This thesis proposes a large eddy simulation reduced order model (LES-ROM) framework for the numerical simulation of realistic flows. In this LES-ROM framework, the proper orthogonal decomposition (POD) is used to define the ROM basis and a POD differential filter is used to define the large ROM structures. An approximate deconvolution (AD) approach is used to solve the ROM closure problem and develop a new AD-ROM. This AD-ROM is tested in the numerical simulation of the one-dimensional Burgers equation with a small diffusion coefficient ( ν= 10⁻³). / Master of Science

Proper Orthogonal Decomposition (POD), combined with the Method of Snapshots and Galerkin projection, is a popular method for the model order reduction of nonlinear PDEs. The POD requires the left singular vectors from the singular value decomposition (SVD) of an n-by-m "snapshot matrix" S, each column of which represents the computed state of the system at a given time. However, the direct computation of this decomposition can be computationally expensive, particularly for snapshot matrices that are too large to fit in memory. Hierarchical Approximate POD (HAPOD) (Himpe 2016) is a recent method for the approximate truncated SVD that requires only a single pass over S, is easily parallelizable, and can be computationally cheaper than direct SVD, all while guaranteeing the requested accuracy for the resulting basis. This method processes the columns of S in blocks based on a predefined rooted tree of processors, concatenating the outputs from each stage to form the inputs for the next. However, depending on the selected parameter values and the properties of S, the performance of HAPOD may be no better than that of direct SVD. In this work, we numerically explore the parameter values and snapshot matrix properties for which HAPOD is computationally advantageous over the full SVD and compare its performance to that of a parallelized incremental SVD method (Brand 2002, Brand 2003, and Arrighi2015). In particular, in addition to the two major processor tree structures detailed in the initial publication of HAPOD (Himpe2016), we explore the viability of a new structure designed with an MPI implementation in mind. / Master of Science

In this thesis, I examine filtering based stabilization methods to design new regularized reduced order models (ROMs) for under-resolved simulations of unsteady, nonlinear, convection-dominated systems. The new ROMs proposed are variable delta filtering applied to the evolve-filter-relax ROM (V-EFR ROM), variable delta filtering applied to the Leray ROM, and approximate deconvolution Leray ROM (ADL-ROM). They are tested in the numerical setting of Burgers equation, a nonlinear, time dependent problem with one spatial dimension. Regularization is considered for the low viscosity, convection dominated setting. / Master of Science / Numerical solutions of partial differential equations may not be able to be efficiently computed in a way that fully captures the true behavior of the underlying model or differential equation, especially if significant changes in the solution to the differential equation occur over a very small spatial area. In this case, non-physical numerical artifacts may appear in the computed solution. We discuss methods of treating these calculations with a goal of improving the fidelity of numerical solutions with respect to the original model.

