Enhancing the resilience of the bridge network by identifying key performance indicators through survival analysis and evaluation of maintenance prioritisation and budgeting methods

Abstract

In recent years there has been a rapid deterioration in the condition of bridges in the UK and globally. Climate change and increasing traffic have undoubtedly contributed to this rapid growth. However, consultation with bridge owners has confirmed outdated management and maintenance methods are also attributing to the deterioration of our road networks. Bridge Management Systems (BMS) have been introduced across the world with the goal of aiding the decisions regarding maintenance, rehabilitation and replacement (MR\&R) of bridges. In order to implement deterioration and maintenance rehabilitation models, consistent bridge inspection information is required. Due to the changing of bridge information storage and recording methods a change in how bridge inspections are conducted has occurred. This has resulted in many years of inspection information being in a different format to current inspections. This results in a loss of data value however taking steps to convert this data would lead to a longer bridge inspection history and a consistent bridge condition metric across over 20 years. The Bridge Condition Index (BCI) has been introduced to facilitate a uniform national assessment method which is less subjective. Ultimately BCI will facilitate significant improvement in the predictions of future bridge deterioration. However, in the short term the lack of consistency between the methods means condition deterioration is no longer directly comparable over long periods of time leading to uncertainty in the true condition of many bridges across strategic road networks. Machine learning and statistical approaches have transformed the management of infrastructure systems such as water, energy and modern transport networks. Artificial Intelligence-based solutions allow asset owners to predict future performance and optimize maintenance routines through the use of historic performance and real-time sensor data. The industrial adoption of such methods has been limited in the management of bridges within aging transport networks. Predictive maintenance at bridge network level is particularly complex due to the considerable level of heterogeneity encompassed across various bridge types and functions.

This thesis will begin with a literature review showing why bridge management systems have been used and how they have been implemented across the world. It will show how deterioration modelling and maintenance prioritisation are essential to maintain the safety and resilience of road bridges. Several gaps have been identified and the work in the chapters which follow will show firstly how survival analysis can be used to identify factors affecting bridge deterioration highlighting that the bridge’s function affects all condition states. Secondly, a review of how bridges are currently being prioritised for maintenance across the UK concluding that there is no consistency in what bridge managers determine to be the factors that assign a bridge to be the highest of priority. Furthermore, engineering bias and the effects of strict budgets on short- and long-term maintenance will be explored. A chapter discussing the key findings of this thesis and introducing the use of Bayesian Networks (BN) to show the connection between bridge condition, river levels and environmental data will be added. This thesis will close with drawing final conclusions and outlining some points of further work.

Date of Award	Dec 2024
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	RAEng (Royal Academy of Engineering)
Supervisor	Myra Lydon (Supervisor), Adele Marshall (Supervisor) & Su Taylor (Supervisor)