Yang Wang

Affordable large-scale and long-scale Grid-Scale Electrical and Thermal Energy Storage

• Compressed Air Energy Storage (CAES) is the cheapest large-scale energy storage technology.
• Thermal Energy Storage (TES) integrated adiabatic CAES (A-CAES) creates the new generation carbon-free storage technology.

• The developed CAES-TES toolbox enables dynamic modeling of all multi-physical components, and track the time-variant system responses arising from various mechanical (compressors and expanders), thermal (heat exchangers and thermal reservoirs), and electrical (motor and generator) components.

• PBTES tank with phase change material (PCM) fillings
• The time-resolved simulation indicated a higher steady-state cycle efficiency of 56.5% for the system with the PCM-filled PBTES, versus 53.2% of the system with the rock-filled PBTES.

Features: large-scale energy storage, fluid/thermal dynamics, dynamic modeling/analysis,multi-physics, toolbox development

Numerical Modelling in Power Plant System

• The liquid and gas phases of water disappear during ultra supercritical (USC) steam generation, which results in a more efficient steam cycle. USC power plant is friendly to the environment with less fuel/reagent consumption, less solid waste and water use.
• Circulating fluidized bed (CFB) boiler significantly increases the fuel efficiency of a power plant to generate electricity.
• The power plant system can be decoupled as the combustion system of the boiler and water-steam system based on Rankine cycle of other downstream components.

• To optimize and control the CFB, a dynamic CFB model is developed. The CFB modeling is built on a multi-scale framework by simulating the internal flow and temperature fields of every sub-system (combustor, water-wall, etc.), and assembling the multi-physics to the system-level power plant performance matrix.
• This Zone-based model is used to optimize process operation strategies for higher efficiencies and better fuel flexibility.
• Convection: 1) gas-wall (furnace side); 2) wall-water (water side)
• Radiation: gas-gas and gas-wall (inside boiler)

Features: multi-scale modeling, multi-phase flow, combustion kinetics, heat transfer, energy conservation

Membrane processes in water desalination and osmotic energy harvest

• Water desalination creates an alternative drinking water source to solve the global water challenges, but its high energy consumption leads to a high cost of drinking water production.
• Reverse osmosis (RO) is the most utilized technology at the moment around the world, forward osmosis (FO) and pressure retarded osmosis (PRO) are promising technologies to significantly reduce the energy consumption in desalination for a lower water cost.

Features: clean/green energy technology, sustainable energy system, energy storage

Modeling in Membrane Process

The performance of a membrane module is determined by the complex flow inside the membrane flow channels and the mass transfers across the membranes. The coupled effects build the difficulty to either predict the performance or optimize the desalination system. What my research focus on include both system-level optimization and the detailed flow dynamics and mass transfer inside the module.

1. Membrane module/stack model

• Membrane module/stack models are powerful tools to optimize the mechanical design of the module/stack and the system.
• The coupled effects between the nonlinear feed flow and the mass transfer of both water and salts across the membrane are considered in a coupled CFD-based modeling. These models that are validated with experimental data.
• Concentration polarisation (CP) effects negatively the mass transport across the membrane, which can be diminished by reducing the thickness of the boundary layer.

2. Computational fluid dynamics (CFD) modelling of spacer-promoted channel flow

I developed a CFD spacer-promoted flow model in OpenFOAM and in-house code GPDE, and verified the codes with experimental data. The developed CFD model will precisely simulate the complex flow and mass transfer coupling to improve the scaled-up system model and the understanding of the scaled-down micro-channel flow dynamics.

• Spacers/filaments improve the permeation by interrupting the boundary layer and enhancing the flow mixing, but also result in additional pressure loss.
• Solute concentration (passive scalar in the flow field) is balanced by its convection, diffusion and source.

3. Gradient-based optimization of spacer shape

Current optimisation methods are mainly based on try-and-error. I developed a new framework of gradient-based shape optimization of spacer design loop via using discrete adjoint approach.

• The trade-off between pressure loss and water permeation (indicated in the sensitivity distribution) increase challenges of the spacer design in membrane channels.
• The discrete adjoint method (for sensitivity computation) will significantly reduce the convergence iteration time to guarantee an optimal design.

• Compared to the standard zigzag spacer design, the optimal spacer reduces pressure loss by 24% with a permeation drop in the flow channel of only 0.43%.
• Gradient-based optimisation using discrete adjoint approach allows spacer shape design with more complex geometries.

Features: multi-physics modeling, Computational Fluid Dynamics, mass/species transfer, gradient-based optimisation

CAD-based shape optimisation via discrete adjoint solver derived from CFD codes

Solver development of in-house CFD-Adjoint tools:

• Accuracy improvement: residual reduced by 1 order of magnitude
• Fast running speed: CPU time reduced by ~60%
• Robustness improvement: 1) wider parameter space achieved; 2) skewer mesh case affordable
• Stabilisation of both of the flow solvers and counterpart adjoint solvers

CAD-based shape optimisation

• CAD-based parametrisation via Non-uniform rational basis spline (NURBS) patches

• The air duct in Volkswagen Golf vehicle is optimised to reduce pressure drop/loss.
• The NURBS boundary representation (BRep) is writen in STEP standard.
• In-house CAD tool, NsPCC (NURBS-based parametrisa-tion with continuity constraints) guarantees G1 (1th order) and G2 (2st order) geometry continuity at interfaces.

• Dean vortices in curvature duct lead to energy loss.

• The optimised shape suppresses the secondary flow motion.
• Pressure drop from the optimised S-bend duct is reduced by ~20% compared with base-line design.

Features: CFD software development, Automatic Differentation, CAD-based parametrisation, shape optimisation

Spectral Element Method (SEM) for acoustic propagation in non-uniform flow

Spectral Element Method is a combination of Finite Element Method and Spectral Method.

The pressure pulse (Gaussian Pulse) transports in the computational domain with absorbing boundaries: Clayton-Engquist-Majda (CEM) and Eliane-Dan-Thomas (EDT).

In non-uniform flow field, sound wave from dipole source convected by the flow shows irregular pattern.

Features: Computational Aero Acoustics (CAA), High accuracy solution, Discontinuous Galerkin Method, Chebyshev polynomials, implicit Newmark method