Investigation of behaviors of the composite plates during static and dynamic loads

Composite materials have been used for many years. The reason why people prefer composite materials since very old years is that they have important advantages. The advantages of composite materials resulted in an increase of use in structural design of the vehicles and have revealed the need for a better understanding of behavior under crash loads. The purpose of the studies on the safety and crash behavior of composite materials showed that they have similar or better capabilities compared to conventional materials. Studies on the crashworthiness of composite materials have increased significantly in parallel to the knowledge in this field last 30 years. With the development of technology, increased knowledge can be used more efficiently. The calculations required to predict the crash behavior of composite materials are much more difficult than the manual calculations. For this reason it is necessary to make analysis with the help of computers. Performing accurate simulations depends on accuracy of damage criterias. From past to present, numerous composite failure criterias were developed such as, maximum stress, maximum strain, Tsai-Hill, Tsai-Wu, Hoffman, Hashin, Hash-Rotem, Puck, Chang-Chang, and computer-aided analysis and tests in different scenarios with different composite materials were performed by researchers. Damage criterias generally work as a result of using an appropriate formula for the stress values of an element in different directions and analyzing the resulting stress. This thesis examined the failure crtierias developed in the past and many damage criterias not mentioned in this study are available in the literature. In the selection of damage criterias that will be examined, finite element softwares used in this thesis were taken into consideration. Purpose of failure analysis studies are to predict the damage behavior, damage evolution during crush and absorbed energy end of the crush of every structure using the material datas. In order for a damage criterion to be regarded as successful, the correct result must be provided in every case. Studies on this subject suggest that our knowledge is not enough for damage analysis of composite materials. Uniaxial carbon fiber reinforced composite materials are preferred commonly at crash scenarios due to their high strength and high capability of energy dissipation. Studies in the literature in order to examine the crushing behavior of composite materials, flat plate, pipe, tubing, corrugated, and a hat-shaped specimen geometries were used. The damage scenario and samples used in this study were selected from the axial crushing of composite plates studies published in the literature. The sample used in this thesis is a carbon fiber-reinforced, uniaxial, 16 layered, [(0°/90°)4]s layer placement, called as T700/M21 composite flat plate. The thickness of each layer is 0.26 mm. This composite material exhibits very high strength properties in the fiber direction and is used mainly in aeronautical industry as the wing box or the fuselage. The effect of holes on the material strength to be opened during assembly is another research topic that is frequently studied. The performed tests used in this thesis were taken from literature. In the reference works, quasi-static and dynamic tests were carried out to examine the damage in low-speed behavior of the composite plates. Test setup was designed for testing of composite plate specimens. Quasi-static tests carried out with a constant speed of 6 mm/min using a hydraulic testing machine in compression and dynamic tests carried out with 9 m/s initial speed with a falling load of 36 kg on the specimen. The unsupported length of specimen was 30 mm at the beginning of the quasi-static test and decreases as the displacement of piston increases. The unsupported length of specimen was fixed to 20 mm and is constant during dynamic tests. The specimens were crushed 10 mm in static tests and 20 mm in dynamic tests. The size of the specimens are 60 mm x 10 mm in quasi-static tests and 160 mm x 60 mm in dynamic test. Thickness of the plates is 4,16 mm. In order to trigger the damage and create progressive damage, 45ᵒ chamfer was created on one end of the specimen. The tests were repeated twice and were followed by high-speed cameras and data collection devices. Force-displacement results and damage behavior images obtained from the tests were reported. There are many commercial finite elements softwares to perform failure analysis of composite materials. In this study, commonly used ABAQUS 6.14-4, LS-DYNA R7 and RADIOSS 14.0 finite element softwares, material models and failure criterias offered by this softwares were used for the same scenarios to make comparisons between them. After creating 3D geometry of the specimen using CATIA, HyperMesh and HyperCrash were also used for creating the finite element model. HyperMesh and HyperCrash are very useful softwares for defining element and material properties, contact parameters and boundary conditions. ABAQUS user interface is used for analysis used ABAQUS solver and HyperCrash user interface is used for LS-DYNA and RADIOSS analysis. The softwares that are used on the date of this study selected as the final versions, improvements that are implemented in softwares are considered to make up-to-date