Research Classification
Research Interests
Relevant Thesis-Based Degree Programs
Affiliations to Research Centres, Institutes & Clusters
Graduate Student Supervision
Doctoral Student Supervision
Dissertations completed in 2010 or later are listed below. Please note that there is a 6-12 month delay to add the latest dissertations.
A fully Eulerian phase-field framework for contact in multiphase fluid–structure interaction (2025)
Multiphase fluid-structure interaction (FSI), defined as the interaction of multiple fluid phases with several structural components, is a fundamental physical phenomenon observed in applications ranging from natural processes to engineering systems. One particularly important and timely example is the interaction between ships and sea ice in the Arctic region. As the polar ice caps continue to melt due to climate change, the Arctic is becoming increasingly navigable, spurring interest in marine transportation, resource extraction, and scientific exploration in the region. Understanding this complex system of ship-ice interaction involves modeling the multiphase coupling between ship, ice, water, and air and capturing multiscale effects such as drifting and rotation of ice floes, free-surface phenomena, and contact dynamics. The development of a generalized FSI-contact framework to capture such dynamic interactions is the focus of the present dissertation.In the current work, we employ a fully Eulerian approach to model the evolution and interaction of the different phases. We use the phase-field model to capture the interfaces between the different physical systems, which naturally allow for large displacements and/or deformations of the solid structures. The strains and stresses in the solid bodies are modeled by evolving the left Cauchy-Green deformation tensor. To address the challenges associated with interface smearing in convection dominated regimes, an interface-preserving adaptivity technique is proposed to minimize the losses in interface accuracy while simultaneously reducing the computational cost. Two distinct contact models are developed to simulate the different interactions inherent in the coupled ship-ice dynamics. To model ice-ice contact, we propose a novel monofield interface advancing scheme for collision detection that is capable of efficiently modeling contact between multiple submerged solids with identical physical properties. To model ship-ice contact, we develop a 3D sliding contact formulation based on the overlap of the diffuse interfaces of the respective colliding solids. We verify the solver after each new addition with relevant test cases and benchmark problems. We also demonstrate the efficacy and robustness of the framework by analyzing quantities of interest in each case. Finally, we present a simplified ship-ice interaction problem by passing a representative container ship through floating ice floes.
View record
A graph neural network framework for simulating unsteady fluid flow and fluid-structure interaction (2025)
Modern iterative design optimization and real-time active control of marine vessels require fast modeling of the fluid flow around its various sub-systems. Traditional numerical simulations, however, can be too slow for such purposes. In this dissertation, we provide a comprehensive hypergraph neural network framework that can be applied to fast surrogate modeling of various fluid and fluid-structure systems around marine vessels. Revolving around this target, this dissertation can be split into two parts. In the first part, we develop a hypergraph neural network architecture that serves as the backbone of the framework. Inspired by the data connectivity in the finite element method, we construct a hypergraph by connecting the nodes by elements. A hypergraph message-passing network that mimics the calculation process of local stiffness matrices is defined on such a node-element hypergraph. We verify the efficiency of the network on fluid flow around fixed body benchmark data sets, and compare its performance with baseline model MeshGraphNet.In the second part, we equip the framework with additional components that enable it to model various fluid and fluid-structure systems of different sizes. To model fluid flow around vibrating and deforming bodies, we embrace an arbitrary Lagrangian-Eulerian formulation and use a sub-network to model the mesh and solid movements. To model fluid flow around rotating structures, we employ a co-rotating domain under the inspiration of the sliding mesh method. We further enforce a series of geometric and physical priors in the framework to enhance its generalization capability. We also design partitioning and buffering schemes that enable training and inference for large three-dimensional cases on single and distributed machines. The completed framework is tested on a series of benchmark problems involving both periodic and chaotic fluid systems and fluid-structure systems in both two-dimensional and three-dimensional setups, and demonstrates ability to generate stabilized and accurate roll-out predictions over a long time horizon. It is expected that the framework proposed in this dissertation can serve as the physics simulation component in a digital twin framework used in iterative design optimization or real-time active control of marine vessels.
