Skip to content

MachineStateEvents¤

Detect run/idle transitions and intervals from a boolean state signal.

Module: ts_shape.events.production.machine_state Guide: Production Monitoring

When to Use¤

Use as the foundational production module — machine state drives downtime, OEE, and throughput analysis. Requires a boolean run/idle signal from the PLC. Every production analytics workflow typically starts with machine state detection before layering on throughput, quality, or downtime analysis.

Quick Example¤

from ts_shape.events.production.machine_state import MachineStateEvents
import pandas as pd

df = pd.DataFrame({
    "timestamp": pd.date_range("2024-01-01 06:00", periods=120, freq="30s"),
    "machine_running": [True]*20 + [False]*5 + [True]*40 + [False]*10 + [True]*45,
})

events = MachineStateEvents(df, state_column="machine_running")
intervals = events.detect_run_idle(min_duration="0s")
transitions = events.transition_events()
print(intervals.head())

Key Methods¤

Method Purpose Returns
detect_run_idle(min_duration='0s') Identify contiguous run and idle intervals, optionally filtering by minimum duration DataFrame of intervals with start, end, duration, and state
transition_events() Generate point events at each run-to-idle or idle-to-run transition DataFrame of transition timestamps and direction
detect_rapid_transitions(threshold='5s', min_count=3) Flag suspicious rapid state changes that may indicate sensor noise or PLC faults DataFrame of rapid-transition windows
state_quality_metrics() Compute data quality metrics such as gap count, coverage, and signal stability Dict of quality metric values

Tips & Notes¤

Filter short idle gaps

Use min_duration in detect_run_idle() to ignore brief sensor glitches — a 2-5 second threshold works well for most PLCs.

Related modules

See Also¤