From the Lab to the Cognitive platform: CAPRI sensors for process industry
CAPRI project is approaching the design and implementation of its Cognitive Automation Platform (CAP) for process industry, thanks to results achieved so far by the three CAPRI use cases (asphalt, steel and pharma) at laboratory level.
Nineteen Cognitive Solutions in the three domains saw the light at industrial Labs and they are now almost ready to be installed in the plant and to be integrated into the CAPRI CAP. Technological requirements deriving from the lab implementation allowed to define the CAPRI CAP Reference Architecture (see our recent press release).
The Cognitive Solutions encompass four Smart Cognitive layers:
Smart IoT connection, to capture and acquire Data in Motion;
Smart events processing, thanks to highly configurable edge processors;
Smart knowledge modelling, for semantic interoperability and for implementing cognitive process plants;
Smart decision support, with the collection and implementation of algorithms and tools for decision support and business intelligence.
CAPRI project is not developing traditional sensors: the word “Cognitive Solution” does not indicate simply physical hardware, but any cognitive components able to increase our consciousness of the world.
Cognitive Solutions encompassing the Smart IoT Connection layer are physically installed in the plant and they generate data to be collected in the CAP and further processed; other Cognitive Solutions are the output of cognitive algorithm and mathematical models, running in the CAP, that discover insight from the plant’s data.
Digital Twin Architecture – Steel domain
The Plant Cognitive Digital Twin implemented in Steel domain is the digital representation of both machines and products, including billets, bars and aggregates. It is powered by incoming data from all sensors in the steel plant and the related common process data: innovative Cognitive Solutions have been developed in CAPRI for this purpose (Sensor for product tracking, Sensor for Solidification, Sensor for Product temperature and Scale sensor for scale build-up).
The Digital Twin encompasses three layers and it is input for the Cognitive System for risk and anomalies, developed as well in CAPRI.
Control of the asphalt drum – Asphalt domain
The Cognitive Control of the asphalt drum implemented in Asphalt domain is conceived for obtaining a dry product at the optimum temperature possible, by maintaining the combustion gases temperature at the lowest possible level. The Cognitive Solution automatically combines several indicators coming from the asphalt plant and re-elaborates them according to a cognitive algorithm. Three new Cognitive Solutions developed in CAPRI directly powers it, the Sensor for particle size, the Predictive Maintenance of baghouse system and the Solution of Planning and control of asphalt production, that itself takes data from the new Sensor for bitumen content.
Cognitive Operation Concept – Pharma domain
The Cognitive Operation Concept implemented in Pharma domain informs the operator about potential failures and suggests to him/her potential actions to be performed in order to prevent the occurrence of a failure.
The cognitive algorithm is based on data captured from several sensors available in the pharma plant; it is directly powered by two innovative Cognitive Solutions developed in CAPRI: the Cognitive System for fault detection and the Cognitive prediction of dissolution system, that re-elaborate data coming from either the Sensor for blend uniformity or the Sensor for granule quality and the Sensor for product moisture, implemented in CAPRI as well.