View record
Physics-Guided Deep Learning for Dynamical Systems: Applications to Fluid Flow and Ocean Acoustics (2025)
Real-time prediction of complex dynamical systems is critical in many scientific and engineering domains, where traditional numerical solvers can be computationally expensive and limited in generalization across changing environments. This dissertation presents a physics-guided deep learning framework for reduced-order modeling, generalization, and uncertainty quantification in nonlinear dynamical systems, with particular focus on underwater acoustics as a representative application.The dissertation is structured around five core contributions. First, we develop a reduced-order model based on deep learning that combines convolutional encoders with an attention-based recurrent neural network to learn latent dynamics from high-fidelity simulations. Second, we introduce a multistep integration-inspired attention mechanism that connects to and generalizes linear multistep methods for learning latent space dynamics, improving numerical stability and interpretability. Third, we propose a space-time coupled deep learning model based on 3D convolutional neural networks, enabling simultaneous learning of spatial and temporal correlations in unsteady flow data.Fourth, we address the challenge of domain generalization by developing a range-dependent conditional convolutional neural network with a continual learning framework. This allows adaptation across varying underwater ocean environments without performance degradation. Finally, we present a sparse variational Gaussian process model for uncertainty-aware three-dimensional acoustic field prediction in real time.These methods are applied to benchmark dynamical systems as well as realistic underwater acoustic scenarios, including transmission loss prediction over range-dependent ocean bathymetry. The results demonstrate high predictive accuracy, generalizability and computational efficiency, which support the deployment of digital twins in marine applications.This dissertation advances the development of physics-guided learning architectures for dynamical systems, offering new tools for real-time prediction, interpretability, and domain adaptation. The underwater acoustics is considered as a representative use case, highlighting the broader potential of these methods in science and engineering.
View record
A unified Eulerian variational framework for multiphase fluid-structure interaction (2023)
Multiphase fluid-structure interaction (FSI) involving multiphase flow and contact between immersed solids is omnipresent in numerous processes in nature, biology, and engineering applications. Examples include bio-inspired avian-aquatic vehicles, aneurysm and cardiovascular diseases in biomedical engineering, and marine vessels in ocean engineering. Of particular interest to the present study is the ice-going ships in the Arctic environment. Numerical simulation of this complex system involves modeling dynamics of disparate materials, evolving multiphase interfaces, and collisions between solids. The development of a novel three-dimensional multiphase and multiphysics computational framework based on unified continuum mechanics laws is the focus of the present dissertation.In the proposed numerical framework, we employ a fully Eulerian description for the continua of different phases, which facilitates topological changes of their interfaces during the evolution and contact processes. The interfaces and phase components are captured by the phase-field-based diffuse interface description. While the diffuse interface description circumvents the complexity of explicit interface reconstruction, it poses challenges in consistent interface transition and accurate geometric representation. To address these challenges, we developed an interface and geometry preserving phase-field method, which is the key contribution of this dissertation and lays the foundation for the success of the current framework in handling multiphase interfaces. With the phase components, we unify the mass and momentum conservations by phase-dependent interpolation. The kinematics of solid phases in an Eulerian frame of reference is resolved by evolving the left Cauchy-Green tensor.The unified Eulerian framework for two-phase and multiphase FSI is implemented in a partitioned-block iterative manner within the in-house 3D parallel variational multiphysics solver. The solver is systematically explored for a variety of cases of two-phase flow with surface tension effect, single-phase FSI, multiphase FSI, and contact of immersed deformable solids. The study is concluded by a 3D demonstration of ice-going ships sailing across floating ice floes. The unified Eulerian variational framework with a parallel implementation based on the novel interface and geometry preserving phase-field method provides a general and robust approach for investigating a wide range of multiphase FSI problems.
View record
Master's Student Supervision
Theses completed in 2010 or later are listed below. Please note that there is a 6-12 month delay to add the latest theses.
