Natasa Trisovic

Reliability Estimation of Stochastic Dynamical Systems with Fractional Order PID Controller

In this paper, the reliability of stochastic dynamical systems under Gaussian white noise excitations with fractional order proportional–inegral–derivative (FOPID) controller is estimated. First, the FOPID controller is approximated by a set of combination of displacement and velocity based on the generalized van der Pol transformation. Then, the stochastic averaging method of energy envelope is applied to obtain a diffusive differential equation, from which the Backward Kolmogorov equation, governing the conditional reliability function, and the Generalized Pontryagin equation, governing the statistical moments of first-passage time, are derived from the averaged equation and solved numerically. Finally, in the two examples, the critical parameters in the FOPID controller are shown to be capable of improving the reliability of the stochastic dynamical system apparently, and all numerical results are verified to be efficient and correct by the Monte Carlo simulation.

Modal Strain Based Method for Dynamic Design of Plate-Like Structures

Design optimization of dynamic properties, for example, modal frequencies, can be of much importance when structures are exposed to the shock and/or vibration environments. A modal strain based method is proposed for fast design of natural frequencies of plate-like structures. The basic theory of modal strains of thin plates is reviewed. The capability of determining the highly sensitive elements by means of modal strain analysis is theoretically demonstrated. Finite element models were constructed in numerical simulations. Firstly, the application of the proposed method is conducted on a central-massed flat plate which was topologically optimized by the Reference. The results of modal strain analysis at the first mode have good agreement with the results from the topology optimization. Furthermore, some features of the strain mode shapes (SMSs) of the flat plate are investigated. Finally, the SMSs are applied to the optimization of a stiffened plate. Attention is focused on the distributions of the SMSs of the stiffeners, which also shows good agreement with the results from the topology optimization in the previous study. Several higher orders of SMSs are extracted, which can visualize the most sensitive elements to the corresponding modal frequency. In summary, both the theory and simulations validate the correctness and convenience of applying SMSs to dynamic design of plate-like structures.

The stationary solution of a random dynamical model

This paper studies the stationary probability density function (PDF) solution of a nonlinear business cycle model subjected to random shocks of Gaussian white-noise type. The PDF solution is controlled by a Fokker–Planck–Kolmogorov (FPK) equation, and we use exponential polynomial closure (EPC) method to derive an approximate solution for the FPK equation. Numerical results obtained from EPC method, better than those from Gaussian closure method, show good agreement with the probability distribution obtained with Monte Carlo simulation including the tail regions.