COMPOSITE CRUSH SIMULATION EMERGING TECHNOLOGIES AND METHODOLOGIES

Transcription

1 COMPOSITE CRUSH SIMULATION EMERGING TECHNOLOGIES Graham Barnes Engenuity Limited, UK Stuart Nixon SIMULIA, UK Marc Schrank SIMULIA, USA THEME Composites KEYWORDS Composites, Carbon, Laminate, Layup, Crush, Crashworthiness, Occupant Safety, Vehicle Crash, Collapse, Energy Absorption, Energy Management, Composite Impact, Crash Simulation, Lightweight, Minimum Weight, Low Emissions, Failure, Sled, Aerospace, Automotive SUMMARY In the quest to lower environmental impact while maintaining vehicle performance, automakers and aerospace companies are knocking on the same door that is, increasing use of composite materials in order to reduce structural mass. It can be expected that material costs will drop considerably over the next few to several years, as the capacity to produce such materials begins to catch up with the growing demand. The benefits of using these materials are well-documented, including their substantial capacity to absorb energy in an impact scenario. Composite structures generally behave much differently than their metallic equivalents in a crash event when subjected to large compressive and dynamic forces. Whereas a metal structure will typically undergo large plastic deformations to absorb the kinetic energy of the event, a well-engineered composite member subjected to large axial compression will pulverize in a progressive manner from one end to the other as the crush front moves along

2 its length. The crush front is characterized by numerous microscopic interactions between fiber and matrix, and efficient energy absorption can be achieved when the moving crush front becomes essentially a continuous event, traversing steadily along the length of the composite member. However, it is also this type of crushing failure which can prove difficult to simulate using conventional finite element methods. CZone technology has been developed to bridge the gap between experimental observations regarding a material s ability to absorb energy in an impact and the need to understand complex structural interaction and stability in a largerscale crash event. CZone technology is being incorporated as an add-on product for Abaqus/Explicit in order to provide a predictive capability for composite crush scenarios the CZone functionality handles the material s behavior at the crush front, while the damage and failure models in Abaqus/Explicit address the integrity of the backup structure. This paper will demonstrate the technology in action on some automotive scale structures and will identify the additional material characterization required to successfully predict their performance in crash. 1: Introduction Worldwide energy consumption is projected to increase by 50% between 2005 and 2030, and the rate of increase in the transportation sector during this period is expected to be more rapid than in any other end-use sector [1]. Such projections focus growing attention on potential energy conservation measures for transport vehicles of various types, including rail, aircraft and automotive vehicles. As in most transport vehicles, the rail industry includes weight as an important design criterion. A growing emphasis by rolling stock manufacturers on providing greater operational flexibility has the potential negative consequence of heavier vehicles. This not only increases energy consumption, but also reduces track life and quality. Hence, greater usage of composites is an active area of investigation in the rail industry [2,3]. At the same time, aggressive targets are being set in Europe to increase railway usage substantially by year 2020, and such targets are accompanied by recommendations to improve energy efficiency and reduce weight [4]. Composite-intensive aircraft have been common for military purposes for many years, where for instance, the high stiffness- and strength-to-weight ratios afforded by composite materials contribute considerably to enabling the high performance characteristics required in fighter jets. However, more recently, the wider incorporation of composites for aircraft has moved also to commercial aviation (Figure 1) [5].

3 Figure 1: Usage of composite materials in commercial aircraft [5]. Probably the most recognized example of this shift is the Boeing 787 Dreamliner, a new passenger jet that is 50% composite material construction (by weight), and is projected to provide a 20% improvement in fuel economy over similar sized planes that are of primarily aluminum construction [6,7]. Airbus is also introducing a similar level of composites content into the A350 aircraft [8]. The automotive industry is confronted with significant challenges in improving fuel economy and reducing carbon emissions over the coming years. In the United States, legislation passed in 2007 is mandating a 40% improvement in corporate average fuel economy (CAFE) standards by 2020, to 35 miles per gallon. In Europe, stringent requirements on carbon dioxide emissions are being established which effectively infer substantial improvements in fuel economy. Automotive OEMs have recognized the enormous opportunities that composite materials present. In addition to the high stiffness- and strength-to-weight ratios for composites already being leveraged in the aerospace industry, composite materials provide tremendous potential for energy absorption, such as absorbing the energy of a crash event and mitigating the accompanying accelerations and forces. Figure 2, from a 2007 report by the National Highway Traffic Safety Administration (NHTSA), shows a plot of the specific energy absorption capacity of various materials [9]. Clearly apparent is the appreciably greater capacity for several types of composite materials to absorb energy on a per unit mass basis, as compared to more commonly used metals, such as steel and aluminum. In some cases, the difference is nearly an order of magnitude.

