AbstractAsset owning organisations worldwide are responsible for the operation and maintenance of deteriorating stocks of bridge structures. Many of these structures are vital links within strategic infrastructure networks. Bridge managers must design and implement programmes of maintenance, repair and replacement which ensure structures remain safe and serviceable, while operating within limited budgets.
Making these decisions is complex - the data on which decisions are to be based is often incomplete, is costly to acquire and in many cases can only indirectly measure the on-going deterioration processes. Structural deterioration progresses over time in a non-linear fashion and due to a range of mechanisms. A variety of interventions are available, depending on structure condition, however, the times and condition levels at which different intervention options are viable are difficult to predict. There is huge potential for data to assist with these decisions, however this must not be blind to the sources, and inherent uncertainties, of this data.
Decisions must be made by teams and individuals within organisations and their supply chains. The design of processes and systems for asset decision making needs to recognise the roles stakeholders play, and the value of their experience.
This thesis uses a series of interviews and workshops to understand and model current bridge management practice within the United Kingdom. A large study into the reliability of visual inspection data is presented, along with demonstrations of the ways in which this data can be analysed using modern data science techniques to add value to bridge owners. Necessary developments to asset decision making processes are demonstrated through modelling of systemic behaviours, and workshops with stakeholders in existing processes.
|Date of Award||25 Jun 2019|
|Supervisor||Paul J Vardanega (Supervisor), Colin Anthony Taylor (Supervisor) & Stephen R Denton (Supervisor)|