Automotive

Prediction and prevention of component failures based on sensor data

Starting point

  • Real-time analytics are required to effectively utilise the vast amount of machine data from the relevant production machines
  • The goal is to make production more efficient and failure-resistant both in the short and long term
  • A feasibility study should, in the first step, validate the results and solution models using a prototype

Procedure

  • Documentation of the gathered requirements in the form of user stories
  • Implementation of the user stories in the agile process model using SCRUM within sprints
  • Setting up the infrastructure as an OpenShift environment based on pre-configured images in MS Azure
  • Visualisation of the analysis results in Splunk dashboards directly in the production environment

Funktionen/Projektergebnis

  • Near-Realtime Bearbeitung aller Sensordaten in der Azure Cloud über Kafka zur splunk>- Visualisierung
  • Near-Realtime Dashboards steuern unmittelbar die laufenden Produktions- und Maintenance-Prozesse
  • Validierung der Ergebnisse direkt in der Produktion

Customer benefits

  • Increase in productivity through real-time analytics
  • Prevention of production machine or component failures using predictive analytics
  • Improved product quality through targeted interventions in the production process
  • Low cost of the analytics solution through the use of MS Azure and OpenShift
Joachim Kirschner | Senior Manager / Authorised Officer