Font Size: a A A

Gradient-Based Aero-Structural Multidisciplinary Optimization Method And Application Research

Posted on:2019-07-26Degree:DoctorType:Dissertation
Country:ChinaCandidate:T H YangFull Text:PDF
GTID:1360330623453254Subject:Aircraft design
Abstract/Summary:PDF Full Text Request
The high-aspect-ratio wing with light structural weight has become one of the most promising technologies for advanced civil passenger aircraft and High-Altitude Long-Endurance(HALE)Unmanned Aerial Vehicle(UAV)due to its characteristics of low induced drag.Such wings feature in much more severe aeroelastic problems compared with conventional aircraft,including static aeroelasticity,flutter and so on.How to comprehensively considering the influence of static aeroelasticity and flutter on the aero-structural multidisciplinary design is a crucial problem for fully exploiting its technical potential.The introduction of static aeroelasticity and flutter causes great challenges to the development of efficient and robust multidisciplinary design method.Moreover,the tight coupling feature among aerodynamics,structure and aeroelastic prompts the need to explore design principles and characteristics of high-aspect-ratio wings with light structural weight from a multidisciplinary perspective.In response to these problems,based on adjoint theory and improved Chebyshev spectral method,an aero-structural multidisciplinary optimization design method is well-developed.The built design method can consider the effects of static aeroelasticity and linear flutter characteristics by using aeroelastic analysis model with different precision.Then,optimizations on typical flexible high-aspect-ratio wings are performed to research the design characteristics of drag reduction,structural mass and stiffness distribution.The main research works and problems solved in this dissertation are as follows:1.For static problems,based on high-credible solvers and gradient-based optimization algorithm,an aero-structural multidisciplinary optimization design system with the adjoint theory is constructed.The optimization design system can deal with optimization problems with large-scale design variables.First,An high-credible static aeroelastic analysis framework is established.Interpolation method through rigid links is used for data transfer of aerodynamic forces and structural deformations between RANS-based CFD solver and plate-shell finite element-based structural solver TACS.IDW mesh deformation method is utilized to update aerodynamic calculation meshes automatically.Then,aero-structural adjoint equations are derived with the adjoint theory.At last,by introducing FFD geometric parameterization based on B-spline as well as gradient algorithm optimizer SNOPT,a gradient-based aero-structural multidisciplinary optimization design system for static problems is constructed.The built design system can consider the influence of aeroelastic deformation.2.A new flutter calculation algorithm is developed based on the improved Chebyshev spectral method.The underlying mathematics properties of Chebyshev spectral method is researched first.In order to develop an efficient calculation method for the derivative of flutter object function with respect to design variables based on adjoint theory,some modification methods are utilized.A mapping transformation is introduced to improve the calculation precision of the Chebyshev spectral method by reconstructing the distribution of Chebyshev collocation points.Additional,the “secondary transformation” improved method is proposed for the Chebyshev operator,which eliminates the dependence of the calculation precision of the Chebyshev spectral method on the discrete time step.After this,for high-aspect-ratio wings at low subsonic speeds,a new flutter calculation method is established by combing the improved Chebyshev spectral with the unsteady panel method and geometrical-exact beam model.The new method converts the time domain calculation problem for flutter into a coupled problem calculation at a finite number of time nodes.The calculation process for the coupled problem is similar to that of static aeroelasticity.Verification cases on Goland and HALE wings verify the calculation accuracy of the built calculation method for flutter.The new calculation algorithm based on the inproved Chebyshev spectral method lays the theoretical foundation for the development of an efficient gradient calculation method for the derivative of flutter characteristics with respect to design variables.3.An efficient gradient calculation method for the derivative of linear flutter characteristics with respect to structural design variables is proposed.And,by combing this method with aerostructural multidisciplinary optimization system for static problems,a new gradient-based aerostructural multidisciplinary optimization method is established.This optimization method can consider influences of both static aeroelasticity and linear flutter characteristic.The damping ratio under given conditions is selected as the parameter to describe the linear flutter characteristics of the whole system.Utilizing the developed calculation method for flutter,the derivative of the damping ratio to design variables can be obtained by constructing and solving adjoint equations at a finite number of chosen time nodes.In this way,compared to the traditional unsteady time domain adjoint method,for each gradient calculation of the damping ratio with respect to structural design variable,the times for constructing and solving adjoint equations is reduced from hundreds to once.As a result,the gradient calculation efficiency is improved significantly.Cases about Goland and HALE wings verify the calculation precision of gradients.Then,the effect of flutter is introduced into the aero-structural multidisciplinary optimization design system for static problems.The classic theory of engineering beam is utilized as the conversion bridge between complex finite element model based on shell elements and simple finite element model based on beam elements.Finally,the new gradient-based aerostructural multidisciplinary optimization design system is established,which can consider the influence of both static aeroelastic deformation and linear flutter characteristics.4.For typical aircraft with very high-aspect-ratio flexible wings like ‘Helios',the impact of distributions about wing geometrical twist and distributed engines on the induced drag and the wing linear flutter speed is researched from a multidisciplinary perspective.Additional,the principle of induced drag reduction is revealed and the traditional formula for estimating the induced drag is modified.Research results indicate that for flexible wings,in addition to changing the circulation distribution of wings to make it closer to the ideal elliptical circulation distribution,improve the aerodynamic efficiency to reduce the lift loss is another way to decrease induce darg.The negative geometrical twists at the wingtip as well as the outward movement of engines in spanwise direction can suppress excessive bending deformation of the wing and avoid the generation of large lateral forces that do not benefit to lift.As a result,the aerodynamic efficiency of the wing is improved.The maximum 14.5% induced drag reduction gain is obtained.Due to the effect of static aeroelastic deformation,the induced drag of flexible wings is no longer linearly related to the square of the lift coefficient,but is approximately linear with the ratio of the square of the aerodynamic load coefficient perpendicular to the wing to the actual aspect ratio of the wing.Fix the structural weight and stiffness property of wings,a reasonable position distribution of engines can not only reduce the induced drag but also improve the linear flutter speed of wings.Design results show that linear flutter speed is increased by 12.6%.5.For typical HALE UAVs like ‘Global Hawk',an aero-structural multidisciplinary optimization design research is performed.The aerodynamic shape and the thickness of structural finite element are selected as the design variables.The influence of different objective functions and disciplinary coupling relationships on design results is discussed.In contrast to the initial model,the lift-drag ratio of the multidisciplinary design result is improved by 4.9%,and the structure weight of the wing is reduced obviously.Taking the influence of flutter into account during optimization,compared with the multidisciplinary optimization design result considering the impact of static aeroelastic deformation single,the linear flutter critical speed is improved by 11.2% at the cost of 12.3% structural weight gain.Design results verify that the built gradient-based multidisciplinary optimization design method has the ability to deal with aero-structural multidisciplinary optimization design problems with large-scale design variables.
Keywords/Search Tags:Multidisciplinary optimization design, Adjoint method, Spectral method, Unsteady, Computational aeroelasticity, Flutter
PDF Full Text Request
Related items