Software tool piaAI provides plant prophylaxis

Published on Aug 1, 2019


In order to improve the OEE (overall equipment efficiency) of automated production plants, it is necessary to increase availability, performance and quality. Industrial analytics helps to predict and prevent production problems. PIA Automation has integrated the software tool piaAI (artificial intelligence) into its assembly lines, significantly improving OEE by establishing plant prophylaxis in three steps.

PIA Automation’s digital strategy for customers’ automated assembly systems consists of three steps, as Claude Eisenmann, Chief Digital Officer of the PIA Group, explains: "First, you need global production visibility. All plant components worldwide are connected to a platform for the analysis and evaluation of data, either as a Fog constellation, i.e. within the customer’s local intranet, or as a cloud solution. Both are possible variants for PIA." The decisive aspect here is to establish worldwide data connectivity and to guarantee access to all collected data.

The second step is global production transparency, in which root cause analysis is used to check which combination of factors or events in the past has led to the current deterioration of OEE. This analysis can be very fine-grained and requires software with highly complex algorithms that can generate this information.

Claude Eisenmann (CDO, PIA Holding)

"piaAI is the successful combination of software engineering and special mechanical engineering."

The third step is a forward-looking analysis - plant prophylaxis. This illustrates what will happen in the plant in the future and how this can be prevented. Stage four, which is still theory for the time being, contains machines that learn from each other and optimise themselves. "This industrialization 5.0 may still be a vision of the future," says Claude Eisenmann, "but PIA is already thinking about it."

Digital production twin with piaAI

piaAI is a new software tool that implements this three-part digital strategy, developed by PIA together with a partner specialising in process-based industrial artificial intelligence. piaAI is based on a digital production twin in which all stations and modules are visible.

In the software display, three tabs are used for the three steps of visibility, transparency and prediction (see illustration):

_ "Overview"

_ "Loss Analysis"

_ "Alerts"

Software engineering meets special mechanical engineering

For Claude Eisenmann, piaAI is "the successful combination of software engineering and PIA's expertise in special mechanical engineering." Test runs conducted so far on a large PIA system show the analytical potential of the tool. It is important to collect a critical mass of data so that the algorithms behind the software have enough input for precise information and predictions, especially in the "Alerts" tab. The more data is stored, the more accurate the predictions.

piaAI can be used not only to equip new PIA production facilities, but also to retrofit existing ones. It is also possible to equip existing systems from other manufacturers. The possibilities offered by piaAI are highly flexible in terms of data connectivity.