机械监测诊断中的理论与方法

章节摘录

插图：INTRODUCTIONNowadays，the grinding operation is one of the mostimportant metal processing methods.The per.formance and reliability of modem machine productsare seriously influenced by the surface perfection.dimensiona！and geometrical accuracies of their ele.ments.the majority of which are finished bv digfferentkinds of grinding operations.In modem automaticgrinders it is quite necessary to adopt the on.1inesurveillance via dynamic signal recognition in orderto prevent the degradation and deterioration ofworkpieces as the result of wheel wear andlOSS of itscutting ability.In general，the wheel life between twosucceeding turning operations is characterized by thevariation of grinding sound，increase of grindingforces and the violence of grinding chatter。which canbe easily recognized by experienced operators.Butwith regard to the automatic in-process recognitionof these signals the closest attention js now solicitedin order to monitor the performance of wheelandgrinder under different grinding conditions and toprognosticate the remaining wheel lift.In this paper.a new criterion called the Kullback-Leibler informa.tion humber based on time.series analysis and infor.mation theory is suggested to monitor the per.formance of whee！and grinder under difierentworking conditions and to prognosticate the remain.ing wheel life.The computer flow chart for supc-vision of the grinding cycles is also designed andtested.PROCEDURE  The main aspccts for on.1ine surveillance of agrinding process by means of the Kullback-Leiblerinformation humber are as follows：（1）measurementof the dynamic signals emitted in difierent stages ofgrinding process between two succeeding trueingoperations；（2）A／D conversion.of each signal to formthe corresponding discrete time series；（3）establish.ement of the reference models.These models can beautoregressive moving average（ARMA）models orsimplified autoregressive（AR）models[1]；（4）storageof the information about these reference modelsin computer memory.i.e.the values of modelcoefficients as well as the residua！varianoe of eachmodel.The more reference models are stored，thebetter recognition capability the computer will pos.sess；（5）modelling the current tested signal in grind.ing operation and calculation of the Kullback-Leibler information numbers between this new modeland each reference model already stored in computer；（6）assignment of the tested signal to the category ofone of the reference models with minimum valuesof the KullbacLLeibler number.The computerflowchart is shown in Fig.l.Its performance is todecide which reference model the tested sequenceshould belong to，and to establish new referencemodels.That is to say，the software pOSSeSSeS theself-learning ability and can improve the recognitioncapability of computer.ExPERHMENTATION  The experiments were carded out using an ordi.nary cylindrica！grinder with hydrostatic main bear.ings.The materials taken for testpieces were hard.ened carbon steels marked 45.55 and T8 withhardness from R=40 to 60 and diameters from 035to 85 mm.A grinding whee！of GB46ZR2SP.e500 x 40 x 305 mm and ordinary emulsion fluidwere selected.In experiments.the testpieces were fedradially to the wheel without transverse feed.Thcnormal and tangential grinding forces，vibrationaldisplacement 0f a specia！designed dead centre andgrinding sound were measured by the eddy currenttransducer，  dynamic strain gauges and precisionsound level meter accordingly as shown in Fig.

前言

屈梁生是中国工程院院士、西安交通大学教授、博士生导师。他工952年于交通大学机械系毕业，留校任教；后随校迁往西安，长期任教于西安交通大学。屈梁生院士长期从事机械质量控制与监测诊断领域的基础性与开拓性研究。他把信息论、系统论的原理融入到机械故障诊断和状态监测中去，从整体上提升了这一传统学科。屉梁生院士的研究课题来自生产实践，在提升为理论后，再放回到实践中去考验，然后又反过来再推动理论进一步完善提高。这种不断反复的精神贯穿在他整个科研工作中。他在科研方法论中不受传统与成见的束缚，独具慧眼，不断开拓出新的领域。典型的一个实例是，在已成熟并已广泛应用的FFT技术中加入相位信息，将多个传感器的振动信号在频域中合成，提出了这个由中国人始创的“全息谱”理论，大大地提高了FFT的水平，扩大了其工作范围。屈梁生教授在注意提高学生的理论知识与独立工作能力的同时，也十分注意培养他们的道德修养。他认为，作为新一代的科技人员，首先应该是一个具有高尚品德的人。记得有一次我与他谈起科研工作中出现的浮躁情绪及带来的种种弊端时，都感到不断提高自身的修养，才能耐得住寂寞，长期坚持，出真正的成果。由此还谈到李叔同的名言：“事能知足心常乐，人若无求品自高”，咏之感叹再三，相互勉励。这本论文集是屈梁生教授一生科研工作的总结，是留给后人的一笔财富。科技发展无止境，在他的基础上进一步发展提高，才是出本论文集的更高期望。

