神经网络在电路板无损检测系统中的应用

提出一种基于人工神经网络(ANN)的电路板无损检测系统的故障诊断方法.红外热成像技术是增强电路板检测系统能力的方法之一,但系统自动诊断能力比较弱,在电路板红外热像数据处理过程中,引入人工神经网络形成故障诊断规则,有效提高了系统故障诊断能力.     

改进的BP神经网络与Snort系统的接入设计

Snort是一个轻量级的入侵检测系统（包含4个模块：解码模块,预处理模块,检测分析模块,输出插件）。文中针对BP神经网络的不足,对其算法进行了改进,并通过Snort系统的预处理插件,实现BP神经网络的接入;接入后用于对网络数据包的异常检测,实现误用与异常相融合的Snort系统。     

人工神经网络在压缩机故障检测中的应用

应用人工神经网络技术和多个特征参数实现对全封闭式压缩机的故障检测。经一定数量典型样本训练过的神经网络系统，对压缩机的故障具有较强的识别能力。利用声发射传感器进行信号提取和ＳＴＤ５０８８工控机进行信号处理，对压缩机运行状态进行在线检测，证实了该方法在压缩机故障检测中的正确性，可行性和实时性     

深度对比一类分类在轴承异常检测中的应用

针对机械设备的异常数据难以获取和传统异常检测方法容易误报的问题,提出了一种基于深度对比一类分类的无监督异常检测框架,用于检测轴承等关键部件产生异常的时间点。所提出的框架分为2部分。第1部分,针对深度一类分类方法的分布未知与模型坍缩问题,提出一种改进的深度对比一类分类损失,修改了相似度度量方式,并添加了增强样本对之间的相似度约束。在训练过程中,选取4种备选的数据增强方案进行实验和分析,并选取了最佳的数据增强组合,使模型学习得到了更加均匀的正常数据分布。第2部分,采用极值理论在检测过程中不断拟合分布尾部的极值分布,动态更新异常样本阈值进而避免误报。最后,在辛辛那提轴承寿命数据集上验证了提出的异常检测框架在特征分布的均匀性、异常样本的分类准确性与故障起始点检测的精准性方面都具有优越性。     

基于神经网络的入侵检测方法研究

在传统BP算法的基础上,提出对BP算法的改进方法,并对KDDCup99数据集进行预处理,以达到提高检测效率、降低误检率的效果。利用决策树ID3算法在筛选属性方面的优势,对KDDCup99数据集属性进行筛选,并将结果进行标准化与归一化处理,使训练数据和测试数据更加符合网络训练的数值特征。最后通过Matlab进行编程,将结果与传统BP算法加以比较,证实了该方法的有效性。     

模糊逻辑及遗传算法在入侵检测中的应用

在入侵检测系统中,使用模糊逻辑与数据挖掘相结合的方法挖掘关联规则,利用遗传算法调整模糊隶属函数并且选择一套合适的特征,以此提高入侵检测系统的检测性能.     

数据挖掘技术在入侵检测系统中的应用研究

扼要阐述了入侵检测和数据挖掘技术,并建立了采用数据挖掘技术的入侵检测系统模型.     

FANUC异常负载检测在加工中心的应用

本文介绍了利用发那科PMC窗口功能实时读取主轴负载信息,并通过PMC程序进行数据的处理,来判断主轴刀具负载是否异常的检测方法,达到有效保护刀具、工件和设备的目的。     

神经网络在语音识别中的应用研究

随着计算机技术的飞速发展。电脑已经走入千家万户,信息时代的到来使人们对计算机的智化要求也越来越高。人机对话让“机器”听懂人的语言越来越被人关注,语音识别技术正是实现这一功能。简单叙述了人工神经网络技术进行语音识别的原理。给出了其中的关键技术,求解语音特征参数和典型神经网络的学习过程。然后通过一个具体应用实例。展示了如何使这一技术实用化。     

往复压缩机相空间LDA模型在异常检测中的应用

针对往复压缩机异常检测不及时、漏报、误报的问题,提出一种多特征融合的相空间LDA(Latent Dirichlet Allocation)模型的异常检测方法。为了全面描述波形特征,提取往复压缩机正常运行数据和实时运行数据的特征集,对特征集进行预处理后,运用LDA方法计算正常状态和当前状态相空间分布模型,并用JS(Jensen Shannon divergence)距离计算两者差异度,若差异度超过设定值则认为发生故障。实验验证了该方法能有效实现往复压缩机异常检测,并能大幅提前往复压缩机典型故障异常检测报警时间点。     

Object Detection on Spatially Inhomogeneous Backgrounds Using Neural Networks

Optoelectronics, Instrumentation and Data Processing - Several approaches to the use of neural networks for object detection on spatially inhomogeneous backgrounds are considered. A method for...     

基于支持向量回归机的航空发动机异常检测研究

准确判断发动机的工作状态,预知发动机的性能变化,为预防和排除故障提供充足的时间和决策依据。开发出利用(quick access recorder,QAR)数据的发动机异常检测系统,该系统基于发动机的QAR数据,采用支持向量回归机(support vector regression,SVR)算法,建立监控参数与相关参数的回归模型。利用健康回归模型,监控后续航班参数是否出现异常,从而实现发动机故障检测。采用该系统监控航空发动机低压转子转速N1及高压转子转速N2,及时发现发动机运行异常,证明了系统的可行性和有效性。     

Using Artificial Neural Networks for Energy Regulation Based Variable-speed Electrohydraulic Drive