Fluid-structure interaction modeling of ice-going ships using a multiphase fully Eulerian framework (2026)
In the current work, we present fully coupled 3D numerical models of a hierarchy of problems pertaining to Arctic shipping, predicting hydrodynamic resistance on marine structures interacting with ice. A previously developed fully Eulerian phase-field framework is utilized to model the presented problems, spanning local to large scale ice-structure interactions. The present work establishes the versatility of the framework for ship-ice interaction problems, taking a step toward guiding safe and efficient navigation in ice-covered waters.The first part of the study assesses the framework to accurately predict resistance on a ship model in 3D with free surface effects, by tuning numerical parameters. The section begins by verifying the force computation routine and numerical setup with the framework for flow past a stationary cylinder. Resistance on a bluff body and on a container ship, partially submerged in water, are compared with analogous experimental results. We identify coupled requirements for both the mesh and diffuse interface resolutions to accurately predict resistance by adequately capturing complex fluid-structure interaction (FSI) and free surface effects.In the second part of the study, we present hydrodynamically mediated ice-structure interactions of two different regimes, highlighting the versatility of the framework. We first model the force induced on a marine structure, such as a thruster or a bulbous bow, when it collides with a heavy ice block under water. The peak force and momentum transfer during collision compare reasonably well with an experiment conducted using real sea ice. We utilize the setup to investigate the effect of sea ice shear modulus and collision speed on the collision force and impulse. Finally, we present our technique using the framework to model continuous ship interaction with a drifting ice floe field, mimicking long range transit of an ice-going vessel. The setup captures the rich interactions that are fully coupled between ship, ice, water and air, with implications on the stress distribution on the hull. We examine the effect of ice field concentration and flow speed on the characteristics of the temporally oscillatory resistance, highlighting their importance in constraining operability margins in a given ice field.
View record
Hydroelastic vibrations and acoustic radiation from ship hull structures near the free surface (2026)
Underwater radiated noise from ship hull structures operating near the free surface represents an important and unresolved problem in naval engineering, driven by growing environmental concern and evolving regulatory requirements. Wave-induced structural excitation is a recognized noise source, yet the coupled mechanisms by which hydrodynamic loading, hydroelastic vibration, and free-surface acoustic interference jointly govern radiated noise are not fully characterized, a gap that has prevented the development of physics-informed noise-reduction strategies.This thesis establishes the hydroelastic mechanisms of underwater noise generation from ship hull panels near the free surface and demonstrates how these mechanisms can be exploited in structural design optimization. A partitioned fluid-structure interaction solver coupled to a Ffowcs-Williams-Hawkings acoustic analogy with image-source free-surface correction decomposes radiated noise into loading-controlled dipole and deformation-controlled monopole contributions. Immersion depth governs modal energy distribution, hydroelastic frequency compression, and the transition from narrowband to broadband acoustic behaviour, with free-surface interference suppressing near-field radiation by up to 18 dB at shallow immersion.A surrogate-assisted multi-objective optimization framework trades structural mass against radiated noise for stiffened panels. Applied to a stiffened square panel and a representative ORCA-class curved hull segment, optimized designs achieve up to 41 dB and 13 dB of sound exposure level reduction, respectively, with Pareto front analysis identifying a practical mass budget beyond which acoustic return per unit mass collapses.The deformation-controlled monopole regime in the 20-250 Hz band dominates biologically weighted acoustic exposure for three of four marine mammal functional hearing groups by 30 to 35 dB over the loading-controlled dipole regime, and targeted stiffener placement informed by the immersion-dependent modal activation sequence is both necessary and sufficient to achieve meaningful noise reduction. Immersion ratio emerges as the primary design axis: it governs modal participation, free-surface suppression, and the priority order of structural interventions simultaneously. This thesis provides the physical characterization, computational framework, and design methodology needed to treat underwater radiated noise from near-waterline hull panels as a predictive, physics-informed structural design problem at the concept stage.
