Skip to content

ChangeoverEvents¤

Detect product/recipe changes and compute changeover windows.

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


When to Use¤

Use to track product/recipe changes on production lines. Detects when the product signal changes value and computes changeover duration. Essential for understanding SMED (Single-Minute Exchange of Die) improvements and tracking setup time as a component of planned downtime.


Quick Example¤

from ts_shape.events.production.changeover import ChangeoverEvents
import pandas as pd

df = pd.DataFrame({
    "timestamp": pd.date_range("2024-01-01 06:00", periods=360, freq="1min"),
    "product_id": ["SKU-A"]*120 + ["SKU-B"]*100 + ["SKU-A"]*140,
})

co = ChangeoverEvents(df, timestamp_column="timestamp")
changes = co.detect_changeover(product_uuid="product_id")
windows = co.changeover_window(product_uuid="product_id", until="fixed_window")
print(changes)

Key Methods¤

Method Purpose Returns
detect_changeover(product_uuid) Generate point events at each product/recipe change DataFrame with timestamp, previous product, and new product
changeover_window(product_uuid, until='fixed_window') Compute changeover duration windows from change point to next stable production DataFrame with start, end, duration, and product pair
changeover_quality_metrics(product_uuid) Compute quality metrics around changeover events such as first-article pass rate DataFrame with changeover quality statistics

Tips & Notes¤

Combine with machine state

Pair changeover detection with MachineStateEvents to distinguish changeover downtime from unplanned downtime — changeovers typically show a planned idle pattern.

Related modules

  • Machine State — run/idle detection to separate changeover idle from unplanned stops
  • Batch Tracking — similar concept for batch-ID signals in process industries
  • Downtime Tracking — categorize changeover as planned downtime

See Also¤