Mandelli: from data to smart maintenance

Our collaboration originated from Mandelli’s need to plan maintenance activities through the accurate collection and processing of data. This optimization aimed to provide end users with useful insights into the operating conditions of a system and the need to schedule a maintenance intervention in the near future.

We conducted an analysis on a Spark 2.100 machining center equipped with state-of-the-art sensors, examining the available sensors and service reports. Based on this, we divided the machine into eight main groups on which to focus data collection and analysis activities. This subdivision of the system immediately delivered tangible results, later confirmed by subsequent tests.

Thanks to trials carried out on different machines and across a variety of applications, we gathered a large amount of data and experience, allowing us to refine our algorithms and make the results increasingly reliable.

Our approach is based on three pillars:

• accurate definition of the infrastructure for data collection and availability
• data collection, filtering, and analysis activities
• development of physics-based models and statistically driven algorithms

Within iPum@, we have integrated more than 25 algorithms that continuously monitor machine parameters. Each algorithm is based on 4 or 5 correlated measurements, collected 24/7 and processed at a frequency defined together with Mandelli, according to risks and actual operational needs.

We have thus created a simple and intuitive tool that does not require specific IT skills. Customers who choose iPum@predict access a customized and secure web page, where they can find all the information on the status of the machines in the workshop.

Thanks to this system, Mandelli technicians periodically analyze the data and report any anomalies, while the customer simply needs to read the notification and adjust the parameters, preventing wear and malfunctions of machine components.