Browsing by Subject "morphing"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item A Morphing-based Approach for the Verification of Precipitation Forecasts(2014-11-11) Han, FanThis thesis described a morphing-based precipitation verification strategy inspired by Keil and Craig. This strategy is based on an optical flow algorithm to morph the image (field) of the forecast precipitation into an image that resembles the image (field) of the observed (analyzed) precipitation. This method treats the precipitation as a passive scalar and carries out the morphing by computing a vector field, called the optical flow, which is then used to advect the original forecast precipitation field. The information provided by the optical flow and the morphed image of the forecast precipitation field is used to define the measures of the displacement error and residual error. There are two novel aspects of our strategy. First, it imposes a constrain on the morphing process in order to prevent the over-convergence of pixels during morphing to a few locations of large errors. Second, it uses a new definition of the displacement error and provides a new interpretation of the other error terms. By applying the new morphing-based precipitation strategy to a schematic idealized example and a real hurricane example, we demonstrate that the constrain imposed largely reduces the risk of over-convergence and the error measures we derive from the morphing process accurately measure the corresponding error components.Item Discretization and Approximation Methods for Reinforcement Learning of Highly Reconfigurable Systems(2010-07-14) Lampton, Amanda K.There are a number of techniques that are used to solve reinforcement learning problems, but very few that have been developed for and tested on highly reconfigurable systems cast as reinforcement learning problems. Reconfigurable systems refers to a vehicle (air, ground, or water) or collection of vehicles that can change its geometrical features, i.e. shape or formation, to perform tasks that the vehicle could not otherwise accomplish. These systems tend to be optimized for several operating conditions, and then controllers are designed to reconfigure the system from one operating condition to another. Q-learning, an unsupervised episodic learning technique that solves the reinforcement learning problem, is an attractive control methodology for reconfigurable systems. It has been successfully applied to a myriad of control problems, and there are a number of variations that were developed to avoid or alleviate some limitations in earlier version of this approach. This dissertation describes the development of three modular enhancements to the Q-learning algorithm that solve some of the unique problems that arise when working with this class of systems, such as the complex interaction of reconfigurable parameters and computationally intensive models of the systems. A multi-resolution state-space discretization method is developed that adaptively rediscretizes the state-space by progressively finer grids around one or more distinct Regions Of Interest within the state or learning space. A genetic algorithm that autonomously selects the basis functions to be used in the approximation of the action-value function is applied periodically throughout the learning process. Policy comparison is added to monitor the state of the policy encoded in the action-value function to prevent unnecessary episodes at each level of discretization. This approach is validated on several problems including an inverted pendulum, reconfigurable airfoil, and reconfigurable wing. Results show that the multi-resolution state-space discretization method reduces the number of state-action pairs, often by an order of magnitude, required to achieve a specific goal and the policy comparison prevents unnecessary episodes once the policy has converged to a usable policy. Results also show that the genetic algorithm is a promising candidate for the selection of basis functions for function approximation of the action-value function.Item Six Degree of Freedom Morphing Aircraft Dynamical Model with Aerodynamics(2010-01-14) Niksch, AdamMorphing aircraft are envisioned to have multirole capability where the ability to change shape allows for adaptation to a changing mission environment. In order to calculate the properties of many wing configurations efficiently and rapidly, a model of a morphing aircraft is needed. This paper develops an aerodynamic model and a dynamic model of a morphing flying wing aircraft. The dynamic model includes realistic aerodynamic forces, consisting of lift, drag, and pitching moment about the leading edge, calculated using a constant strength source doublet panel method. The panel method allows for the calculation of aerodynamic forces due to large scale shape changing effects. The aerodynamic model allows for asymmetric configurations in order to generate rolling and yawing moments. The dynamic model calculates state information for the morphing wing based on the aerodynamic forces from the panel method. The model allows for multiple shape changing degrees-of-freedom for the wing, including thickness, sweep, dihedral angle, and chord length. Results show the model provides a versatile and computationally efficient tool for calculating the aerodynamic forces on the morphing aircraft and using these forces to show the associated states.Item Structural and Aerodynamic Interaction Computational Tool for Highly Reconfigurable Wings(2011-10-21) Eisenbeis, Brian JosephMorphing air vehicles enable more efficient and capable multi-role aircraft by adapting their shape to reach an ideal configuration in an ever-changing environment. Morphing capability is envisioned to have a profound impact on the future of the aerospace industry, and a reconfigurable wing is a significant element of a morphing aircraft. This thesis develops two tools for analyzing wing configurations with multiple geometric degrees-of-freedom: the structural tool and the aerodynamic and structural interaction tool. Linear Space Frame Finite Element Analysis with Euler-Bernoulli beam theory is used to develop the structural analysis morphing tool for modeling a given wing structure with variable geometric parameters including wing span, aspect ratio, sweep angle, dihedral angle, chord length, thickness, incidence angle, and twist angle. The structural tool is validated with linear Euler-Bernoulli beam models using a commercial finite element software program, and the tool is shown to match within 1% compared to all test cases. The verification of the structural tool uses linear and nonlinear Timoshenko beam models, 3D brick element wing models at various sweep angles, and a complex wing structural model of an existing aircraft. The beam model verification demonstrated the tool matches the Timoshenko models within 3%, but the comparisons to complex wing models show the limitations of modeling a wing structure using beam elements. The aerodynamic and structural interaction tool is developed to integrate a constant strength source doublet panel method aerodynamic tool, developed externally to this work, with the structural tool. The load results provided by the aerodynamic tool are used as inputs to the structural tool, giving a quasi-static aeroelastically deflected wing shape. An iterative version of the interaction tool uses the deflected wing shape results from the structural tool as new inputs for the aerodynamic tool in order to investigate the geometric convergence of an aeroelastically deflected wing shape. The findings presented in this thesis show that geometric convergence of the deflected wing shape is not attained using the chosen iterative method, but other potential methods are proposed for future work. The tools presented in the thesis are capable of modeling a wide range of wing configurations, and they may ultimately be utilized by Machine Learning algorithms to learn the ideal wing configuration for given flight conditions and develop control laws for a flyable morphing air vehicle.