书籍目录

屈梁生院士简介序机床自动停刀机构精度的研究热变形对导轨磨削精度的影响计量光栅的几何、光学特性评几项超微米技术的进展自回归谱在机器故障诊断中的应用时间序列分析在机器状态识别中的应用On—line Surveillance of Process Equipment Using Autoregressive Feature Extraction and Walsh TransformationKULLBACK—LEIBLER信息数在识别机器振动信号功率谱中的应用An ExDerimental Research on Optical Fiber Sensor and Its Application to the Grinding Process Supervision机器故障诊断技术及其现状与展望The Evaluation of Low—Frequency Knock of Precise Gearbox Via Spectra Distance Measure Autoregressive Cepstrum and Its Application in Noise Diagnosis of GearboxesOn—line Surveillance of a Grinding Process Via a Kullback—Leibler Information Number Statistical Control of Low Frequency Knock in Machine Tool Gearbox Production光纤传感器及其在磨削过程监测中的应用The Thermal Behavior of Machine Tool GuidwaysThe Application of Maximum Entropy Spectrum in Analysis of the Accuracy 0．f Precision Gear Transmission The Evaluation of Low—frequency Knock in the Gearbox of Modern Machine Tools Friction Failure Diagnosis of Steam Turb0—Generator Sets Via Principal Components andAutoregressive TechniqueGrinding Process Supervision Via Information Distance Measure A New Approach to Computer—aided Vibration Surveillance of Rotating Machinery Discovering the HolospectrumThe H010spectrum：A New Method for Rotor Surveillance and Diagnosis 专家系统中知识一致性和完备性检验的一种方法The Holospectrum：A New FFT Based Rotor Diagnostic MethodA New Me!thod Evaluating Operating Condition for Rotating MachineryTime—frequencv Distributions of Vibration Signals in Rotating Machinery A Rule—based ExPert System for Real—Time Gear—manufacturing Process Control 回转机械振动频谱的模糊分类及应用The Phase Information in the Rotor Vibration Behaviour Analysis Investigation of the Special Trending Method for Rotating Machinery Monitorin9Rotating MachineRy Fault Diagnosis Using wigner Distri bution全息谱分析方法的原理和应用神经网络在大型回转机械故障诊断中的应用全息谱技术用于化工厂机械故障诊断机械监测诊断中的谱距分类与预报大型机组诊断信息的深层次处理问题IntelugentConntro1 oftheGear—Shaving Process转子径向摩擦故障诊断技术的研究HomoSourCe ANC and Its Application to the Machinery DiagnosisStudV and PerformanCe Evaluation of Some Nonlinear Diagnostic：Methods for．Large Rotating Machineryorbit Corrlplexity：A New Criterion for Evaluating the Dynamk：QuantyofRotorsystems齿轮联轴节对中不良振动信息研究Fault Prognosis for Large Rotating Machinery Using Neural Net work基于二维全息谱的回转机械亚同步振动分析Vibrational Diagnosis of Machine Parts Using the wavelet．：Packet Technique转子横向裂纹振动诊断技术的研究谈谈机组故障的可诊断性问题谈谈机组故障的可诊断性问题(续)小波包原理及其在机械故障诊断中的应用Somc Analvtical Problems of High Performance Flexm．e Hinge and Micro—mot：ion Stage Design The Prognosis of Vibration Condition for a 200 MW Turbo—Generator Set Using Anificial NeuralNetworkThe Noise Suppression Chain—A New Approach to Time Series Preprocessmg共轭梯度神经网络的研究机械故障诊断技术与当代前沿科学(一)机械故障诊断技术与当代前沿科学(二)机械故障诊断技术与当代前沿科学(三)小波分析的工程理解及其在机械诊断中的应用小波包的移频算法与振动信号处理柔性转子键相信号初始相位及振动信号相位的确定TeSt SCquencing and Diagnosis in Electronic System withDecision Table大型回转机械故障的组合网络识别方法基于神经网络的喘振早期发现Predicting Grinding Burn Using Artificial Neural Networks非线性动力系统理论在机械故障诊断中的应用基于信息优化的前馈神经网络及其应用人工神经网络与机械工程中的智能化问题全息谱的分解及其在机械诊断中的应用多分辨小波网络的理论及应用小波分析及其在机械诊断中的应用The Fauh Recognition Problem in Engineering Diagnost icsFractal Geometry Used for Chartactcrization of Grinding wheel Profiles全息动平衡技术：原理与实践多平面平衡中平衡面相关问题的处理全息谱力、力偶分解法在全息动平衡中的应用Optimization of the Measuring Path on a Coordinate Measuring Machine Using Genetic Algorithms机械诊断中的故障识别问题平稳熵：一种新的机组运行瞬时稳定性定量化监测指标转子横向裂纹故障的诊断信息提取大机组振动信号复杂性的定量描述截尾奇异值分解技术在动平衡中的应用三维全息谱分解在回转机械诊断中的应用研究回转机械诊断信息的集成：全息谱技术十年基于主分量分析的噪声压缩技术研究全息谱十年：回顾与展望转子动平衡中的相关平衡面问题基于扭振信号的齿轮故障诊断研究Intelligent