View record
A decision support system for minimizing underwater radiated noise from ships (2025)
Underwater radiated noise from ships threatens marine mammals, which rely on sound for navigation, foraging, and communication. Noise levels depend on vessel-specific characteristics and operational parameters. While research on voyage optimization of operational parameters prioritizes fuel efficiency, emissions, and safety, acoustic impact is often overlooked. This thesis introduces a novel decision support system for voyage planning to minimize a vessel’s acoustic footprint.The first contribution is a multi-objective optimization framework for fixed-path voyage scheduling, integrating two competing objectives: minimizing noise levels and fuel consumption. The problem, constrained by voyage parameters, is solved using a non-dominated sorting genetic algorithm. A two-dimensional ocean acoustic environment, incorporating marine mammals from diverse audiogram groups and realistic oceanographic conditions, is simulated. Effectiveness is demonstrated through real-world case studies of a large container vessel.The second contribution extends the framework to dynamic route planning and speed optimization, allowing ships to adapt trajectories while minimizing their acoustic footprint. This approach comprises three components:• Modeling: Near-field noise levels are estimated using a regression-based reference spectrum model (JOMOPANS-ECHO), while far-field propagation losses are computed with a Gaussian radial basis function model for real-time 3D underwater noise modeling. A data-informed density distribution of Southern Resident killer whales models environmental interaction.• Optimization: Route planning is performed using the Batch Informed Trees algorithm, integrating graph-based and sample-based methods. Speed optimization uses genetic algorithms to ensure noise-aware navigation under voyage constraints.• Simulation: A ROS-based simulation models adaptive ship-mammal interactions in a realistic oceanic setting with 3D visualization. To evaluate the proposed system, real-world case studies are simulated using AIS data from vessels operating between the Strait of Georgia and the Strait of Juan de Fuca. A comparative analysis of noise exposure levels under optimized and unoptimized voyage conditions is conducted, demonstrating the practical applicability and effectiveness of the proposed system in mitigating noise impact on marine ecosystems.
View record
Tip vortex cavitation in marine propellers: mechanism and morphing concept (2025)
Tip vortex cavitation (TVC) is a complex phenomenon that occurs during the operation of marine propellers, contributing to the underwater radiated noise (URN) of marine vessels. The urgent need for alleviation of the URN levels of marine vessels due to its detrimental impact on marine ecosystems necessitates a thorough investigation of this phenomenon and developmentof mitigation strategies.In this thesis, due to the challenges associated with the numerical simulation of this phenomenon, a comprehensive validation of large eddy simulation of tip vortex cavitating flow past a hydrofoil is conducted, followed by a detailed analysis of the oscillatory dynamics of TVC and its contributionto the near-field pressure fluctuations.In the next stage, geometric modifications of bending and twisting types are proposed and studied as a strategy for the mitigation of tip vortex cavitation. An investigation of modified hydrofoils in various bending and twisting configurations is conducted demonstrating the potential of these strategies for TVC mitigation without adversely affecting the hydrodynamic performance.A detailed analysis of the flow field in various modified configurations is carried out to reveal the mechanisms involved in the TVC mitigation achieved through these geometric modifications.The thesis is concluded with the development of a toolbox for the application of chord-normal bending and twisting modifications to marine propellers. Preliminary simulations of flow past a conventional and a modified propeller are conducted, which demonstrate that the TVC mitigation strategies proposed in the present work are promising methods for reducing the footprints of human activities on the ocean soundscape.
View record
A computational framework for flow-induced vibrations of propeller blades in cavitating flows (2024)
In the current work, we present a finite-element computational framework for Fluid-StructureInteraction problems subjected to the coupling of unsteady cavitating flows with flexible structures.We build upon a previously developed stabilized variational framework for multiphase FSI byincorporating a dynamical structural solver based on the modal decomposition of the structure.In the first part of the work, we present the validation of the current framework for conductingLarge Eddy Simulations (LES) of cavitating flows past rigid structures. We identify the resolutioncriteria for dynamic subgrid-scale LES based on the re-entrant jet momentum - a prominent phe-nomenon associated with cavity-shedding in the wake of immersed bodies. The validated frameworkis then used to elucidate the features of cavitating flow past a rigid hydrofoil. In particular, we iden-tify features of sheet-cavitating flow which enable the transition to cloud cavitation. Further, weevaluate the instabilities driving sheet-cavity breakdown and establish the vortical structures whichdrive cloud cavity collapse, and quantify the frequencies observed over the course of a cavitationcycle.In the second part of the work, we present the validation of the framework for LES of cavitatingflows past flexible structures. Based on the validation study conducted over a flexible NACA66rectangular hydrofoil, we elucidate the role of cavity and vortex shedding in the structural dynamicsat three different cavitation numbers. We identify a broad spectrum frequency band whose centralpeak does not correlate to the frequency content of the cavitation dynamics or the natural fre-quencies of the structure, indicating the induction of unsteady flow patterns around the hydrofoil.Finally, we discuss the coupled fluid-structure dynamics during a cavitation cycle associated withthe promotion and mitigation of cavitation.
