Please enable JS

Maintenance cost reduces and production reliability improves through Predictive Maintenance development in a chemical plant

Client situation

The client is a global leader in Materials, Solutions and Chemicals, ranked in the top 10 of the European chemical companies. EFESO performed a Proof of Value in one of the flagship plants of the company, a plant which manufactures half-finished products from raw materials.

The client was facing high maintenance costs as 43% of maintenance jobs were unplanned and urgent and therefore often had to be scheduled for night shifts or weekends.

Over a six months period, we reviewed the relevance of predictive maintenance across all equipment families in the plant. We worked intensively with maintenance, process experts, IT and their data science teams on developing a system to detect and avoid failures for 17 different equipment families and brought online a total of 33 equipment.

The main objectives were to prove the business value of predictive maintenance and to develop the client’s in-house capability to autonomously select, develop and implement predictive maintenance solutions for their assets.


EFESO support was focused on:

  • Working intensively together with the client to build a Predictive Maintenance (PdM) system, which is run entirely on an open-source platform, while simultaneously developing the capability at the client to run future PdM projects on their own
  • Capabilities were built on:
    • Capturing maintenance and process expertise into algorithms
    • Technical Platform selection for Predictive Maintenance
    • Integrating the methodology into day-to-day work
    • Calculating business cases for further equipment per site and across GBU


  • Developed data architecture for automated processing of sensor, MES & ERP data and generating classifications of equipment failures
  • Created dashboards for exploration data, alert system for output of the prediction models and machine learning model performance
  • Proved the business case to roll out predictive maintenance to 8 other plants of the Business Unit, and prepared the roll-out approach and plan. By applying the right priorities, the roll-out finances itself.
  • Implemented Knowledge transfer on applied Data Science methodologies.