The Experts below are selected from a list of 105 Experts worldwide ranked by ideXlab platform
K W Wang - One of the best experts on this subject based on the ideXlab platform.
-
active vibration isolation via simultaneous Left right Eigenvector assignment
Smart Materials and Structures, 2008Co-Authors: K W WangAbstract:The objective of this research is to synthesize a simultaneous Left and right Eigenvector assignment (SLREA) method for active vibration isolation. It is a pioneering effort to utilize such an Eigenvector assignment concept for active isolator design, where the approach can provide good physical insight into the problem. In this investigation, a new algorithm for the synthesis of the desired Left Eigenvectors is developed, which is an improvement over the classical methods. The purpose of the right Eigenvector assignment method is to alter the closed-loop system modes such that the modal components corresponding to the concerned region (isolation area of the isolator) have relatively small vibration amplitude. Correspondingly, the design goal of the Left Eigenvector assignment is to alter the Left Eigenvectors of the closed-loop system so that they are as closely orthogonal to the system's forcing vectors as possible. With the proposed approach, one can achieve both disturbance rejection and modal confinement concurrently for the purpose of vibration isolation. In this research, a new formulation is developed so that the desired Left Eigenvectors of this integrated system are selected through solving a generalized eigenvalue problem, where the orthogonality indices between the forcing vectors and the Left Eigenvectors are minimized. The components of the right Eigenvectors corresponding to the concerned region are minimized concurrently. It is shown that, with the SLREA technique, both disturbance rejection and modal confinement can be achieved, and thus vibration amplitude in the isolated region can be suppressed significantly.
-
vibration control via disturbance rejection through Left Eigenvector assignment
Proceedings of SPIE - The International Society for Optical Engineering, 2005Co-Authors: K W WangAbstract:The objective of this research is to investigate the feasibility of utilizing Left Eigenvector assignment for vibration disturbance rejection. In the previous study, it has been shown that through right Eigenvector assignment and modal confinement, one can enhance the performance of periodic vibration isolators. However, it was also recognized that since vibration mode confinement is based on the concept of modal response, it does not guarantee that vibration will always be reduced in a forced excitation scenario. In this research, the Left Eigenvector assignment technique is utilized to achieve vibration suppression throughout a broad frequency range. The principle is to alter the Left Eigenvectors of the closed-loop system so that the system's forcing vectors are as closely orthogonal to each Left Eigenvector as possible. With such an approach, one can directly attack the forced response problem. A new formulation is developed so that the desired Left Eigenvectors of this integrated system are selected through solving a generalized eigenvalue problem, where the orthogonality indices between the forcing vector and the Left Eigenvectors are minimized. The integrated system with assigned Left Eigenvectors achieves to reject external disturbance of the complete electromechanical system. An integrated closed-loop system with state estimator is also developed so that the algorithm can be implemented realistically. Numerical simulations are performed to evaluate the effectiveness of the proposed method on disturbance rejection for an isolator design example. Frequency responses of the isolator in the selected frequency range are illustrated. It is shown that with the Left Eigenvector assignment technique, the system"s external disturbances are rejected and vibration amplitude of the isolated regions can be effectively suppressed.© (2005) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.
