Dynovy Intelligence Platform
One model for battery health, today and tomorrow.
The platform turns short, controlled tests and other available evidence into a state-of-health result with a stated confidence, and a recommended next step for each cell or module.
How it fits together
Several inputs, one model, clear results.
Each input adds information to the same model instead of producing a separate verdict. The Core Model turns them into one answer for each unit.
Inputs
Results
Concept view. The internal structure of the Core Model is not shown. Abnormal-unit detection, forecasting and battery-passport records are still in development.
Two value propositions
What the platform tells you, now and next.
Understanding battery health today
- Capacity-based state of health and usable capacity
- Internal resistance and power capability
- Degradation indicators and diagnostic status
- A grade, a referral to a full test, or a recycling recommendation
Every result states the confidence range and the evidence used.
Predicting battery behaviour tomorrow
- Degradation trajectory under a stated second-life use profile
- Remaining useful life to an agreed end point, always as a range
- Matching units to applications such as stationary storage
Not a delivered output. A forecast depends on the future use profile and is never presented as a guarantee.
Leaner assessment
Less test time, less data, the same decision.
Because the model combines evidence and knows its own uncertainty, most units can be decided from a short test, and the slow reference test is kept for the cases that need it.
Fewer full reference cycles
Only reference units and uncertain cases need one.
Shorter tester time per unit
From hours towards minutes for most units.
Less data for each new family
A few reference units instead of large characterisation batches.
Targets until demonstrated in blind validation on physical units.
Product preview
Every result shows its working.
Each unit result shows the estimate, its range against your grade limits and the grade that follows.
Unit result · M-0417
B · Second lifeThe whole range sits inside one band, so the unit is graded without a full cycle. Grade limits are set by the user.
Scope and status
Cells first, then modules, then packs.
Cells
One battery family and test set-up first, with physical tests and reference measurements.
Modules
Group-level results where sense leads allow, with the weakest group reported.
Packs
A separate step with its own calibration and validation, once modules are proven.
Where we are: the core software is implemented and verified on synthetic data, and physical tests on cells and modules are used to develop the model. Blind validation on a first battery family is the next phase.
Questions
Frequently asked
Which battery chemistries does the platform support?
Each battery family is calibrated and released separately. The first family is chosen together with the partners of the first validation project.
Cells, modules or packs?
Cells are the current focus and modules are in development. Packs are on the roadmap as a separate step with their own calibration and validation.
How long does a test take?
The target is minutes for most units, with a full reference test only where the model asks for one. Durations are confirmed per family during validation.
Do you need access to the battery management system?
Not for cells and modules: the short electrical test is the core evidence, and BMS data is added where available. Pack assessment will need access through the pack connector.
What happens when the model is unsure?
It says so, and refers the unit to more evidence or a full test instead of reporting a guess.
How is test data handled?
Confidentiality, ownership and use of test data are agreed in writing with each validation partner before any data is exchanged.
See what the platform can do with your battery family.
Tell us about your cells or modules and the decision you need to make. We will reply with how a validation project could work for you.
Discuss a validation project