4 Hence, composites offer the potential to be used as very efficient energyabsorbing crash structures in automotive vehicles. Figure 2: Specific energy absorption capacities for various materials [9]. 2: Composite Crushing and Crashworthiness Figure 3 shows a sequence of images of the dynamic crushing of a simple cone structure made of a laminated composite. A sled mass of 324 kg with an initial velocity of 10 m/s impacts the left end of the cone while the right end is held fixed. The cone responds much differently than a metallic structure might be expected to behave. While a steel or aluminum structure could be expected to buckle and fold, and perhaps fracture in various locations as it absorbs the kinetic energy of the sled mass, the composite cone behaves much differently. It essentially pulverizes into very small fragments as the crush front, or contact interface between the cone and rigid impactor progresses down the length of the cone structure. The specific energy absorption achieved in this particular test is 47.5 kj/kg. Figure 3: Experimental dynamic crush testing of a simple composite cone structure.

5 The micromechanics of what transpires at the moving crush front are quite complex. There are a number of phenomena that evolve and interact, and which then lead to the crushing behaviour observed at the global level. These phenomena can include: localized buckling, delamination, matrix cracking and fiber breakage. Seeking to account for such phenomena in detail both accurately and robustly in composite crushing simulations is challenging, with no generally accepted approach yet established. This present difficulty in being able to carry out design for composite crash structures that can be largely based on predictive, robust simulations is in contrast to the pervasiveness of crashworthiness and occupant safety simulation in the automotive industry today. Over the past 20 years or so, such simulation has evolved into having a prominent role in being able to design modern automobiles to meet increasingly stringent crashworthiness regulations. However, without reliable capabilities for composite crushing applications, simulation-based design for composite crash structures is less feasible, often necessitating an expensive, time-consuming, and often unacceptable approach of Design-Build-Test-Redesign. Various efforts have been initiated to support the greater incorporation of composite crash structures for crashworthiness purposes. For example, the Composite Materials Handbook (CMH-17) has within the last few years established a separate Crashworthiness Working Group, to develop a new, self-contained section of the handbook on composite Crashworthiness and Energy Management for vehicle safety certification, and also address the needs of the composites and vehicle safety community at large, and to provide a unique forum of discussion for those working in industry, research institutions, and government agencies. [10] Likewise, the U.S. Congress has funded NHTSA to undertake research and continue development of a program to examine possible safety benefits of Lightweight Plastic and Composite Intensive Vehicles (PCIV). [9] 3: CZone Technology CZone technology has been developed by Engenuity Limited as a way to enable robust simulation of composite crushing so that simulation-based design of composite structures for crashworthiness can be carried out with a similar level of predictability as is presently possible for metallic structures. The basis of CZone recognizes crush or crush stress as a distinctive mechanical property of a composite material. Crush is essentially defined as the ability of a material to progressively absorb energy through destruction and disintegration. Like other mechanical properties, it needs to be measured, just as fiber compression stress needs to be measured in order to be

6 implemented in a finite element analysis simulation. Crush stress provides the ongoing resistance when a portion of the composite structure is destroyed against an impacting body, and it is this resistance that effectively becomes the input force into the rest of the finite element model. CZone technology is being made available commercially within the Abaqus finite element software suite, in particular Abaqus/Explicit, an explicit dynamics solver used across several industries for various applications, including crashworthiness and occupant safety. The integration of CZone within the framework of a commercial finite element software suite has the noteworthy benefit that the wide range of features and functionality already available in the commercial software package can be used in conjunction with the specialized crushing simulation capabilities of CZone. 4: Calibration of Crush Stress The crush stress mechanical property required for CZone usage can be obtained through various test methods. The most feasible and cost-effective is crush testing of coupons extracted from flat plaques or sheets of the candidate laminate. These coupons can readily be cut from the flat plaque using a water jet, and they have a saw-tooth shape cut into one end in order to induce the initiation of crushing (Figure 4). Crushing is carried out with a specially designed test fixture that can be placed in any suitable loading apparatus, such as a high-rate Instron machine. The fixture provides lateral restraint against buckling for the test coupon during crushing, while also providing a means for the crush debris to freely exit the fixture and not influence further crushing response. Figure 4: Crush test coupons cut from flat plaque laminate and crush test fixture. Crush stress can be dependent on several factors, including temperature, direction of crush, and crush velocity, as well as a potential dependency on ply stack sequence. The cost-effectiveness of the flat coupon testing method allows for these dependencies to be accounted for in the calibration process.

