Structural Health Monitoring and Predictive Maintenance for Ensuring Long-Term Structural Health

Summary: This research proposal addresses the challenges posed by infrastructure in the UK's construction industry, focusing on sustainability and safety. Infrastructure, including bridges and rail structures, often experiences shorter service lives than predicted, leading to increased risks and costs. To mitigate these issues, this study aims to develop machine learning models that combines data from similar infrastructure and integrates information from previous inspections to predict potential problems. The solutions will also account for the impact of changing climate conditions on these structures. In this respect, Large Language Models (LLMs) will be considered as a key approach in addressing the challenges. By leveraging the predictive capabilities of LLMs, the project aims to provide actionable insights that can inform maintenance schedules, thereby enhancing the longevity and safety of critical infrastructure. Furthermore, the research will engage with industry stakeholders to ensure that the developed models are practical, scalable, and aligned with the current needs and future challenges of the UK's construction industry.

Project Background: The construction industry in the UK faces a dual challenge: ensuring the longevity and safety of infrastructure while adapting to a changing climate. Infrastructure encompassing bridges, rail structures, and more frequently experiences a shorter service life than anticipated. This results in elevated maintenance costs and safety risks that are further exacerbated by the increasing complexity of infrastructure networks and the scarcity of skilled engineers.

Furthermore, monitoring and inspecting these structures have become increasingly complex and costly due to the scarcity of skilled engineers and the inherent risks. The demand for expert structural inspectors often outpaces the available workforce, leading to inspection delays and increased expenses. Additionally, manual inspections in hazardous or hard-to-reach locations pose significant safety concerns for both inspectors and the infrastructure.

Climate change presents an additional layer of complexity. Projections indicate that the UK is set to experience a rise in average summer temperatures of 3-4°C and a potential decrease in summer rainfall by 11-27% by the 2080s. These changes have profound implications for the structural integrity of infrastructure assets, as they challenge the design codes originally based on historical climate data. As climate-related stressors intensify, the urgency of adapting infrastructure management practices becomes evident.

Large Language Models (LLMs) present a promising solution in this challenging context. LLMs can process vast amounts of textual data, allowing them to extract and analyse patterns from historical inspection reports, research papers, and other relevant documents. Their ability to understand and generate human-like text enables them to provide insights and recommendations based on accumulated knowledge, bridging the gap between the scarcity of human experts and the increasing need for advanced analysis.

This proposal aims to address these multifaceted challenges by leveraging machine learning, LLMs, and past inspection data. The research will develop a predictive maintenance algorithm that considers the changing climate conditions in the UK, enhancing infrastructure resilience, sustainability, and cost-efficiency. It will streamline monitoring processes, reducing the complexity and costs associated with inspections while ensuring critical infrastructure's long-term safety and sustainability.

Achieving these objectives is crucial not only for the safety and well-being of communities but also for reducing the environmental footprint of the construction industry. Adapting infrastructure to the projected climate changes is imperative for the UK's long-term economic and environmental sustainability.

  1. How can a machine learning algorithm effectively blend data from infrastructure in the UK, including information from previous inspections, to predict potential problems, while considering changing climate conditions?
  2. Can the algorithm synthesise diverse data sources, including image, time series, previous inspections, and real-time data, to enhance infrastructure management and adapt to climate change?
  3. What is the accuracy and reliability of the predictive algorithm in anticipating potential problems under changing climate conditions in the UK?
  4. What are the environmental benefits of implementing the algorithm in infrastructure maintenance and inspection practices, considering climate resilience and leveraging historical data to inform decision-making?


  1. Data Collection: Gather a comprehensive dataset from infrastructure in the UK, including structural parameters, environmental factors, historical inspection records, and real-time sensor data. Incorporate historical inspection data from the same infrastructure to analyze trends and changes over time, particularly under changing climate conditions.
  2. Data Preprocessing: Clean and preprocess the collected data, addressing missing data and outliers. Integrate climate change-related variables and historical inspection data.
  3. Machine Learning Model Development: Design and implement a machine learning model that effectively blends and analyzes data, including information from previous inspections and climate resilience considerations.
  4. Model Validation: Validate the predictive model using appropriate metrics and cross-validation techniques, focusing on its performance under changing climate conditions and leveraging historical inspection data.
  5. Implementation and Evaluation: Assess the algorithm's practical applicability by implementing it in real-world infrastructure maintenance and inspection processes in the UK, with a focus on climate adaptation and leveraging historical data to inform decision-making.








Further Information: 



Further Information: The outcomes of this research hold the potential to revolutionize infrastructure management in the UK. By leveraging machine learning, historical inspection data, and accounting for changing climate conditions, this study aims to enhance the sustainability, safety, and economic efficiency of infrastructure, contributing to the broader goal of reducing the environmental impact of the construction industry and achieving climate resilience.


The University of Edinburgh is committed to equality of opportunity for all its staff and students, and promotes a culture of inclusivity. Please see details here:

Closing Date: 

Thursday, October 10, 2024

Principal Supervisor: 


Applicants for the PhD position should possess an Honours degree (2:1 or above) or its international equivalent in a relevant science or engineering discipline; an MSc degree is advantageous. They should have a basic understanding of machine learning or infrastructure engineering and be familiar with programming or scripting.

Further information on English language requirements for EU/Overseas applicants.


Tuition fees + stipend are available for applicants who qualify as Home applicants (International students can apply, but the funding only covers the Home fee rate)

To qualify as a Home student, you must fulfill one of the following criteria:

• You are a UK student

• You are an EU student with settled/pre-settled status who also has 3 years residency in the UK/EEA/Gibraltar/Switzerland immediately before the start of your Programme. (International students not eligible.)

Further information and other funding options.

Informal Enquiries: