基于多值关联矩阵扩展的诊断策略设计

对于复杂系统,在实际测试中测试不通过的情况可能有多种,如果依然采用基于确定性关联矩阵的测试序列生成算法,将损失大量信息,导致测试精度和效率降低．提出了基于关联矩阵扩展的诊断策略设计方法,将多值关联矩阵扩展为二值关联矩阵,然后利用AO*算法得出诊断策略．实例验证表明,算法提高了故障隔离精度,为不确定系统的诊断策略设计提供一种高效算法．     

基于解析模型的气制动系统泄漏故障诊断研究

针对客车气压制动系统中存在的泄漏故障的问题,提出了一种基于解析模型的客车气压制动系统泄漏故障诊断的方法。通过对气制动系统中的关键部件制动总阀及制动气室的工作状态进行分析,建立了正常状态下的客车气制动系统制动气室压力特性的理论模型,该模型能够预测制动过程中制动气室中的瞬时气压值。搭建了整车气压制动系统模拟试验台,该试验台可针对整车制动系统进行制动性能模拟试验。在模拟制动过程中对制动气室中气压变化量进行测量,将归一化后的试验值与理论值进行比较,对气制动系统有无泄漏故障的存在进行了诊断。试验结果表明,该方法能很好地对气制动系统泄漏进行故障诊断。     

应用概率神经网络诊断自行火炮发动机的故障

目的　研究概率神经网络模型 ,并应用于故障诊断 .方法　对基于概率统计思想和 Bayes分类规则的概率神经网络模型、网络结构、算法及其特点进行分析 ,利用其进行故障诊断 ,并提出一种优化估计平滑因子的方法 .结果　概率神经网络可很好地诊断自行火炮发动机进行中油路和气路的故障 .结论　概率神经网络在模式识别和故障诊断领域中可取得良好地应用效果     

基于多Agent的远程协作故障诊断模型研究与设计

随着Agent技术的发展,出现了多种基于多Agent远程协作故障诊断模型,但它们不能很好地满足实时性要求较高系统的需求,因此本文以诊断Agent为基础,构建了基于多Agent的远程协作故障诊断模型,并对模型采用了层次划分,根据各层Agent的功能,采用合同网协议对Agent间任务进行分配协作,提高了系统故障诊断速度,从而为远程故障诊断的实现提供了支持。     

应用概率神经网络诊断自行火炮发动机的故障

目的 研究概率神经网络模型,并应用于故障诊断。方法 对基于概率统计思想和Ｂａｙｅｓ分类规则的概率神经网络模型、网络结构、算法及其特点进行分析,利用其进行故障诊断,并提出一种优化估计平滑因子的方法。结果 概率神经网络可很好地诊断自行火炮发动机进行中油路和气路的故障。结论 概率神经网络在模式识别和故障诊断领域中可取得良好的应用效果。     

基于多Agent的远程协作故障诊断模型研究与设计

随着Agent技术的发展,出现了多种基于多Agent远程协作故障诊断模型,但它们不能很好地满足实时性要求较高系统的需求,因此本文以诊断Agent为基础,构建了基于多Agent的远程协作故障诊断模型,并对模型采用了层次划分,根据各层Agem的功能,采用合同网协议对Agent间任务进行分配协作,提高了系统故障诊断速度,从而为远程故障诊断的实现提供了支持.     

基于小波神经网络的水下航行器传感器故障诊断

针对水下航行器系统的传感器故障诊断问题,提出了一种基于小波神经网络的传感器故障诊断方法．在对水下航行器系统的传感器故障信号进行特征提取时,发现其大部分能量都集中在低频部分．若直接以此能量分布来区分正常与故障、故障与故障的信号,将导致神经网络训练时间和分辨时间都会很长,因而不能实时地监控系统．为了很好地进行区分,凸显其差异,将低频部分能量舍去,只保留其余部分,并将其归一化,再利用径向基神经网络进行分类．利用小波分解的节点能量差异与特征提取特点以及神经网络的自我学习能力,通过大量的样本训练后,使神经网络很好地分辨出5类故障信号及正常信号．仿真结果表明：此方法简单、易于实现,适于水下航行器的传感器故障的诊断．     