In the energy regulation based varibable-speed electrohydraulic drive system, the supply energy and the demanded energy, which will affect the control performance greatly, are crucial. However, they are hard to be obtained via conventional methods for some reasons. This paper tries to a new route: the definitive numerical values of the supply energy and the demanded energy are not required, except for their relationship which is called energy state. A three-layer back propagation(BP) neural network was built up to act as an energy analysis unit to deduce the energy state. The neural network has three inputs: the reference displacement, the actual displacement of cylinder rod and the system flowrate supply. The output of the neural network is energy state. A Chebyshev type II filter was designed to calculate the cylinder speed for the estimation of system flowrate supply. The training and testing samples of neural network were collected by the system accurate simulation model. After off-line training, the neural network was tested by the testing data. And the testing result demonstrates that the designed neural network was successful. Then, the neural network acts as the energy analysis unit in real-time experiments of cylinder position control, where it works efficiently under square-wave and sine-wave reference displacement. The experimental results validate its feasibility and adaptability. Only a position sensor and some pressure sensors, which are cheap and have quick dynamic response, are necessary for the system control. And the neural network plays the role of identifying the energy state.     

粒子群优化算法训练模糊神经网络

研究合适的神经网络学习算法是令人感兴趣的问题.提出一种用粒子群优化(PSO)算法训练模糊神经网络的方法.PSO的位置向量对应模糊神经网络的权值向量,而PSO的适应函数对应模糊神经网络的目标函数,然后,通过演化PSO达到训练模糊神经网络的目的.用PSO算法训练模糊神经网络预测混沌时间序列的实验结果表明PSO算法性能优良,适合训练模糊神经网络.     

神经网络与回归相结合实现传感器特性线性化

本文提出了一种用多个改进的BP神经网络与回归相结合的技术去实现传感器非线性误差校正的方法。该方法将传感器的特性曲线分为一个直线段和两个非线性低,并自适应的确定线性段的区间,在线性段,用回归方法拟合出直线方程。在非线性段则用两个改进的BP神经网络分别映射其反函数作为校正环节,从而实现非线性误差校正。仿真和试用表明,这种方法可使传感器的非线性误差减小近十倍。最后,给出了一些仿真实验及其结果。     

基于神经网络分析的鲜蛋破损检测

为了提高禽蛋生产中检测自动化程度,提出一种替代人工检测鲜蛋的自动化方法。该方法根据鲜蛋受到专用仪器轻轻敲击所发出的声音的频谱谱形识别鲜蛋的状态。谱形通过本文改进的BP神经网络进行识别．测试结果表明,本方法比较可靠。     

Modelling Deformation of Hydroformed Circular Plates Using Neural Networks

The process of applying fluid pressure to form metal sheets into desired shapes is widely used in the industry and is known as hydroforming. Similar to most other metal forming processes, hydroforming leads to non-homogeneous plastic deformation of the workpiece. Predicting the amount of deformation caused by any sheet metal forming process leads to better products. In this paper, a model is developed to predict the amount of deformation caused by hydroforming using an artificial intelligence technique known as neural networks. The data used to design the neural network model is collected from an apparatus that was designed and built in our laboratory. The neural network model has a feedforward architecture and uses Powell’s optimisation techniques in the training process. Single- and two-hidden-layer feedforward neural network models are used to capture the nonlinear correlations between the input and output data. The neural network model was able to predict the centre deflection, the thickness variation, and the deformed shape of circular plate specimens with good accuracy. ID="A1"Correspondance and offprint requests to: Dr M. Karkoub, Mechanical and Industrial Engineering Department, College of Engineering and Petroleum, Kuwait University, PO Box 5969, Safat 13060, Kuwait     

Adaptive Control of Multivariable Process Using Recurrent Neural Networks

Abstract

Industrial processes are naturally multivariable in nature, which also exhibit non-linear behavior and complex dynamic properties. The multivariable four-tank system has attracted recent attention, as it illustrates many concepts in multivariable control, particularly interaction, transmission zero, and non-minimum phase characteristics that emerge from a simple cascade of tanks. So, the multivariable laboratory process of four interconnected water tanks is considered for modeling and control. For processes which show nonlinear and multivariable characteristics, classical control strategies like PIDs have performance limitations. Hence, intelligent approaches like Neural Networks (NN) is an important term in this juncture. The use of Recurrent Neural Network (RNN) is apt for modeling and control of nonlinear dynamic processes as it contains the past information about the process. The objective of the current study is to design and implement an adaptive control system using RNN for a nonlinear multivariable process.

The proposed adaptive design comprises an estimator based on RNN, which adapts online and predicts one step ahead output. A Recursive Least Square (RLS) based back propagation algorithm is used for training the network. The controller used is also a RNN, which minimizes the difference between the predicted output and reference trajectory. The objective function is minimized using a steepest descent algorithm which gives the optimum control input. Desired performance of the system is ensured by the parallel operation of both. The proposed control strategy is implemented in a laboratory scale four tank system. The trajectory tracking and disturbance rejection response obtained are compared with the response obtained by using a well designed decoupled, decentralized IMC controller.     

基于自适应模糊神经网络的噪声抵消器

讨论了基于自适应模糊神经网络的噪声抵消器的设计方法。自适应模糊神经网络系统具有非线性映射和自学习能力，能够用于噪声信号的非线性建模。它不仅能够获取信号的最佳估计，并互能够克服信号处理中存在的模型和噪声的不确定性、不完备性。该法设计的滤波器效果良好，并可以用于多路信号和复杂信号的噪声消除。     

RBF网络在混凝土强度研究中的应用

本文以混凝土抗压强度建模为例,介绍一种自组织学习的RBF算法,并与广受欢迎的BP算法比较.仿真结果表明,RBF网络的学习速度显著加快,并具有较好的泛化能力,能有效地应用于混凝土领域.