Skip to content

StartupDetectionEvents¤

Detect equipment startup intervals based on threshold crossings or sustained positive slope.

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

When to Use¤

Use to detect and characterize equipment startup events for motors, ovens, pumps, and similar assets. Supports multiple detection strategies including threshold-based, slope-based, and adaptive approaches. Also provides startup quality assessment and failed startup detection.

Quick Example¤

from ts_shape.events.engineering.startup_events import StartupDetectionEvents

detector = StartupDetectionEvents(
    df=motor_data,
    uuid="motor_current_01"
)

# Detect startups when current crosses 5A with hysteresis
startups = detector.detect_startup_by_threshold(
    threshold=5.0,
    hysteresis=0.5
)

# Assess quality of detected startups
quality = detector.assess_startup_quality(startup_events=startups)

# Detect startups that failed to reach operating conditions
failed = detector.detect_failed_startups(
    threshold=5.0,
    min_rise_duration="5s"
)

Key Methods¤

Method Purpose Returns
detect_startup_by_threshold(threshold, hysteresis=None) Threshold-based startup detection DataFrame of startup intervals
detect_startup_by_slope(min_slope, slope_window='0s') Slope-based startup detection DataFrame of startup intervals
detect_startup_multi_signal(signals, logic='all') Multi-signal startup detection DataFrame of startup intervals
detect_startup_adaptive(baseline_window='1h', sensitivity=2.0) Adaptive baseline startup detection DataFrame of startup intervals
assess_startup_quality(startup_events) Quality assessment of startups DataFrame with quality metrics
track_startup_phases(phases) Phase tracking within startups DataFrame with phase breakdowns
detect_failed_startups(threshold, min_rise_duration='5s') Detect failed startup attempts DataFrame of failed startups

Tips & Notes¤

Use adaptive detection for varying baselines

When the idle-state baseline drifts over time (e.g., ambient temperature changes), use detect_startup_adaptive() instead of a fixed threshold. It recalculates the baseline from a rolling window and adapts automatically.

Related modules

See Also¤