https://godot-bloggy.xyz/categories/research/ Recent content in Research on ｢ALL FICTION｣ Hugo -- gohugo.io en-uk Mon, 01 Jan 2024 00:00:00 +0800 https://godot-bloggy.xyz/project/aircraft-design/ Sun, 10 Jan 2021 00:00:00 +0800 https://godot-bloggy.xyz/project/aircraft-design/ This work involved performing multiphase computational fluid dynamics (CFD) analyses of supercavitating hydrofoils and hulls to compute their hydrodynamic coefficients. Multiphase computational fluid dynamics simulation of an amphibious aircraft hull The CFD sweeps were performed over speed and angle of attack to produce surrogate models for the hydrodynamic coefficients of a supercavitating hydrofoil, capturing the water volume fraction around the wetted section. Water volume fraction around a super-cavitating hydrofoil section Surrogate contours of hydrofoil force and moment coefficients These surrogates were used in a water-takeoff analysis that coupled the hydrofoil, hull, and aircraft aerodynamic loads, over which the hydrofoil incidence and span were optimized to minimise takeoff distance. https://godot-bloggy.xyz/project/flight-performance/ Tue, 01 Mar 2022 00:00:00 +0800 https://godot-bloggy.xyz/project/flight-performance/ Data-enhanced flight simulations with machine learning for performance analysis. Derived nonlinear formulations for incorporating airline data into a flight simulation framework. Performed analyses of short, medium, and long-haul flights with validation against airline data from Cathay Pacific Airways. Prepared visualization and reporting tools for post-processing. The framework simulates each flight phase — takeoff, climb, cruise, descent, and landing — and blends recorded Quick Access Recorder (QAR) data into the dynamics to correct fuel and time predictions against real operations. https://godot-bloggy.xyz/project/aeroacoustics/ Mon, 01 Jan 2024 00:00:00 +0800 https://godot-bloggy.xyz/project/aeroacoustics/ Shape optimization of airfoils with multi-fidelity CFD and noise prediction models. Implemented differentiable boundary layer data extraction from the CFD solvers ADflow and CMPLXFOIL. Developed components of semi-empirical noise models for constructing coupled analyses with MPhys. Currently researching multi-objective gradient-based constrained optimization. Boundary layer quantities are extracted differentiably from the flow field, validated against reference CFD and experiment, then fed into semi-empirical wall-pressure-spectrum models that predict trailing-edge noise. Flow field around the airfoil Boundary layer profile evaluation against reference data Wall pressure spectrum predictions Related publications Yuan Lyu, Arjit Seth, and Rhea P. https://godot-bloggy.xyz/project/aerodynamics-vii/ Thu, 01 Jun 2023 00:00:00 +0800 https://godot-bloggy.xyz/project/aerodynamics-vii/ Viscous-inviscid interacting (VII) boundary layer methods for aerodynamic analyses. Implemented a modular, strongly-coupled VII integral boundary layer method. Incorporated automatic differentiation for generalizations to coupled analyses (e.g. aero-structural). Researching generic optimization capabilities within various MDO architectures (MDF, SAND). The solver strongly couples an inviscid panel solution with the integral boundary layer equations over the upper surface, lower surface, and wake, in a single differentiable system, implemented in ViscFoil.jl (currently private). https://godot-bloggy.xyz/project/abstract-algebra/ Sun, 01 Jan 2023 00:00:00 +0800 https://godot-bloggy.xyz/project/abstract-algebra/ New constructs for mappings between group and graph theory. Investigated a new definition for certain types of endomorphisms of groups called “descending” endomorphisms. Implemented programs in SageMath to compute descending endomorphisms of groups for proof validation. Analyzed representations of descending endomorphisms in graph theory. A descending endomorphism $\delta$ of a group $G$ is one that descends to every quotient: for each normal subgroup $N \trianglelefteq G$, the quotient map $\varphi : G \to G/N$ satisfies $\varphi \circ \delta = \bar{\delta} \circ \varphi$ for a well-defined induced map $\bar{\delta}$ on $G/N$.