基于ISSA⁃SVM的露点测量系统电路故障诊断方法研究

针对高精度谐振式露点测量系统中电路故障诊断问题,提出了一种基于改进的麻雀搜索算法（Improved Sparrow Search Algorithm, ISSA）优化智能分类器参数的电路故障诊断模型,采用测前仿真故障诊断方法中的智能诊断方法,选择适用于小样本、非线性问题的支持向量机（Support Vector Machine, SVM）作为智能分类器,针对麻雀搜索算法中收敛速度慢、易陷入局部最优等问题进行改进,并将改进后的优化算法用于SVM参数寻优,构建ISSA?SVM故障诊断模型用于谐振电路故障诊断。实验结果显示,ISSA?SVM模型在建立的电路上能够达到88.9%的故障诊断率,可靠性较强,能够作为高精度谐振式露点传感器电路的故障诊断方法。     

故障诊断Agent研究

设备故障诊断技术在现代社会的应用越来越广泛 ,其应用对象多处于动态多变环境。现有的诊断软件尚不能适应这样的复杂环境。智能 agent是目前人工智能领域的研究热点之一。随着网络技术的成熟 ,实用型的智能 agent系统越来越多地应用于各个领域。本文根据故障诊断系统的实际要求 ,提出并初步实现了基于 agent的故障诊断系统。这进一步提高了诊断系统的自动化程度 ,较好地解决了诊断软件的通用性和适应性之间的矛盾。本文提出的诊断 agent系统能很方便地应用于各种网络环境 ,从而可以更充分地利用已有的诊断资源 ,便于构造大型诊断系统。     

基于灰色理论的电站机组状态监测与故障诊断方法

本文主要研究了灰色理论在大型旋转机械的状态监测与故障诊断中的方法与应用问题。鉴于一般关联度分析方法在故障诊断应用中没有考虑到特征参数对诊断结果的不同影响,本文提出了加权关联度这一新概念,并给出了加权关联度用于旋转机械诊断中的方法与步骤。这种诊断方法能根据诊断结果采用自学习修正方法来修正关联度参数,提高诊断的准确度,并有很好的前景。     

A numerical method for coupled differential equations

This paper describes a numerical method for solving first-order coupled matrix differential equations. Recursive equations are used to find ‘reflection’, ‘transmission’ and ‘source’ coefficients; these coefficients are then used to construct the vector solution to the differential equations. The method can be used to solve linear and non-linear differential equations with specified initial or two-point boundary values. Numerical results for several initial value problems are given in the paper.     

基于Woodbury非线性方法的迭代算法对比分析

在结构局部非线性求解过程中,刚度矩阵仅部分元素发生改变,此时切线刚度矩阵可写成初始刚度矩阵与其低秩修正矩阵和的形式,每个增量步的位移响应可用数学中快速求矩阵逆的Woodbury公式高效求解,但通常情况下迭代计算在结构非线性分析中是不可避免的,因此迭代算法的计算性能也对分析效率有重要影响。该文以基于Woodbury非线性方法为基础,分别采用Newton-Raphson （N-R）法、修正牛顿法、3阶两点法、4阶两点法及三点法求解其非线性平衡方程,并对比分析5种迭代算法的计算性能。利用算法时间复杂度理论,得到了5种迭代算法求解基于Woodbury非线性方法平衡方程的时间复杂度分析模型,定量对比了5种迭代算法的计算效率。通过2个数值算例,从收敛速度、时间复杂度和误差等方面对比了各迭代算法的计算性能,分析了各算法适用的非线性问题。最后,计算了5种算法求解基于Woodbury非线性方法平衡方程的综合性能指标。     

On updating the structure of sparse matrix factors

Structural analysis often involves the solution of a sparse system of linear equations using matrix factorization. For structural reanalysis problems, the system of equations is progressively changing and matrix factorization is required at each iteration step. In this paper, we study the problem of updating the structure of sparse matrix factors using an ordered-tree model. The ordered-tree model provides the information needed to symbolically compute the non-zero structure of the matrix factor of a sparse matrix. Furthermore, for matrix factor modification problems, this tree model can be used to determine the numerically modified entries in the matrix factor. Examples for modification of a structure and h-adaptive refinement of a finite element model are given to illustrate the potential application of the procedures developed in this study.     