View record
A unified variational framework and applications for flow-induced vibration in cavitating flows (2022)
In this work, we present the development of a new computational framework based on the stabilized variational finite element methods for unsteady cavitating flows and application to flow-induced vibrations of freely oscillating hydrofoils. The ultimate goal is to build a robust and accurate high-fidelity framework for the computational study of the coupled multiphase fluid-structure dynamics and noise reduction of marine propellers. The first part of this work involves a delineation of the systematic development and testing of the new computational framework. We propose novel linearizations of the governing cavitation partial differential equations (PDEs) for numerical modeling. Numerical challenges arising due to the particular characteristics of two-phase cavitating flows and fluid-structure interaction are addressed. We demonstrate the ability of the numerical implementation to accurately capture prominent features of cavitating flows such as bubble collapse, steady and unsteady partial cavitation, cavity shedding, and re-entrant jet formation. The second part focuses on the application of the developed computational framework for flow-induced vibrations of freely oscillating hydrofoils with unsteady partial cavitating conditions. The interaction dynamics of the hydrofoil with the fluid forces are represented as an elastically mounted rigid body. A frequency lock-in mechanism of the unsteady cavity and vortex shedding to a sub-harmonic of the structural natural frequency is observed to sustain high-amplitude transverse oscillations of the hydrofoil. This exploratory work paves the way for the coupled multiphase hydroelastic interaction of propellers with the target of noise mitigation by active or passive control mechanisms.
View record
Deep learning-based reduced order modeling for unsteady flow dynamics and fluid-structure interaction (2022)
This work presents data-driven predictions of nonlinear dynamical systems involving unsteady flow and fluid-structure interaction. Of particular interest is to develop a new simulation framework integrating high-fidelity models with deep learning towards Digital Twin. The final goal is to learn and predict the coupled dynamics via the digital twin of ship vessels and propellers. End-to-end deep learning-based reduced order models (DL-ROMs) are presented for digital twin development.The first part of this study develops an overall framework for DL-ROMs. The emphasis is to investigate the predictive performance of the hybrid DL-ROMs, which vary in obtaining the low-dimensional features, i.e., proper orthogonal decomposition (POD) and convolutional autoencoders. The low-dimensional features are evolved in time using recurrent neural networks (RNNs). This leads to the formulation of two DL-ROM frameworks: the POD-RNN and the convolutional recurrent autoencoder network (CRAN). To assess data-driven predictions, POD-RNN and CRAN are applied to predict unsteady flows and instantaneous forces for flow past static bluff bodies. We perform flow prediction analysis for a configuration of side-by-side cylinders with wake interference. For systems with moving interfaces and three-dimensional (3D) geometries, we develop modular DL-ROM techniques.The second part of this study includes model reduction strategies to predict vortex-induced vibration and 3D unsteady flows. The knowledge gained in the previous parts is utilized to develop partitioned and scalable DL-ROMs for unsteady flows with moving interfaces and parametric effects. We first develop a partitioned DL-ROM framework for fluid-structure interaction. The novel multi-level DL-ROM combines the effect of POD-RNN and CRAN by modular learning of two physical fields independently. While POD-RNN provides extraction of the fluid-structure interface, the CRAN enables the prediction of flow fields. For time series prediction of 3D flows, we present a 3D CRAN-based framework for predicting the fluid forces and vortex shedding patterns. We provide an assessment of improving learning capabilities using transfer learning for complex 3D flows with variable Reynolds numbers. The simplicity and computational efficiency of the proposed DL-ROMs allow investigation for various geometries and physical parameters. This research opens ways for digital twin development for near real-time prediction of unsteady flows and fluid-structure interaction.
