VALIDATION
Discover the Power of UNDERPIN Data Space through Use Cases
Our pilots serve as real-world demonstrators in dynamic asset management and predictive maintenance, spotlighting the full potential of UNDERPIN Data Space. With a focus on refineries and wind farms, these pilots highlight cross-domain data utilization, management methods, analytics, and valuable lessons learned. By exemplifying best practices, they pave the way for future implementation, fostering the expansion of Data Space mechanisms.
These initiatives validate the benefits of industrial data sharing, particularly enhancing operations for SMEs, the primary stakeholders in these sectors. Join us as we showcase innovation and drive impactful change.
Join us as we showcase innovation and drive impactful change.
USE CASE: Smart Predictive Maintenance for Refineries
Modern refineries like Motor Oil Hellas (MOH) generate massive amounts of operational data every day from sensors monitoring pressure, temperature, vibration, and more. But the real value comes not just from collecting data, but from turning it into actionable insight.
This use case focuses on using machine learning to anticipate equipment failures before they happen. Instead of relying on fixed maintenance schedules or reacting to unexpected breakdowns, the system continuously analyzes sensor data to detect early signs of wear, degradation, or abnormal behavior across the entire production chain.
By learning from more than 2 terabytes of data per year, the system delivers:
- Early detection of threshold violations, reducing downtime and production impact
- Proactive maintenance recommendations, improving planning and resource use
- Continuous adaptation to operational realities, thanks to integration with vendor data and on-site observations
- Clean and anonymized datasets to support algorithm development by external partners
The result is a refinery operation that becomes smarter over time, with better resilience, fewer interruptions, and more informed decision-making.
Boosting Uptime. Reducing Risk. Empowering Decisions.
USE CASE: WIND FARMS
Within Wind Turbine Generators (WTGs), sensitive electrical systems and critical mechanical components installed inside the nacelle and the tower are prone to failures. While electrical failures are more frequent, mechanical failures, such as those in the gearbox, bearings, shafts, yaw, and pitch systems, are costly and have a significant impact on WTG performance, leading to prolonged downtimes, high maintenance costs, and production losses.
Leading example in the field of renewable energy production
This use case focuses on implementing a robust, proactive fault detection and predictive maintenance system for wind turbines, driven by advanced machine learning (ML) models. These models are designed to forecast equipment failures and identify emerging patterns of abnormal behavior. To achieve this, data is aggregated from multiple sources – including SCADA systems and maintenance logs.
Anomaly detection
The collected data undergoes pre-processing and quality control procedures to ensure accuracy and consistency. This cleansed data is then used to train machine learning models on historical failure cases and normal operating conditions. Once deployed, the models will monitor turbine operations in real time, continuously updating with incoming data to improve predictive accuracy and adapt to performance drifts in turbine components over time.
Another focus of the use case is the development of a predictive maintenance algorithm specifically targeting the health status of wind turbine blades. In addition to utilizing SCADA data, this approach incorporates external climate and weather data – such as temperature fluctuations, wind patterns, humidity levels, and lightning activity -to enhance the accuracy of health predictions. Integrating environmental data is expected to improve model performance by capturing external stressors that traditional SCADA data alone may overlook. For instance, lightning strikes pose a substantial threat to blade integrity, often leading to internal damage that is not immediately detectable through standard sensor readings. By accounting for such events, the algorithm aims to identify early signs of blade degradation and enable timely maintenance actions, ultimately increasing turbine reliability and reducing long-term repair costs.
The UNDERPIN Data Space is meticulously crafted to be durable, secure, and fully compliant with EU standards. Our goal is to establish new benchmarks for performance, cost reduction, and durability in industrial applications, paving the way for innovation and excellence.