基于线性空间投影的计算Volterra级数高阶核的方法

研究了对非线性动态系统作任意精度逼近的Volterra级数高阶核的全新估计方法。该方法在核函数理论基础上构造特殊线性空间,将求解Volterra级数的各阶核的问题转换为求用输出观测向量在希尔伯特空间中某一子空间上的投影问题,使原本复杂、难以计算的非线性系统的Volterra级数的逼近问题在所构建的线性空间中巧妙地以向量内积的方式解决,并给出了具体算法。相比于其他时域或频域估计Volterra核的方法,该算法的优点在于理论体系严密、计算量不会随着阶数增高而呈几何级数增加,辨识精度高,理论上能够辨识出任意阶的核,弥补了迄今现有的各种估计Volterra核的方法难以估计超过四阶或更高阶核的缺点,特别能够应用在对动态系统和强非线性系统的建模上。仿真研究的结果证明了该方法的有效性。     

基于Volterra泛函级数的涡轮建模研究

给出了对非线性动态系统做任意精度逼近的Volterra级数高阶核的全新估算方法并将其应用在涡喷发动机的转速控制上。该方法在核函数理论基础上,构造线性空间,将求解Volterra级数各阶核的问题转化为求输出观测向量在希尔伯特空间中某一子空间上的投影的问题,使原本复杂的非线性系统的Volterra级数的逼近问题在线性空间中以向量内积的方式得到解决。与其他时域或频域估算Volterra核的方法相比较,该算法的优点在于理论体系严密、计算量不随阶数增高而成几何级数增加、辨识精度高。该方法理论上能够估算任意阶核,弥补了现有方法难以估算四阶以上核的缺点,可应用于动态系统和强非线性系统的建模。将发动机动态过程描述为四阶的Volterra级数模型。     

Volterra高阶核的新算法在航空涡轮转速控制上的应用

针对航空发动机的转速控制这个难题,提出了对非线性动态系统做建模研究的Volterra泛函方法的任意高阶核估计方法;该方法在核(kernel method)理论基础上,构造线性空间,将求解Volterra泛函各阶核的问题转化为求输出观测向量在希尔伯特空间(Hilbert space)子空间上的分量,利用线性空间中向量内积的求解而间接辨识出复杂的非线性动态系统;相对于其它在时域或频域估计Volterra核的理论,该方法数学基础牢固、计算量不随辨识精度增高而大量增加、理论上能够对任意高阶核进行估计,可对强非线性动态系统进行辨识。     

基于Volterra频域核辨识的非线性模拟电路故障诊断

基于Volterra级数时域频域混合模型,提出了辨识非线性模拟电路频域核的故障诊断方法.利用混合模型辨识算法和范德蒙特法估计各种故障状态下电路响应的前3阶频域核,提取故障特征并与相应的故障模式一起构成特征样本集,借助于支持向量机多分类器进行分类识别,实现非线性模拟电路的故障诊断.阐述了诊断原理及诊断步骤,并给出了诊断实例.仿真结果表明,该方法的故障识别率较高,便于计算机计算.     

非线性系统广义脉冲响应函数的盲辨识

探讨减少非线性系统广义脉冲响应函数（GIRF）盲辨识所需计算量问题。 基于线性MIMO模型，应用多项式矩阵理论和子空间盲辨识技术，研究使用部分噪声向量对非线性Volterra系统的GIRF盲辨识方法。该方法的优点是能有效减少GIRF盲辨识所需的计算量。这对GIRF盲辩识方法的在线应用是有利的。仿真结果说明了这一方法的有效性。     

基于空间投影算法的汽轮机轴建模与故障识别

针对电厂的汽轮机在临近故障状态时无法有效监测和预警这个问题,提出了再生核希尔伯特空间理论的估计Volterr级数核系数的新方法,利用空间投影理论对汽轮机轴非线性动态系统进行辨识和故障建模仿真,还提出了一种指数化再生核,对指数化核不需要计算无穷级数,因此避免了按照原始构建方法计算普通再生核时的近似处理过程,避免误差在这一阶段出现.通过在不同主蒸汽温度、不同记忆长度的广义高阶频率响应的仿真图形分析,说明了汽轮机运行稳定程度与主蒸汽温度之间的关系,对汽轮机在临近故障状态时的广义高阶频率响应图的特征做了研究.     

