ORIGINAL COMPANY-CONTEXT PRACTICE
Compare battery-test curves with unequal observation windows
Original practice scenario: A hypothetical battery research dataset contains full and partial charge-discharge tests collected under different protocols. How would you compare degradation signals without treating every trace as equivalent?
Written for independent practice using company context; this is not a verified question from a company interview.
A REASONED APPROACH
Model answer.
Align the comparison with a defined protocol and exposure measure. A short partial trace cannot be assumed to represent the same cycle or operating conditions as a complete test.
Work through it
- Retain cell or pack identity, protocol version, temperature and measurement coverage, distinguishing independent units from repeated traces of the same unit.
- Define comparable windows and reference calculations, marking unsupported extrapolation and missing endpoints rather than stretching every trace to a standard shape.
- Evaluate trends on held-out units and relevant protocol slices, reporting uncertainty and sensitivity to censoring or unequal test duration.
Tradeoffs to discuss
- Restricting to complete comparable tests improves interpretability but reduces coverage; broader modeling needs explicit assumptions about missing portions.
Common pitfalls
- Do not leak later-life measurements into an early-life prediction.
- This synthetic analysis exercise does not establish battery safety, lifetime certification or suitability for an application.