Configurable Distributed Data Collection and Programmable Analytics: Netcompany-Intrasoft’s Approach to Modular Manufacturing in the Horizon Europe Modul4R project
For over a decade, the advent of Industry 4.0 and the digital transformation of manufacturing operations have opened new horizons to manufacturing flexibility. One of the most prominent approaches to increasing the flexibility of production operations is the so-called “modular manufacturing”, which is a novel manufacturing process that involves production systems that are can be divided into smaller, self-contained modules which can be flexibly combined and deployed into different configurations to support different production processes. These modules can be easily interchanged or replaced, which boosts production flexibility and customization.
One of the main goals of the Horizon Europe Modul4R project is to produce a novel platform that lowers the barrier for manufacturers and providers of industrial automation solutions to develop, deploy and operate modular manufacturing solutions. The project breaks down complex cyber physical production systems into many different building blocks including networks, automation devices/protocols, data management systems, internet-connected devices, sensors, machine learning and artificial intelligence systems, as well as enterprise systems (e.g., Enterprise Resource Planning (ERP)). Netcompany-Intrasoft is one of the technical partners of the project, which contributes to the development of the Modul4R platform. Out of the many different levels of modularity and different types of building blocks, Netcompany-Intrasoft’s contribution is focused on providing solutions for modular, configurable, plug-n’-play data collection, data analytics and data mining based on machine learning. In this direction, Netcompany-Intrasoft has been designing and developing a configurable solution that integrate three different modules (“building blocks”), including:
Configurable data collection and routing of data streams from diverse sources. This module collects streams for different data sources and routes them to different consumers according to its configuration. By properly configuring this module, one can access streams/data from different data sources and route them to different consumers (e.g., industrial applications or analytics modules).
Configurable (“plug n’ play”) distributed data analytics over the heterogeneous data streams. This module applies different analytics functions over one or more data sources. The configuration of the module is based on a Domain Specific Language (DSL). By properly configuring this module, it is possible to apply different types of data analysis in-line with the requirements of the digital manufacturing application.
Configurable machine learning modes over data raw or pre-processed data streams. This module applies different machine learning models over data streams or analytics results (i.e., results of the configurable data analytics functions). By properly configuring this module, providers of industrial solutions can deploy, use and evaluate the effectiveness of different machine learning algorithms in terms of their ability to support industrial functionalities like defect detection, Remaining Useful Life (RUL) calculation and more.
These three types of modules enable a three-tier modular approach that uses three different “lego-like” plug-n-play modules to support diverse data-driven production configurations stemming from different requirements.
The work has leveraged and extended background work of the partners in configurable distributed data analytics, which are incorporated on the DataCrop platform for Industrial Internet of Things (IIoT) data management (see [Kefalakis19] and [Christou22]). A prototype of the modular solution is already available for integration in the Modul4R platform and deployment/use in the pilots of the project.
References
[Christou22] Ioannis T. Christou, Nikos Kefalakis, John K. Soldatos, Angela-Maria Despotopoulou, End-to-end industrial IoT platform for Quality 4.0 applications, Computers in Industry, Volume 137, 2022, 103591, ISSN 0166-3615,
[Kefalakis19] Kefalakis, N.; Roukounaki, A.; Soldatos, J. Configurable Distributed Data Management for the Internet of the Things. Information 2019, 10, 360. https://doi.org/10.3390/info10120360
