Resumen de Study of structural joints with composite materials to enhance the mechanical response of bus superstructures
The present work focuses on the study of multi-material joints for incorporation into bus structures to enhance their rollover crashworthiness. There are numerous vehicular accidents every year with unfortunate and lethal consequences. Of these accidents, bus rollovers are relatively rare; however, they can be life-threatening. For this reason, there are specific regulations for the crashworthiness of bus structures, with UN/ECE Regulation 66 (or R66) being the best known and most commonly adopted. The design of bus vehicle structures requires not only that the superstructure be capable of supporting the various components of the vehicle, but also that it be capable of resisting the loads of an impact while deforming and dissipating the kinetic energy of the impact. This latter function is vital to maximizing the survivability of passengers. R66 defines a survival space, also called residual space, within the bus cabin. To evaluate the rollover impact resistance, a test is performed in which the structure is overturned in an 800 mm deep trench. The structure is deemed adequate and R66 compliant if the deformed structure has not penetrated or come into contact with the previously defined residual space. Due to the high cost and extensive preparation required for performing a rollover test, there are alternatives contemplated within R66, such as calculations based on experimental bending results of the joints or structural nodes. If the calculations show that the energy dissipated by the structure, when applying a load equivalent to the rollover impact, is greater than the gravitational potential energy of the bus before falling into the trench, then the structure is deemed to have passed the test by equivalent calculation. It should be noted that the required experimental information refers to the bending moment - rotation angle curves of the zones where the plastic deformation of the structure is concentrated, also called plastic zones or plastic hinges. The characterization of these plastic hinges is, to this day, a problem in the engineering of automotive and lightweight structures. Generally, the superstructure of a bus is manufactured using rectangular tubular shapes. The use of shapes with larger thicknesses can improve energy absorption, however, it can also significantly increase the weight of the structure, since the thickness of the shape is constant along its entire length. This implies that non-critical areas will have a cross-section with a much greater thickness than necessary, which implies a considerable oversizing. In addition, the existing theories regarding bending collapse, developed in the early 1980s, have only been applicable to structural tubes with very thin walls, and are no longer effective when applied to commercial sections or when applied to reinforced shapes.
On the other hand, the use of composite materials or composites, specifically fiber-reinforced polymer composites, has been growing rapidly in different engineering sectors, such as the automotive and aerospace industries. This is due to their excellent specific properties such as specific strength and specific stiffness, which translates into remarkable materials with low weight. The vehicle industry has taken advantage of the low-weight aspect of composites, since it means fewer materials and, above all, lower fuel consumption and emissions. One of the most popular fiber composite materials is CFRP (carbon fiber reinforced polymer), which has been quite successful as it is currently being used for the manufacture of secondary structural elements, parts that do not withstand high loads, and even cosmetic elements; however, despite its excellent properties, it is not yet used as a material for main structural elements. The non-linear behavior of composites also represents a challenge even for engineering; unlike metals such as steel and aluminum, which are linear and isotropic, composites are generally anisotropic materials and the properties depend heavily on the manufacturing process. Also, the material and manufacturing costs of composites can be quite high when compared to metals, not to mention the costs associated with engineering and modifying production lines. For example, instead of joining components by welding, composites are generally bonded with adhesives to other elements, since using other joining mechanisms, such as bolted joints, can have a disastrous effect on the strength of the material. Thus, when adhesives are used, the entire manufacturing line must be modified to take into account the curing period of the adhesive in order to guarantee maximum adhesion. This limitation has made past projects focused on redesigning vehicle structures solely using composites technically feasible, but not economically feasible.
This thesis proposes that, for the design of the superstructure of a bus, instead of using only tubular sections of a single material (steel or CFRP), steel should be used as the base material of the structure, and the critical areas should be reinforced with CFRP elements. These critical areas are precisely the joints or structural nodes that suffer collapse due to bending. In this way, it is proposed that the objective of having superstructures with adequate and R66 compliant rollover crashworthiness, and low weight can be achieved. The localized use of CFRP also means that the cost associated with this material is kept to a minimum and only where it is needed, whereas the rest of the structure is kept with steel. The study of this type of structural nodes requires then a new and more general collapse model for steel sections with localized reinforcement.
