No abstract

The formation and accretion of ice on the leading edge of a wing can be detrimental to airplane performance. Complicating this reality is the fact that even a small amount of uncertainty in the shape of the accreted ice may result in a large amount of uncertainty in aerodynamic performance metrics (e.g., stall angle of attack). The main focus of this work concerns using the techniques of Polynomial Chaos Expansions (PCE) to quantify icing uncertainty much more quickly than traditional methods (e.g., Monte Carlo).

This paper focuses on the problem of quantifying the effects of model-structure uncertainty in the context of timeevolving dynamical systems. This is motivated by multi-model uncertainty in computer physics simulations: developers often make different modeling choices in numerical approximations and process simplifications, leading to different numerical codes that ostensibly represent the same underlying dynamics. We consider model-structure inference as a two-step methodology: the first step is to perform system identification on numerical codes for which it is possible to observe the full state; the second step is structural uncertainty quantification (UQ), in which the goal is to search candidate models "close" to the numerical code surrogates for those that best match a quantity-of-interest (QOI) from some empirical dataset. Specifically, we: (1) define a discrete, local representation of the structure of a partial differential equation, which we refer to as the "local dynamical operator" (LDO); (2) identify model structure non-intrusively from numerical code output; (3) non-intrusively construct a reduced order model (ROM) of the numerical model through POD-DEIM-Galerkin projection; (4) perturb the ROM dynamics to approximate the behavior of alternate model structures; and (5) apply Bayesian inference and energy conservation laws to calibrate a LDO to a given QOI. We demonstrate these techniques using the two-dimensional rotating shallow water (RSW) equations as an example system. operator (LDO), which is simply a functional relationship between spatially-local field values that approximates the discretized governing field dynamics at a spatial point. For example, if the governing equations are hyperbolic, then the LDO is a function that takes field values in a spatially-local neighborhood of a center point and outputs the field value at that center point, one time step forward in time. Note that there is an attractive consequence of our assumptions of locality and spatio-temporal invariance with respect to system identification: if we wish to infer a LDO from numerical/experimental data, access to the full global state vector is not strictly required. We may simply collect data from a subset of spatial points (together with the appropriate surrounding local neighborhoods). This is a notable advantage relative to a system identification technique that would require the full global state vector (e.g., POD).As all of the dynamics are encoded in the LDO, any structural uncertainties are as well. Furthermore, if we can design the LDO to be a weighted sum of different elementary functions of the local field values, then the relevant structural uncertainties manifest themselves as uncertainties in the values of the weights, which are simply parameters. This is a sketch of the process by which we convert structural uncertainties to parametric ones.Having formalized a means to parameterize model structure, all of the machinery of parametric uncertainty quantification (UQ) is available to study the structural uncert...

Like other experimental techniques, X-ray photon correlation spectroscopy is subject to various kinds of noise. Random and correlated fluctuations and heterogeneities can be present in a two-time correlation function and obscure the information about the intrinsic dynamics of a sample. Simultaneously addressing the disparate origins of noise in the experimental data is challenging. We propose a computational approach for improving the signal-to-noise ratio in two-time correlation functions that is based on convolutional neural network encoder–decoder (CNN-ED) models. Such models extract features from an image via convolutional layers, project them to a low dimensional space and then reconstruct a clean image from this reduced representation via transposed convolutional layers. Not only are ED models a general tool for random noise removal, but their application to low signal-to-noise data can enhance the data’s quantitative usage since they are able to learn the functional form of the signal. We demonstrate that the CNN-ED models trained on real-world experimental data help to effectively extract equilibrium dynamics’ parameters from two-time correlation functions, containing statistical noise and dynamic heterogeneities. Strategies for optimizing the models’ performance and their applicability limits are discussed.

The formation and accretion of ice on the leading edge of an airfoil can be detrimental to aerodynamic performance. Furthermore, the geometric shape of leading edge ice profiles can vary significantly depending on a wide range of physical parameters, which can translate into a wide variability in aerodynamic performance. The purpose of this work is to explore the variability in airfoil aerodynamic performance that results from variability in leading edge ice shape profile. First, we demonstrate how to identify a low-dimensional set of parameters that governs ice shape from a database of ice shapes using Proper Orthogonal Decomposition (POD). Then, we investigate the effects of uncertainty in the POD coefficients. This is done by building a global response surface surrogate using Polynomial Chaos Expansions (PCE). To construct this surrogate efficiently, we use adaptive sparse grid sampling of the POD parameter space. We then analyze the data from a statistical standpoint.