Ocean University of China, in Collaboration with AECC South Industry, Pioneers Knowledge Graphs for Digital Twins for manufacturing
Ocean University of China, in Collaboration with AECC South Industry, Pioneers Knowledge Graphs for Digital Twins for manufacturing
As Industry 4.0 in Manufacturing demands ever-greater precision and efficiency, a transformative approach to manufacturing is emerging from the Ocean University of China, in collaboration with AECC South Industry. Their study, published in Scientific Reports on April 14, 2025, introduces a three-layer knowledge graph architecture that enhances digital twin systems, delivering tangible improvements in quality and productivity through integrated data and intelligent decision-making. Tested in the rigorous production of aero-engine blades, this innovation from the Ocean University of China, in collaboration with AECC South Industry, addresses critical challenges in data management and real-time processing, pointing to broader possibilities for data-driven industries.
The Manufacturing Digital Twin Challenge: Mastering Complexity
Digital twins—virtual replicas of physical systems—are vital to smart manufacturing, enabling real-time monitoring, predictive maintenance, and process optimization. However, their potential is often limited by the difficulty of unifying diverse data from sensors, controllers, and legacy systems. As researchers from the Ocean University of China, in collaboration with AECC South Industry, note, “Collecting and integrating heterogeneous data from diverse sensors, controllers, and devices makes constructing an accurate digital model exceedingly complex.” This issue is particularly acute in high-precision sectors like aerospace, where rapid data updates are essential to align virtual models with physical operations.
Copyright images: Source
In fields like aero-engine blade production, where errors measured in micrometers can lead to significant costs or safety concerns, the stakes are immense. Traditional digital twin systems often struggle to process real-time data or adapt to changing production conditions, creating a need for a more robust solution. The Ocean University of China, in collaboration with AECC South Industry, has stepped into this gap with a system that bridges the physical and virtual with precision and agility.
A Breakthrough Framework: The Three-Layer Knowledge Graph
The Ocean University of China, in collaboration with AECC South Industry, has developed a three-layer knowledge graph architecture that strengthens digital twin operations in manufacturing. This model organizes information into three interconnected tiers—concept, model, and decision layers—each addressing specific needs to create a dynamic, intelligent system capable of meeting modern manufacturing demands.
The Concept Layer: A Foundation of Knowledge
The concept layer serves as the system’s cornerstone, building a universal knowledge framework for manufacturing. It relies on ontology libraries—structured collections of domain-specific knowledge—developed with expert insights and technical manuals. By defining concepts, rules, and relationships, this layer ensures standardization and consistency. In the aero-engine blade case study, the Ocean University of China, in collaboration with AECC South Industry, used this layer to establish critical parameters like machining protocols and material specifications, creating a robust foundation for the digital twin.
This layer’s ability to consolidate varied data into a cohesive network is a key strength. “The concept layer provides all the foundational knowledge and frameworks necessary for the model, ensuring the standardization and logical consistency of the knowledge,” the researchers from the Ocean University of China, in collaboration with AECC South Industry, explain. This approach streamlines data management and supports scalable, adaptable digital twins.
The Model Layer: From Concepts to Reality
The model layer translates abstract knowledge into detailed digital representations of physical products. By applying rules from the concept layer, it generates simulations that capture both static attributes, such as standardized machining steps, and dynamic data, like real-time operational conditions. In the aero-engine blade production line, this layer, developed by the Ocean University of China, in collaboration with AECC South Industry, enabled precise modeling of blade contours, ensuring virtual simulations mirrored physical outcomes.
Copyright images: Source
This layer’s fidelity between virtual and physical systems is crucial. By integrating data from IoT sensors and computer-aided engineering tools, it supports accurate predictions and anomaly detection, allowing manufacturers to address potential issues proactively.
The Decision Layer: Intelligence in Action
The decision layer is where the system’s intelligence comes to the fore. In the aero-engine blade case, the Ocean University of China, in collaboration with AECC South Industry, demonstrated its capabilities in dynamic optimization, predictive maintenance, and risk management using real-time data and advanced algorithms. The layer supported tasks like rapid geometric modeling, process simulation, and recommending optimal machining tools based on live conditions.
