Draft:Data-Centric Engineering (DCE)

Data-centric engineering emphasises systematic sensing and measurement, data capture and management, and the application of statistics, optimisation, and artificial intelligence to support engineering decision-making (e.g., design optimisation, process control, condition monitoring, and predictive maintenance). In high-stakes and safety-critical settings, it is often framed as a “hybrid” modelling paradigm that integrates mechanistic simulations with data-driven methods to balance predictive performance with interpretability and physical consistency.

• The term “data-centric engineering” has been promoted in the UK research landscape through work associated with The Alan Turing Institute and support from Lloyd’s Register Foundation. This partnership led to the foundation of the Data-Centric Engineering programme at the Turing in 2015. The term itself was coined in a foresight review where experts identified how big data could advance safety in engineering. In 2020, a dedicated open-access journal, Data-Centric Engineering, was launched by Cambridge University Press to publish peer-reviewed research and translational case studies at the intersection of engineering sciences and data sciences. Subsequently, the Data-Centric Engineering summit (www.turing.ac.uk/dceng-summit) took place in 2021. On a more global stage, the Data-Centric Engineering workshop took place as part of the 2024 International Conference on Artificial Intelligence CAI2024, in Singapore.

Commonly cited elements of data-centric engineering include: • Data acquisition and sensing: systematic measurement, instrumentation, and condition monitoring that generate operational and experimental data streams. • Data management and engineering: data pipelines, storage, metadata, and integration practices to make data usable across teams and lifecycle stages. • Statistical inference and uncertainty: calibration, validation, and uncertainty quantification for models and decisions. • Hybrid modelling: integration of mechanistic simulations (e.g., CFD/FEM and other first-principles models) with machine learning and statistical methods. • Governance and responsible practice: attention to ethics, privacy, inclusion, and professional responsibilities around engineering data and AI use.

Data-centric engineering is applied across engineering disciplines, with reported use cases including: • Infrastructure: The development of the world's first 3D printed steel smart bridge in Amsterdam and the optimisation of subterranean farms in London. • Marine Engineering: Enhancing ship performance and reducing emissions through air lubrication systems that create bubble carpets to reduce hull resistance. • Environment: Applying AI to combat biodiversity loss and improve renewable energy generation. • Telecommunications: Utilising machine learning to improve the reliability of wireless systems like Wi-Fi. • structural asset monitoring in civil engineering using distributed sensing and cloud computing; • performance-based engine design enabled by measurement-driven datasets; • autonomy and guidance/control systems reliant on multi-sensor data; • maritime asset management and autonomous shipping enabled by sensor and data technologies; • accelerated materials discovery and design using data-driven modelling.

In the UK higher-education context, DCE is described as a cross-cutting capability rather than a single standalone discipline, spanning data acquisition, management, analytics, optimisation, and dissemination across engineering programmes. The Royal Academy of Engineering and The Alan Turing Institute with the support from Lloyd’s Register Foundation have launched a national Skills initiative aiming to embed data-centric engineering in UK engineering degrees. In the UK, there are example of DCE specific university courses:

• Data-Centric Engineering | King's College London • Data Centric Engineering | Institute for Data Science and Artificial Intelligence | University of Exeter • Data-Centric Engineering - Data-Centric Engineering Master's Programme in Data-Centric Engineering | LUT University

Data-centric engineering overlaps with data-driven engineering, digital engineering, industrial data science, and digital twin programmes. A common distinguishing emphasis is the explicit integration of domain physics and engineering knowledge with data-centric practices (data quality, governance, and lifecycle integration) to support decisions in risk-averse or safety-critical regimes.