Any manufacturer working in a highly competitive and environmentally friendly world must pursue process optimization to ensure that no resources are wasted in their production processes.

To achieve this purpose, it is increasingly necessary to rely on information exchange platforms such as ZDMP where artificial intelligence algorithms are constantly monitoring the process parameters to ensure that the quality of the next part to be produced will perfectly meet the required parameters.

With this goal in mind, 3 applications have been developed and tested in different scenarios: zAnomalyDetection, zDigitalTwin and zAlarm. Thanks to these applications, it is possible to monitor the process, detecting and reporting anomalies long before a quality defect occurs.

The three scenarios chosen to test these applications within the production process of a combustion engine have been: the foundry of the engine block, in our case MartinRea Honsel, the manufacturer of the machinery with which they are machined, in this case Etxetar and Ford as the company that finalizes the blocks that are later assembled into the engine.


App able to monitor a high number of variables and identify and collect anomalies in an automatic manner, offering operators the possibility to, given a detected anomaly, have an easy-to-understand view of the manufacturing process:


  • Anomaly detection based on unsupervised learning algorithms
  • Quality control metrics
  • Explanaibility


The mode of operation of this application is: given a sequential production process (A + B), the app can optimise quality variables in process B based on the current values of process A


This functionality is used to be able to predict and recommend the best parameters based on the results of the simulations carried out with the parameters received as inputs from the process.


Application and equipment ready to send alarms/ events to wearables supporting LoRaWAN messaging:


  • LoRaWAN communications in the sub-GHz band more robust against interference, low consumption, and long range. Private network deployment.
  • Improvement of machine uptime: alert operators in real-time when an anomaly is detected
  • Efficient maintenance coordination and collaboration: direct assignment of repair orders to the correct technician
  • Increase maintenance productivity: avoidance of unnecessary work steps