Our partners

FPS generates the project's high-fidelity simulation data and develops the physics-informed surrogate models — PINNs, operator learning and reduced order models — that make it fast enough to use, targeting speed-ups above 100×.

It assembles these into a containerised multi-scale simulation pipeline exposed as a service, integrates the platform's components, and builds the natural-language interface through which researchers query it.

FPS releases the platform open-source and hosts the public demo instance on SimuPort for at least three years after the project. As Exploitation Manager, it leads IPR and sustainability planning.

Agency of European Innovations leads dissemination, communication, and outreach work packages that connect SimuLingua with policy stakeholders, research communities, and the wider public.

AEI developed the project brand, public website, and social presence, and coordinates scientific dissemination as validation results and open deliverables become available.

The partner works closely with FLOWPHYS on exploitation narratives and with Vytautas Magnus University on responsible-innovation messaging and ELSI-aligned outreach.

Dycotec Materials Ltd contributes industrial formulation expertise for printable sensor inks and functional coatings used in the digital olfaction use case.

The team supports inkjet and ESJET process windows, analyte-sensitive layer design, and the experimental datasets that feed UC5 surrogate training and validation.

Dycotec bridges lab-scale formulation work with device-level requirements from the olfactory-array integration path led by Oliveris.

KTH Royal Institute of Technology brings DFT and CALPHAD heritage to SimuLingua, supplying atomistic and thermodynamic datasets and simulation-backed labels for alloy-focused workflows.

KTH leads substantial contributions to high-entropy and refractory alloy design (UC3) and supports the wire-cladding digital twin (UC4) with multi-physics simulation data.

The partner contributes DFT-derived descriptors and structured metadata that enrich the consortium knowledge graph and FAIR release pipeline.

Oliveris Tech Incubator develops and integrates advanced sensing and industrial intelligence systems, combining Oliveris-owned intellectual property and engineering know-how with sensor hardware, data acquisition, control, SCADA and AI-enabled process monitoring.

Within SimuLingua, Oliveris leads UC6 and translates AI-designed sensing materials into functional digital olfactory systems. The work includes the architecture and integration of a 64×128 multiplexed interdigitated-electrode sensing array, with 8,192 independently addressable sensing pixels, multi-band impedance interrogation and real-time generation of high-dimensional chemical fingerprints.

Oliveris also designs the SimuLingua Closed-Loop Training Platform (CLTP), a controlled experimental environment used to generate labelled multimodal datasets under programmed operating disturbances. The platform combines the sensing array with calibrated industrial sensors, controlled fluid or VOC inputs, temperature and pressure variation, PLC/MIMO data capture and supervisory control.

The CLTP creates the bridge between laboratory materials development and industrial deployment. It allows AI models to learn the difference between true chemical and process signatures and environmental or operational variability, supporting recognition of chemical states, faults, scaling and corrosion conditions, process efficiency and remaining useful life.

This capability builds on Oliveris’s wider digital-twin and process-monitoring work, where sensing, thermodynamic modelling, AI, SCADA and decision support are integrated into a single engineering workflow.

Thermo Challenges Limited provides metal synthesis and advanced characterisation capabilities that ground UC3 alloy predictions in experimental reality.

The SME supports preparation of validation alloys, microstructural analysis, and property measurements used to score SciFM surrogate rankings.

Thermo Challenges helps close the loop between in-silico screening and lab confirmation for complex composition spaces.

TU Eindhoven architects the core SciFM platform stack: multimodal ingestion, knowledge-graph integration, model-order reduction, and physics-informed generative modelling.

The team defines uncertainty-quantification workflows that let downstream use cases compare surrogate confidence before committing to expensive experiments.

TU/e coordinates technical interfaces between simulation, literature, and experimental modalities so partners can train and evaluate models consistently.

The University of Sheffield leads electroceramics and durable-glasses use cases (UC1/UC2), contributing domain datasets, expert labels, and validation protocols.

Sheffield also supports UC5 metrology and cross-use-case validation, ensuring that SciFM predictions remain interpretable for ceramic and glass property targets.

The partner publishes structured experimental outcomes that become reference benchmarks for consortium-wide model evaluation.

Vytautas Magnus University embeds ethical, legal, and societal impact (ELSI) considerations across SimuLingua from proposal stage through final open-data release.

VDU maintains the FAIR Data Management Plan, advises on responsible AI practices, and coordinates policy feedback that informs standardisation activities.

The partner hosts ethics-advisory processes and ensures that public releases respect consortium agreements and regulatory expectations.