直角坐标系下层状地基力学计算中的传递矩阵技术

给出了在直角坐标系下计算层状地基力学问题的传递矩阵技术,改变了过去只能在柱坐标系下求解层状地基力学问题的状况,大大地简化了任意荷载作用下层状地基力学问题的计算。从直角坐标系下弹性问题的基本方程出发,通过改变坐标系的常规位置,成功地利用Laplace变换及其微分性质,首次推导出直角坐标系下的状态控制方程和对应的传递矩阵,并给出了利用该方法的求解过程。     

基于完全笛卡尔坐标的多体系统微分-代数方程符号线性化方法

多体系统动力学方程分为两类形式,即微分方程和微分-代数方程。这两类方程都是针对大位移系统,并且方程呈强非线性。为研究多体系统小位移或振动问题,从多体系统动力学方程出发,讨论微分-代数方程线性化计算机代数问题。利用完全笛卡尔坐标描述多刚体系统,建立多刚体系统动力学微分-代数方程。利用逐步线性化方法和计算机代数,分别对多体系统微分-代数方程的广义质量阵,约束方程和广义力阵在平衡位置附近进行Taylor展开。给出一种基于完全笛卡尔坐标的多体系统动力学微分-代数方程符号线性化方法。最后通过两个算例验证该方法的有效性。     

Inverse damage prediction in structures using nonlinear dynamic perturbation theory

A non-linear perturbation theory which furnishes an exact relationship between the perturbation of structural parameters and the perturbation of modal parameters is presented. A system of governing equations is derived, where the information about incomplete modal data can be directly adopted. The Direct Iteration and the Gauss–Newton Least Squares techniques for an inverse prediction of structural damage are discussed, where both the location and the extent of structural damage can be correctly determined using only a limited amount of incomplete modal measurements data. Structural damage is assumed to be associated with a proportional reduction of the original element stiffness matrix or with a proportional reduction of the contribution of a Gauss point to the element stiffness matrix, which characterises a structure at an element level or at a Gauss point level. Finally, a damaged cantilever beam is considered using different model problems to demonstrate the effectiveness of the proposed techniques.     

Research on the Correlation Between Oil Menitoring and Vibration Monitoring in Information Collecting and Processing Monitoring

Oil monitoring and vibration monitoring are two principal techniques for mechanical fault diagnosis and condition monitoring at present. They monitor the mechanical condition by different approaches, nevertheless, oil and vibration monitoring are related in information collecting and processing. In the same mechanical system, the information obtained from the same information source can be described with the same expression form. The expressions are constituted of a structure matrix, a relative matrix and a system matrix. For oil and vibration monitoring, the information source is correlation and the collection is independent and complementary. And oil monitoring and vibration monitoring have the same process method when they yield their information. This research has provided a reasonable and useful approach to combine oil monitoring and vibration monitoring.     

Galerkin's method and the variational procedure

Galerkin's method and the variational procedure, when applied to most practical problems in electromagnetics, lead to matrix equations of the same form. Variational procedures for self-adjoint and nonself-adjoint operators also result in the same form of matrix equations for a large subclass of problems. However, the three cases may yield different matrix equations in general. This paper examines the subclass of problems for which these methods result in the same matrix equation and provides systematic ways for classification of problems for which two or all three of the cases lead to the same matrix equation. It also describes properties of the coefficient matrix in the matrix equation     

3H-2100型CCD高速摄像电视系统的应用

浙江大学研制成功的“3H-2100型CCD高速摄像电视系统”具有摄像频率高、冻结能力强、在计算机控制下,所有摄像参数均可在大范围内按需控制的特点,解决了动态测试中快速、在线、实时和全自动问题,具有广泛的用途。本文介绍系统在飞行子弹和高速旋转体试验中的应用。     

A noninvasive, remote and precise method for temperature and concentration estimation using magnetic nanoparticles

This study describes an approach for remote measuring of on-site temperature and particle concentration using magnetic nanoparticles (MNPs) via simulation and also experimentally. The sensor model indicates that under different applied magnetic fields, the magnetization equation of the MNPs can be discretized to give a higher-order nonlinear equation in two variables that consequently separates information regarding temperature and particle concentration. As a result, on-site tissue temperature or nanoparticle concentration can be determined using remote detection of the magnetization. In order to address key issues in the higher-order equation we propose a new solution method of the first-order model from the perspective of the generalized inverse matrix. Simulations for solving the equation, as well as to optimize the solution of higher equations, were carried out. In the final section we describe a prototype experiment used to investigate the measurement of the temperature in which we used a superconducting magnetometer and commercial MNPs. The overall error after nine repeated measurements was found to be less than 0.57?K within 310-350?K, with a corresponding root mean square of less than 0.55?K. A linear relationship was also found between the estimated concentration of MNPs and the sample's mass.