Methods you can
inspect.
We explicitly separate a functioning calculation, a synthetically exercised demo, an independent benchmark and a product validated for industrial decisions.
What the calculations show.
Where evidence is limited.
Numerical checks test calculations against specified examples or references. Independent experimental studies address how a method performs on measured data. Passing numerical checks does not certify a factory workflow.
| MODULE | CURRENT EVIDENCE | EVIDENCE STILL NEEDED |
|---|---|---|
| EIS Audit Studio | Parsing, unit/sign normalization, separate sweep grouping and numerical consistency checks. The public measured-spectrum demo has source attribution and is distinct from synthetic references. | Further exporter coverage, false-warning studies and independent experimental evidence. ECM/Lin-KK are outside the browser core and are implemented separately in the Python research engine. |
| FormationIQ | Unit-tested signed-current and power integration on synthetic cases; CSV preview | Real cross-cycler comparisons, missing-data bias, sign convention, field pilot |
| FactoryIQ | Unit-tested descriptive group statistics and spec comparisons on synthetic cases | Sampling plan, gauge R&R, process stability, industrial benchmarks |
| MaterialsIQ | Unit-tested growth and mass-loss calculations on synthetic cases | Controlled replicate datasets, calibrated measurement uncertainty, mass balance and comparable exposure conditions |
Clear scientific boundaries.
Our design choices deliberately keep interpretation separate from computation.
Traceability.
Source columns and units must be retained in the audit record.
Validity.
KK validation and ECM fitting serve different scientific questions.
Statistics.
Descriptive Pp/Ppk do not imply a capable or stable process.
Operations.
No automatic safety-critical formation control or battery rejection is offered.
Methods and scope of the current tools
Results depend on the supplied data, assumptions and software version. Repeating a calculation with the same code, checking independent equations, comparing reference libraries and testing against independent measurements provide different kinds of evidence. HTML report export is unavailable. The browser core does not perform ECM fitting or Kramers-Kronig analysis; those functions are in a separate Python research workflow. Numerical agreement alone does not establish industrial fitness.
Keep each conclusion
within its evidence.
Passing a calculation check is one step. It does not automatically establish a physical interpretation or an operational decision.
| Layer | Purpose | Boundary |
|---|---|---|
| A - Measurement and QC | Retain source records, units, grouping and quality flags. | A warning requests review; it does not prove physical failure. |
| B - Numerical analysis | Calculate explicitly defined quantities with reproducible inputs. | Agreement with reference arithmetic is not independent experimental validation. |
| C - Physical interpretation | Assess models, assumptions and parameter identifiability. | A circuit fit is not a unique electrochemical explanation. Lin-KK consistency is conditional on its assumptions. |
| D - Industrial workflow | Integrate a reviewed method into a defined operating process. | Needs plant-specific evidence, measurement-system review, security, support and acceptance criteria. |
Capabilities and limitations
| Capability | Current implementation and delivery | Scope and evidence limits |
|---|---|---|
| Four core tools | Implemented browser-local calculations; public evaluations and versioned local packages. | Defined calculations with numerical checks; no general experimental or industrial certification. Online purchasing is not yet available. |
| Advanced ECM and Lin-KK | Implemented in separate local Python research engine; constrained multistart fitting, residuals, conditional scores and local diagnostics. | Separate Python installation and numerical checks. Not included in the core tools; a fit does not uniquely identify physical processes. |
| Batch / temporal EIS and calibrated uncertainty | Further research scope; existing ingestion/grouping does not establish the full workflow. | Not included in the current tools; additional implementation and independent evidence are needed. |
| Formation segmentation / efficiencies | Future scope beyond currently supported signed throughput arithmetic. | No SEI, capacity, health or charge-control claims. |
| Process analysis / measurement systems | Current descriptive overall Pp/Ppk; stability, Cp/Cpk and gauge studies separately scoped. | Plant-specific sampling and measurement evidence required. |
| Materials uncertainty / degradation | Current matched-context descriptive endpoints; uncertainty calibration and predictive degradation research deferred. | No lifetime or superiority claim without controlled evidence. |
| Experimental design / active learning | Research direction outside the current core tools. | No available optimizer or promised autonomous experiments. |