Rotordynamic instability problems in turbomachinery have become more important in recent years due to rotordynamic components with higher speeds and higher power densities. These features typically lead to increased instability risk in rotor dynamic components as fluids-structure interactions take place. In addition, critical damage of rotordynamic components can result from high level vibrations of supporting bearing system, where the reduced rotor speed can lead to system operating near the rotor critical speed. Therefore, increased accuracy in modeling of rotordynamic components is required to predict the potential instability issues in high performance rotordynamic design. The instability issue may potentially be eliminated in design stage by varying the characteristics of the unstable components. One such turbomachinery component is the annular pressure seal. The annular pressure seals are specifically designed to prevent the fluid leakage from high pressure stage to low pressure stage in turbomachinery. Typical annular pressure seals have two different flow regions, an annular jet-flow region between the rotor and stator, and cylindrical or circumferential indentions on the stator/rotor surface that serve as cavities where flow recirculation occurs. As the working fluid enters the cavities and recirculates, the kinetic energy is reduced, resulting in a reduction of leakage flow. The current challenge is to model with higher precision the interaction between the rotordynamic components and the working fluid. In this dissertation, this challenge was overcome by developing a hybrid Bulk Flow/CFD method to compute rotordynamic responses for the annular pressure seals. In addition, design of experiments studies were performed to relate the surface patterning with the resulting rotordynamic response for the annular pressure seals, in which several different geometry specifications were investigated. This study on annular pressure seal design generated regression models for rotordynamic coefficients that can be used as optimization guidelines. Research topics related to the annular pressure seals were presented in this dissertation as well. The reduced order model of both hole-pattern seals and labyrinth seals were investigated. The results showed that the flow field representing the flow dynamics in annular pressure seals can be expressed as a combination of first three proper orthogonal decomposition modes. In addition, supercritical state of carbon dioxide (sCO2) process fluid was examined as the working fluid in a preliminary study to better understand the effects on annular pressure seals. The results showed that the performance and stability in the annular pressure seals using sCO2 as process fluid can both be improved. / Doctor of Philosophy / This dissertation focused on understanding the correlations between surface patterning and rotordynamic responses in the annular pressure seals. The annular pressure seals are a specific type of rotordynamic component that was designed to prevent the fluid leakage from high pressure stage to low pressure stage in turbomachinery. As the working fluid enters the cavities and recirculates, the kinetic energy is reduced, resulting in a reduction of leakage flow through the annular pressure seals. Rotordynamic instability becomes an issue that may be related to the annular pressure seals in some cases. In recent years, rotordynamic components with higher rotor speeds and higher power densities are commonly used in industrial applications. These features could lead to increased instability risk in rotor-bearing systems as fluids-structure interactions take place. Therefore, high precision modeling of the rotodynamic components is required to predict the instability issues in high performance rotordynamic design. The instability issue may potentially be eliminated in design stage by varying the characteristics of the potentially unstable components. In this study, the surface patterning and rotordynamic responses were investigated for several different annular pressure seal models with a hybrid Bulk Flow/Computational Fluid Dynamics method. This dissertation provides for the first time regression models for rotordynamic coefficients that can be used as optimization guidelines. Research topics related to the annular pressure seals were presented in this dissertation as well. The reduced order model of both hole-pattern seals and labyrinth seals were investigated. The results showed that the flow field representing the flow dynamics in annular pressure seals can be expressed as a combination of first three proper orthogonal decomposition modes. In addition, supercritical state of carbon dioxide (sCO2) process fluid was examined to better understand the effects of working fluid on annular pressure seals. The results showed that the performance and stability in the annular pressure seals using sCO2 as process fluid can both be improved.

We investigate the effect of spatial filtering on the recently proposed data-driven correction reduced order model (DDC-ROM). We compare two filters: the ROM projection, which was originally used to develop the DDC-ROM, and the ROM differential filter, which uses a Helmholtz operator to attenuate the small scales in the input signal. We focus on the following questions: ``Do filtering and differentiation with respect to space variable commute, when filtering is applied to the diffusion term?'' or in other words ``Do we have commutation error (CE) in the diffusion term?" and ``If so, is the commutation error data-driven correction ROM (CE-DDC-ROM) more accurate than the original DDC-ROM?'' If the CE exists, the DDC-ROM has two different correction terms: one comes from the diffusion term and the other from the nonlinear convection term. We investigate the DDC-ROM and the CE-DDC-ROM equipped with the two ROM spatial filters in the numerical simulation of the Burgers equation with different diffusion coefficients and two different initial conditions (smooth and non-smooth). / M.S. / We propose reduced order models (ROMs) for an efficient and relatively accurate numerical simulation of nonlinear systems. We use the ROM projection and the ROM differential filters to construct a novel data-driven correction ROM (DDC-ROM). We show that the ROM spatial filtering and differentiation do not commute for the diffusion operator. Furthermore, we show that the resulting commutation error has an important effect on the ROM, especially for low viscosity values. As a mathematical model for our numerical study, we use the one-dimensional Burgers equations with smooth and non-smooth initial conditions.

