In this paper, a type of easily-prepared adhesive multi-cell tube is investigated by experimental and numerical methods. The deformation and specific energy absorption of the multi-cell tubes under three-point bending condition are analyzed. Quasi-static three-point bending tests show that adhesive multi-cell tubes perform generally higher energy absorption capacity than their constituent single-cell tubes, and in some case, up to 70% can be achieved due to the effect of adhesive. Numerical simulations of the three-point bending tests are conducted by LS-DYNA. The deformed patterns and punch force-displacement curves are consistent with test results. In addition, three different interface conditions and foam-filled tubes which are the same weight with adhesive multi-cell tubes are simulated to analyze and compare the response of adhesive multi-cell tubes. Results show that energy absorption characteristics of adhesive multi-cell tubes are comparable to traditional multi-cell tubes if there is no obvious adhesive detachment. Otherwise the performances of adhesive multi-cell tubes will be seriously weakened.
Damping is an effective way of suppressing mechanical vibrations. Passive layer damping patches are often used in the industry as damping solution for vibrations and acoustics problems. In this study, topology optimization methods were applied to the finite element models (FEM) of plates and damping patches in order to determine the critical locations for the frequency range of interest. Correlation of the finite element model with experimental data is performed both on modal analysis and frequency response functions. The optimized damping patch topology is compared to the base plate with full application of damping patch on frequency response functions. Finally, the thickness of the damping pads is varied to see its effect on the frequency response functions.
Topology optimization considering dynamic problems currently is a challenging topic. It demonstrates wide prospects and great values in engineering application covering high-end equipment, aeronautical and aerospace structure designs. This work is focused on the structural topology optimization methods related to dynamic responses under harmonic base acceleration excitations for large-scale problems. We propose using the large mass method (LMM) in which artificial large mass values are attributed to each driven nodal degree of freedom (DOF), which can thus transform the base acceleration excitations into force excitations. Mode acceleration method (MAM) is then used to calculate the harmonic displacement response and the accuracy of dynamic response is verified by several numerical examples. A density based optimization model is established, which the relative displacement amplitude at the concerned location of a structure is defined as the objective function subjected to the volume constraint. Numerical examples are finally presented to demonstrate the validity of the proposed method in improving dynamic performance.
The present research concerns the dynamical mistuning analysis of a rotating bladed-disk for which nonlinear geometrical effects can occur, in which the mistuning phenomenon is taken into account using a probabilistic approach of uncertainties. An alternative strategy to [1], where the mistuning is described by using a nonparametric probabilistic approach based on random matrix theory for modeling the random operators of a reduced-order computational model, is proposed. It is based on the use of a new nonparametric probabilistic approach of model-form uncertainties [2]. The deterministic vector basis, which is obtained from the Proper Orthogonalization Decomposition method, is replaced by a stochastic reduced-order basis (SROB). Each realization of the SROB respects some mathematical properties linked to the available information under constraints concerning the specified boundary conditions and the usual orthogonality properties. With such strategy, it is necessary to compute the stochastic nonlinear reduced internal forces combining the use of the SROB with the finite element method. There are hyper-parameters that control the uncertainties in the structure and its calibration with respect to experimental data is investigated. The numerical application is a rotating mistuned bladed-disk subjected to a load for which geometrical nonlinearities effects occur. [1] E. Capiez-Lernout, C. Soize, and M. Mbaye. Mistuning analysis and uncertainty quantification of an industrial bladed disk with geometrical nonlinearity. Journal of Sound and Vibration, 356(11):124-143, 2015. [2] C. Soize and C. Farhat. A nonparametric probabilistic approach for quantifying uncertainties in low dimensional and high-dimensional nonlinear models. International Journal for Numerical Methods in Engineering, 109(6):837-888, 2017.
