PhD: Intelligent Dynamic Flood Response and Recovery Strategy

Flood recovery response in urban areas depends on the type of flood (e.g., pluvial, fluvial, groundwater) and the type of water (i.e., clean, grey or black water) affecting the area. In certain areas medium and high intensity rainfall events can cause combined clean and dirty water drainage systems to surcharge and cause contamination of river and flood waters. In addition, high water levels in rivers can cause hydraulic locking of sewage treatment works causing a backup within the associated drainage systems again causing discharges of sewage contaminated water into streets and properties. This contaminated water poses significant risks to human health and requires intensive post event cleaning activities.

Flood response services require detailed and dynamic information to determine the extent and continued risk of each flood and water type to better target the nature and deployment of recovery resources.

The data is critical to rescue and response teams, but also infrastructure management organisations  responding to protect critical infrastructure such as electricity supply, gas supply and water as well as critical roads that are the escape routes for people and their belongings.  The appropriate dynamic choices not only have the potential to save lives and infrastructure but also may improve community recovery and enhance the recovery of businesses driving localised economic recovery.

Current remote sensing methodologies used for that purpose are curtailed by several limitations:

•           Spatio-temporal coverage may not be available for the required zone and period,

•           Optical imagery cannot provide information if there is low cloud cover, and

•           Satellite (SAR) data, which can penetrate cloud cover, has an oblique viewing angle which makes it difficult to discriminate the water signature from other urban features.

•           Recent advances in the use of emerging technologies carried out at Cranfield University [EP/P02839X/1; EP/N010329/1; NE/N020316/1; NE/P018890/1] have enabled the development of novel monitoring solutions to address this gap in knowledge

This project will contribute to address current gaps in knowledge via the extensive use and testing of remote sensing emerging technologies and novel machine learning methods.

The aim of this project is to develop a combined UAV-AI based framework to identify the impacts of flooding from different sources and facilitate clean up and recovery activities. This will be achieved through the following four objectives:

– Objective 1. To develop a set of algorithms for the detection of flood impact to infrastructure (electricity supply, gas supply, water networks, water treatment plants and transport networks, amongst others) from high resolution UAV imagery.

– Objective 2. To quantify the robustness of key indicators for the identification of damage caused by different types of flood to critical infrastructure.

– Objective 3. To optimise decision making for a set of scenarios that look at maximising the saving of lives, community recovery, localised economic recovery and infrastructure protection.

– Objective 4. To develop a set of operational guidelines (and associated risk communication strategy) for flood managers and first emergency responders based on the results from objectives 1 to 3.

This PhD is sponsored by EPSRC, Atkins and Cranfield University.  EPSRC is the British Research Council that funds research in engineering and the physical sciences. Atkins was established in 1938 by Sir William Atkins (London). Originally, the company specialised in civil and structural design. Nowadays, Atkins offers a varied range of services, including planning, engineering sciences, architecture and project management. More information about Atkins can be found here.

What will the student gain from experience (transferable skills, employability): The successful applicant will gain the necessary skills to be able to interpret and analyse high-resolution drone imagery within the context of flood impact to infrastructure using machine learning techniques. The successful applicant will have access to a wide range of training activities already in place for PhD students at Cranfield University plus additional training offered via other channels. The student will work in close relationship with the industrial sponsor.