An optimal solution of vortex-induced vibrations of structures would be a time-domain numerical simulation that simultaneously solves the fluid flow and structural response. Yet, the requirements in terms of computing power remains a major obstacle for implementing such a simulation. On the other hand, lower- or reduced-order models provide an alternative for determining structural response to forcing by fluid flow. The objective of this thesis is to provide a consistent approach for the development of reduced-order models for the lift and drag on oscillating cylinders and the identification of their parameters. Amplitudes and phases of higher-order spectral moments of the lift and drag coefficients data are combined with approximate solutions of the representative models to determine their parameters. The results show that the amplitude and phase of the trispectrum could be used to model the lift on the oscillating cylinder under different excitation conditions. Moreover, the amplitude and phase of the cross-bispectrum could be used to establish the lift-drag relation for oscillating cylinders. A forced van der Pol equation is used to represent the lift on a transversely oscillating cylinder, and a parametrically excited van der Pol equation is used to model the lift coefficient on an inline oscillating cylinder. All cases of excitations lead to close values for the damping and nonlinear parameters in the van der Pol equation. Consequently, and as shown in this thesis, different excitation cases could be used to identify the parameters in the governing equations. Moreover, the results show that the drag coefficient could be derived from the lift coefficient through a square relation that takes into account the effects of the forced motions. / Ph. D.

When designing a control for a physical system described by a PDE, it is often necessary to reduce the size of the controller for the PDE system. This is done so that real time control can be achieved. One approach often taken by engineers is to reduce the approximating finite-dimensional system using a balanced reduction method known as balanced truncation and then design a control for the lower order system. The unsettling idea about this method is that it involves discarding information and then designing a control. What if valuable physical information were lost that would have allowed a more effective control to be designed? This paper will explore an alternate balanced reduction method called LQG balancing. This approach allows for the designing of a control on the full order approximating system and then reducing the control. Along the way, the basic ideas of feedback control design will be discussed, including system balancing and model reduction. Following, there will be mention of the linear Klein-Gordon equation and the development of the one-dimensional finite element approximation of the PDE. Finally, simulations and numerical experiments are used to discuss the differences between the two balanced reduction methods. / Master of Science

This thesis uses Proper Orthogonal Decomposition to model parametric and coupled systems. First, Proper Orthogonal Decomposition and its properties are introduced as well as how to numerically compute the decomposition. Next, a test case was used to show how well POD can be used to simulate and control a system. Finally, techniques for modeling a parametric system over a given range and a coupled system split into subdomains were explored, as well as numerical results. / Master of Science