7 Some recent experimental evidence suggests that crush stress for certain composite materials exhibits an influence of geometry. In relatively flat regions of the structure ahead of the crush front (prior to crushing), partial delamination between plies can occur, influencing the subsequent crush stress. Whereas in geometries that have sufficient curvature normal to the crushing direction, this delamination is suppressed, thereby enhancing the measured crush stress. This difference in measured crush stress can be accommodated for in crushing simulations by assigning appropriate crush stress values to regions of the original composite structure that are either considered flat or curved. Recent work to enhance the flat coupon test fixture by incorporating a pin restraint device for the coupon seems to reasonably induce the same type of delamination suppression as due to curved geometry. Not all composite materials exhibit good crushing behavior, and CZone technology is not intended for those materials that do not crush well. The flat coupon test method can be used effectively to not only calibrate crush stress, but also to screen candidate materials in order to assess their ability to sustain crush. Two primary characteristics are monitored: the average crush stress over the length or duration of crush, as well as the stability or consistency of the crush stress. Three examples of Good, Fair, and Poor crushability are shown in Figure 5. For the example showing Poor crush characteristics, the large variations in crush stress over the length or duration of the test indicate large fragments of the coupon that break away during the test, with no continuous crush front ever really being established, similar to the test coupon also shown in this figure. Figure 5: Examples of crushing characteristics exhibited by different composite materials, along with test coupon exhibiting poor crush characteristics.

8 5: Examples The simple cone crush test previously shown in Figure 3 is simulated using CZone and Abaqus. The experimental cone is fabricated with a layup sequence that establishes three different thickness regions (thinnest at the left end and thickest at the right end in Figure 3). Initial impact at the thinnest end causes local buckling in this region, with some fragments of the cone breaking away due to bending stresses that develop. Figure 6 shows a comparison of these fragments produced in the experimental test versus those predicted in the simulation (dark gray shaded areas plotted on the undeformed cone geometry). Figure 6: Comparison of fragments from simple cone crush test (left) against those predicted through simulation (right). As further impact proceeds, the increasing thickness of the cone leads to progressive crushing of the cone without further buckling and bending failure, along with increasing resistance against the moving sled mass, as shown in the comparison of measured and predicted sled acceleration in Figure 7. Figure 7: Comparison of measured and predicted sled acceleration for simple cone test. Figure 8 shows a more complex test cone specimen, similar in size and shape to an automotive structural rail, and including a through-hole feature also common in such geometries. The cone has an attached mass of 1150 kg and is moving at an initial velocity of 9.1 m/s prior to impacting a rigid fixed barrier. Crushing of the cone progresses to a point where a large fracture develops suddenly in the transition region between the cone and its backup structure (denoted by the red oval in the last frame).

9 Figure 8: Experimental crushing of a complex cone structure in a sled test, with sudden fracture subsequently occurring away from the crush front. Figure 9 shows simulation results for this test, indicating good correlation both for the prediction of crushing response as well as the sudden fracture in the transition region between the cone and its backup structure (outside the crush front). Figure 9: Simulation results for crushing of complex cone structure, comparing acceleration history against experimental data, and prediction of fracture away from the crush front. This example demonstrates that accounting for the possibility of structural damage or failure away from the crush front is equally important in the design of crashworthy composite structures. If this damage or failure occurs, the energy absorbing capacity of the structure can be substantially reduced or compromised. In this example, the sudden fracture that develops away from the crush front causes any further energy absorption potential to be largely lost, as is evident in the abrupt change in acceleration history shortly after 30 ms when the fracture occurs.

