Skip to main content
European Commission logo
English English
CORDIS - EU research results
CORDIS
CORDIS Web 30th anniversary CORDIS Web 30th anniversary

MOdel based coNtrol framework for Site-wide OptmizatiON of data-intensive processes

Article Category

Article available in the following languages:

An innovative AI platform to boost resource and energy efficiency

A data-driven lab enables experts to collaborate, model, develop and evaluate novel predictive control functions in a rapid and cost-effective way.

Process industries, which manufacture and process large batches of material, draw on the intensive use of raw materials and energy. As such, even small efficiency improvements can lead to large savings, in terms of both economic and environmental costs. This is particularly the case for specific industrial processes such as the smelting of primary aluminium, in which the production is organised in different areas including the anode production plant, potlines and casthouse, characterised overall by significant consumption of energy and raw material. One way to drastically improve performance is using predictive modelling, exploiting widespread sensors and data acquisition, and using AI techniques. Yet, the monitoring of production processes remains challenging since it requires the collection and evaluation of large amounts of data; and the adoption of model-based predictive functions is not always feasible at a sustainable cost or with sufficient reliability. The EU-funded MONSOON (MOdel based coNtrol framework for Site-wide OptmizatiON of data-intensive processes) project aimed to rectify this problem, by establishing a data-driven framework and suite of tools to allow experts from across multiple disciplines to work together effectively and enhance the production efficiency of the European process industry. MONSOON uses AI-based predictive approaches to perform plant- and site-wide optimisation of production processes.

Data and development

The MONSOON framework is split into two major components: the ‘Real Time Plant Operation Platform’ and the ‘Cross Sectorial Data Lab’. The Real Time Plant Operation Platform is designed to be deployed onsite, and helps with various functions including: data collection and monitoring of plant-wide resources; dealing with interoperability issues and the management of data flows; detection of problems and faults that could compromise MONSOON platform operations; and carrying out the designed predictive functions – predicting, for example, abnormal behaviours in specific processes. The Data Lab is a collaborative environment, where data flows in from multiple sources to be analysed at scale. This allows: collaboration among data scientists and process experts; advanced modelling; simplified access to data collected from the field; and the design, development and testing of new predictive functions. “Thanks to the MONSOON solution, data scientists involved in collaboration with process experts can develop, evaluate and deploy novel predictive control solutions in a rapid and cost-effective way, create predictive functions with the help of machine learning algorithms and make simulations, possibly using online data from the connected production environment,” says Claudio Pastrone, Head of IoT and the Pervasive Technologies Research Area at the Links Foundation.

Bringing the MONSOON

The MONSOON solution was deployed and tested in two pilot sites, proving near-real-time prediction for improving product quality and machine maintenance. At a GLN manufacturing plant in Portugal, MONSOON was used to improve injection moulding. The other pilot site was focused on carbon and electrolysis areas of primary aluminium production and involved Aluminium Pechiney and a plant from Liberty House in France. Overall, the project addressed aspects related to the digitisation of the plants, data collection, and the creation of specific predictive functions, and fostered teamwork between data scientists and process experts. Under the ongoing digital revolution of Industry 4.0 the MONSOON project could contribute to a wider adoption in European process industries of plant-wide and site-wide monitoring and control. Due to the decrease in process time and resource and energy consumption, the project could also help boost European industry in the worldwide race for competitiveness and sustainability. “I would like to highlight and thank all the MONSOON partners for the great collaboration within the project and also beyond,” Pastrone adds. “I’ve had the opportunity to work with great professionals from all around Europe,” he says.

Keywords

MONSOON, process, industries, manufacturing, tools, sensors, AI

Discover other articles in the same domain of application