ABEL PARDO, JOSE CARLOS FERNANDEZ STRUCTURAL ANALYSIS ENGINEERING GRUPO TAM
(Published on June-July 2010 - JEC Magazine #58)
The Boeing 787 Dreamliner will be the first commercial jet to be more than 50% composite by weight. The Airbus A380, with deliveries that started in 2008, is also increasing its reliance on composites. With the material paradigm shifting in aerospace, it was predicted at the CompositesWorld 2008 conference that demand for composites would grow 30% in the general aviation market over the next three years.
Carbon-fibre-reinforced polymer (CFRP) is the most common composite used in the aerospace industry. Carbon fibres have a micro-graphite crystalline structure and a pattern similar to chicken wire; they derive their strength from layering, or sandwiching, multiple sheets in a polymer matrix. With their attractive strength-to-weight ratio and other attractive material properties such as high tensile strength, high elastic modulus, heat resistance, low thermal expansion and chemical stability, CFRP composites are highly desirable in high-performance aerospace and automotive applications. They are also used widely in sailboats, canoes, bicycles, tennis rackets and golf clubs, as well as consumer goods such as laptops and stringed instrument bodies.
Like any material, composites have their own set of manufacturing, assembly and lifespan challenges that must be fully understood to make their use in critical applications, such as commercial flying, acceptable and safe.
To fully understand the performance of these composite components, as well as assembly and maintenance challenges, a Grupo TAM structural analysis engineering team conducted a series of in-depth analyses of components including a curved, stiffened composite panel typical of a fuselage or fan cowls (see Figure 1). The panel and stiffeners are made of uniaxial and biaxial carbon fibres that are bonded with adhesive. The team focused on the composite manufacturing variables and tolerances for the panel, including material properties, geometric tolerances, thicknesses and lay-up alignment axes, as well as the delamination and disbonding that can occur during the manufacture, assembly and service life of the composite structure. The objective of these analyses was to identify the influence of deviations, defects and damage and to consider it during the initial design phase. In this way, non-conforming parts would be minimized, with associated cost savings.
Abaqus FEA creates baseline for composite analysis
For the intact panel analysis, the Grupo TAM engineers chose Abaqus FEA from SIMULIA, the Dassault Systèmes brand for realistic simulation, largely for its ability to handle both implicit and explicit non-linear analysis. The team needed more than their in-house tools for their analysis. They chose Abaqus for its extensive composite capabilities and to meet the high quality standards required by customers. They also chose Isight from SIMULIA for its Monte Carlo and Stochastic Design Improvement components, its sampling capability and the ease with which it can interface with inhouse software. Isight allowed the team to conduct trade-off studies with their Abaqus models and achieve rapid design optimization.
To carry out their finiteelement analysis (FEA) of the intact panel, the team started with nominal values typical of the aeronautics industry for all variables. They considered three load cases – two with a uniform aerodynamic pressure on the panel (one directed towards the inside of the structure, the other directed out) and one with a shear load directed axially across the face of the panel (see Figures 2 and 3). The team then performed two additional analyses of damaged panels – one with delamination in the middle of the panel, the other with two disbondings under the panel stiffeners.
The team constructed their geometry model in CATIA v5 from Dassault Systèmes using the following: S4R planar elements for the skin and stiffeners; the C3D8R element for the adhesive; shell composite with a single ply for delamination analysis; and a homogeneous solid in which mechanical properties were reduced six orders of magnitude for disbonding analysis. The model included approximately 49,500 elements, 45,400 nodes, and 272,600 variables.
The analysis was run on a Windows server with four Intel Xeon processors, each with 64 bits and 8GB of RAM. The team conducted multiple Monte Carlo simulations, with each full cycle analysis – including nominal, delamination, and disbonding analysis – taking 30 minutes.
The results of all the FEA analyses – both for intact and damaged components – provided the baseline data used to optimize the design and build of the composite panel using Isight.
Isight helps optimize composite performance
As with many materials and structures, there are numerous variables to consider when designing a composite panel for an airplane, and it is difficult to determine which variables might be key to improving structural strength and performance. In such instances, a stochastic approach is useful for managing the enormous amount of data inherent in composite analysis. Isight streamlines this iterative solution process by using an interactive graphical interface and aeronautics features built into the software to enable tools like Monte Carlo, Design of Experiments, and/or Six Sigma for optimization. In this case, the Grupo TAM team chose the Monte Carlo method, which is particularly useful when there is significant uncertainty on the variables and input. The Isight solver allowed the team to quickly evaluate a large number of design possibilities and identify those meeting the required parameters.
To begin the stochastic analysis in Isight, the Grupo TAM engineering team looked at the manufacturing variables and tolerances, as well as the range of damage during the component life cycle, determining that there were 58 important input variables. Statistical distributions for each variable were taken from either the baseline analysis data described above or standard industry values. The team then built a calculation flow chart by using Isight’s intuitive graphical tools and icons (see Figure 4). Isight then automatically ran this analysis string repeatedly without the need for individual manual FEA analyses, with each Monte Carlo simulation including between 100 to 800 samples. Descriptive sampling was chosen because it has better convergence with the statistical distributions and requires fewer iterations. In the end, this powerful computational process identified the most critical tolerances and variables for them.
With the results of this study, the team now have a clear understanding of which variables are most critical to the manufacture of composite panels that will meet the stringent quality and safety criteria.
|The manufacturing of aircraft is evolving into a process in which a variety of specialized manufacturers are contracted to produce structures or sub-assemblies that are then assembled into a finished aircraft by the parent company. One such specialty firm is Grupo TAM, headquartered near Madrid, Spain, which manufactures auxiliary components with state-ofthe- art CNC tools. About 40% of its business is in aeronautics, including the design and manufacture of composite structures.|
Optimization leads to cost reduction
While the goal of optimizing the composite panel with Abaqus FEA and Isight was to increase panel strength and ultimately performance, the analysis process also provided insight into associated manufacturing, assembly and maintenance costs.
The engineering team reached a number of interesting conclusions. They found that buckling pressure was the most critical factor and that a stricter material tolerance would lead to improved performance along with lower quality control and maintenance costs. They also determined that other less critical tolerances could be relaxed, resulting in both material cost savings for the carbon fibre sheets and manufacturing cost savings where lay-up tolerances are involved. In addition, the analysis demonstrated that delamination damage had a high impact on performance, while disbonding could be tolerated, especially with a new lay-up procedure. All of these results lead to resource optimization, with a quality and maintenance plan focused on the most influential inputs.
Looking to the future, the Grupo TAM structural analysis department identified a number of developments that will further improve the overall cost evaluation process. For instance, parallel computing in Isight will cut computing time in half (see Figure 5). A design-to-cost strategy will also be employed, in which costing functionality will be added to the analysis using software that is currently under development in the industry and can be incorporated into Isight.
This analysis process would lead to what Grupo TAM calls robust design, robust, because it takes into consideration the entire product life cycle as well as all associated costs. Incorporating such cost considerations within stochastic analyses will undoubtedly provide tremendous value to manufacturers in any industry.