Linear state representations for identification of bilinear discrete-time models by interaction matrices

Bilinear systems can be viewed as a bridge between linear and nonlinear systems, providing a promising approach to handle various nonlinear identification and control problems. This paper provides a formal justification for the extension of interaction matrices to bilinear systems and uses them to express the bilinear state as a linear function of input–output data. Multiple representations of this kind are derived, making it possible to develop an intersection subspace algorithm for the identification of discrete-time bilinear models. The technique first recovers the bilinear state by intersecting two vector spaces that are defined solely in terms of input–output data. The new input–output-to-state relationships are also used to extend the equivalent linear model method for bilinear system identification. Among the benefits of the proposed approach, it does not require data from multiple experiments, and it does not impose specific restrictions on the form of input excitation.     

Excursion probabilities of non-linear systems

This paper suggests a procedure for estimating excursion probabilities for linear and non-linear systems subjected to Gaussian excitation processes. In this paper, the focus is on non-linear systems which might also have stochastic properties. The approach is based on the so-called “averaged excursion probability flow” which allows for a simple solution for the interaction in excursion problems. Simplifying, the dynamic reliability problem can be reduced to a simpler “static” problem by considering the probability flow at fixed time instances. The proposed approach is very general and can be applied to both linear and non-linear systems of which the response can be determined by deterministic methods. Hence, the procedure applies to arbitrary structures and any suitable mathematical model including large FE-models solved by deterministic FE-codes.     

On the use of neural networks for identification of linear and nonlinear systems

A procedure based on neural networks for the classification of linear and nonlinear systems is presented, using excitation and response data under swept sine excitation. Special attention is paid to the classification and identification of linear and bilinear systems, the latter being considered since they exhibit typical characteristics of cracked systems. The computer simulations show that: (1) using the procedure presented in this paper the trained classification network can reliably classify a linear system and different nonlinear systems; (2) the output of the trained identification neural network for a linear system and a bilinear system can be used as a quantitative indicator of characteristics of bilinear systems having different stiffness ratios (k _(x>0)/k _(x<0)) with respect to the bilinear system used in the training stage; (3) for two-degree-of-freedom systems, the trained network can not only determine the existence of a bilinear stiffness and the magnitude of its stiffness ratio, but also specify which stiffness is bilinear, i.e. indicate its position. These results provide a possibility of using the trained neural networks to detect and locate structural cracks which have the characteristics of bilinear systems.Visiting scholar, from People's Republic of China.     

Identification and parameter estimation of a bilinear oscillator using Volterra series with harmonic probing

Identification of non-linear systems is mainly limited to polynomial form non-linearities. Among the non-polynomial forms, bilinear oscillator constitutes an important class of non-linear systems and it has been used for modeling of various physical systems, particularly for structural elements with a breathing crack. An identification procedure is presented here for the class of bilinear oscillator, using higher order FRFs derived from Volterra series under harmonic excitation. The procedure addresses the problem of both; identification of the non-linearity structure as well as estimation of the bilinear parameter, which can be correlated to the crack severity and structural degradation. The procedure is illustrated with numerical simulation and the estimation results indicate that even a weakly bilinear state introduced by a small crack size can be accurately identified and measured.     

A method for solving the factorized vorticity-stream function equations by finite elements

A new finite element method for solving the time-dependent incompressible Navier-Stokes equations with general boundary conditions is presented. The two second-order partial differential equations for the vorticity and the stream function are factorized, apart from the non-linear advection term, by eliminating the coupling due to the double specification on the stream function at (a part of) the boundary. This is achieved by reducing the no-slip boundary conditions to projection integral conditions for the vorticity field and by evaluating the relevant quantities involved according to an extension of the method of Glowinski and Pironneau for the biharmonic problem. Time integration schemes and iterative algorithms are introduced which require the solution only of banded linear systems of symmetric type. The proposed finite element formulation is compared with its finite difference equivalent by means of a few numerical examples. The results obtained using 4-noded bilinear elements provide an illustration of the superiority of the finite element based spatial discretization.     

