Case Study: Mondi achieves predictive maintenance and €50,000+ annual savings with MathWorks MATLAB

Mondi Implements Statistics-Based Health Monitoring and Predictive Maintenance for Manufacturing Processes with Machine Learning

Mondi Gronau, a Mondi plastic production plant that makes about 18 million tons of plastic and thin film products annually, faced costly machine failures and wasted raw materials from frequent downtime. The plant’s machines generate 300–400 parameter values per minute (about 7 GB/day), but personnel had limited statistical and machine‑learning expertise and needed an operator‑friendly, production‑ready predictive maintenance solution. Mondi engaged MathWorks (MathWorks Consulting) to develop a MATLAB‑based approach.

MathWorks Consulting and Mondi (with academic support) built a MATLAB health‑monitoring and predictive‑maintenance application using Database Toolbox, Statistics and Machine Learning Toolbox, Deep Learning Toolbox, and MATLAB Compiler. The team cleaned and visualized sensor data, added statistical process control alerts, evaluated multiple algorithms and chose an ensemble of bagged decision trees, and deployed a standalone executable running 24/7. The MathWorks solution delivered a working prototype in six months, saves more than €50,000 per year across eight machines (with expected fourfold increase as more machines are covered), and runs continuously in production.

Mondi

Michael Kohlert

Head of Information Management and Process Automation