Skip to main content
Use case · Quality

Quality control with statistical limits

Attach control and warning limits to a measurement, get an alert the moment a value breaks them, and spot a slow drift before a part goes out of spec.

Time-series chart of a measurement
Apache StreamPipes, demo data

The situation

A diameter, a torque, a fill weight: the measurement exists, the limits exist, and still the first sign of trouble is a rejected batch. Control charts on paper or in a spreadsheet are checked once a shift, if at all.

StreamPipes puts the limits next to the live value. Every measurement carries its control and warning limits, a breach becomes an event that can notify someone, and a change detector flags the drift that a fixed limit would miss.

How it works in StreamPipes

  1. 1

    Connect the measurement

    Read the value from the controller or gauge with an OPC UA, S7 or Modbus adapter, or from a file. Set the unit in the adapter so charts show it correctly.

  2. 2

    Attach the limits

    The Quality Control Limits Enrichment processor adds upper and lower control limits and warning limits to each event. The limits are configured once and travel with the data.

  3. 3

    Alert on a breach

    The Sensor Limit Alert processor emits an event only when a value crosses a limit, including which limit and whether it was a warning or a control breach. Route those events to Teams, Slack or email.

  4. 4

    Watch for drift

    Welford Change Detection tracks mean and standard deviation of the stream and raises a flag when the process shifts, even while every single value is still within its limits.

What it looks like

Pipeline enriching measurements with control limits, alerting to Teams and storing a dataset
Limits, alerts and storage in one pipeline.
Configuration of the Quality Control Limits Enrichment processor
Control and warning limits are set once in the processor.
Configuration of the Sensor Limit Alert processor
The alert checks the value against the limits that travel with it.

Try it on your machine.

One command starts StreamPipes with Docker. The quickstart takes you from there to a first dashboard.

docker composerelease 0.98.0
mkdir -p streampipes-0.98.0 && cd streampipes-0.98.0 \
&& curl -fsSLO https://raw.githubusercontent.com/apache/streampipes/0.98.0/installer/compose/docker-compose.yml \
&& curl -fsSLO https://raw.githubusercontent.com/apache/streampipes/0.98.0/installer/compose/.env \
&& docker compose up -d