Comparative Study of Evolutionary Algorithms for Parameter Identification of an Impact Oscillator

The use of non-classical evolutionary optimization techniques such as genetic algorithms, differential evolution, swarm optimization and genetic programming to solve the inverse problem of parameter identification of dynamical systems leading to chaotic states has been gaining popularity in recent years. In this paper, three popular evolutionary algorithms — differential evolution, particle swarm optimization and the firefly algorithm are used for parameter identification of a clearance-coupled-impact oscillator system. The behavior of impacting systems is highly nonlinear exhibiting a myriad of harmonic, low order and high order sub-harmonic resonances, as well as chaotic vibrations. The time-history simulations of the single-degree-of-freedom impact oscillator were obtained by the Neumark-β numerical integration algorithm. The results are illustrated by bifurcation graphs, state space portraits and Poincare’ maps which gives valuable insights on the dynamics of the impact system. The parameter identification problem relates to finding one set of system parameters given a chaotic or periodic system response as a set of Poincaré points and a different but known set of system parameters. The three evolutionary algorithms are compared over a set of parameter identification problems. The algorithms are compared based on solution quality to evaluate the efficacy of using one algorithm over another.

Dynamics of Bladed Disks With Frictional Coupling and Alternate Mistuning Pattern

In the field of turbo machinery design frictional coupling has been found to be a low cost method to increase the mechanical damping of bladed disks. Underplatform dampers (UPD’s) are commonly used which are metal devices pressed against the blades by centrifugal forces. The main task is to find the optimum value of the contact normal force to maximize energy dissipation. This optimum strongly depends on the excitation of the structure. Traveling waves are excited by engine order excitation and flutter. Flutter caused by fluid structure interaction can be reduced by intentional mistuning of the bladed disk whereas forced response levels will be typically increased by mistuning. A compromise is alternate mistuning. The present paper deals with the influence of alternate mistuning on frictional coupling of blisks. Firstly, the dynamics of a tuned blisk are explained with a simplified lumped mass cyclic oscillator model. It is pointed out that eigenfrequencies of traveling waves around the blisk are influenced by structural coupling. Alternate mistuning leads to mode coupling with the possibility of energy transfer. The performance of friction coupling strongly depends on the nodal diameter mode shape of vibration which is stated analytically for pure Coulomb sliding contact. Following this, a simplified blisk model with underplatform dampers is developed to analyze alternate mistuning and frictional coupling. The simulation results show a significant influence of the mistuning on the damping performance.

Multi-Scale Fluid-Structure Interaction Simulations Based on Mesoscopic Approaches

Many challenging fluid-structure interaction problems in nuclear engineering remain unresolved because current CFD methodologies are unable to manage the number of computational cells needed and/or the difficulties associated with meshing changing geometries. One of the most promising recent methodologies for fluid dynamics modeling is the lattice-Boltzmann method — an approach that offers significant advantages over classical CFD methodologies by 1) greatly reducing meshing requirements, 2) offering great scalability, and 3) through relative ease of code parallelization. While LBM often requires increased numerical effort compared to other methods, this can be dramatically reduced by combining LBM with Adaptive Mesh Refinement (LB-AMR). This study describes an ongoing collaboration investigating nuclear fuel-assembly spacer grid performance. The LB-AMR method, used to simulate the flow field around a specific spacer grid design, is capable of describing turbulent flows for high Reynolds numbers, revealing rich flow dynamics in good qualitative agreement with experimental results. Prepared by LLNL under Contract DE-AC52-07NA27344.

Two-Way Fluid-Structure Coupling in Vibration and Damping Analysis of an Oscillating Hydrofoil

Fluid-structure interaction (FSI) and unavoidable vibrations are important characteristics in the operation of hydropower structures and must be taken into account in the analysis and design of such equipment. Hydrodynamic damping influences the amplitude of vibrations and is directly related to fatigue problems in hydraulic machines which are of great importance. The aim of this study is to investigate the coupled effects of flowing fluid on a simplified hydrofoil by using three-dimensional two-way fluid-structure interaction modeling, in order to determine its importance in predicting vibration amplitudes and damping. The effect of considering different flow velocities is also investigated in the present study. The results of this research are compared with those obtained from experiments done by ANDRITZ [1]. The influences of mesh size and time step are also studied. Our results indicate that considering FSI in predicting the frequencies of the fluctuating fluid forces in practical problems might be ignored if the main concern of the analysis is to check the possibility of resonance. However, FSI must be included in the modeling when we aim to predict the influence of the fluid on the damping behavior in the hydrofoil vibration.