-
vibration isolation control via simultaneous Left and right Eigenvector assignment
DETC2005: ASME International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, 2005Co-Authors: K W WangAbstract:The objective of this research is to investigate the feasibility of utilizing the simultaneous Left and right Eigenvector assignment concept for vibration isolation feedback control design. The purpose of the right Eigenvector assignment method is to alter the closed-loop system modes such that the modal components corresponding to the concerned regions (isolation end of an isolator) have relatively small amplitude. Correspondently, the design goal of Left Eigenvector assignment is to alter the Left Eigenvectors of the closed-loop system so that they are as closely orthogonal to the system’s forcing vectors as possible. With this approach, one can achieve both disturbance rejection and modal confinement concurrently for the purpose of vibration isolation. In this research, a new formulation is developed so that the desired Left Eigenvectors of this integrated system are selected through solving a generalized eigenvalue problem, where the orthogonality indices between the forcing vector and the Left Eigenvectors are minimized. The components of right Eigenvectors corresponding to the concerned regions are minimized concurrently. To realistically implement the algorithm, an integrated closed-loop system with state estimator is developed. Numerical simulations are performed to evaluate the effectiveness of the proposed method on concurrent disturbance rejection and modal confinement for a isolator rod design. Frequency responses of the isolator in the selected frequency range are illustrated. It is shown that with the simultaneous Left-right Eigenvector assignment technique, both disturbance rejection and modal confinement can be achieved, and thus the vibration amplitude in the isolated regions can be suppressed significantly.Copyright © 2005 by ASME
Hong Wang - One of the best experts on this subject based on the ideXlab platform.
-
disturbance attenuation in fault detection of gas turbine engines a discrete robust observer design
Systems Man and Cybernetics, 2009Co-Authors: Xuewu Dai, Zhiwei Gao, Timofei Breikin, Hong WangAbstract:This study is motivated by the onboard fault detection of gas turbine engines (GTEs), where the computation resources are limited and the disturbance is assumed to be band-limited. A fast Fourier transformation (FFT)-based disturbance frequency estimation approach is proposed and performance indexes are improved by integrating such frequency information. Furthermore, in the Left Eigenvector assignment, both eigenvalues and free parameters are optimized. As illustrated in the application to the actuator fault detection of a GTE, significant improvements are achieved compared to the existing methods. By combining the frequency estimation and eigenvalue optimization, the main contribution of the paper is the reduction of the computation complexity and the avoidance of the local optimal solution due to fixed eigenvalues.
Christoforos N Hadjicostis - One of the best experts on this subject based on the ideXlab platform.
-
distributed finite time computation of digraph parameters Left Eigenvector out degree and spectrum
IEEE Transactions on Control of Network Systems, 2016Co-Authors: Themistoklis Charalambous, Michael G Rabbat, Mikael Johansson, Christoforos N HadjicostisAbstract:Many of the algorithms that have been proposed in the field of distributed computation rely on assumptions that require nodes to be aware of some global parameters. In this paper, we propose algorithms to compute some network parameters in a distributed fashion and in a finite number of steps. More specifically, given an arbitrary strongly connected network of interconnected nodes, by adapting a distributed finite-time approach, we develop distributed strategies that enable nodes to compute the following network parameters: the Left-Eigenvector, the out-degree, and the spectrum of weighted adjacency matrices.
Gustavo E Scuseria - One of the best experts on this subject based on the ideXlab platform.
-
communication projected hartree fock theory as a polynomial similarity transformation theory of single excitations
Journal of Chemical Physics, 2016Co-Authors: Yiheng Qiu, Thomas M Henderson, Gustavo E ScuseriaAbstract:Spin-projected Hartree-Fock is written as a particle-hole excitation ansatz over a symmetry-adapted reference determinant. Remarkably, this expansion has an analytic expression that we were able to decipher. While the form of the polynomial expansion is universal, the excitation amplitudes need to be optimized. This is equivalent to the optimization of orbitals in the conventional projected Hartree-Fock framework of non-orthogonal determinants. Using the inverse of the particle-hole expansion, we similarity transform the Hamiltonian in a coupled-cluster style theory. The Left Eigenvector of the non-Hermitian Hamiltonian is constructed in a similar particle-hole expansion fashion, and we show that to numerically reproduce variational projected Hartree-Fock results, one needs as many pair excitations in the bra as the number of strongly correlated entangled pairs in the system. This single-excitation polynomial similarity transformation theory is an alternative to our recently presented double excitation th...
Bernhard O. Palsson - One of the best experts on this subject based on the ideXlab platform.
-
Analysis of complicated mode structure through power iteration with modified Jacobian matrix.