A key innovation is its real-time data processing architecture, which combines edge and cloud computing to manage high-frequency updates. “The system employs an integrated approach that combines edge computing and cloud computing resources to manage real-time data effectively,” the researchers note. Edge devices handle initial tasks like data filtering, while cloud platforms perform deeper analytics. Tools like Apache Kafka and Apache Spark, referenced in broader cloud computing contexts, were not detailed in this study but underscore the system’s reliance on scalable frameworks. This hybrid model ensures low latency and robust performance, keeping the digital twin aligned with the physical environment.
Real-World Impact: Aero-Engine Blade Production
The architecture’s effectiveness was proven over five months in 2024 on an aero-engine blade production line. Across 200 batches, the maximum contour error precision improved from 0.073 mm to 0.062 mm, reflecting enhanced accuracy. The product qualification rate also rose from 81.3% to 85.2%, marking a significant quality improvement.
These results stem from the system’s ability to integrate multi-source data, detect anomalies, and optimize processes in real time. The decision layer’s predictive analytics, for example, identified potential bottlenecks, enabling proactive adjustments. “The significant improvements in these two key indicators show that the application of our designed system in the aero engine blade production line was very successful, greatly increasing production efficiency and product quality,” says Yong Han, a corresponding author from the Ocean University of China, in collaboration with AECC South Industry.
Copyright images: Source
The system’s versatility was evident across the manufacturing lifecycle, from design to evaluation. It equipped engineers with tools for geometric modeling, real-time monitoring, and post-production analysis, fostering continuous improvement by feeding insights back into the process.
Beyond Aerospace: Exploring New Possibilities
The aero-engine blade case study highlights the system’s strength in high-precision manufacturing, but its framework could theoretically align with other verticals where data-driven architectures are explored, such as smart cities and logistics. In smart cities, digital twins powered by knowledge graphs might optimize infrastructure like traffic or energy systems by integrating IoT sensor data. In logistics, the system could enhance supply chain efficiency by predicting disruptions and refining routes.
Its scalability and adaptability make it well-suited for such applications. As manufacturing evolves and emerging standards like 6G networks gain traction, the knowledge graph’s modular design, developed by the Ocean University of China, in collaboration with AECC South Industry, allows it to incorporate new data sources and algorithms without compromising performance. This flexibility positions it as a foundation for future-ready industries.
Challenges and Future Directions
Despite its promise, the architecture faces challenges. Integrating heterogeneous data sources risks errors or information loss, and real-time processing requires significant computational resources. The system’s reliance on expert knowledge may also limit its flexibility, necessitating advancements in automated knowledge extraction through AI or natural language processing.
The Ocean University of China, in collaboration with AECC South Industry, is exploring solutions. Semantic data modeling and ontology alignment could improve data integration, while edge-cloud hybrids might enhance real-time capabilities. Automated knowledge extraction could reduce dependence on manual input, broadening the system’s applicability.
Speculatively, integrating quantum computing could further enhance this architecture, accelerating the processing of complex knowledge graphs and boosting predictive accuracy. While theoretical, such possibilities highlight the system’s potential to reshape manufacturing and related fields.
A New Era for Smart Manufacturing
The three-layer knowledge graph architecture, pioneered by the Ocean University of China, in collaboration with AECC South Industry, marks a significant step toward unlocking the full potential of digital twins in manufacturing. By addressing challenges in data integration, real-time processing, and decision-making, it offers a blueprint for intelligent, responsive production systems. Its success in aero-engine blade manufacturing validates its effectiveness, while its adaptability suggests broader applications in data-driven sectors.
As Industry 4.0 advances, this innovation charts a course for greater efficiency, precision, and sustainability. “The three-layer architecture knowledge graph provides a powerful methodological framework for digital twins, making operations more efficient and precise during the digital transformation process,” the researchers conclude. For manufacturers and policymakers, it’s a compelling call to reimagine how we design, manage, and optimize the systems driving our industrial future.
More info here – To submit a story to us address to “the Editor” here
Featured image copyright arstechnica.net
Disclaimer
The information provided in this article is for general informational purposes only and from publicly available sources. While we strive for accuracy, we do not make any representations or warranties, express or implied, regarding the completeness, reliability, or validity of the content. This article does not make any direct claims about specific companies, individuals, or organizations. Any references to reports or external sources are for context and do not imply endorsement or verification of any specific allegations. Readers are encouraged to conduct their own research and seek professional advice before making business decisions. We disclaim any liability for any losses or damages incurred as a result of reliance on the information provided.