神经网络

Aimed at the characteristic of multiple types of faults possibly happened in nuclear power plant, large scale of training sample, and requirement of quick and accurate diagnosing, after the analysis of large sample number generated by large fault type exceeds limitation of 64K paragraph in DOS and large number of local minmum in error surface, measurements of same format of data file, Win95/NT operating system platform, resetting weight learning rate, dynamic training set in quick learning algorithm and improvement on quick learning algorithm using homotopy method which can avoid local minmum points in error surface have been adopted to ensure quickly and effective process of the course of neural network's training and testing.

针对核动力装置可能发生的故障种类多，训练样本规模大，故障诊断需要快速准确的特点等，分析了样本多和故障种类多产生的训练样本量超出DOS段大小和网络误差曲面上局部极小值多的情况，采取了使用相同格式数据文件、Win95/NT操作系统平台，对快速学习算法采用重置神经网络权值学习率和动态训练集、并采用能够有效克服网络误差曲面上局部极小点的同伦方法对学习算法进行改进等措施保证神经网络训练测试过程的快速、有效地进行；针对装置发生的故障须快速、准确诊断的需要，分析了故障的产生对装置参数变化的影响及操纵员对故障诊断的基础，在核动力装置发生故障时参数曲线的变化量与正常运行时参数曲线的变化量存在明显差异的基础上，提出采用参数曲线的变化量作为神经网络的输入，并围绕参数变化量的方法采用二次曲线拟合滤波求变化量和阈值技术来保证神经网络得到精确的装置参数变化量，从而得到准确的诊断结果。

Randomized orthogonal table is used to produce the specimens for radical basis function neural network.Self-organizing feature map is used to classify the specimens.

选取模具参数和工艺参数等作为影响冲压成形结果的因素，用正交表和随机法产生径向基函数神经网络的学习样本；利用自组织神经网络对样本进行分类，用有限元网格法反算的毛坯的长度作为神经网络的输出；设计了神经网络流程，定义了神经网络输出与有限元分析数据的相对误差。

Aiming at the disadvantage of conventional BP neural network, such as selecting parameter values by the empirical method, slow convergence speed and easy trap into local minimum points, this paper designs an improved BP neural network system. In order to improve the network convergence rate and reduce the training error, this paper optimizes the initial connecting weight value of neural network by the use of ant colony algorithm and trains artificial neural network by Levenberg-Marquardt algorithm.

针对常规BP神经网络参数的经验式取值方法以及收敛速度慢，容易陷入局部最小点等缺陷，设计了一种改进的神经网络系统，利用蚁群算法优化神经网络连接权初值，并采用LM算法对人工神经网络进行训练，提高了网络的收敛速度，降低了训练误差。

In chapter 4, basing theories and methods of scientific visualization, and artificial neural network BP algorithm, we integrate the Visual C++, OpenGL graphics library and Excel VBA technique to develop the program of artificial neural network and to make the BP algorithm visually, this program works can be divided into four parts: Using C language to develop program about BP algorithm; Using Visual C++, develop the GUI Interface, make input parameter visually; Using OpenGL graphic technique to display the training sample point in three dimension; at last using Excel DDE technique display the error graphic tables in Excel system In chapter 5, on the view of engineering application, we establish new method of surface reconstruction basing artificial neural network, develop interface program between module and commercial CAD/CAM system, meantime deeply discuss some key problems, for example, setting up the base plane, using the API technique, cutting and editing surface boundary, and also discuss the more compliant problem: how to intersect surface, at end we finish the work of translation from our surface reconstruction module to commercial CAD/CAM system, then make reverse engineering system basing artificial neural network more useful.

第四章基于科学计算可视化理论，依据人工神经网络BP算法理论模型，综合Visual C++，OpenGL图形库以及Excel VBA等多项软件开发技术，编制了人工神经网络程序，实现了BP算法的可视化映射。具体工作分为四部分：利用C语言实现人工神经网络BP算法；利用VisualC++的GUI技术开发图形用户界面，实现参数设置可视化；利用OpenGL图形技术进行三维映射，显示学习样本及训练样本点；利用微软电子表格DDE动态数据交换技术，在Excel上动态显示学习误差曲线图。第五章从工程应用的角度出发，提出了一种新的基于人工神经网络算法的曲面裁剪重构方法，完成了曲面重建模块与通用CAD/CAM系统的接口设计工作，对其中的若干关键问题进行了深入讨论，例如基平面设定、API技术的应用、边界裁剪等问题，同时，对曲面计算中较为困难的曲面相交问题也进行的专门探讨，最终完成了曲面重建模块向CAD/CAM系统的数据传输工作，使人工神经网络逆向工程系统趋向实用。

First, the Lyapunov functional and variation of constants method are adopted to study the effect that Sigmoid function and the relation of resistance, capacitance and current in Hopfield neural networks have on the stability of networks. The stability criterion constructed by physics parameters is obtained. Thus how the constrained relation of physics parameters affects the stability of Hopfielf neural networks is clear. Based on the study above, the perturbation model of recurrent neural networks is constructed. And the theorems of the existence of solution of perturbation model are presented.