10 This example also demonstrates CZone technology working in concert with existing features available in Abaqus/Explicit. CZone addresses the crushing response at the interface between the cone structure and rigid wall, while Abaqus/Explicit addresses all other aspects of the simulation, including the potential for damage and fracture away from the crush front, in this case utilizing the Tsai-Wu failure criterion [11]. The last two examples are similar complex cone structures as the previous one; however, the impact angle against the fixed rigid barrier is 30 degrees rather than a normal or zero degree impact as in the previous example. The mass attached to the cone is 1164 kg and the impact velocity is 9.0 m/s. The sequence of images in Figure 10 shows a direct comparison of physical test and simulation over the progression of the impact event. In both the physical test and simulation, virtually no crushing response occurs due to a catastrophic fracture that develops very early between the cone and its backup structure. Figure 10: Complex cone sled test at impact angle of 30 degrees; comparison of experimental and predicted cone deformations. Figure 11 shows a similar sequence of images, this time for a cone that is reinforced with additional plies in the backup structure transition region. This cone also includes additional features, such as indentations, and a bonded rigid insert, that could be necessary to accommodate when such a structure is incorporated into a complex vehicle assembly. While the impact angle again is 30 degrees, substantial crushing is evident, both in the physical test as well as in the simulation. Figure 11: Sled test at impact angle of 30 degrees for reinforced complex cone; comparison of experimental and predicted cone deformations.

11 6: Summary Projections of substantial increases in energy consumption by the transportation sector in the coming years, along with directives to reduce carbon emissions, are leading to various energy conservation measures for transport vehicles. Reducing vehicle weight through greater incorporation of composite materials is one such measure, and commercial aviation is beginning to adopt composites in a considerable way for certain new passenger airplanes. From a crashworthiness perspective, composite structures offer the potential for far greater specific energy absorption capacity, through a crushing mechanism, than their metallic counterparts. However, the present ability to accurately and robustly simulate such crushing response with commercial finite element software is lacking, and this is contributing to the relatively slow adoption of composites in the automotive industry to date. CZone technology has been developed as a means to provide more accurate and robust composite crushing simulation, and has been implemented within the Abaqus finite element software. CZone utilizes the concept of crush stress (a measured property) to address the crushing response at the contact interface between the composite structure and other impacting body. Examples of different composite cones and load cases show good correlation between simulation and experimental test data for prediction of the crushing response, as well as for prediction of gross fracture away from the crush front. CZone technology provides an effective approach to modeling composite structures for crash that differs from the more common degradation models implemented in existing explicit finite element methods. When the appropriate crushing loads derived from the CZone technology are applied to the reinforcing structure within the Abaqus finite element software, it is shown to be proficient at reliably predicting the generation and evolution of damage away from the crush front. The ability to predict the detailed damage response to relevant impact events provides increased confidence in the development of large scale structures to meet their intended requirements, without having to resort to a protracted and expensive Design-Build-Test-Redesign approach. Faced with the current challenging economic environment, organizations are ever more determined to avoid the risk of test failure and associated damages and costs. Industry is increasingly relying on analysis and simulation tools to design right the first time, and the CZone technology enhancement to Abaqus now extends these capabilities to include the development of composite structures for crashworthiness.

12 REFERENCES 1. Energy Information Administration, International Energy Outlook 2008, Report No. DOE/EIA-0484(2008), September INGLETON, S., FOUND, M.S., and ROBINSON, A.M. Design of composite vehicle end structures in railway rolling stock. Experimental Techniques and Design in Composite Materials 4, Ed. Found, M.S., Taylor & Francis, NewRail, The Research Requirements of the Transport Sectors to Facilitate an Increased Usage of Composite Materials. Part III: The Composite Material Research Requirements of the Rail Industry, COMPOSIT Thematic Network, June European Rail Research Advisory Council, Strategic Rail Research Agenda 2020, May EADS Deutschland GmbH, The Research Requirements of the Transport Sectors to Facilitate an Increased Usage of Composite Materials. Part I: The Composite Material Research Requirements of the Aerospace Industry, COMPOSIT Thematic Network, June Boeing Company, 7. SMOCK, D., Boeing 787 Dreamliner Represents Composites Revolution, Design News, June 4, EADS/Airbus, 9. National Highway Traffic Safety Administration, A Safety Roadmap for Future Plastics and Composites Intensive Vehicles, Report No. DOT HS , November Composite Materials Handbook, TSAI, S.W. and WU, E.M. A general theory of strength for anisotropic materials. Journal of Composite Materials. Vol. 5, pp , 1971