View record
Flow-induced vibration of flexible cantilever cylinders at low Reynolds number (2022)
Whiskers in some mammals, such as rats and seals, have a mysterious level of sensing ability. A whisker interacting with the fluid flow can sense minuscule aero/hydrodynamic information and turn this information into an understanding of the environment. Our present work investigates the fluid-structure interaction of a flexible cantilever cylinder, as a canonical model of a whisker, to help understand how a rat or a seal whisker vibrates in response to low-speed air or water flow. We employ a fully-coupled fluid-structure solver based on the three-dimensional Navier-Stokes and structural equations to examine the dynamics of the cylinder. Of particular interest is to explore the possibility of flow-induced vibrations at laminar subcritical Reynolds numbers, where no periodic vortex shedding pattern is present. We show that the flexible cantilever cylinder could undergo sustained oscillations in this Reynolds regime when certain conditions are satisfied. The vibration frequencies are shown to match the cylinder's first- or second-mode natural frequency. The range of the frequency match, known as the lock-in regime, is found to have a strong dependence on the Reynolds number and mass ratio. Unlike the steady wake behind a stationary rigid cylinder, the wake of the flexible cantilever cylinder in the water flow is shown to become unstable at Reynolds numbers as low as 22 for a particular range of system parameters. We find that the cylinder could also experience sustained oscillations when positioned in the wake of a rigid stationary cylinder in a tandem configuration. For the cylinder in airflow, we show that a wavy pattern in the shear layer is the dominant feature of the wake. These findings provide a unified understanding of the flow-induced vibration phenomenon in flexible cantilever cylinders and lay the foundation for designing novel flow-measurement sensors for the next-generation underwater and aerial vehicles.
View record
System identification and deep learning for stability analysis of flow-induced vibration (2021)
In this work, we present the coupled dynamics and stability predictions of marine vessels in the ocean environment, with particular focus on the synergy of physics-based and data-driven models towards Digital Twin. The ultimate goal is to predict and control the coupled dynamics and stability in normal and extreme conditions via the digital twin of marine vessels and propellers.The first part of this study includes a high-dimensional representation of multiphase fluid-structure interaction via the nonlinear system of partial differential equations. The second part of this study includes the model reduction of flow-induced vibrations (FIV) and the application of the knowledge gained in the previous parts in efficient parametric design optimization and control of marine tugboats. Towards this goal, two advanced physics-based system identification approaches are considered via projection-based and deep-learning-based reduced-order models. The projection-based approach includes a linear reduced-order model (ROM) for stability prediction using the eigensystem realization algorithm (ERA), which provides a low-order approximation of unsteady flow dynamics in the neighbourhood of equilibrium steady state. We perform a systematic ROM-based stability analysis to understand the frequency lock-in mechanism and self-sustained FIV phenomenon by examining eigenvalue trajectories.However, for high Reynolds number flows and near real-time feedback control, this goal can only be achieved through the recent advances in nonlinear model reduction and deep learning (DL) algorithms. To demonstrate this idea, we have developed a data-driven coupling for predicting unsteady forces and vortex-induced vibration (VIV) lock-in by using a long short-term memory network (LSTM) as a DL-based ROM technique. The structure of the LSTM has the format of a nonlinear state-space model (NLSS) and provides a nonlinear mapping of input-output dynamics that can potentially predict the dynamics for a longer horizon utilized for the stability predictions. The simplicity and computational efficiency of the proposed ROMs allow investigation of the FIV mechanism for a variety of geometries and parameters, and open ways for the development of control devices and on-board and in real-time predictions.
View record
If this is your researcher profile you can log in to the Faculty & Staff portal to update your details and provide recruitment preferences.
Membership Status
Program Affiliations
Academic Unit(s)