The first step taken in the thesis is to confirm, by means of tests and mainly numerical simulations, that it is precisely the floor-column and column-roof joints of the bus that concentrate most of the plastic deformation during the rollover test, and that it is precisely the bending collapse that is the mechanism responsible for most energy absorption. Next, possible CFRP reinforcement geometries are proposed and evaluated according to technical and economic criteria. The main technical criterion to be followed is the ease of fabrication and how feasible it is to include the reinforcements in the production line. Additionally, reinforcements that can also be applied to existing structures are preferred. On the other hand, the economic criterion is associated with the amount of material required for each proposed geometry, and the possible associated costs, such as tooling and dies for manufacturing, as these items can add significant costs and require engineering design on their own.
The second step in the thesis is the study of the bending collapse of rectangular shapes. This step is carried out in two stages, both previously published in a high-impact journal.
The first stage corresponds to a study of the bending collapse of medium-thin-walled sections of a single material, in this case, steel. In the literature, it was found that the collapse theory developed by D. Kecman in 1980s is the most widely used in industry and research. This method consists of the definition of hinge yield lines, on which all the plastic deformation of the plastic collapse zone is concentrated. Then, the energy required to bring these lines to the plastic flow stress is calculated. It should be noted that this energy is totally dependent on the geometry, and especially on the angle of rotation of the entire plastic zone. Then, by numerical differentiation, the bending moment required to obtain the previously calculated energies is obtained. Subsequent investigations showed that this first theory has a number of limitations, which were gradually resolved by several authors. However, one important limitation remained, and that was that the existing theories did not adequately incorporate thicker-walled sections, namely, medium-thin-walled shapes. These shapes are typically commercially available, and their thickness is not large enough to consider them thick-walled shapes. Commercial rectangular tubular shapes or tubes, used in buses and other lightweight structures, tend to be thicker than those used in research. The discrepancy becomes evident when observing that thicker tubes do not have well-defined yield lines, and at greater thicknesses, the definition of these lines becomes more complicated. Moreover, from a mathematical point of view, the difference also becomes evident when comparing the ratios between thickness and side of the tube, since it is precisely on the basis of these geometrical ratios that proposals for modifications to Kecman's theory are made. The first modification consists of defining parameters that redefine the values of yield stress and flow stress, based on the geometric ratios. In other words, greater thicknesses translate into higher effective yield and flow stresses. The main reason for making this change is that greater thicknesses generate bending normal stress distributions in the through-the-thickness directions, which is not considered by Kecman and other models. On the other hand, the models published so far only consider the plastic rotation angles, i.e., the zone starting from the maximum plastic collapse moment, and do not consider the elastic and elastoplastic deformation portion. This supposition is made since it is assumed that the elastic and elasto-plastic portion of the moment-angle collapse curves are negligible compared to the plastic deformation. However, due to the additional thickness of commercial pipes, the elasto-plastic deformation zone is no longer negligible for this type of tube. For this reason, a regression is performed to obtain a relationship between the angle corresponding to the maximum collapse bending moment and the geometric ratios. The results of this first part of the thesis are corroborated by three-point bending tests performed on commercial tubes, as well as by numerical FEM simulations to explore cases and variants not seen experimentally. In these tests, collapse occurs in the central zone, with the highest bending moment, which acts as a critical zone. It is precisely in this zone that the plastic hinge is formed in the form of a bulge of the side walls. These results show that the proposed modifications to Kecman's theory better capture the bending collapse in medium-thin-walled rectangular cross-sections adequately and more accurately than previous models. It should be noted that this part of the research was published in Thin-Walled Structures. A few months after publication, other researchers (outside the thesis development) came to a very similar conclusion about the collapse of rectangular hollow sections with large thicknesses.