基于Volterra级数的一类非线性系统性能评估

化工过程多具有非线性特征,针对用线性系统性能评估方法处理非线性系统会存在过估计的情况,研究了一类叠加线性干扰的非线性系统的性能评估问题。通过使用Volterra级数近似非线性环节,把最小方差性能评估问题转化成一类模型辨识问题,并从辨识误差中得到非线性系统的最小方差估计值。通过数值仿真,将新方法所得结果和现有线性性能评估方法进行了比较,验证了设计算法的优越性。     

一种基于Volterra模型的非线性系统具有全解耦结构的递阶式自适应辨识方法

研究了在输入输出观测数据均含有噪声时如何对基于Volterra级数描述的非线性系统进行解耦自适应辨识的问题. 按照Volterra级数模型的伪线性组合结构, 采用总体最小二乘辨识技术的原理, 导出了一种总体全解耦辨识的思想. 从而建立了一种具有全解耦结构的递阶式自适应辨识算法, 给出了该算法的结构图. 相比于部分解耦辨识算法, 该算法的优点在于它能够在全噪声数据环境下得到更高的收敛速度和精度. 仿真研究的结果证明了本文方法的有效性.     

Volterra模型辨识及其在蒸馏塔预测控制中的应用

本文研究一类单输入单输出非线性系统的预测函数控制问题，这类系统能用有限阶离散Volterra级数模型表示，采用最小二乘法进行参数辨识，并通过求解高次方程得到控制律。针对化工过程蒸馏塔控制系统，通过仿真计算验证了该方法的有效性。     

基于双阶跃信号输入的Volterra模型辨识

Volterra模型作为非线性领域的一种非线性模型,由于其对工业过程可以以任意精度逼近,使得该模型有很广泛的应用研究意义。在将该模型运用到实际控制系统中之前,模型的高精度辨识显得尤为重要。在以往针对Volterra模型的辨识算法中,基本上主要是采用通用辨识算法识别模型参数,比如最小二乘法及各种改进的最小二乘法。这些通用的辨识算法在辨识Volterra模型时,不能充分考虑其非线性特点,同样不能在辨识过程中充分利用该特点。本文在充分考虑Volterra模型非线性的前提下,提出了一种基于双阶跃信号输入的Volterra模型辨识算法,该算法辨识原理简单,计算量较小,论文最后将该辨识算法应用到典型非线性CSTR系统的的辨识中,辨识结果证明了算法的有效性。     

A spatio-temporal Volterra modeling approach for a class of distributed industrial processes

It is difficult to model a distributed parameter system (DPS) due to the infinite-dimensional time/space nature and unknown nonlinear uncertainties. A low-dimensional and simple nonlinear model is often required for practical applications. In this paper, a spatio-temporal Volterra model is proposed with a series of spatio-temporal kernels for modeling unknown nonlinear DPS. To estimate these kernels, they are expanded onto spatial and temporal bases with unknown coefficients. To reduce the model dimension and parametric complexity in the spatial domain, the Karhunen–Loève (KL) method is used to find the dominant spatial bases. To reduce the parametric complexity in the temporal domain, the Laguerre polynomials are selected as temporal bases. Next, using the Galerkin method, this spatio-temporal modeling becomes a linear regression problem. Then unknown parameters can be easily estimated using the least-squares method in the temporal domain. After the time/space synthesis, the spatio-temporal Volterra model can be constructed. The convergence of parameter estimation can be guaranteed under certain conditions. This model has a low-dimensional and simple nonlinear structure, which is useful for the prediction and control of the DPS. The simulation and experiment demonstrate the effectiveness of the proposed modeling method.     

A unifying view of wiener and volterra theory and polynomial kernel regression

Volterra and Wiener series are perhaps the best-understood nonlinear system representations in signal processing. Although both approaches have enjoyed a certain popularity in the past, their application has been limited to rather low-dimensional and weakly nonlinear systems due to the exponential growth of the number of terms that have to be estimated. We show that Volterra and Wiener series can be represented implicitly as elements of a reproducing kernel Hilbert space by using polynomial kernels. The estimation complexity of the implicit representation is linear in the input dimensionality and independent of the degree of nonlinearity. Experiments show performance advantages in terms of convergence, interpretability, and system sizes that can be handled.     