2017Co-Authors: Daniel C. Zielinski, Bernhard O. PalssonAbstract:We divide the vector multiplication with the Jacobian matrix into multiple steps. First of all, each row of the Jacobian matrix is multiplied by every element of the starting vector (Panel B solid black circles). We then sum up each column of the second matrix to obtain the resulting vector (Panel B dash black circles), which is normalized to give the ending vector. (A) The original Jacobian matrix and its leading Left Eigenvector. The matrix and the Eigenvector are the same as in Fig 3 and will be used for comparison with later panels. (B) Starting vector multiplied with the modified Jacobian matrix. We modified the Jacobian element at position (4, 4) to be the same value as the element at position (2, 2). The ending vector has a smaller ratio between the 2nd and 4th elements than that of the original Eigenvector, as would be expected with a larger absolute value at position (4, 4). The Eigenvector of this modified matrix is shown in the upper right of the panel. (C) Starting vector multiplied with a different modified Jacobian matrix. We further changed the modified Jacobian matrix in panel A to create a more symmetric structure, where the element at position (2, 4) is same as the element at position (4, 2). The ending vector has the same absolute values at the 2nd and 4th positions, showing that a fully symmetric Jacobian structure will create an equally weighted structure in Eigenvector. The Eigenvector of this modified matrix is shown in upper right. Overall, we demonstrate that changing the Jacobian element at either diagonal or off-diagonal position can alter the Eigenvector of the matrix in a predictable manner, based on the topological pattern of the key elements determining the Eigenvector structure. For clear demonstration purposes, the comparison of relative colors only works for individual box (surrounded by black stroke) itself, but not across different boxes.
-
The power iteration algorithm demonstrates how complicated dynamic structures arise from topologically connected elements of similar magnitude within the Jacobian matrix.
2017Co-Authors: Daniel C. Zielinski, Bernhard O. PalssonAbstract:(A) Power iteration can be used to calculate the dominant Left Eigenvector of the Jacobian matrix. The Left Eigenvectors are the modes of the metabolic network. The algorithm Left multiplies the Jacobian matrix by a random vector (ui), normalizes the resulting vector and repeats the process until the vector converges to the Eigenvector. (B) Topologically connected Jacobian elements of similar magnitude determine complicated Eigenvector structure. In this case study, we extracted a submatrix of J that corresponds to the nonzero elements of a certain Eigenvector, which contains G6PDH enzyme forms. The four Jacobian elements (also the largest) that are key in determining this Eigenvector structure are located in the 2nd and 4th rows, circled in black. Specifically, the structure of 2nd or 4th rows matches closely with that of the Eigenvector, with similar ratios at the 2nd and 4th positions. Multiplying the Jacobian matrix by any non-orthogonal starting vector (u1), for example the one shown, results in a vector (u2) that has a structure more similar to the Eigenvector. The contribution of those rows individually to Eigenvector formation are further shown in Fig 4 and S4 Fig. For clear demonstration purposes, the comparison of relative colors only works for individual box (surrounded by black stroke) itself, but not across different boxes. (C) Principal component analysis on all power iteration vectors starting with 1000 different random vectors. We randomly picked 1000 starting vectors and multiplied them with the full Jacobian matrix (292 × 292). The starting vector is multiplied through several iterations (10 ~ 20) until it converges to the Eigenvector (the dot product of the ending vector and the Eigenvector is no greater than 1.0001 and no less than 0.9999). We then performed principal component analysis on all iteration vectors (including the starting vectors) and plotted each vector in terms of the contribution from the first two principal components. The first principal component corresponds to the leading Eigenvector of the Jacobian matrix while the rest of components (less than 1% contribution each, only component 2 shown here) together explain the variation of the vector from the Eigenvector. Ideally, the contribution of the rest of components will be 0 when the ending vector becomes the Eigenvector. However, due to large order of magnitude differences between elements in J and the cutoff we set when comparing the ending vector with the Eigenvector, we ended up with variations from the Eigenvector (nonzero contribution of component 2 in the inset plot).