study. When the previous studies were examined, it was seen that the studies were usually done using ABAQUS. The lack of studies using LS-DYNA and RADIOSS is noteworthy. In addition, lack of benchmarking studies performed in the literature by using a combination of finite element programs, forced the study of this thesis on this issue. Finite element softwares, which are almost mandatory in the study of damage thus will be better recognized and this work may be a helpful resource for future works. There are various methods of modeling with finite elements of composite materials. Among these methods, the most commonly used technique is the modeling of layer groups within shell elements. However, it is not possible to predict delamination by using this method. While the damage analysis is performed with the composites and when the effect of delamination can not be ignored in the damaged cases, delamination must be ensured in the modeling. Especially in case of damages where the plies are separated from each other after the damage and the other plies are affected, it is very important that the delamination damage is correctly modeled. There are also methods to do this. Three-dimensional solid modeling for each layer was used to observe the delamination to be more realistic. Between layers that were modeled using solid elements, three-dimensional adhesive layer were defined. In this case, test results more compatible with experimental studies can be obtained. When elements that are evaluated using various damage criterias reach limit, they are damaged and layers separated from each other. With ABAQUS and LS-DYNA it is possible to define the adhesive region between the layers using contacts. According to the traction-separation law, the surface-to-surface relationship is deleted after the damage, and the layers are separated from each other. In this case, realistic results can be obtained and analysis time is shortened. However, because RADIOSS does not support, this method is not used. Also, damage criterias used for the adhesive elements are included in this thesis. Force-displacement graphs of damages were obtained using ABAQUS, LS-DYNA and RADIOSS softwares with different material models and failure criterias. Using these graphs, the maximum reaction force and the absorbed energy amount are calculated. Damage behaviors are examined and the final damage status is shared for each analysis. The main purpose of the study is to show that the obtained force-displacement graphs agree with the experimental data. Analysis was repeated with different coefficients of frictions. Thus, it is also aimed to predict the correct friction coefficient. For some damage criterias that require failure strain, results were compared obtained from different values of failure strains. The fact that there is no data related to failure strain in the experimental studies made necessitates analyzing with different values in this study. Element types, contact types and parameters, boundary conditions, and other required parameters that are used during the creation of finite element models are described in detail. Thus, comparable results were obtained as compared with each other within the softwares themselves. As a result; capabilities of finite element analysis and damage criterias that are used for damage analysis of composite plates with three-dimensional elements are showed. While examining the obtained results from analysis, force-displacement curves and damage behaviors are compared with the results obtained from the experimental studies. In order to be able to accept that an analysis result is correct, the force-displacement curve, the maximum reaction force, the amount of absorbed energy, and the damage behavior must be compatible with the experimental study. In addition, the accuracy of the damage behavior of the composite plate is considered using the images obtained with the cameras in the experimental study. ABAQUS and LS-DYNA softwares are generally estimate the behavior of damage more accurately. Although LS-DYNA provides various composite material models and damage criterias, ABAQUS is limited at this subject. Using ABAQUS software with different damage criteria needs to identify the material models or the damage criterias by writing codes. It is seen in the literature that many researchers use their own material model codes for their studies. In this way, desired element types and contact definitions can be used easily. Composite material models in LS-DYNA generally require more experimental work, and if they are done, more realistic analysis results will be obtained. Although damage criterias offered to the users with RADIOSS succeeds in other two softwares, RADIOSS is found as the most unsuccessful in this study than the other two. Although the material models are easy to define, the disadvantages of the element formulations and contact definitions caused this situation. When the damage criterias used were examined, it was found that the Hashin-Rotem and Hashin damage criterias were successful. Besides, the combination of Chang-Chang and Tsai-Wu damage criteria in MAT 55 material model could be successful in LS-DYNA. In this case, it is seen that the damage criteria related to the damage modes give more realistic results. However, when the same damage criteria are examined with a single finite elements software, the result will be more accurate.