Development of fast flood inundation emulators using data-driven models and spatial reduction and reconstruction methods
Wenyan Wu
Melbourne University
DEIB - Alario Room (Building 21)
On Line by Zoom
September 27th, 2022
2.30 pm
Contacts:
Matteo Giuliani
Research Line:
Planning and management of environmental systems
Melbourne University
DEIB - Alario Room (Building 21)
On Line by Zoom
September 27th, 2022
2.30 pm
Contacts:
Matteo Giuliani
Research Line:
Planning and management of environmental systems
Abstract
On September 27th, 2022 at 2.30 pm, Dr Wenyan Wu, University of Melbourne, will hold a seminar on "Development of fast flood inundation emulators using data-driven models and spatial reduction and reconstruction methods" in DEIB Alario Room and in live streaming by Zoom.
Fast flood inundation modelling is a key to applications including, but not limited to, large-scale flood simulation, real-time ensemble flood forecasting, and uncertainty analysis for flood risk management. The computational intensity of traditional 2D hydrodynamic models always leads to prolonged run time and limits their applications. Flood inundation emulators which can be developed based on 2D model results have gained attention in the recent years and are found suitable for fast modelling of inundation. This work presents the development of a fast flood inundation emulator using data-driven models (including deep learning models) and spatial reduction and reconstruction methods. The inundation emulator translates streamflow discharge and water level boundary conditions into flood inundation evolution with time with high accuracy and efficiency. It can be used in applications such as real-time flood forecasting, flood infrastructure design and operation considering uncertainty.
Fast flood inundation modelling is a key to applications including, but not limited to, large-scale flood simulation, real-time ensemble flood forecasting, and uncertainty analysis for flood risk management. The computational intensity of traditional 2D hydrodynamic models always leads to prolonged run time and limits their applications. Flood inundation emulators which can be developed based on 2D model results have gained attention in the recent years and are found suitable for fast modelling of inundation. This work presents the development of a fast flood inundation emulator using data-driven models (including deep learning models) and spatial reduction and reconstruction methods. The inundation emulator translates streamflow discharge and water level boundary conditions into flood inundation evolution with time with high accuracy and efficiency. It can be used in applications such as real-time flood forecasting, flood infrastructure design and operation considering uncertainty.
Short Bio
Dr Wu is a Senior Lecturer in Environmental Hydrology & Water Resources in the Department of Infrastructure Engineering, the University of Melbourne, Australia. She has over ten years’ experience across academia and industry. Dr Wu has a background in environmental modelling and optimisation using numerical methods, with applications in optimal design and operation of water distribution systems and management of water resources systems. She also worked as an Environmental Engineer, when she conducted flood risk assessment and developed flood mitigation strategies with the aid of 2D hydrodynamic models.
Her current research focuses on the use of numerical and statistical modelling tools to solve environmental engineering problems, including:
• Water resource/supply system planning, design and operation
• Machine learning techniques for environmental modelling
• Multiobjective optimisation and its application in decision-support
• Flood modelling, forecasting and risk analysis
The event will be held online by Zoom
ID riunione: 982 0072 3672
Passcode: 389290
Her current research focuses on the use of numerical and statistical modelling tools to solve environmental engineering problems, including:
• Water resource/supply system planning, design and operation
• Machine learning techniques for environmental modelling
• Multiobjective optimisation and its application in decision-support
• Flood modelling, forecasting and risk analysis
The event will be held online by Zoom
ID riunione: 982 0072 3672
Passcode: 389290