High-fidelity computer simulations of childbirth are time consuming, making them impractical for guiding decision-making during obstetric emergencies. The complex geometry, micro-structure, and large finite deformations undergone by the vagina during childbirth result in material and geometric nonlinearities, complicated boundary conditions, and nonhomogeneities within finite element (FE) simulations. Such nonlinearities pose a significant challenge for numerical solvers, increasing the computational time. Simplifying assumptions can reduce the computational time significantly, but this usually comes at the expense of simulation accuracy. The work herein proposed the use of reduced order modeling (ROM) techniques to create surrogate models that capture experimentally-measured displacement fields of rat vaginal tissue during inflation testing in order to attain both the accuracy of higher-fidelity models and the speed of lower-fidelity simulations. The proper orthogonal decomposition (POD) method was used to extract the significant information from FE simulations generated by varying the luminal pressure and the parameters that introduce the anisotropy in the selected constitutive model. In our first study, a new data-driven (DD) variational multiscale (VMS) ROM framework was extended to obtain the displacement fields of rat vaginal tissue subjected to ramping luminal pressure. For comparison purposes, we also investigated the classical Galerkin ROM (G-ROM). In our numerical study, both the G-ROM and the DD-VMS-ROM decreased the FE computational cost by orders of magnitude without a significant decrease in numerical accuracy. Furthermore, the DD-VMS-ROM improved the G-ROM accuracy at a modest computational overhead. Our numerical investigation showed that ROM had the potential to provide efficient and accurate computational tools to describe vaginal deformations, with the ultimate goal of improving maternal health. Our second study compared two common computational strategies for surrogate modeling, physics-based G-ROM and data-driven machine learning (ML), for decreasing the cost of FE simulations of the ex vivo deformations of rat vaginal tissue subjected to inflation testing to study the effect of a pre-imposed tear. Since there are many methods associated with each modeling approach, to provide a fair and natural comparison, we selected a basic model from each category. From the ROM strategies, we considered a simplified G-ROM that is based on the linearization of the underlying nonlinear FE equations. From the ML strategies, we selected a feed-forward dense neural network (DNN) to create mappings from constitutive model parameters and luminal pressure values to either the FE displacement history (in which case we denote the resulting model ML) or the POD coefficients of the displacement history (in which case we denote the resulting model POD-ML). The numerical comparisons of G-ROM, ML, and POD-ML took place in the reconstructive regime. The numerical results showed that the G-ROM outperformed the ML model in terms of offline central processing unit (CPU) time for model training, online CPU time required to generate approximations, and relative error with respect to the FE models. The POD-ML model improved on the speed performance of the ML, having online CPU times comparable to those of the G-ROM given the same size of POD bases. However, the POD-ML model did not improve on the error performance of the ML. In our last study, we expanded our investigation of ML methods for surrogate modeling by comparing the performance of a DNN similar to what was used previously to that of a convolutional neural network (CNN) using 1-D convolution on the input parameters from FE simulations of active vaginal tearing. The new FE simulations utilized a custom continuum damage model that provided material damage and failure properties to an existing anisotropic hyperelastic constitutive model to replicate experimentally-observed tear propagation behaviors. We employed our DNN and CNN models to create mappings from constitutive model parameters, geometric properties of the propagating tear, and luminal pressure values to either the full FE displacement history or the POD coefficients of the displacement history. The root-mean-square error (RMSE) with respect to the FE displacement history achieved by full order output ML predictions was reproducible with POD-ML using a basis of only dimension l=10. Additionally, an order of magnitude reduction in offline time was observed using POD-ML over full-order ML with minimal difference between DNN and CNN architectures. Differences in online computational costs between ML and POD-ML were found to be negligible, but the DNNs produced predictions slightly faster than the CNNs, though both online times were on the same order of magnitude. While convolution did not significantly aid the regression task at hand, POD-ML was demonstrated to be an efficient and effective approach for surrogate modeling of the FE tear propagation model, approximating the displacement history with RMSE less than 0.1 mm and generating results 7 orders of magnitude faster than the FE model. This set of baseline numerical investigations serves as a starting point for future computer simulations that consider state-of-the-art G-ROM and ML strategies, and the in vivo geometry, boundary conditions, material properties, and tissue damage mechanics of the human vagina, as well as their changes during labor. / Doctor of Philosophy / Computer simulations of childbirth are extremely time-consuming, making them impractical for guiding decision-making by obstetricians when a patient is entering labor. The complex geometry, material microstructure, and large deformations undergone by the vagina during childbirth result in material and geometric properties that are challenging to mathematically model. Consequently, numerical solver methods (e.g., finite elements) require large amounts of time to simulate childbirth. Simplifying assumptions can reduce computational time, but this simplification usually comes at the expense of simulation accuracy. The work of this dissertation proposes the use of several techniques to reduce model complexity and create accurate approximations and predictions of results from full-order models (FOMs) with profound reductions in computational time. Our first study used reduced order models (ROMs) to extract the significant information from a FOM of the rat vagina subjected to inflation. We compared a basic ROM and an advanced, data-driven ROM. Our second study compared the basic ROM to a basic machine learning (ML) technique for approximating a FOM that simulated inflation of the rat vagina with a pre-imposed tear. A hybrid technique incorporating elements of both ROM and ML to approximate FOM results was also considered. Our final study made use of ML and hybrid techniques using a more advanced neural network (a convolutional neural network). These ML models were used to predict the results of a FOM simulation of vaginal tear propagation. These numerical investigations serve as a starting point for future development of computer simulations using state-of-the-art ROM and ML strategies as well as more realistic models for the mechanics of the human vagina during childbirth.

