Fedrigoni

Fedrigoni produces many different types of special paper products: some of their plants are focused on a few of them, while others often vary production. This aspect hugely impacts their productivity. 

Besides, each plant’s machinery adopts a different system for data collection. Nevertheless, Fedrigoni needs to be able to analyse plants’ performance, and increase and simplify its control on production, by:

• Integrating and normalising all data coming from different plants in one digital platform
• Extracting the same key indicators from all plants, to optimise production performance’s control

For this project, 13 machines (10 of which are for continuous production), distributed in 7 plants throughout Italy have been involved. 

Each plant exposes from 3.000 to 5.000 signals; a subset of these tags per each plant has been verified, historicised, analysed and normalised to create a common labelling and collection system.

As a matter of fact, each machine has its control system, but none of them clearly expose the relation between its indicators and the machine signals, which hinders any control and objective comparison with any other machine’s indicator, as well as any root-cause analysis. Tags, on the other hand, are punctual and, once understood, by comparing to the data extracted by the integrated machine control system, and normalised (cleaning data, managing different scales or different signals’ frequency), allowing to extract uniform data and calculate trustable indicators.

The machinery data were then integrated by the possibility for the plant operators to manually add production information (such as the type of production, the grammage of paper, and the reason for each production stop…), that is necessary for any in-depth ex-post analysis.

Thanks to this elaboration, we could identify and translate each tag meaning related to the production process and create a common ontology to elaborate a general vision of all involved plants in one unique platform.

The key focus for Fedrigoni was to extract OEE key indicators: time efficiency, material efficiency, and speed efficiency.

In addition to specific algorithms developed per machinery (to obtain comparable indicators starting from different types of tags), the statistical process control module was integrated by a custom algorithm allowing to distinguish in the production signals’ time series the end of production data from the tags related to the new paper reel.

This attribution was key to generating the analysis per type of product as well as per type of production (and not only per time stamp), enabling the analysis of production fluctuations.

Hence performance control in Fedrigoni became the key to process optimisation.