Sandwich damping treatments with multi-layer and multi-material viscoelastic cores

Paper FUNDAÇÃO CALOUSTE GULBENKIAN II ECCOMAS THEMATIC CONFERENCE ON SMART STRUCTURES AND MATERIALS C.A. Mota Soares et al. (Eds.) Lisbon, Portugal, July 18 - 21, 2005 SANDWICH DAMPING TREATMENTS WITH MULTI-LAYER AND MULTI-MATERIAL VISCOELASTIC CORES Rui A.S. Moreira∗and José Dias Rodrigues† ∗Departamento de Engenharia Mecânica. Universidade de Aveiro Campus Santiago. 3810-193 Aveiro. Portugal e-mail: [email protected], web page: http://www.mec.ua.pt †Departamento de Engenharia Mecânica e Gestão Industrial Faculdade de Engenharia da Universidade do Porto R. Dr. Roberto Frias, s/n. 4200-465 Porto. Portugal e-mail: [email protected], web page: http://www.fe.up.pt Keywords: Viscoelastic Damping Treatments, Multi-Layer Sandwich, Passive Dynamic Control Abstract. The application of a viscoelastic core inside a sandwich plate provides an efficient dissipative mechanism able to introduce high levels of damping in light and flexible structures. Due to the decoupling effect produced by the soft viscoelastic core, the flexural stiffness of these sandwich panels can be significantly reduced and, therefore, can compromise the behavior of the damped structure. In this work, it is proposed a multi-layer core configuration which proved to be able to reduce the decoupling effect, maintaining though the treatment efficiency. A multi-material configuration is also proposed providing a good solution to enlarge the efficient temperature range of the damping treatment. A numerical simulation and an experimental study were performed on multi-layer and multi-material specimens. The achieved results verify and sustain the above assumptions and confirm that, under particular temperature conditions, a multi-layer configuration can be more advantageous than using a single layer. The overall flexural stiffness and damping efficiency increase of the sandwich structure are, thus, the most important benefits when adopting a multi-layer configuration. 1 INTRODUCTION The application of thin viscoelastic layers in the core of sandwich plates provides an effective passive damping mechanism broadly applied on light structures such as aeronautic fuselage panels and satellite panels. In fact, the viscoelastic core is cyclically shear deformed due to the relativemotion of the external skins of the sandwich, leading to an important thermal dissipative effect and, thus, to a considerable reduction of the vibration energy of the structure. Usually, very thin viscoelastic layers (0.02-0.10mm) are efficiently applied due to the high Rui A.S. Moreira and José Dias Rodrigues shear deformation that is imposed by the adjacent stiff layers. The application of thick viscoelastic layers, which increase the viscoelastic deformation energy, strongly reduces the overall flexural stiffness of the sandwich panel due to the reduced skin stiffness coupling provided by the soft and thick viscoelastic core. Moreover, the relative shear deformation developed within thin viscoelastic layers is significantly higher than the one developed within thick layers. To solve these restrictions it is proposed to apply several thin viscoelastic layers separated by interlaminar constraining layers. With this configuration, it is possible to increase the amount of viscoelastic material maintaining the flexural stiffness of the sandwich plate. Moreover, by using a multi-layer scheme, it is possible to apply viscoelastic materials with different transition temperatures, which can be useful to enlarge the efficient temperature range of the damping treatment. The aim of this work is to test and simulate several multi-layer and multi-material specimens in order to verify and validate the feasibility of the proposed treatment configurations. 2 MULTI-LAYER AND MULTI-MATERIAL CORE DAMPING TREATMENTS The application of multiple viscoelastic layers in free or constrained surface treatments was initially proposed by Jones [1,2] as a promising procedure to increase the treatment efficiency. This multi-layer configuration was also reported as a solution to enlarge the narrow efficient range of the viscoelastic treatments by employing layers of materials with different transition temperatures. The application of this technique to the integrated layer configuration in sandwich structures, pursuing the same benefits, isn’t, however, straightforward. In fact, the number of layers, the relative dimensions of the layers, the material properties and the layering sequence of multi-material configurations are important design parameters and play an important role in the structure behavior. The influence of these variables is considerably of extreme importance when compared to the multi-layer surface treatments, where the damping layer do not greatly modify the flexural stiffness of the structure. Contrary, the core of the integrated treatment is simultaneously responsible for the dissipative effect and the stiffness coupling between the outside layers, which defines the global flexural stiffness of the structure. 3 EXPERIMENTAL STUDY The first step of this study was the experimental verification of the feasibility of the multilayer concept. In this initial part of the study, an experimental work on a set of representative specimens was developed to evaluate the variation of the fundamental natural frequency and corresponding modal loss factor when adopting a multi-layer configuration. The experimental results obtained were also used to validate the model adopted in the numerical analysis of this study. 