Nonlinear Vibration of Nanobeam With Quadratic Rational Bezier Arc Curvature

Nonlinear vibration of nanobeam with the quadratic rational Bezier arc curvature is investigated. The governing equation of motion of the nanobeam based on the Euler-Bernoulli beam theory is developed. Accordingly, the non-uniform rational B-spline (NURBS) is implemented in order to write the implicit form of the governing equation of the structure. The simply-supported boundary conditions are assumed and the Galerkin procedure is utilized to find the nonlinear ordinary differential equation of the system. The nonlinear natural frequency of the system is found and the effects of different parameters, namely, the waviness amplitude, oscillation amplitude, aspect ratio, curvature shape and the Pasternak foundation coefficient are fully investigated. The hardening and softening responses of the natural frequency of structure are detected for variations of the shape and amplitude of the curvature waviness. It is revealed that the ratio of nonlinear to linear frequency increases with an increase in the oscillation amplitudes. It is found that by increasing the Pasternak foundation coefficient, the ratio of nonlinear to linear frequency decreases.

Actuator Array Manipulation Using Low Resolution Local Sensing

An actuator array is a planar distributed manipulation system that uses multiple two degree-of-freedom actuators to manipulate objects with three degrees of freedom (x, y and θ). This paper presents an accurate method of estimating position and orientation of an object using local sensing and communication. In this method, each of the distributed modules contains a number of binary sensors, weight sensors, and two planar actuators. The binary sensors combined together give a binary image and analog sensors in each module combined together form a grayscale image representation of the weight distribution of the object under manipulation. Additive normalization has been used to combine binary and grayscale distributed sensing images together to come up with increased precision estimates of the position and orientation of an object. A distributed sensing simulation has been developed in Simulink and the effectiveness of this method has been verified for rectangular and circular objects using the Simulink model.

Dynamic Modeling and Slippage Analysis in Object Manipulation by Soft Fingers

In this paper, dynamic modeling and slippage analysis of a three-link soft finger manipulating a rigid object on a horizontal surface is studied. In order to integrate the dynamics of soft tip with the finger linkage, power-law model and a linear viscous damper are used to model the elastic behavior and damping effect of soft tip respectively. Because of the enlarged contact area in the soft contact, a frictional moment can be exerted at the contact interface along with the normal and tangential forces. Furthermore, because of planar motion of object, frictional forces and moment are applied in the contact of object and ground. Therefore, friction limit surface is used as a mapping between contact forces/moment and sliding motions in both contacts. Instead of using equality and inequality equations of frictional contact conditions, a method is proposed to describe different states of the contact forces and moment by a single second-order differential equation with variable coefficients. This kind of formulation of the system dynamics facilitates the design of a controller to cancel the undesired slippage occurs between the soft tip and object during the manipulation.

LQG Controller Design for Identified Wind Turbine Systems

In this research, it is developed to design LQG controller for wind turbine systems which are identified with Predictor-Based System Identification (PBSID) technique. The PBSID technique works well under closed-loop condition, which is useful for a system requiring closed-loop operation due to safety reason. First, a wind turbine system is identified using PBSID technique in full range of wind speed. Afterwards, using the identified system matrices, 1-DOF LQG controller is designed. The controller enables power generation to track the optimal power trajectory of a system. Simulation is used to demonstrate its usefulness.

Modeling and Trajectory Control of a Forklift-Like Wheeled Robot

In this paper, kinematic and dynamic models are derived for a forklift-like four-wheeled mobile robot, and then, based on the models, a new trajectory control scheme is designed and evaluated for the robot. The dynamic model, exhibiting non-minimum-phase characteristics, is derived by applying Lagrange’s equations and then the control law is design by using Lyapunov stability theorem and the loop shaping method. The proposed control scheme consists of a trajectory generator, a motion control law, and a steering control law. First, a real-time trajectory generator is designed based on the nonholonomic kinematic constraints of the robot, in which the reference driving speed and time rate of heading angle are computed in real time for a given desired trajectory of the robot. The proposed trajectory generator guarantees a local asymptotic stability. Next, motion and steering control laws are designed based on the dynamic model of the robot. The motion and steering control laws are used to control the robot speed and steering angle. The proposed control guarantees asymptotic stability of the trajectory control while keeping all internal signals bounded. Finally, the validity of the proposed control scheme is shown by realistic computer simulations with one sampling-time delay in the control loop.