Open models
Synthetic-data modelOpen · synthetic

Lithium Impurity Risk Model

Estimates the risk of impurity breakthrough to battery-grade product from feed profile and purification conditions.

LithiumLithium Processing IntelligenceImpurity predictionv0.3.0Apache-2.0
7.2k 205 44

This model is a reference implementation. Industrial use requires validation on plant-specific data. Customer-specific fine-tuning remains private unless explicitly approved for publication.

Input schema

FieldTypeDescription
feed_impuritiesfloat[] · ppmMg, Ca, B, Na in feed
reagent_dosingfloat[]Precipitation / IX dosing
phfloatPurification pH

Output schema

FieldTypeDescription
impurity_riskfloatBreakthrough risk 0 to 1
driver_rankingstring[]Top contributing impurities

Training data & evaluation

Training data type

Synthetic impurity-control dataset

Evaluation metrics

ROC-AUC 0.90Brier 0.08

Limitations

  • Battery-grade thresholds are configurable; defaults are illustrative.

Discussion

  • s.

    s.iyer

    ML researcher · 3d

    Validation note

    Reproduced the reported metric on the public split with a fixed seed. Numbers line up within noise, so it's a good baseline to build on.

    18
  • l.

    l.zhang

    Process engineer · 1w

    Improvement

    Adding an uncertainty head would make this far more useful for prioritization. Happy to open a PR against the reference repo.

    11
Reference, not production-proven

Validate on your data, keep the result private

Reference models are a starting point. We can fine-tune privately on plant-specific data, with customer-owned IP and no public disclosure.