02 · Technology
A decision-intelligence platform for battery recycling built on Scientific AI: physics-based models combined with machine learning, so operators get predictions they can trust and decisions they can act on in real time.
Scientific AI
Digital twins
Generative AI
Industrial IoT
How it works
Scientific AI is an objective-driven approach pioneered by IntelliSense. It combines mechanistic process models with machine learning to provide trusted predictive intelligence for operational decisions, in a way that resembles how experienced engineers plan and evaluate.
Its strength is adaptability. Objectives can be changed in real time without exhaustive retraining on new data: the models understand their environment and optimise actions in line with the defined targets, enabling autonomous decision execution.
Powered by Scientific AI and the Industrial Internet of Things (IIoT), BRACE brings process optimisation, material tracking, strategic simulation and augmented intelligence together in one closed-loop system.
The decision loop
01
Sense
Live plant and laboratory data streams in through IIoT integrations.
02
Predict
Scientific AI models forecast metallurgical and process behaviour.
03
Optimise
Set-points and recommendations are calculated against your objectives.
04
Act
Closed-loop control and decisions feed back into the plant, and the loop starts again.
↻ Continuous, closed-loop improvement
Components
Two new applications, MTI and brains.VOS, join IntelliSense’s proven optimisation apps and augmented-intelligence tools to form a closed-loop system for real-time control and strategic planning.
Component 01
TRL 9 in mining
TRL 7 → 9 in battery recycling
Optimisation applications proven in live mining deployments, extended and upscaled for every step of the battery recycling flowsheet: sorting, pyrolysis, crushing, flotation, leaching and solid–liquid separation.
→
Crushing & grinding: reach the target particle size while optimising energy and water use
→
Flotation: achieve target recovery and grades while balancing production
→
Leaching: maximise metal recovery rates
→
Solid–liquid separation (thickener): maximise underflow density and water recycling
Component 02
TRL 9 in mining
TRL 7 → 9 in battery recycling
Tracks and predicts the metallurgical and physical properties of the material as it moves through the plant (Material Transport) and their effect on the process in real time (Material Influence).
→
Continuous optimisation of process flowsheets, from grinding to refining
→
Integrated with KPIs to maximise net metal production per tonne
→
End-to-end traceability, real-time mass reconciliation and reporting of operational emissions
Component 03
TRL 8 in mining
TRL 6 → 9 in battery recycling
A strategic digital twin of the entire value chain, for planning plant designs, scale-ups and training through simulation.
→
Greenfield design: simulate multiple plant or flowsheet designs to find the highest net present value (NPV)
→
Simulated states: test new equipment, such as a grinding circuit, against the baseline to get a true ROI
→
Planning simulations under varying market conditions, including five-year plans
Component 04
Generative AI
Agentic workflows
A generative AI tool for augmented operational intelligence, onboarding and workforce upskilling, helping make the recycling workforce “AI ready”.
→
Natural-language conversations in the local language
→
Fast access to deep operational knowledge and context
→
Upskilling on both operational details and new technologies
Component 05
Leading KPIs
AI analytics
LLM workflows
Leading-KPI monitoring, AI analytics, and LLM-based information retrieval and workflow automation, so teams can act before problems escalate.
→
Track continuously changing feeds
→
Predict accidents and unwanted stops
→
Access reliable data for every decision
Across the flowsheet
BRACE is configured to each plant’s flowsheet. At Sunlight it covers the full chain, from incoming feed to the recovery of battery metals.
01
Sorting
Feed preparation
02
Pyrolysis
Thermal pre-treatment
03
Crushing & grinding
Target size with less energy and water
04
Flotation
Target recovery and grades
05
Leaching
Maximum metal recovery
06
Solid–liquid separation
Denser underflow, more water recycled
MTI · tracks material, properties and KPIs across every step
brains.VOS · simulates the whole plant for strategic planning
brains.Expert · supports the people who run the plant
Pilot deployment
Sunlight Group has commissioned a fully operational lithium battery recycling pilot plant in Xanthi, Greece, with a nominal capacity of 500 tonnes per year. Its modular, highly versatile units handle cells, modules, packs and manufacturing scrap regardless of chemistry or physical characteristics.
In the first year, IntelliSense maps the plant’s operational, data and workflow requirements, with TU Bergakademie Freiberg contributing hydrometallurgical expertise. Models and user interfaces are then configured, tested and refined with the operators, through user acceptance testing (TRL 8) to full deployment on the live flowsheet (TRL 9).
The outcome is a scalable model for replication in other recycling plants, which is the basis for BRACE’s commercial roll-out.
Sunlight pilot plant
Location
Xanthi, Greece
Capacity
500 tonnes per year (nominal)
Inputs
Cells, modules, packs and manufacturing scrap
Chemistries
LFP · NMC · LTO · LMFP
M18 · Jun 2027
Digital twins and interfaces demonstrated
M21 · Sep 2027
Closed-loop automation implemented
Technology readiness
Technology readiness levels (TRL) for the BRACE components, based on the project plan. TRL 8 is reached through user acceptance testing at system completion; TRL 9 through full deployment at Sunlight.
Component
Mining (today)
Battery recycling (start)
End of project
Process optimisation apps
TRL 9
TRL 7
TRL 9
Materials Tracking & Influence (MTI)
TRL 9
TRL 7
TRL 9
brains.VOS
TRL 8
TRL 6
TRL 9
BRACE platform (overall)
n/a
TRL 5–6
TRL 9
Interested in a pilot demonstration or in deploying BRACE at your recycling plant? Get in touch with the team.