Classification and regression models of audio and vibration signals for machine state monitoring in precision machining systems

We present a data-driven method for monitoring machine status in manufacturing processes. Audio and vibration data from precision machining are used for inference in two operating scenarios: (a) variable machine health states (anomaly detection); and (b) settings of machine operation (state estimation). Audio and vibration signals are first processed through Fast Fourier Transform and Principal Component Analysis to extract transformed and informative features. These features are then used in the training of classification and regression models for machine state monitoring. Specifically, three classifiers (K-nearest neighbors, convolutional neural networks and support vector machines) and two regressors (support vector regression and neural network regression) were explored, in terms of their accuracy in machine state prediction. It is shown that the audio and vibration signals are sufficiently rich in information about the machine that 100% state classification accuracy could be accomplished. Data fusion was also explored, showing overall superior accuracy of data-driven regression models.     

基于子波变换的涡街流量传感器信号分析

长期以来 ,如何提取潜在噪声下的涡街流量信号一直是个问题。流体流速脉动 ,局部阻力 ,随机振动———所有这些因素都给解决这一问题带来难度。文章应用子波变换消噪理论 ,从软件滤波的角度分析了强噪声干扰下的涡街流量信号 ,并提出了单支重构计数方法。分析结果表明 ,这种方法对低流速流量计量效果很好 ,能够有效地扩展量程下限     

1.8 V千兆以太网接收器模拟前端预均衡电路

介绍了一种用于1.8 V电源电压下的千兆以太网接收器的模拟前端预均衡电路。电路分为三个部分:预处理电路、基带漂移补偿电路和可变增益放大电路,主要实现回波消除、基带漂移补偿和电路增益自动控制等功能。为了与百兆模式兼容,提出了一种预处理电路。仿真结果表明,该电路可以很好地实现回波消除的功能,能够对由于基带漂移引起的信号失真给以补偿,可以提供16级不同的增益,并进行频率补偿。电路采用0.18μm标准CMOS工艺实现。     

脉冲序列的快速搜索法及可信度分析

对于周期脉冲序列的信号分选识别，可采用序列搜索的时域方法[1，2]。本文将大周期内完全重复的脉冲序列分解为多个“单周期信号”．并把它们均定义为周期性信号．同时，提出了序列搜索的一种软件实现方法及密集视频脉冲信号分选的变步长、信步长快速搜索法．在此基础上，本文针对序列搜索分选法的可信度计算公式进行了推导和分析，证实了这种方法的可行性和有效性．     

超低频信号的记录与回放

介绍一种经济实用的用普通磁带记录与回放超低频信号的方法，该方法在地质，建筑工程和医疗等行业中有一定的实用价值。     

Driver detection and recognition of lineside signals and signs at different approach speeds

A study was carried out using simulation to investigate driver responses to lineside signals and signs at various approach speeds. The objectives of the study were: (1) to find out whether train speed would significantly affect signal/sign reading; (2) to examine at which point certain types of signs or signals could be detected or recognised, and (3) to determine a speed cut-off level above which certain types of signs or signals are no longer recognisable or detectable. Fifty-seven train drivers from 12 Train Operating Companies in the UK participated in the trials. Twenty different types of lineside signs and ten types of signals were tested under six different approach speeds ranging from 100 to 350 km/h (62–218 mph). Driver performance measures were ‘time remaining to the signal/sign’ at the point of detection or recognition, and reading error rate. The results showed a significant influence of train speed on driver responses to lineside signals/signs and demonstrated a non-linear relationship between driver responses to signals/signs and approach speed. This has been used to estimate a maximum approach speed limit within which a specific signal or sign can be correctly detected or recognised. The findings and implications of the study are discussed in the paper.     

用于机器人轨迹定位纠正系统的回声状态神经网络研究

循环神经网络(RecurrentNeuralNetworks)是人工神经网络(ArtificialNeuralNetworks)中重要的分支,与前馈神经网络(ForwardNeuralNetworks)相比具有更好的时间序列学习能力。但长期以来其学习法一直不能脱离前馈神经网络而自成一体,回声状态神经网络(EchoStateNetworks(ESN))是打破这一局面的全新学习方法。其独特的结构,良好的短期记忆能力,方便的学习方法,不俗的非线性特性是以前循环神经网络所不可比的。本文在介绍了回声状态神经网络之后将其用于四轮机器人的位置测量系统中,有良好的表现。     

带限信号高频重建的一种新的正则外推算法

本文提出生种新的带限信号的外推方法，该方法能够很好地抑制噪声干扰，且计算复杂性较低，本文给出算法，证明了收敛性，最后给出了误差计算机实验结果。     

A damage identification technique for CF/EP composite laminates using distributed piezoelectric transducers

In this study, a damage identification approach was developed for carbon fibre/epoxy composite laminates with localized internal delamination. Propagation of the Lamb wave in laminates and its interaction with the delamination were examined. The fundamental symmetric Lamb wave mode, S₀, and the lowest order shear wave mode, S₀^′, were chosen to predict damage location. A real-time active diagnosis system was therefore established. This technique uses distributed piezoelectric transducers to generate and monitor the ultrasonic Lamb wave with narrowband frequency. The two-way switches were employed to minimize the number of transducers. A signal-processing scheme based on the time–frequency spectrographic analysis was utilised to extract useful diagnostic information. Also, an optimal identification method was applied on damage searching procedure to reduce errors and obtain the diagnostic results promptly. Experiments were conducted on [0/−45/45/90]_s CF/EP laminates to verify this diagnosis system. The results obtained show that satisfactory detection accuracy could be achieved.     

交变光源型位敏器件应用技术

在利用直流光源型半导体位敏器件（ＰＳＤ）进行测量时，常因环境光强的改变而引起误差。本文介绍的交变调制光源及其相应的ＰＳＤ光电信号处理电路，可大大地降低和消除杂散光对测量的影响。