|
|
|
| Linear dynamic system identification using proper orthogonal decomposition and Padé approximants |
|
|
|
|
Abstract An identification method for linear dynamical system is proposed. The Padé approximants is used to fit the curve of system's dynamic stiffness, and the coefficient matrices in the Padé polynomial are determined by the least squares method, in addition, genetic algorithms is adopted to optimize the parameters in Padé polynomial. By comparing the Padé approximants with the theory s of the system's dynamic stiffness matrices, the mass, damping and stiffness matrices can be obtained. When the order of the system is high, the POD reduced-order technology can be applied, and it means that the present method is valid for full and reduced-order model. Numerical examples illustrate the present method with good accuracy and robust.
|
|
Received: 10 December 2015
Published: 02 November 2016
|
|
|
|
|
|
| [1] |
. THE CHARACTERISTICS OF OPTIMAL TWO-DIMENSIONAL SOLID/SOLID PHONONIC CRYSTALS[J]. , 2015, 36(4): 283-289. |
| [2] |
. THERMAL BENDING ANALYSIS OF SIMPLY SUPPORTED THIN PLATE BY THE HYBRID BOUNDARY NODE METHOD[J]. , 2012, 33(6): 617-622. |
| [3] |
. LOCATION OPTIMAL OF PIEZOELECTRIC PATCH USING GENETIC ALGORITHM[J]. , 2011, 32(4): 398-404. |
| [4] |
. An Inverse Technique for Identification of Dynamic Constitutive Damage Parameters of Brittle Material[J]. , 2011, 32(3): 242-248. |
| [5] |
. A reduced order modeling of nonlinear electrostatic-mechanical systems based on energy ID approach[J]. , 2010, 31(3): 244-251. |
|
|
|
|
2016, Vol. 37 