Quadrupedal robots play a vital role in various applications, from search and rescue operations to exploration in challenging terrains. However, locomotion tasks involving unknown payload transportation on rough terrains pose significant challenges, requiring adaptive control strategies to ensure stability and performance. This dissertation contributes to the advancement of adaptive motion planning and control solutions that enable quadrupedal robots to traverse unknown rough environments while tasked with transporting unknown payloads. In the first project, a novel hierarchical planning and control framework for robust payload transportation by quadrupedal robots is developed. This framework integrates an adaptive model predictive control (AMPC) algorithm with a gradient-descent-based adaptive updating law applied to reduced-order locomotion (i.e., template) models. At the high level of the control hierarchy, an indirect adaptive law estimates unknown parameters of the reduced-order locomotion model under varying payloads, ensuring stability during trajectory planning. The optimal trajectories generated by the AMPC are then passed to a low-level and full-order nonlinear whole-body controller (WBC) for tracking. Extensive numerical investigations and hardware experiments on the A1 quadru[pedal robot validate the framework's capabilities, showcasing significant improvements in payload transportation on both flat and rough terrains compared to conventional MPC strategies. Specifically, the robot demonstrates proficiency in transporting unmodeled, unknown static payloads up to 109% of its own mass in experiments on flat terrains and 91% on rough experimental terrains. Moreover, the robot successfully manages dynamic payloads with 73% of its mass on rough terrains. Adaptive controllers must also address external disturbances inherent in real-world environments. Therefore, the second project introduces a hierarchical planning and control scheme with an adaptive L1 nonlinear model predictive control (ANMPC) at the high level, which integrates nonlinear MPC (NMPC) with an L1 adaptive controller. The prescribed optimal state and control input profiles generated by the ANMPC are then fed to the low-level nonlinear WBC. This approach aims to stabilize locomotion gaits in the presence of parametric uncertainties and external disturbances. The proposed controller is analyzed to accommodate uncertainties and external disturbances. Comprehensive numerical simulations and experimental validations on the A1 quadrupedal robot demonstrate its effectiveness on rough terrains. Numerical results suggest that ANMPC significantly improves the stability of the gaits in the presence of uncertainties and external disturbances compared to NMPC and AMPC. The robot can carry payloads up to 109% of its own mass on its trunk on flat and rough terrains. Simulation results show that the robot achieves a maximum payload capacity of 26.3 (kg), which is equivalent to 211% of its own mass on rough terrains with uncertainties and disturbances. / Doctor of Philosophy / In the rapidly advancing domain of robotics, there is a growing demand for intelligent robotic systems capable of adeptly addressing novel and unforeseen scenarios, such as uneven paths or external forces applied to the robots, like kicks and hits. This necessitates robots with the capability to handle diverse tasks with precision, particularly in the domains of object transportation and navigation through unknown terrains in applications such as search and rescue operations or cargo handling. This dissertation introduces innovative motion planning and control frameworks designed to imbue robots with adaptive capabilities, enabling them to adapt to real-world unanticipated scenarios and uncertainties during their movement, particularly when carrying unknown payloads. In the first project, a new framework is developed to enhance payload transportation by quadrupedal robots. This framework integrates an adaptive model predictive control (AMPC) algorithm with a gradient-descent-based adaptive updating law. Through extensive experiments and simulations, the framework shows remarkable improvements in payload transportation on both flat and rough terrains. The robot successfully transports payloads exceeding its own mass by up to 109% on flat terrains and 91% on rough terrains. Recognizing the need to address uncertainties in real-world environments, the second project introduces a hierarchical planning and control scheme with adaptive L1 nonlinear model predictive control (ANMPC). This approach stabilizes legged locomotion in the presence of uncertainties and disturbances. Results demonstrate that ANMPC significantly improves gait stability compared to existing methods. The robot achieves a payload capacity of up to 109% of its own mass on both experimental flat and rough terrains and reaches a maximum of 26.3 kg (around 212% of its own mass) on rough terrain simulations with uncertainties and disturbances.