What problem does it solve?
Analyzing battery degradation requires extracting meaningful features from raw cycling test data, but hand-coding segmentation, differential curves, and peak detection is error-prone and slow. This Skill guides an AI agent to use the Predictive Maintenance Toolbox battery functions correctly, avoiding silent failures like missing IC features or inconsistent protocol phases.
Core Features & Use Cases
- Full Cycling Pipeline: Parse multi-cycle test data with batteryTestDataParser and extract 165+ features per cycle with batteryTestFeatureExtractor, covering statistics, cumulative capacity, CC/CV/CCCV segments, and IC/DV/DT differential curve peaks.
- Single-Segment Analysis: Compute statistical and cumulative features from any partial charge or discharge segment with batteryMeasurementFeatures, or generate IC/DV/DT curves from constant-current data with batteryDifferentialCurves.
- Protocol Consistency Checking: Run the bundled checkCyclingProtocol script to select the correct CyclingPhase, detect anomalous cycles, and avoid extracting features from inconsistent test protocols.
- Use Case: Given a table of multi-cycle battery test data, extract IC peak and capacity features per cycle to track state-of-health degradation and support RUL prediction.
Quick Start
Ask your agent to extract degradation features from your battery cycling data file, for example: "Extract SOH features from my battery cycling test data in cycling_data.mat using the charge phase."