Many safety-critical control systems adopt multiple redundant sensors to estimate the same control signal. If the sensors were to operate perfectly, only a subset of them would need to be used for the estimation. In practice, however, the sensors are subject to uncertainty, minor or major faults and their operation may be nonlinear. Therefore, it is crucial to reliably estimate the controlled signal under these conditions, assessing the degree of confidence with which each sensor should be treated. This paper investigates the virtual sensing of wheel position in ground-steering systems for aircraft using the output of four linear variable differential transformer (LVDT) sensors. The sensors are arranged to monitor the wheel position, which is calculated based on the nonlinear geometry of their alignment. A digital twin is first developed of the ground-steering mechanism. Even if each of the sensors is working without noise, there is ambiguity associated with all of the sensors that is seen as double solutions in the estimation. A kinematic analysis of position is presented, which relates the measurement of each LVDT with the actual steering angle. A bounded type of uncertainty is then introduced for each sensor output and is propagated through the model in order to calculate the maximum error in the steering angle estimation. The variation of such error with the wheel position is also presented. The best estimate is considered as the nearest solution between a set of measurements and the actual steering angle curve for the system without uncertainty. However, it is crucial that the estimate of the controlled signal does not change discontinuously, for example, if a sensor fails.
In the present study, stochastic natural frequency for laminated composite plates is presented by using Kriging model approach. The universal Kriging model is employed as a surrogate model constructed by using Latin hypercube sampling. A trigonometric higher order shear deformation theory is assessed for uncertainty quantification by finite element formulation. The influence of random variation of material input parameters on the natural frequencies is presented. The accuracy of present approach is ensured by convergence studies and error analysis. The present method is computationally efficient and requires reduced sampling size compared to direct Monte Carlo simulation. The frequency response function and stochastic mode shapes are also depicted for a laminated configuration.
A structural optimization method of subsystems to realize desired SEA parameters was proposed by the authors in the past studies. This method is based on a combination of SEA and FEM calculation, calculating repeatedly until satisfying the value of objective functions under arbitrary constraints. As a result of applying the proposed method to a simple structure consisting of two flat plates connected in an L shaped configuration, the design variable is taken as the thickness of the FEM element, a subsystem structure with the desired value of the CLF or power flow between subsystems for the one frequency band or multi frequency bands were constructed. However, it is difficult to adjust or process the thickness of the FEM element, so the optimization results is not realistic. In this paper, the method is also validated through numerical analyses, using a finite element method, of a flat plate and an L shaped plate, the subsystem is grouped into a plural elements, and each grouped element is set as a design variable, which should take a discrete value (in the form of either a damping rubber or original thickness), the total mass is taken as a constraint function in order to minimize the subsystem energy or CLF12 at one frequency band under each condition to realize utilization of this optimization method. As a result, in comparison with experimental data, numerical analysis results are qualitatively accurate.
In this work, the optimal design of an intermediate support of a beam structure is investigated for the purpose of controlling the reaction forces such that the excessive vibration loads transferred by the restraint supports can be reduced considerably during the vibration situation. By use of the differential transformation method, it is convenient to solve the differential vibration equation of the beam and obtain the restraint reaction forces of the supports. First, the influences of the intermediate support position and stiffness on the restraint reaction forces are studied. Then, the optimum design of the support position and stiffness of a beam structure is performed based on balancing the reaction forces in the restraint supports under the foundation excitations. Finally, an illustrative example is provided to show the effects of the optimal support design for balance of the restraint reaction forces under different excitation frequencies.
Common practice in vibration serviceability assessment of floors is to use a single stationary pedestrian loading scenario, corresponding with a deterministic walking force as suggested by contemporary design guidance. However, slender and lightweight floor structures are susceptible to excessive vibrations origi-nating typically from multiple pedestrians walking across floors with random walking paths. Along each walking path occupants excite a structure with a range of potential walking forces and as such produce different levels of vibration response. This study presents analysis of vibration responses under different loading scenarios, i.e. single and mul-ti-person. A typical new floor structure is used to study a range of walking paths under a number of con-trolled walking tests. A multiple pedestrian loading scenario is also implemented to investigate vibration responses at various locations on the floor. The study shows that vibration responses under multiple pe-destrian loading scenario produce a higher vibration responses than those of a single person and as such they should be considered at the design stage.
