Find on this page

Climate adaptation and risk reduction of Swedish railway infrastructure (AdaptRail)

Future transport infrastructure is expected to be more interconnected and complex leading to different types of vulnerabilities and risks affected by changes in climate conditions.

The aim of the project is to improve the resilience of railway infrastructure from adverse future climate conditions through implementing climate adaptation strategy in design, operation and maintenance of infrastructure.

Climate change and its associated impacts are the most critical global problems of our time. A growing body of scientific literature suggests that "climate change will continue for many decades, and even centuries, regardless of the success of global initiatives to reduce greenhouse gas emissions." In addition, climate change and digital transformation including Artificial Intelligence (AI) are the two most significant trends of the century. There is a need to create pathways to combine climate and digital transformation to support our health, safety, socio-economy values and critical infrastructure.

Scientific leader: Amir Garmabaki | Researchers: Ulla Juntti, Ahmad Kasraei, Johan Odelius

WP 1: Project preliminary study

The potential modelling approaches e.g., mathematical, statistical and machine learning algorithm will be explored for developing DSS.

For instance, different machine learning algorithms will be utilized for developing decision support systems including anomaly detection and dynamic risk assessment and maintenance optimization. In addition, we aim to identify climate adaptation measures and related KPI for selected assets. For instance, to demonstrate the added value of the tools in terms of economic efficiency for stakeholders (infrastructure managers, maintainers, railway undertakings, passengers), a KPI model will be developed for quantitative assessment of climate adaptation options.

WP 2: Development of decision support system for climate adaptation

The first task for this WP is Data acquisition/collection tasks includes

(i) Maintenance and operation data collection (inspection, repair, failure, etc.) for each case (Rail, S&C, and drainage system) (ii) Meteorological and Satellite data (iii) sustainable railway infrastructure Impacts and its associated KPIs.

In addition we will explore the potential machine learning algorithms, statistic-based learning approaches (e.g., Bayesian methods, Multivariate Analysis), and expert-based systems to reach the project goal. Furthermore, there may need to use ensemble models that consist of combinations of these algorithms. Ensembles will be fused with model-based approaches (physical knowledge) to create a hybrid model and increase the accuracy of the predictions with minimum uncertainty.

WP3: Validation and full-scale demonstration of the developed methods

To demonstrate the proposed modelling frameworks in full scale, various demonstration scenarios in terms of time horizon and location (Urban, Rural) and under various RCPs, RCP 2.6, RCP 4.5 and RCP 8.5 will be considered.

For instance, the first scenario will be the assessment of flooding in the cities considering for the next 30 years on urban rail infrastructure, and the second scenario will be the assessment of track and S&C buckling (high-temperature impact) on the iron ore line from Luleå till Kiruna city.

WP4: Dissemination Communication, and competence building

The main goal of this working package is to ensure effective collaboration for sustainable competence building through knowledge sharing and exchange of training/education materials across participating institutes and within the quadruple helix.

Results and outcomes of this multidisciplinary research project will be communicated via organizing workshops, seminars, and webinars, and project homepage.