Method for Online Vibration MonitoringA Soft Computing Based Approach for Multisensor Data FusionFeaturc Extraction Using Continuous WaveletTransform and Its Application for Mechanical Fault Diagnosis基于概率神经网络的机组状态多步预报方法A Difference Resonator for Detecting weak SignalsA Synergetic Approach to Genetic Algorithms for Solving Traveling Salesman ProblemA Genetic Algorithm Based Balancing Framework for Flexible RotorsOne Decade of Holospectral Technique：Review and Prospect小波分析及其在压缩机气阀故障检测中的应用研究改进的决策树生成算法及条件决策表的创建基于概率神经网络的大机组组合故障识别一种改进的随机减量信号提取方法Feature Extraction Based on Morlet Wavelet and Its Application for Mechanical Fault Diagnosis机械诊断中的几个基本问题遗传算法进化截止代数分布规律的研究非对称转子的全息动平衡技术The OptimizationTechnique－Based Balancing of Flexible Rotors Without Test Runs基于影响系数法的柔性转子无试重平衡法研究遗传算法优化效率的定量评价A Reduction Method of Rough Set Model遗传算法在故障特征选择中的应用研究应用连续小波变换提取机械故障的特征基于连续小波变换的信号检测技术与故障诊断神经网络在转子动平衡中应用的几个关键问题A New Practical Modal Method for Rotor Balancing柔性弯曲转子的特征识别与诊断Instantaneous Purified Orbit：A New TOol for Analysis of Nonstationary Vibration of Rotor SystemUsing Continuous Wavelet Transform to Detect Impact Component in Machine Fault Diagnosis机器振动诊断中信号处理方法的研究Extracting the Characteristic Frequency from Vibration Signal of Rolling Bearing with Self—OrganizinMetbodPartially Blind Source Separation of the Diagnostic Signals with Prior Knowledge提高故障诊断质量的几种方法Machine Diagnosis with Independent Component Analysis and Envelope Analysis应用Bootstrap方法构造机械故障特征库机械信号连续小波系数的统计特征研究提取机械信号中弱冲击成分的研究遗传编程在无量纲指标构建中的应用Some Applications of Statistical Simulation in Engineering Diagnostics基于小波变换和支持向量机的人脸检测系统自适应阈值选择和小波消噪方法研究应用独立分量分析提取机器的状态特征汽车发动机诊断的统计模拟方法故障诊断中多传感器信息冗余性的研究循环统计量方法在滚动轴承故障诊断中的应用Application of Wavelet Packet Analysis for Fault Detection in Electro—Mechanical Systems Based on Torsional Vibration Measurement基于冗余信息量的设备行为可预测性的研究Enhanced Diagnostic Certainty Using Information Entropy TheoryCyclic Statistics in Rolling Bearing Diagnosis一种贝叶斯诊断网络的拓扑结构贝叶斯网络推理的一种仿真算法基于遗传编程和支持向量机的故障诊断模型二代小波消噪在数字信号处理器中的实时实现应用全息谱技术诊断热变形不均匀引起的振动故障回转机械故障诊断中的三维全息谱技术机械产品的信息化——面向机械装备的信息技术全息谱技术在现场动平衡前故障诊断中的应用机械故障诊断的推理规律研究Translation—Invariant Based Adaptive Threshold Denoising for Impact Signal基于混沌和符号序列统计的滚动轴承故障诊断Discussion Authors ReplyA Rapid Response Intelligent Diagnosis Network Using Radial Basis Function NetworkApplication of Adaptive Neuro—Fuzzy Inference System in Field BalancingA nonlinear Diagnosis Method of Gear Early Fatigue CrackSigmoid Model：A Simulation for Inference Process of Engineering Diagnosis齿轮早期疲劳裂纹的混沌检测方法An Improved Independent Component Analysis algorithm and Its Application in Preprocessing of Bearing soundsA New Time Series Forecasting Approach Based on Bayesian Least Risk PrincipleA genetic Algorithm Based Balancing Framework for Flexible RotorsApplications of Chaotic Oscillator in Machinery Fault Diagnosisz息动平衡原理A New Field Balancing Method of Rotor Systems Based on Holospectrum and Genetic AlgorithmDiagnosis of Subharmonic Faults of Large Rotating Machinery Based on EMD……附录

编辑推荐

《机械监测诊断中的理论与方法》适合大学相关专业的教师、研究生，以及从事机械故障诊断研究的科研人员阅读。

作者简介

《机械监测诊断中的理论与方法》全文收录了中国工程院院士屈梁生的166篇高水平论文。内容涉及全息谱、全息动平衡、独立分量／主分量、质量保障与优化、小波分析、遗传算法、贝叶斯网络、支持向量机、模糊理论、神经网络、专家系统、监测预报、Wigner分布、噪声抑制、回归分析、信息化生产、粗糙集、决策表／决策树、循环统计、统计模拟、EMD、故障诊断理论与方法等。另附有345篇论文的总篇目、论文主题索引和研究对象索引三个附录。

图书封面

---

 机械监测诊断中的理论与方法下载

---