双线性滞回结构滤波问题的分段线性化法

提出了一种双线性滞回特性的分段微分表达式，根据伊藤随机微分方程和连续Markov向量过程的理论，导出了无须进行积分计算的双线性滞回结构的分段线性滤波器，简化了计算．算例表明，该法比随机等效线性化法和数字模拟的结果更为吻合．     

H 2 optimal filtering for bilinear systems

The filtering problem is among the fundamental issues in control and signal processing. Several approaches such as H ₂ optimal filtering and H _?? optimal filtering have been developed to address this issue. While the optimal H ₂ filtering problem has been extensively studied in the past for linear systems, to the best of our knowledge, it has not been studied for bilinear systems. This is indeed surprising, since bilinear systems are important class of nonlinear systems with well-established theories and applications in various fields. The problem of H ₂ optimal filtering for both discrete-time and continuous bilinear systems is addressed in this paper. The filter design problem is formulated in the optimization framework. The problem for the discrete-time case is expressed in terms of linear matrix inequalities which can be efficiently solved. The results are used for the optimal filtering of a bilinear model of an electro-hydraulic drive.     

Averaging methods for numerical simulations of flows through heterogeneous porous media

Many natural rock systems contain small patches of different permeability which affect the flow of fluids through them. As these heterogeneities become smaller and more numerous, they become harder to model numerically. We consider how to reduce the computational effort required in simulations by incorporating their effects in the boundary conditions at the edges of each grid block. This is in contrast with current methods which involve often arbitrary changes in the fluid properties. The method is restricted to the case of widely-spaced patches, which simplifies interaction effects. The system then reduces to an array of dipoles, and two averaging methods are proposed for finite grid blocks. Several infinite systems, including vertical and horizontal bands, are also considered as further approximations. There is a great wealth of existing results from different fields which lead to identical mathematical problems and which can be used in these cases. Finally, we consider how to use these techniques when the precise configuration of the grid block is not known, but only its statistical properties. This can lead to results which are very different from the deterministic case.     

A switching fractional calculus-based controller for normal non-linear dynamical systems

In this paper, a fractional calculus-based terminal sliding mode controller is introduced for finite-time control of non-autonomous non-linear dynamical systems in the canonical form. A fractional terminal switching manifold which is appropriate for canonical integer-order systems is firstly designed. Then some conditions are provided to avoid the inherent singularities of the conventional terminal sliding manifolds. A non-smooth Lyapunov function is adopted to prove the finite time stability and convergence of the sliding mode dynamics. Afterward, based on the sliding mode control theory, an equivalent control and a discontinuous control law are designed to guarantee the occurrence of the sliding motion in finite time. The proposed control scheme uses only one control input to stabilize the system. The proposed controller is also robust against system uncertainties and external disturbances. Two illustrative examples show the effectiveness and applicability of the proposed fractional finite-time control strategy. It is worth noting that the proposed sliding mode controller can be applied for control and stabilization of a large class of non-autonomous non-linear uncertain canonical systems.     

Discrete bright–dark soliton solutions and parameters controlling for the coupled Ablowitz–Ladik equation

Analysis of piecewise-linear nonlinear dynamical systems is critical for a variety of civil, mechanical, and aerospace structures that contain gaps or prestress that are caused by cracks, delamination, joints or interfaces among components. Recently, a technique referred to as bilinear amplitude approximation (BAA) was developed to estimate the response of bilinear systems that have no gap or prestress. The method is based on an idea that the dynamics of a bilinear system can be treated as a combination of linear responses in two time intervals both of which the system behaves as a distinct linear system: (1) the open state and (2) the closed or sliding state. Both geometric and momentum constraints are then applied as compatibility conditions between the states to couple the linear vibrational response for each time interval. In order to estimate the response for more general cases where there are either gaps or prestress in the system, a generalized BAA method is proposed in this paper. The new method requires inclusion of contact stiffness and damping to model contact behavior in the sliding state, and new equilibrium positions for each state to establish proper coordinates. The new method also finds the bilinear frequency of the system, which cannot be computed using the bilinear frequency approximation method previously developed since that method is only accurate for the zero gap and no prestress case. The generalized BAA method is demonstrated on a single degree of freedom system, a three degree of freedom system, and a cracked cantilever beam model for various gap sizes and prestress levels.     

