Parametric fuselage design

Integration of mechanics and acoustic & thermal insulation

More Info
expand_more

Abstract

Designing a fuselage is a very complex process, which involves many different aspects like strength and stability, fatigue, damage tolerance, fire resistance, thermal and acoustic insulation but also inspection, maintenance, production and repair aspects. It is difficult to include all design aspects from the start of the design process. In this research it is tried to find an answer to the question whether a Multidisciplinary Design and Optimisation (MDO) approach will lead to a better (lighter) fuselage design compared to the normally practiced sequential approach. To find an answer to this question, a multidisciplinary design method is developed for fuselage design. Since fuselage design involves many aspects this research is restricted to the mechanical and the thermal and acoustic insulation aspects. Also the fuselage is simplified by excluding detail parts like windows and interior parts. In future research this could easily be extended by including more detail parts and considering other design aspects. The multidisciplinary design method of a fuselage is given shape in the form of a Design and Engineering Engine (DEE). The DEE is an engineering tool that consists out of different computer tools linked together. The central heart of the DEE is the Multi Model Generator (MMG), which is a flexible parametric description of the fuselage geometry. The MMG consists of building blocks from which any type of fuselage configuration can be constructed. The input to the MMG is the definition of the input parameters that define the fuselage configuration. The user can specify which input parameters will be used as design variables. The outputs of the MMG are the different input models and/or data for the different analysis modules. The current DEE has four analysis modules that are explained briefly: 1.) The structural analysis module can analyse two different structural concepts; the stiffened skin concept and the sandwich skin concept. The stresses are determined with FEM analysis and are evaluated with strength, buckling and wrinkling criteria depending on the structural concept. 2.) The acoustical analysis module, which determines the TL, consists out of three parts. The first part is based on literature equations translated into a MATLAB script. The literature equations cover all insulation concepts that are involved for a fuselage wall. The sound insulation concepts that are involved are the single skin, influence of frames and stiffeners, circular resonance effects, the double wall principle, insulation blankets and visco-elastic damping layers. The second part consists out of a FEM natural frequency analysis and a FEM steady state dynamic analysis to determine the natural frequencies and sound pressure differences over the fuselage wall. Since the second part is quite time consuming this part is only used for the final optimum solution, while the first part of the acoustic analysis module is used during the optimisation loop of the DEE. The third part consists out of an active noise control module using piezoelectric actuators. This part has been developed in cooperation with TNO TPD within the ‘Smart Panel’ project. Unfortunately this part is not fully operational because this project was terminated before the DEE could be linked to the active noise control, prediction algorithms developed by TNO TPD. 3.) The thermal insulation module performs a FEM transient heat analysis on the fuselage wall. The inside surface of the fuselage wall is heated by a constant heat flux. The FEM solution shows equilibrium for the temperature difference between the inside and outside surface of the fuselage wall after a period of time. The magnitude of the equilibrium temperature difference is a measure for the thermal insulation of the fuselage wall. 4.) The weight module uses the geometric dimensions and material properties to determine the weight of each part of the fuselage and sums these weights to the final fuselage weight. The genetic algorithm concept has been used to optimise the fuselage design. A Design Of Experiments (DOE) method is used to determine, with the DEE, a population of solutions. With this population, response surfaces are created that are used within the optimisation procedures. Validation calculations are performed with the DEE to check the optimum solution. If required a new population is created in the neighbourhood of the optimum solution to perform another optimisation step. Two fuselage concepts for medium sized civil aircraft have been analysed with the developed DEE; the stiffened skin and the sandwich skin concept. In the analyses, both fuselages concepts were exposed to the same load case and boundary conditions. For the given load and boundary conditions, the DEE showed that the sandwich fuselage concept is slightly lighter compared to the stiffened skin fuselage. When considering minimum weight as the design objective the DEE showed that the MDO solution lies close to the sequentially optimised fuselage. By using carbon/epoxy instead of aluminium, the largest weight improvement was achieved for the stiffened skin fuselage. These analyses showed that the multidisciplinary design method did not result in a drastically changed fuselage configuration. The literature equations used in the acoustic module for the influence of the frames and stringers on the sound transmission loss on cylinders suggest that increasing the frame and stringer pitch will improve the sound transmission loss of the fuselage wall. To validate this, sound pressure difference measurements have been performed on cylinder walls with different stiffening. The experiments involved a nonstiffened cylinder, a cylinder with 6 stringers, a cylinder with 12 stringers and a cylinder with 12 stringers and 2 frames. No exact comparison between the measurements and the literature equations could be made. Therefore it is concluded that the literature equations are only indicative. For more accurate predictions more research and experiments are required. The DEE proved to work successfully. It is a flexible knowledge engineering tool that easily can be extended with new analysis modules. In future work this tool could be updated with new analysis tools for more accurate calculations. In conclusion the MDO approach did not deliver spectacular weight savings compared to the normally practiced sequential approach. The reason for this is that the design aspects strength and stiffness and the thermal and acoustic insulation showed little correlation. Perhaps by including more design aspects like impact resistance and fatigue into the DEE more advantage can be achieved with the MDO approach.