结合非线性频谱与核主元分析的复杂系统故障诊断方法

传统非线性频谱分析方法对复杂系统进行故障诊断时,求解出的非线性频谱数据量庞大,不便于直接用于故障检测与分类识别.本文提出了一种非线性频谱特征与核主元分析(KPCA)结合的故障诊断方法,首先通过最小二乘算法估计出前3阶Volterra时域核,由多维傅立叶变换求取出广义频率响应函数,然后利用KPCA方法对谱数据进行压缩与提取谱特征,最后利用多分类最小二乘支持向量机进行多故障检测与识别.考虑到频谱数据具有非线性的特点,KPCA中的核函数选用由多项式函数与径向基函数构成的混合核函数,兼顾了局部特性与全局特性.论文基于非线性频谱数据,给出了核主元模型建立与在线故障诊断的具体算法.对非线性模拟电路和数控机床伺服传动系统进行了仿真实验,结果表明本文方法能够大幅度降低频谱数据维数,故障识别率高,是一种实用的故障诊断方法.     

Integral and integro-differential control plants: Optimality conditions

For the control systems whose dynamics obeys a nonlinear regular integral Volterra equation with additional constraints in the form of equalities, the necessary optimality conditions were established on the basis of the abstract Yakubovich-Matveev theory of optimal control and, in particular, the abstract principle of maximum. Consideration was given to two kinds of the nonlinear controllable singular integral equations with unrestricted multipliers under the integral—with the power kernel of the Cauchy kernel type and with the logarithmic kernel. Attention was paid mostly to the nonlinear controlled dynamic systems obeying an integro-differential Volterra equation of the first order. As before, the study relied on the abstract theory of optimal control. The necessary optimality conditions were established by deriving the corresponding conjugate equation, transversality conditions, and principle of maximum.     

Stability of nonlinear feedback systems in a Volterra representation

Presented in this paper is a stability condition for a class of nonlinear feedback systems where the plant dynamics can be represented by a finite series of Volterra kernels. The class of Volterra kernels are limited to p‐linear stable operators and may contain pure delays. The stability condition requires that the linear kernel is non‐zero and that the closed loop characteristic equation associated with the linearized system is stable. Next, a sufficient condition is developed to upper bound the infinity‐norm of an external disturbance signal thereby guaranteeing that the internal and output signals of the closed loop nonlinear system are contained in L_∞. These results are then demonstrated on a design example. A frequency domain controller design procedure is also developed using these results where the trade‐off between performance and stability are considered for this class of nonlinear feedback systems. Copyright © 2000 John Wiley & Sons, Ltd.     

Synthesis of a Class of Non-linear Systems†

A problem in the practical application of the Volterra series to nonlinear theory is the synthesis of systems with a given set of Volterra kernels. In this article we present a method of determining whether a given Volterra kernel can be synthesized exactly by using only a finite number of multipliers. A method for the synthesis of such kernels is also presented. The use of the techniques discussed here is illustrated by examples.     

大型装备传动系统非线性频谱特征提取与故障诊断

基于Volterra级数的非线性频谱分析方法,建立了大型数控装备传动系统伺服电机的非线性频谱模型,对传动系统两类参数型故障的频谱特征进行了分析.在此基础上,提出一种实用的在线频谱特征提取与故障识别方法,采用自适应辨识算法求解时域Volterra核,用快速多维傅立叶变换获得非线性频谱特征.实验结果表明,该方法实时性好,故障识别率高.     

A multiple-kernel LSSVR method for separable nonlinear system identification

In some nonlinear dynamic systems, the state variables function usually can be separated from the control variables function, which brings much trouble to the identification of such systems. To well solve this problem, an improved least squares support vector regression （LSSVR） model with multiple-kernel is proposed and the model is applied to the nonlinear separable system identification. This method utilizes the excellent nonlinear mapping ability of Morlet wavelet kernel function and combines the state and control variables information into a kernel matrix. Using the composite wavelet kernel, the LSSVR includes two nonlinear functions, whose variables are the state variables and the control ones respectively, in this way, the regression function can gain better nonlinear mapping ability, and it can simulate almost any curve in quadratic continuous integral space. Then, they are used to identify the two functions in the separable nonlinear dynamic system. Simulation results show that the multiple-kernel LSSVR method can greatly improve the identification accuracy than the single kernel method, and the Morlet wavelet kernel is more efficient than the other kernels.     

基于Volterra 级数并行递推AP 算法的陀螺漂移预测

为了预测某导弹陀螺漂移趋势,以该陀螺漂移角速度时间序列为对象,建立基于Volterra级数的非线性时间预测模型,提出了一种基于Volterra级数的并行递推放射投影AP自适应算法.以系统Volterra核向量增量的模与某约束总和为损失函数,按照最陡下降原理导出各阶Volterra核更新公式;再利用矩阵求逆引理递推求取各阶Volterra子系统自相关逆矩阵导出算法.某导弹实测的陀螺漂移数据预测应用研究表明,该算法运算速度快、预测精度高.