Identification of discrete-time bilinear systems through equivalent linear models

This paper presents a novel method for identification of discrete-time, time-invariant state-space models of bilinear dynamical systems using the steady-state portion of a single input/multiple output time-history measurements. These measurements are recorded by exciting the system with a linear combination of sine and cosine functions of user-selected frequencies enriched by a subtle amount of random component. The proposed method relies on conversion of the bilinear system into an equivalent linear model (ELM) by an accurate approximation of the state in the bilinear term using a set of sine and cosine basis functions whose frequencies are obtained as combinations of the input frequencies. Observer/Kalman Filter Identification (OKID), a?linear time invariant (LTI) system identification algorithm, is used to identify the aforementioned ELM from which the original bilinear model is recovered. A?numerical example is also provided.     

Non-linear normal modes,invariance, and modal dynamics approximations of non-linear systems

Non-linear systems are here tackled in a manner directly inherited from linear ones, that is, by using proper normal modes of motion. These are defined in terms of invariant manifolds in the system's phase space, on which the uncoupled system dynamics can be studied. Two different methodologies which were previously developed to derive the non-linear normal modes of continuous systems — one based on a purely continuous approach, and one based on a discretized approach to which the theory developed for discrete systems can be applied-are simultaneously applied to the same study case-an Euler-Bernoulli beam constrained by a non-linear spring-and compared as regards accuracy and reliability. Numerical simulations of pure non-linear modal motions are performed using these approaches, and compared to simulations of equations obtained by a classical projection onto the linear modes. The invariance properties of the non-linear normal modes are demonstrated, and it is also found that, for a pure non-linear modal motion, the invariant manifold approach achieves the same accuracy as that obtained using several linear normal modes, but with significantly reduced computational cost. This is mainly due to the possibility of obtaining high-order accuracy in the dynamics by solving only one non-linear ordinary differential equation.     

几何非线性分析的高效高阶无网格法

准确高效地处理几何非线性对于材料破坏等大变形过程的数值分析至关重要。考虑到无网格法具有易于形成高阶近似函数等诸多优点，本文发展了几何非线性分析的高阶无网格法。采用上一载荷步收敛的构形作为计算的参考构形，位移本质边界条件由罚函数法施加。为提高计算效率，将针对线性问题发展的二阶一致三点积分格式QC3（Quadratically Consistent 3-point integration scheme）拓展到考虑构形变化的几何非线性分析，大幅度减少了所需的积分点数目。数值结果表明，本文发展的高阶无网格法能够准确有效地处理几何非线性问题，而且在计算效率、精度以及应力场光滑性等方面均表现出显著优势。     

Discrete non-linear adaptive data driven control based upon simultaneous perturbation stochastic approximation

A novel adaptive data driven control strategy is proposed for general discrete non-linear systems. The controller is designed based upon the Simultaneous Perturbation Stochastic Approximation (SPSA) method, and is constructed through use of a Function Approximator (FA), which is fixed as a neural network here. In this novel control strategy, the parametric estimation is designed to be adaptive with convergence analysis, and the control ability has been greatly improved. The proposed control method is finally applied into the non-linear tracking problems, as well as near-optimal control problems for discrete-time non-linear systems. Simulation comparison tests were conducted on typical non-linear plants, through which, the convergence and feasibility of the proposed adaptive data driven control strategy are well demonstrated.     

