Short abstract:

Can we simulate our way to sustainable energy futures with digital twins? I followed practitioners involved in digital twins initiatives to find out how they re-invent the notions of scale, truth, and policy advice using computational tools.

Long abstract:

Decarbonising energy systems involves addressing dilemmas between several public values, be it security, affordability, citizen engagement, or public health. What if we could model these dilemmas and anticipate sociotechnical impacts of building a new nuclear station, installation of solar panels on every roof, or a change in energy tariffs? Digital twins – usually defined as advanced virtual representations of social and physical processes and objects – promise to achieve that. As digital twins are still in the ‘demonstrator’ stage (Laurent, 2021), many questions arise about data accuracy, interoperability, security; required expertise, and, most importantly, the epistemic foundations of claims made by the twins.

This presentation investigates the sites of knowledge production related to digital twins of the energy industry, drawing from multi-sited ethnography (Hine, 2007) of relevant projects, government initiatives, and the researcher’s residence at the organisation developing a digital twin demonstrator for energy policy advice.

Thinking through a philosophy of science classic, Goodman’s “Ways of Worldmaking” (1978) as an analytical lens, the paper will investigate the making of multiple ‘scales’ produced by the digital twins. From timescales (e.g. real-time, right-time, forecasting, Coletta, 2020), spatial scales (e.g. resolution, accuracy), to politics of scaling (moving between grand ambitions to pilot projects, Pfotenhauer, 2022), I will explore to what extent digital twinning is a separate analytical category, distinct from modeling, simulations or evidence-based science. I will differentiate between the notions of ‘truth’ (e.g. ‘single source of truth’, ‘ground truth) and ‘rightness’ developed by diverse epistemic communities working with computational tools.