首先，采用Lyapunov泛函法和常数变易法研究Hopfield神经网络中给出的电阻、电容、电流之间的关系以及Sigmoid函数对网络稳定性的影响规律，得出仅由物理模型参数构成的稳定性判据，从而弄清物理模型参数约束关系对Hopfield神经网络稳定性所起的作用，在此基础上，构建了递归神经网络的扰动模型，并通过讨论扰动模型解的存在性问题，给出递归神经网络扰动模型解的存在性定理。

And we used them into the inspection of these polluted gases. We used them because the constringent speed of wavelet neural network is very fast, they are not sensitive to the inputs of the network and they have the characteristic that they also can effectively approach the functions or signals. Classical neural networks mostly train the network with back propagation algorithm.

在目前常用的一维小波神经网络的基础上，我们研究了用来处理多维数据信息的小波神经网络，并将其应用于室内混合气体检测，主要是基于小波神经网络学习收敛速度较快，对网络输入不是很敏感，以及小波神经网络可以有效的进行函数逼近或者信号逼近的特点。

The application of this theory in the attributable optimization of fault diagnosis characteristic parameters is explored. After a brief description of concepts involved in the paper, such as rough set theory and the indiscernibility relation, the methods of how to preprocess the loss value of data and how to discrete the data are discussed.

论文对粗糙集理论及方法在神经网络技术中的应用进行了研究，从而将粗糙集理论与神经网络技术相结合，提出了基于粗糙集理论的神经网络识别系统，利用粗糙集理论来确定神经网络的输入节点数，降低了神经网络的结构的复杂性，给出了该算法的详细步骤。

Based on the differences between the design value and the survey of the bridge deck elevation in constructing the pre-stressed and so on working condition, the model parameters are identified by ANN, and the beam camber of different sections are given. The results demonstrated that the ANN is efficient and the theoretical results fit well with that of forecasting the beam camber in construction.In this thesis, aim to the beam axis shape be exactitude for design, the internal force of the beam be appeased by design request, the concept of modem cybernetics is applied, At the same time, the effecting parameter has collected. The paper build the ANN model in predicting the beam camber an construction and discuss indetail the extraction of inputting nodes information when quoting the fundamental principle of ANN model and introducing research process of the improved BNN. According to them, ANN is applied in the construction control of prestressed concrete continuous bridge. It is successful for used in the No. 4 bridge of Xiangfan bridge over-crossing Hanriver brideg. To predict the construction camber of by ANN method is useful and effective.

本文以大桥成桥线形满足设计要求和成桥内力控制在设计容许范围内为目的，运用现代控制理论的思想，引用神经网络基本原理，详细介绍BP神经网络的研究过程，从而建立预应力混凝土连续梁桥施工预拱度的神经网络模型，并详细讨论了模型的神经网络结点信息的提取，在工程实例的基础上进行了数值计武汉理工大学硕士学位论文算，数值表明，将神经网络应用于预应力混凝土连续梁桥施工控制，获得了满意的结果，在预应力混凝土连续梁桥施工控制技术研究方面，具有方便有效、精度高的优点，具有良好的应用前景。

Coal load in ball tube of pulverizing system is affected by many factors, this paper analyzes these factors and their relations.

由于神经网络技术的发展，神经网络在非线性控制方面显示出强大的优势，同时近年来出现的逆系统方法对于多变量非线性系统的控制指出了一条可算是光明的道路，鉴于以上两项技术的优越性，本文提出对BMPS进行神经网络逆系统控制的方法，首先通过机理分析的方法建立了BMPS的非线性动态数学模型，并依此模型设计了神经网络逆系统控制方案，由于此方案中需要用到球磨机磨筒内负荷的反馈，但是磨内负荷是一个不能直接测量的量，而且其测量方法是多年来人们一直在努力的方向却又一直不能较好地解决，本文提出对其进行神经网络测量的思想和方法，这一方法的仿真试验和现场实测结果均表明其研发造价低廉的优越性和测量较为准确的实用性。

The dissertation recommends some kinds of methods and measures: such as adjusting the network configuration, connecting value and threshold value by the total value of all stylebooks, adding part of the adjusting value of the last time to the current adjusting value, transforming the stylebooks to standard value, optimizing activation function, appending threshold value to the putout of the nerve cell and so on. Simultaneity, the author brings forward a new method to optimize the model of artificial neural network . It is using the automatically adaptive genetic algorithm to make the network configuration, connecting value and threshold value of artificial neural network better. The method can make the model better and improve the simulating effect and forecasting precision.Genetic algorithm is an arithmetic based on evolution and genetics used to search the optimization .

本文针对人工神经网络应用中存在的上述问题，介绍了各种改进方法与措施：如用所有样本的总效果对网络权值矩阵和阈值向量进行调整、调整量中加入动量项、标准化训练样本数据、优化激励函数以及给神经元的输出值添加偏置量等；同时还提出了一种新的优化人工神经网络模型的方法，即采用自适应遗传算法对人工神经网络模型的网络结构和权值阈值进行全局优化搜索，以提高大坝安全监测人工神经网络模型的拟合成果和预测精度。

Singer Leona Lewis and former Led Zeppelin guitarist Jimmy Page emerged as the bus transformed into a grass-covered carnival float, and the pair combined for a rendition of "Whole Lotta Love".

歌手leona刘易斯和前率领的飞艇的吉他手吉米页出现巴士转化为基层所涵盖的嘉年华花车，和一双合并为一移交&整个lotta爱&。

This is Kate, and that's Erin.

这是凯特，那个是爱朗。