Variable knot-based spline approximation recursive Bayesian algorithm for the identification of Wiener systems with process noise

Wiener systems consist of a linear dynamic block in cascade with static nonlinearity. One of the challenging issues in the identification of a process noise disturbed Wiener system is that the influence of noise is difficult to eliminate. For Wiener systems with process noise, traditional algorithms will result in biased estimates. To solve this problem, a novel recursive Bayesian algorithm based on variable knot spline approximation is proposed in this paper. First, a spline function is taken to approximate the inverse function of the nonlinear part, which can achieve excellent extrapolation and eliminate oscillatory behaviors. A knot selection method is then presented to achieve accurate estimates. Furthermore, a knot variation strategy to improve the accuracy of the spline approximation is described. Finally, the proposed algorithm is validated through a numerical simulation.     

基于PETSc的非线性逆向运动并行计算方法研究

在考虑几何非线性的有限元分析中,初始构型和变形构型是严格区分的,并且变形后的构型对结构性能和功能的实现往往具有重要意义.传统的非线性有限元分析主要面向变形前的初始构型为导向的设计问题,而对于变形后构型为导向的设计问题则有较大局限性.针对此问题,引入非线性逆向运动分析方法,为了保证大变形非线性分析迭代的收敛性和计算效率,基于PETSc函数库建立了并行分析框架,并对并行框架中的模块划分、数据并行存储以及通信锁死和负载平衡等进行了详细阐述.在算例部分,首先通过正向运动和逆向运动分析结果对比,阐述了两种分析方法的不同以及逆向运动分析方法对变形前构型求解的准确性;其次,采用不同MPI进程数对并行分析程序的效率进行了测试.结果 表明,合理地选择MPI进程数目可显著提高非线性逆向运动分析的效率.     

Control strategies for timestep selection in finite element simulation of incompressible flows and coupled reaction–convection–diffusion processes

We propose two timestep selection algorithms, based on feedback control theory, for finite element simulation of steady state and transient 2D viscous flow and coupled reaction–convection–diffusion processes. To illustrate performance of the schemes in practice, we solve Rayleigh–Benard–Marangoni flows, flow across a backward‐facing step, unsteady flow around a circular cylinder and chemical reaction systems. Numerical experiments confirm that the feedback controllers produce in some cases a very smooth stepsize variation, suggesting that robust control algorithms are possible. These experiments also show that parameter selection can improve timesteps when co‐ordinated with the convergence control of non‐linear iterations. Further, computational cost of the selection procedures is negligible, since they involve only storing a few extra vectors, computation of norms and evaluation of kinetic energy. Copyright © 2004 John Wiley & Sons, Ltd.     

On the evolution of deterministic non-periodic behavior of an airfoil

This paper describes some interesting transitional behavior in the limit cycle response of a typical section airfoil with a loosely connected flap subject to a fluid flow. The freeplay non-linearity associated with the flap connection has considerable practical importance since wear and maintenance problems inevitably accompany moving mechanical parts. Piecewise linear systems have been studied within the non-linear dynamics community for some time. However, application to a relatively complicated, fluid–structure interaction problem such as this shows the ubiquity of a number of characteristically non-linear features. Special attention is focused on almost- or quasi-periodic behavior, and the correpsonding stability transitions. Brief reference is made to some prior experimental work which suggests the extent to which these responses can be realized in wind tunnel testing.     

Decohesion of a rigid punch from non-linear membrane undergoing finite axisymmetric deformation

Adhesion of flexible thin film or membrane plays an important role in many biological and industrial applications. In many physical systems, large deformation is often expected, especially for biological membranes, which suggests that non-linear analysis might be required. In this paper the adhesion and decohesion between a rigid flat punch and a non-linear membrane undergoing finite axisymmetric deformation is studied. The decohesion is assumed to obey the Griffith fracture criterion. The cases of initially stress-free and prestressed membrane are analyzed. A comparison with published works using the linear theory shows that the non-linear theory substantially deviates from the linear one, qualitatively and quantitatively. These results advocate the use of non-linear consideration when the membrane undergoes large deformation prior to and during the debonding process.