中速磨存煤量软仪表监测

利用混沌运动初值敏感性,能在一定范围内按其自身规律不重复的遍历所有状态获取全局最优值的特点,在电厂现有的DPU中编程实现基于混沌优化BP网络的中速磨存煤量软仪表。离线训练与计算机仿真结果表明,本软仪表具有可行性和有效性,对实际生产具有指导意义。     

METHOD FOR ADAPTIVE MESH GENERATION BASED ON GEOMETRICAL FEATURES OF 3D SOLID

In order to provide a guidance to specify the element size dynamically during adaptive finite element mesh generation, adaptive criteria are firstly defined according to the relationships between the geometrical features and the elements of 3D solid. Various modes based on different datum geometrical elements, such as vertex, curve, surface, and so on, are then designed for generating local refined mesh. With the guidance of the defined criteria, different modes are automatically selected to apply on the appropriate datum objects to program the element size in the local special areas. As a result, the control information of element size is successfully programmed covering the entire domain based on the geometrical features of 3D solid. A new algorithm based on Delaunay triangulation is then developed for generating 3D adaptive finite element mesh, in which the element size is dynamically specified to catch the geometrical features and suitable tetrahedron facets are selected to locate interior nodes continuously. As a result, adaptive mesh with good-quality elements is generated. Examples show that the proposed method can be successfully applied to adaptive finite element mesh automatic generation based on the geometrical features of 3D solid.     

混沌工程学研究的新进展

归纳了混沌工程学的研究分支，评述了混沌工程学近年来的研究成果。指出利用混沌性质解决各种工程实际问题将是混沌工程学今后的主要研究方向。     

应用混沌神经网络诊断发动机磨损故障

利用混沌神经网络,采用混沌动力学中的Logistic映射构造神经元,对发动机的相似早期磨损故障进行了有效诊断。通过故障模拟实验,对不同类型初期磨损故障的振动数据进行了采集和分析,发现其频谱特征具有较大的相似性,且各种初期磨损故障的李雅普诺夫指数大于0,具有混沌特征。用一般方法难以区分不同类型的初期磨损故障,而混沌神经网络能够有效地识别这些相似故障模式,对于发动机磨损故障的早期预防具有积极作用。     

Estimation of length density Lv from vertical slices of unknown thickness

Length density of lineal features, L_v, is an important stereological parameter. The efficient stereological procedure for the estimation of L_v from the counting measurement performed on the projected images of the vertical slices (foils) is modified and improved: L_v can be now estimated from vertical slices of unknown thickness, and the slices need not be of the same thickness. The required assumption-free stereological relationship is is the average number of intersections of straight test lines parallel to the vertical axis with the projected images of the lineal features in the vertical slices, per unit test line length. is the average number of intersections of the lineal features with the parallel planes of the vertical slices, per unit area. Note that there are two parallel planes in a slice, and therefore their total area is twice the area of the observed projected image frame. is the average number of intersections of orientated cycloid-shaped test lines (minor axis perpendicular to vertical axis) with the projected images of the lineal features in the vertical slices, per unit length. For practical applications of this result, it is necessary uniquely to identify the points of intersections of the lineal features with the parallel planes of the slices in the projected images, so that can be estimated unambiguously. However, in practice, this is not a problem in biological, as well as materials microstructures.     

悬辊式旋风磨机的混沌振动分析

对振动超标的悬辊式旋风磨机进行了振动分析，建立了反映磨机运动的非线性微分方程。用非线笥振动理论和Melnikov方法，分析了混沌阈奶，指出磨机工作状态处于混沌区，根据数值计算和现场测试分析，采取了有效措施，改变了磨机的工作状态，减小了磨机的振动和噪声。     

Extraction/conversion of geometric dimensions and tolerances for machining features

It is important for a feature-based system to preserve feature integrity during feature operation, especially when feature interaction occurs. The paper presents a feature conversion approach to convert design features used in a design model into machining features for the downstream applications. This process includes both form features (geometric information) and non-geometric features conversion. Most researchers have concentrated on geometric information extraction and conversion without tackling the important problem of non-geometric feature information. This paper focuses on the extraction and conversion of feature geometric dimensions and tolerances (GD&T) for downstream machining application.The main barrier to the integration of a feature-based CAD/CAPP/CAM system – feature interaction – is discussed in this paper, which alters design features in their geometries and non-geometric information. How to identify and validate these feature dimensions and tolerances is one of the key issues in feature interaction conversion. The development of robust methodologies for preserving feature integrity for use in process planning application is the main thrust of the work reported in this paper.     

基于MODWPT与LFDA的模拟电路故障诊断

在模拟电路故障诊断过程中,存在故障特征信息提取不充分以及特征信息冗余的问题,对此,提出一种基于最大重叠离散小波包变换(MODWPT)与局部Fisher判别(LFDA)的模拟电路故障诊断方法。该方法中,首先利用MODWPT进行模拟电路原始信号处理与故障特征提取;随后,针对高维特征集中存在冗余信息,不利于模式识别与分类,利用LFDA方法进行降维,获取更有益于故障模式识别的低维特征集;最后,支持向量机(SVM)作为故障模式识别分类器,在此基础上构建模拟电路故障诊断模型。电路仿真实验结果表明,所提出方法的最大故障诊断准确率可达99.17%,从而验证了所提方法的有效性。     

优化数学模型特征的提取与识别

介绍了对优化数学模型的特征进行提取与识别的方法,用它可以提取和识别数学模型的各种特征。     

时间序列分维的改进GP算法

在拓扑等价的意义上,证明了系统单变量时间序列混沌吸引子的分维与度量无关,改进了计算分维的ＧＰ算法（ＮＧＰ）,给出了递推ＧＰ算法,并利用此算法计算了Ｈｅｎｏｎ吸引子和Ｌｏｒｅｎｚ吸引子的分维,通过比较发现运算速度显著提高,增强了算法的实用性。