Numerical Methods For Aerodynamic/Kinematic Coupling Problems Based On Dynamic Hybrid Grids

There are numerous flow-structure interaction phenomena in nature, such as the flight vehicle maneuvering, flutter and buffeting of wings, complex multi-body separation, as well as self-propelled swimming of fish. The main characteristics of these phenomena are the non-linear and unsteady effects within fluid-structure interaction. The aeroelastic problems are involved with the coupling of fluid dynamics and structural dynamics, while fluid dynamics, kinematics and kinetics have to be integrated for the problems such as flight vehicle maneuvering, multi-body separation and self-propelled swimming of fish. When the unsteady effects should be considered, the traditional quasi-steady methods are not suitable for these problems, and then the unsteady mutli-disciplinary coupling method is desired.The purpose of this thesis is to develop an integrated method, which couples the dynamic grid generation, unsteady flow simulation and computation of flight mechanics equations into a unified frame to simulate the flow/kinematics coupled problems. The future objective is to carry out the ?Virtual Flight Simulation‘ of realistic flight vehicle.There are six chapters in this thesis.Chapter 1 is the introduction. Firstly, the concept of the ?Virtual Flight Simulation‘ was introduced briefly to present the research motivation of aerodynamic/kinematic coupling simulation methods. Then the main progress in recent years was reviewed, including the dynamic grid generation techniques, unsteady numerical methods and mutli-disciplinary coupling simulations. Finally, the reaseach contents were listed.In Chapter 2, the dynamic grid generation techniques in this thesis were discussed. Based on the dynamic grid generation techniques developed by the author‘s research group, an approach based on radial basis functio ns(RBF) interpolation was achieved in this thesis. According to the greedy algorithms, a basis- function-set reduction approach based on maximum error position was presented to improve the efficiency of RBF method. Associating with the strategy of computational domain decomposition, the dynamic grid generation is more flexible for flow/kinematics coupled problems. The grid generation examples have demonstrated the efficiency and robustness for complex configurations with moving boundary.In Chapter 3, the unsteady flow solver was discussed. The Navier-Stokes equations based on Arbitrary Lagrangian-Eulerian(ALE) system were solved by second-order finite volume method. The dual- time stepping method coupled with the block Lower-Upper Symmetric Gauss-Seidel algorithm(BLU-SGS) were adopted for temporal advancing, and a unified temporal discretization scheme with varying time-steps were achieved in the frame of linear multistep method. For the unsteady simulation on moving grid, the Geometric Conservation Law(GCL) was studied in details. Based on the truncation error analysis, several previous geometric conservation algorithms were classified into two catogeries, i.e. the volume-constrained approaches and the surface-constrained ones. Some typical cases were tested to validate the unsteady solver and the therotical analysis on GCL.In Chapter 4, the flow/kinematics coupling algorithm was presented. Usually, the kinematics system is solved in the integral frame, but the flow system is solved discretely, so it is difficult to solve them coupling. Then we have to solve them in the decoupling algorithm. There are two approaches of decoupling algorithm, i.e. the ?loose coupling‘ and ?strong coupling‘. Both the loose and strong coupling approaches for different temporal accuracy were developed in a unified frame, and the stability of these approaches were analyzed and validated with typical test case.Chapter 5 is the preliminary application of the coupling algorithm, including two examples, i.e. the self-propelled swimming of a two-dimensional fish model, and the simple maneuvering of a morping aircraft. For the first example, the ?loose coupling‘ is unstable since the fish density is close to the density of water(the added- mass is large), then the 1-DOF self-propelled swimming of fish could only be simulated by tiny time-steps, and the 3-DOF simulation will divergent quickly; On the contrary, the ?strong coupling‘ is stable and doesn‘t suffer this problem. The second example, maneuvering of a morping aircraft, is to validate the ability of the completely integrated method discussed in this thesis. The preliminary designed configuration is unstable without control system, so the maneuvering simulated in this thesis is not the real flight condition, and still need further study.Chapter 6 gives the concluding remarks. The deficiencies were pointed out and the possible fields for future work were discussed.