Weidmüller Energy and Injection Molding Data Set

Weidmüller Energy and Injection Molding Data Set

General Description

This data set consists of simulated data using based on real measurements. The sensor measurements in the data set are taken from manufacturing machines. It contains readings from energy meters as well as sensors that monitor the production process of injection molding machines.

Energy related measurements can relate to three different levels:

  • Level 1 is the highest level of aggregation and energy supply areas in the factory. Examples of such areas are dedicated assembly lines, compressed air, heating, cooling, main supply production
  • Level 2 aggregates energy readings of power buses within the area. For production the power bus labels typically contain a number identifying the dedicated power bus. For other aggregation levels such as pressure, heating, cooling the label may contain additional information. For instance the power busses for air pressure may relate to compressors or dryers.
  • Level 3 the most detailed level of energy readings includes energy readings on machine level. Readings are not available for each individual machine. But if they are available there may several meters installed. For instance a mounting machine may have several meters for the respective machine components.

Sensor in an injection molding machine monitor for instance pressure, distances, forces and temperatures. Overall the data set contains timestamped values containing more than 120 dimensions. Additional data includes:

  • alarm logs
  • changes in the settings of the operational parameters of the machine
  • metadata containing general information about the machine
  • metadata about the tools used
  • protocol data from the monitoring system containing information about the quality of the process
  • as well as information about the quality

Size of the Data Set

The data set used in this project encompasses about 120 million measurements points.

Main Use Cases

The data is used in two solutions offered by Weidmüller powered by AGT Analytics:

  1. Peak Load Analytics – in this use case the objective is to save energy costs by predicting and avoiding peaks in energy consumption.
  2. Predictive Maintenance and Anomaly Detection – in this use case the objective is to detect anomalies and predict when machines need maintenance.

Structure of the data set

The data is available in CSV format. Each data point consists of a timestamp and corresponding payload data. Energy readings consists of an additional timestamp marking the end of an interval for aggregated energy readings. Data is recorded with varying temporal resolution varying from 15 minutes intervals to seconds intervals. The reading records contain the following fields:

  • A time string formatted as “dd.mm.yyyy hh:mm:ss” representing the end of the time interval
  • A time string formatted as “dd.mm.yyyy hh:mm:ss” representing the start of the time interval
  • A variable number of field for the respective meters. Values contain the electric power in the unit kilowatt

Data and Resources

Additional Info

Field Value
External Description
Source
Version
Contact Martin Strohbach
Contact Email mstrohbach@agtinternational.com
Benchmark Analysis and Processing