BigData@MA Project
Enhancing maintenance of mechanical assets in the manufacturing sector
| Business | Period | Project Coordinator | Funding Scheme |
|---|---|---|---|
| Industry | 2018 - 2020 | RINA | MANUNET |
Challenge
The BigData@MA Project aimed to develop a specialized Big Data framework tailored for the manufacturing sector. This framework was designed to efficiently process large volumes of structured and unstructured data within a short processing time. Additionally, the project focused on creating online analytics tools and predictive models specifically designed for maintenance of mechanical assets, such as motors. The proposed solutions were successfully implemented and tested in real industrial environments.
Utilizing Big Data technologies, which had already been extensively applied in sectors like finance and telecommunications, allowed for the integration of high-dimensional data (such as pictures, videos, and sensor data) with complex event processing. This integration enabled the acquisition of significant information, detection of anomalies in plants/processes/products, and real-time decision support. Many professionals and companies in manufacturing production, especially those dealing with complex processes, could benefit from these advancements. However, adaptations of existing Big Data technologies were necessary to accommodate the unique characteristics of manufacturing processes, including batch, semi-continuous, and continuous operations, complex production routes, and diverse machinery in different plant settings, while ensuring a wide variety and high quality of products.
Our role in the project
RINA served as the project coordinator, utilizing their extensive knowledge of processes, particularly within the steel industry, and expertise in data analysis and modeling. Based on this foundation, we developed analytic tools and maintenance models specifically for the rolls grinding process, addressing the requirements of the TENOVA use case. These tools and models provided valuable support to TENOVA's customers in their industrial activities.
Conclusions
Through the grinder machines use case, the project successfully developed a strategy to transition from scheduled maintenance to predictive maintenance. This transformation was made possible by establishing a robust data collection infrastructure that incorporated data from the grinding process and various sensors. Furthermore, algorithms based on Artificial Intelligence (AI) were developed and validated for data analysis, and a cloud infrastructure with a Human-Machine Interface (HMI) was implemented for real-time machine status monitoring.
These advanced technologies, in conjunction with real-time monitoring of Key Performance Indicators (KPIs) of the process, enabled the detection of component anomalies before a breakdown occurred. This capability allowed for estimating the residual useful life of machine parts, which was crucial as equipment failures not only incurred repair costs but also led to production losses.
The effectiveness of the work was validated in a pilot plant, where technicians provided machine availability and valuable suggestions for creating a product that could be immediately utilized in the roll shop. As a result of this project, newly acquired machines were equipped with sensors necessary for predictive maintenance, effectively reducing downtime caused by unexpected failures.
