Name: Muhammad Zohaib Sarwar

Title of Project: Automated Structural Condition Assessment for Concrete Bridges

Background and motivation

The maintenance of ageing infrastructure is taking large parts of the total budget available to transport network owners. In particular, the continuously growing stock of bridges is getting old, and many have exceeded their design service life. Additionally, concrete has proven not to be the everlasting, durable material that was once believed. Corrosion, degradation, fatigue and other phenomena are showing their effects now, reducing the performance of concrete. Furthermore, there is often an increase in the required capacity of the infrastructure due to traffic growth. This makes the task of infrastructure maintenance a crucial part of its management. In the last decade, there have been significant developments in fields linked to structural assessment. For instance, sensing technology is introducing energy harvesting and wireless systems, which will make it easier and cheaper to instrument the infrastructure. Furthermore, there is a trend of equipping vehicles with larger number and increasingly more complex sensors. The interconnectivity between vehicle and infrastructure has shown potential benefits to improve traffic and resources management, which is the existing line of work in ITS (Intelligent transport systems). However, that interconnectivity between vehicle sensors and infrastructure sensors has not been fully explored and could also be utilised to benefit infrastructure maintenance.

Objectives

The objective of this thesis is to develop a real-time automatic assessment system for existing concrete bridges. This can be achieved by combining the information provided by sensors installed on the bridges and signals sent from passing vehicles, as shown in Figure 1. The main idea is to obtain as much information as possible from the bridge, from different sources (vehicle and bridges), combine it (data-fusion) and develop automatic procedures (machine learning) to evaluate the current state of the structure.

The main objective of the thesis can be divided into following item

• Review of existing relevant literature on structural deterioration, long-term behaviour, assessment methods, sensor technology, damage indicators, numerical modelling, machine learning.
• Carry out numerical simulation-based studies of vehicle bridge interaction for different damage assessment strategies for bridges using signal processing and machine learning tools and provide proof of concept.
• Analysis of long-term experimental data from bridge using method developed in numerical case studies.
• Benchmarking of different damage assessment method studied using numerical simulation based on their merit and demerits.
• Contribute to extending the current knowledge about structural assessment using structure health monitoring and provide guidelines for efficient and robust strategies for monitoring campaign.

Figure 1: Interconnected vehicle and infrastructure monitoring

Expected Results

This thesis is expected to contribute to overall improve the bridge assessment strategies. The improved assessment method would reduce the maintenance cost of the bridges due to the possibility of timely decision-making, which also results in extending the service life of existing bridges. Furthermore, the following outcome would be expected:

• Literature review of state of art structural assessment methods using direct bridge measurements as well as indirect method.
• Investigation of existing challenges associated with indirect bridge monitoring by using basic and advance machine learning models to overcome some of the issues.
• The thesis would provide a reliable and computationally efficient method to handle long term data for robust analysis.
• The outcome of the thesis would be providing more insight relevant to sensing equipment and damage indicator for practical applications.

Date of start: 15/08/2019