, and 76.7% for the combined grinding conditions if the base wavelet, the decomposition level, and the GA parameters are properly selected.     

Inspection of micro-tools at high rotational speeds

Micro-tools have been widely used in industry, primarily by biomedical and electronic equipment manufacturers. The life of these cutting tools is extremely unrpedictable and much shorter than conventional tools. Also, these miniature tools, with a diameter of less than 1 mm, cannot be inspected by an operator without the aid of magnifying glass.

In this paper, evaluation of the intensity variation of a reflected laser light beam from the cutting tool surfaces is proposed as a method of estimating cutting tool surface conditions. Various encoding methods, including wavelet transformations, were proposed to obtain a small and meaningful set of data from the intensity variation readings of one tool rotation. The encoded data are classified using a simple threshold method, Restricted Coulomb Energy (RCE), and Adaptive Resonance Theory (ART2)-type neural networks. The proposed encoding and classification approaches were tested with over one hundred sets of data.

The threshold method detects only severe tool damage. The RCE neural networks and graphical presentation of the encoded sets demonstrated the feasibility of the proposed monitoring technique and encoding methods. The ART2-type neural networks were found to be the best candidate for tool condition monitoring because of their self learning capability. Wavelet transformation-based encoding and ART2-type neural networks were found to be sensitive enough to recognize wear at the cutting edge.