Digital twin lab

DIGIBAT's Digital Twin Lab connects the facility's automated equipment into one auditable research system.

At its core is the DIGIBAT platform, developed in-house: a lab digital twin and equipment-control layer that brings together the facility's automated research equipment — from materials auto-synthesis to electrode preparation, electrolyte formulation, coin-cell and pouch-cell assembly, the AutoCatX flow-cell rig and online HPLC — and carries out agentic actions to assist research. Vendor orchestration software, such as Chemspeed's ArkSuite, operates the individual robotic platforms beneath it.

Built for trust. Research automation is only useful if its results can be trusted. The platform is designed around pre-registered analyses, guards that are tested to fail, and audit trails a reviewer can follow — so every result can be traced back to the experiment that produced it.

From experiment to dataset. Physical experiments become structured, provenance-tracked datasets, ready for machine learning and closed-loop optimisation (e.g. NIMO and AMlearn).

For researchers. Users work with the DIGIBAT team to translate manual methods into reproducible, automated workflows. Training is provided — no automation experience required.

To discuss access or collaboration: j.feng19@imperial.ac.uk

Arksuite

Arksuite is DIGIBAT's core digital twin platform installed locally. It provides a virtual representation of laboratory workflows, allowing integration between hardware and software for research.

In addition, it allows users to control Flex Catscreen and SWingXL using a workflow diagram.

Data pipeline - Datalab infrastructure

Datalab is the digital backbone of the DIGIBAT facility. Deployed securely on Imperial College London's local servers, this advanced data management platform features dedicated, automated pipelines for every piece of equipment in our lab. By automatically capturing, standardizing, and linking experimental data—from initial synthesis to final electrochemical testing—Datalab eliminates manual data entry and ensures our high-throughput results are instantly ready for AI-driven analysis and closed-loop materials discovery.