We measure response-organisation continuity from battery telemetry: which visible channels carry constrained roles, whether substitutes are thinning, and what the sensor grid can actually resolve. The output is an auditable measurement record, not an asset verdict. Beside your SOH, never instead of it.
Each deliverable starts by freezing the data contract: identity, time spine, sign convention, channel roles, normalisation, missingness, and claim boundary.
One asset or cohort, read against its own history. The report returns bounded organisation readings, a limitations ledger, an evidence boundary sheet, and the next evidence needed to close unresolved questions.
A continuing version of the report: telemetry onboarded once under a frozen measurement contract, then periodic readings against the asset's accumulating history.
Measures what the current telemetry can and cannot resolve: cadence, quantisation, channel coverage, and the minimum organisation change visible under the available sensor grid.
Each service returns the same artifact class: a bounded measurement record where every figure carries its null, its detection floor, and a recompute path. We take no side and issue no verdict; that is what lets every party point at the same record.
A warranty or transaction dispute needs a record both sides can recompute without trusting either side — or us. We measure response organisation and sensor visibility on the agreed data and return the record. We sell the measurement, never the ruling.
A second column beside SOH: what the organisation layer shows, what stayed below the measured floor, and what is honestly marked unobserved because the pack never entered that region. The boundary sheet travels with every figure.
Two packs with matching scores can carry different organisational histories. We compare biography trajectories within the sensor floor: a compatibility reading for pairing decisions, never a health or lifetime verdict.
Data boundary: transfer method, retention, confidentiality, and output scope are agreed before any dataset is received. Discuss a dataset →
What burden is your battery carrying that SOH does not show? The layer asks that in three measurable forms.
Not whether a battery is good or bad. Whether visible burden-bearing roles are distributed, mobile, or persistently concentrated under matched context.
A stable output can be maintained while the number of observable ways to support that output narrows. The report separates surface output from support-path evidence.
If the sensor grid is too coarse, the honest output is not a clean bill of health. It is a visibility floor and an evidence request.
Trace-backed entry points for the audited output folders, kept separate from the visual prototypes below. Each lens keeps its own contract, null gate, and claim boundary.
These pages show the visual language behind the observatory. They remain bounded research surfaces; the product layer is the auditable report, not the animation.
A single controlled cell observed as an organisation rhythm: response arcs, role melody, fragile windows, and memory wake. The reproducible metrology basis for “same SOH, different organisation.”
Open the live view →Scalar-matched cells played side by side. The demo asks whether a single score leaves out visible relation-shape differences; product claims require the audited report layer.
Open the comparison →Matched cell pairs mapped by scalar distance, organisation residual, life stage, and later trajectory. The page is a visual prototype, not a residual-value or warranty claim.
Open the atlas →All prototypes: battery organisation observatory index — public controlled-lab data throughout; no customer or vehicle data appears on any research surface.
| Contract field | Needed before interpretation |
|---|---|
| Identity | Cell, string, module, pack, vehicle, or asset identity; whether channels remain stable across time. |
| Time spine | Timestamp, cycle, odometer, settlement period, or test step; native cadence and gaps preserved. |
| Channels | Voltage, current, temperature, SOC, mode, cycle/odometer, and any protocol or duty markers with units and sign convention. |
| Normalisation | What is preserved raw, what is scaled, and why; no pooling across individuals to make an individual claim. |
| Boundary | Allowed measurement sentences and forbidden interpretations agreed before computation. |