The second stage of the bending collapse study consists of the implementation of localized partial reinforcements, defined previously, as well as the development of a model capable of the prediction of their collapse stage: their maximum collapse moment and the corresponding moment-angle curves. In the previous step of the study, one of the secondary results was that, in any rectangular cross-section, the largest energy dissipation occurs in the hinge lines associated with the webs or lateral walls. In this sense, it is proposed that one way of reinforcing rectangular shapes is to cover the webs with plates made of CFRP, subsequently referred to as LR (left-right) reinforcement. In addition, CFRP plates can also be placed at the top and bottom in order to increase the moment of inertia of the cross-section as much as possible, these reinforcements are hereinafter referred to as UD (up-down) reinforcements. The CFRP plates are bonded to the steel using an epoxy-based structural adhesive, strong enough to guarantee load transfer between both materials. In the study, we avoid completely covering the entire tube with CFRP, as this would significantly slow down the production and manufacturing of the entire structure, since it would be necessary to wait for the CFRP and the adhesive to cure, and there would be no space to continue with the final assembly. Once the two types of reinforcement are proposed, the Kecman model is adapted to consider the effect of these reinforcements in the calculation of the absorbed energy terms, and subsequently in the calculation of the bending moment. To ensure that these relationships are maintained, the adhesive used must be strong enough to transfer the load without fracturing. For the calculation of the maximum collapse moment, a new theoretical model is presented based on the distribution of normal stresses required to produce plastic strain throughout the cross-section. This contribution is of special importance for researchers and structural engineers, since until now there was no approach for the calculation of maximum collapse moment in CFRP cross-sections. Similar to the previous contribution, this theoretical model is compared and calibrated using three-point bending tests on reinforced tubes of different sizes. In this way, not only a model capable of making accurate predictions of the bending collapse phenomenon is obtained but also the improvements associated with reinforcing a tube with UD or LR partial reinforcements are quantified. It has been found that placing localized reinforcements can raise the maximum resistance (shown as the maximum collapse moment) and dissipated energy by up to 57% and 45%, respectively, with a weight increase of up to 14%. When analyzing the different failure modes for tubes with different geometric ratios and reinforcements, it is found that the failure mode is always localized, i.e., it occurs in the area close to the plastic zone. For example, in the case of LR reinforcements, a mixed failure mode I and II occur just in the areas of largest strain, and it happens progressively only in the critical zone of the tube during the test. The rest of the test specimen remains undamaged, neither in the composite nor in the adhesive. In the case of UD reinforcements, the failure mode is mode II, and occurs along the entire length of the reinforcement. However, the analysis reveals that this type of failure is independent of the length of the reinforcement. Therefore, for both types of reinforcement, it is concluded that the reinforcement can be localized to the plastic collapse zone and not necessarily applied in the whole length of the steel shape. This development has also been published in the high-impact journal Thin-Walled Structures.
Finally, the theoretical collapse models developed before are used as inputs for the rollover analysis following the requirements of R66. A FEM model a representative structural loop of the superstructure of the bus is built. The loading comes from a rigid inclined plane that comes into contact and impacts the top corner of the structure, simulating the impact of the rollover test. The analysis and comparison of the rollover crashworthiness is done in terms of the maximum force applied to the structure via the rigid plane, and the energy dissipated by the structure as plastic deformation in the plastic collapse zones. The structural ring of the bus consists of the columns, the roof and the floor, as well as stiffeners that connect the columns, roof and floor, according to design guidelines. The focus of the analysis is on the structural ring in order to isolate the influence of only the dimensions of the tubes that make up the columns and the roof of the structure. Based on this model, different variations of dimensions for the columns and roof, with and without reinforcements, are evaluated to explore the influence of the CFRP reinforcements on the rollover crashworthiness of the structure. The focus is on the structural ring in order to isolate the influence of only the dimensions of the tubes that make up the columns and the roof of the structure. Based on this model, different variations of dimensions for the columns and roof, with and without stiffeners, are evaluated to investigate the influence of the stiffeners on the overturning resistance of the structure. It is demonstrated that the inclusion of partial and localized CFRP reinforcements in steel superstructures, which form multi-material joints, have the potential to significantly increase the rollover resistance of a bus structure without significant weight increase or major alterations to the production line.
