Skip to content

ProcessStabilityIndex¤

Generate a single 0-100 stability score per shift/day for a signal.

Module: ts_shape.events.engineering.process_stability_index
Guide: Process Engineering

When to Use¤

Use as a single-number KPI for process health dashboards. Answers "is my process running well today?" at a glance. Combine with other modules for detailed root cause analysis when the score drops. Ideal for management dashboards, shift handover summaries, and cross-plant benchmarking.

Quick Example¤

from ts_shape.events.engineering.process_stability_index import ProcessStabilityIndex

scorer = ProcessStabilityIndex(
    df=process_data,
    uuid="reactor_temperature_01"
)

# Get a 0-100 stability score per 8-hour shift
scores = scorer.stability_score(window="8h")

# Track score trend over the past 7 shifts
trend = scorer.score_trend(window="8h", n_windows=7)

# Find the 5 worst 1-hour periods for investigation
worst = scorer.worst_periods(window="1h", n=5)

# Compare current performance to historical best
comparison = scorer.stability_comparison(window="8h")

Key Methods¤

Method Purpose Returns
stability_score(window='8h') Composite 0-100 stability score DataFrame with scores per window
score_trend(window='8h', n_windows=7) Score trend over N windows DataFrame with trend data
worst_periods(window='1h', n=5) N worst-scoring windows DataFrame of worst periods
stability_comparison(window='8h') Compare to historical best DataFrame with comparison metrics

Tips & Notes¤

Use worst_periods to drill down from a low score

When the shift-level score drops, call worst_periods(window='1h', n=5) to pinpoint the exact hours that dragged the score down. Then use module-specific tools (disturbance recovery, rate of change, etc.) on those windows for root cause analysis.

Related modules

See Also¤