2 Rui A.S. Moreira and José Dias Rodrigues 3.1 Experimental specimens To develop the experimental study several plate specimens with integrated viscoelastic treatments were produced. Two viscoelastic materials, both provided by the 3M company, were used to produce the viscoelastic layers applied in the core of the sandwich plates. The first material, 3M ISD112 [3], is designed for room temperature applications presenting an efficiency peak between 20 and 30◦C. The other viscoelastic material, 3M ISD110, is designed for higher temperature applications, ranging from 40 to 100◦C. The specimens were produced with aluminum plates (aluminum alloy 1050A H24) with 1mm thickness, 200mm length and 100mm width. A thin aluminum sheet (aluminum alloy 8050 H24) provided the inner constraining layers for the multi-layer and multi-material specimens. Table 1 presents the properties of the materials applied in the study. Material Young’s modulus [Pa] Poisson’s ratio Density [Kg/m3] AA 1050A H24 70E9 0.32 2708 AA 8050 H24 70E9 0.32 2708 3M ISD112 see [3] 0.49 1140 3M ISD110 see [3] 0.49 1140 Table 1. Material properties The viscoelastic layers were applied on the aluminum plates following the manufacturer instructions. While the 3M ISD112 can be easily bounded to the metallic substrate at room temperature, the application of the 3M ISD110 material requires the application of specific temperature and pressure conditions [3]. 3.1.1 Single-layer specimens The single-layer specimens, designated by S1 and S2 (Figure 1), were produced by applying a single core layer of 3M ISD112 and 3M ISD110, respectively, with 0.127mm thickness. These specimens provided the reference to evaluate the multi-layer and multi-material benefits/drawbacks. 3.1.2 Multi-layer specimens Two multi-layer specimens, designated by M1 e M2 (Figure 1), were also produced applying, respectively, 2 and 3 thin layers (0.0508mm) of 3M ISD112 intercalated with thin (0.05mm) aluminum sheets that provided the constraining effect. 3.1.3 Multi-material specimens The specimens M3 and M4 (Figure 1) represent the multi-material configurations. These specimens are dimensionally identical to the multi-layer ones where some 3M ISD112 layers were replaced by identical 3M ISD110 layers, as depicted in Figure 1. The Figure 2 illustrates the experimental specimens produced and tested in this study. 3 Rui A.S. Moreira and José Dias Rodrigues S1 Aluminum 3M ISD112 Aluminum M1 Aluminum 3M ISD112 Aluminum 3M ISD112 Aluminum M2 Aluminum 3M ISD112 Aluminum 3M ISD112 Aluminum 3M ISD112 Aluminum S2 Aluminum 3M ISD110 Aluminum M3 Aluminum 3M ISD112 Aluminum 3M ISD110 Aluminum M4 Aluminum 3M ISD110 Aluminum 3M ISD112 Aluminum 3M ISD110 Aluminum Figure 1. Specimens configuration Figure 2. Specimens analyzed 4 Rui A.S. Moreira and José Dias Rodrigues 3.2 Experimental setup As stated above, the experimental study developed in this work had two distinct purposes: to provide a comparison analysis between the damping efficiency and the flexural stiffness achieved for each treatment, and to validate the model adopted for the numerical analysis presented in the following section. For both purposes, the aim of the experimental study was the determination of a representative set of frequency response functions providing the input data for a modal parameter identification process and, on the other hand, a reliable basis for the numerical layerwise model [5,6] validation for a direct frequency analysis using the complex modulus approach [4]. To obtain free boundary conditions, which minimize the boundary error effects, the experimental specimens were suspended by a thin nylon wire from a rigid frame. A mesh with 15 measuring points, as depicted in Figure 3, was defined for all the tested specimens. 12 3 4 5 67 8 910 11 12 13 1415 ¾ - 100mm ? 6 200mm Figure 3. Measuring mesh and specimen boundary conditions An electrodynamic shaker (Ling Dynamic Systems - model 201), suspended from an independent rigid frame, was utilized to generate a random ([0-800]Hz) excitation in point 5 of each specimen. A thin and flexible stinger was used to link the shaker to the miniature force transducer (Brüel & Kjær - model 8203) attached to the plate surface, which provided the measurement of the applied dynamic force (Figure 4). The specimens responses were evaluated by using a laser vibrometer (Polytec - model OFV303) to measure the velocity of each point of the measuring mesh (Figure 5). The temperature of the measurement was evaluated by a thermocouple located near the specimens. 5 Rui A.S. Moreira and José Dias Rodrigues Figure 4. Experimental setup - specimen excitation Figure 5. Experimental setup - response measurement 6 Rui A.S. Moreira and José Dias Rodrigues 3.3 Experimental validation of the numerical model Fifteen frequency response functions (mobility functions) were determined for each specimen. These experimental frequency response functions were compared to the numerical ones generated by using the finite element model adopted in this study [5], which allowed the validation of the finite element model as well as the complex modulus approach to characterize the viscoelastic material. This model assessment was performed by simple visual comparison of the experimental and numerical frequency response functions, as presented in Figures 6-11, and by using frequency response functions correlation indicators [7]. 0 100 200 300 400 500 600 700 800Hz 0 π −π Phase 10−3 10−2 10−1 100 101 |Mobility| [ms −1/N] Experimental 27.1oC Numerical 27.1oC Figure 6. Direct mobility function for specimen S1 (27.11ºC) 0 100 200 300 400 500 600 700 800Hz 0 π −π Phase 10−4 10−3 10−2 10−1 100 101 |Mobility| [ms −1/N] Experimental 27.1oC Numerical 27.1oC Figure 7. Direct mobility function for specimen S2 (27.11ºC) 7 Rui A.S. Moreira and José Dias Rodrigues 0 100 200 300 400 500 600 700 800Hz 0 π −π Phase 10−3 10−2 10−1 100 101 |Mobility| [ms −1/N] Experimental 27.1oC Numerical 27.1oC Figure 8. Experimental setup - response measurement 0 100 200 300 400 500 600 700 800Hz 0 π −π Phase 10−3 10−2 10−1 100 101 |Mobility| [ms −1/N] Experimental 27.1oC Numerical 27.1oC Figure 9. Experimental setup - response measurement 8