Surrogate models of turbulent diffusive flames could play a strategic role in the design of liquid rocket engine combustion chambers. The present article introduces a method to obtain data-driven surrogate models for coaxial injectors, by leveraging an inductive transfer learning strategy over a U-Net with available multifidelity Large Eddy Simulations (LES) data. The resulting models preserve reasonable accuracy while reducing the offline computational cost of data-generation. First, a database of about 100 low-fidelity LES simulations of shear-coaxial injectors, operating with gaseous oxygen and gaseous methane as propellants, has been created. The design of experiments explores three variables: the chamber radius, the recess-length of the oxidizer post, and the mixture ratio. Subsequently, U-Nets were trained upon this dataset to provide reasonable approximations of the temporal-averaged two-dimensional flow field. Despite the fact that neural networks are efficient non-linear data emulators, in purely data-driven approaches their quality is directly impacted by the precision of the data they are trained upon. Thus, a high-fidelity (HF) dataset has been created, made of about 10 simulations, to a much greater cost per sample. The amalgamation of low and HF data during the the transfer-learning process enables the improvement of the surrogate model’s fidelity without excessive additional cost.

A liquefied natural gas (LNG) facility often incorporates replicate liquefaction trains. The performance of equivalent units across trains, designed using common numerical models, might be expected to be similar. In this article, we discuss statistical analysis of real plant data to validate this assumption. Analysis of operational data for end flash vessels from a pair of replicate trains at an LNG facility indicates that one train produces 2.8%–6.4% more end flash gas than the other. We then develop statistical models for train operation, facilitating reduced flaring and hence a reduction of up to 45% in CO2 equivalent flaring emissions, noting that flaring emissions for a typical LNG facility account for ~4%–8% of the overall facility emissions. We recommend that operational data-driven models be considered generally to improve the performance of LNG facilities and reduce their CO2 footprint, particularly when replica units are present.

Reduction in mobility due to gait impairment is a critical consequence of diseases affecting the neuromusculoskeletal system, making detecting anomalies in a person’s gait a key area of interest. This challenge is compounded by within-subject and between-subject variability, further emphasized in individuals with multiple sclerosis (MS), where gait patterns exhibit significant heterogeneity. This study introduces a novel perspective on modeling kinematic gait patterns, recognizing the inherent hierarchical structure of the data, which is gathered from contralateral limbs, individuals, and groups of individuals comprising a population, using wearable sensors. Rather than summarizing features, this approach models the entire gait cycle functionally, including its variation. A Hierarchical Variational Sparse Heteroscedastic Gaussian Process was used to model the shank angular velocity across 28 MS and 28 healthy individuals. The utility of this methodology was underscored by its granular analysis capabilities. This facilitated a range of quantifiable comparisons, spanning from group-level assessments to patient-specific analyses, addressing the complexity of pathological gait patterns and offering a robust methodology for kinematic pattern characterization for large datasets. The group-level analysis highlighted notable differences during the swing phase and towards the end of the stance phase, aligning with previously established literature findings. Moreover, the study identified the heteroscedastic gait pattern variability as a distinguishing feature of MS gait. Additionally, a novel approach for lower limb gait asymmetry quantification has been proposed. The use of probabilistic hierarchical modeling facilitated a better understanding of the impaired gait pattern, while also expressing potential for extrapolation to other pathological conditions affecting gait.