References

Successes achieved for our customers.

Preventing machine damage through predictive maintenance

 
«Trivadis developed an IoT predictive maintenance model for us that enables us to detect potential machine damage up to 45 days in advance. That makes it easier to plan maintenance work on the machinery and reduces our costs significantly.»

Head of Engineering, Machines & Maintenance Group Operations

 

Challenge and inital situation

An international manufacturer of industrial minerals based in Switzerland, works with machinery that has sensors. Occasionally, the machines break down during production and parts become damaged. That damage as well as the resulting downtime can lead to enormous expenses. Together with Trivadis, a solution had to be found that would make it possible to predict potential machine damage up to 45 days in advance based on the available measurement data.

 

Solution

Within 14 days, a data scientist from Trivadis had analysed the measurement data using the programming language R and had created and verified models. The colleagues from the customer accompanied this process and reviewed the code provided by Trivadis. This enabled them to learn the statistical methods in practise. The infrastructure was then laid out during a joint workshop and defined in the following steps.

With a short period of time, the following was developed:

  • A model that predicts potential machine damage up to 45 days in advance using the available measurement data.
  • A sketch of what a scalable Azure infrastructure must look like in order to collect and archive sensor data and to operationalise the predictions.
  • A plan for the next proof of value to validate the project’s findings and to improve and operationailse the predictions.

 

 
Customer benefits

  • Machine damage can be detected up to 45 days earlier
  • Machine maintenance can be planned more readily
  • Costs and downtimes are reduced

 

Technologies & products

Microsoft Azure, Predictive Maintenance

 

Customer & sector

Industrial minerals/chemicals

 

Success through collaboration

Alongside the original project goal of being able to predict potential machine damage up to 45 days in advance based on the available measurement data, it was also important to include customer employees involved in the entire process and to train them in predictive maintenance. The third goal was to create a shared vision of how the predictions could be operationalised. The excellent cooperation between customer employees and Trivadis was thus an important foundation for the success of the project.

 

Kategorien

  • Branche: Industrie
  • Thema: Internet Of Things
  • Technology: Internet of Things Microsoft Azure Predictive Maintenance