Anna-Karin RyanEfficiency, reliability, and profitability are the cornerstones of success in the renewable energy sector. A high degree of automation is pivotal for delivering swift, high-quality inspection data, but our journey doesn’t end there. Florian Trautweiler, Software Developer and AI specialist at Sulzer Schmid, sheds light on the crucial role of AI in Blade Anomaly Detection and talks about our collaboration with Microsoft and the utilization of Azure Machine Learning Studio to craft a pioneering AI tool.
The Importance of Regular Blade Inspections
Regular inspections of wind turbine rotor blades are imperative for ensuring safety and operational excellence in the renewable energy sector. Early detection of damages enables proactive measures, optimizing performance and reducing costs.
With instant access to this inspection data, wind park owners, operators, and turbine manufacturers can make well-informed decisions regarding the timing of repairs, ultimately improving annual energy production by minimizing downtime.
Automation and AI integration for a Seamless Blade Inspection Process
Florian Trautweiler, Software Developer and AI specialist at Sulzer Schmid, explains how automation and AI comes into play: “Our customers want a seamless inspection process and fast results. They typically have a small window to inspect their assets before the repair season starts. Our commitment to efficiency and cost-effectiveness to meet their needs is underscored by a high degree of automation. This includes the use of autonomous drones, eliminating the need for manual piloting. Additionally, AI-supported software tools process securely uploaded data in the cloud, detecting anomalies efficiently.”
State-of-the-Art Wind Turbine Blade Platform Supported by Microsoft
From the outset, Sulzer Schmid made a strategic choice to harness the capabilities of Microsoft Azure and Power BI for their 3DX™ Blade Platform. Through seamless collaboration between experienced teams and the unwavering support of Microsoft experts, the company has maximized the potential of Azure and Power BI. The result is a state-of-the-art software solution and an intuitive dashboard that epitomizes the 3DX™ Blade Platform.
Elevating Blade Anomaly Detection Capabilities with Azure Machine Learning Studio
Building on this successful relationship, Sulzer Schmid extended its collaboration with Microsoft to elevate its blade anomaly detection capabilities. The company embarked on designing a market-leading AI tool for damage detection that would ensure not only a faster annotation process but also deliver annotations of higher precision and accuracy.
Florian Trautweiler explains the process: “With AutoML in Azure Machine Learning Studio the process of time-consuming, iterative tasks of machine learning model development is automated. Instead, it allows our data scientists to build models with high scale, efficiency, and productivity all while sustaining model quality. Therefore, we could focus on our key differentiators, our platform, and the data. This tremendously shortened the time needed from initial ideation to full production.”
Results: Improved Efficiency, Accuracy and Speed
Florian Trautweiler elaborates further: “The implementation of Sulzer Schmid’s AI-driven anomaly detection system has yielded remarkable improvements in efficiency, accuracy, and speed. Our cutting-edge AI engine boasts an exceptional accuracy rate, identifying over 99% of critical damages automatically. This translates into a significant reduction in manual effort and time spent on reviewing inspection data.”
To guarantee the highest level of precision and excellence in annotating inspection data, Sulzer Schmid complements their AI capabilities with a dedicated team of in-house blade experts. Every inspection is manually reviewed, ensuring unparalleled annotation quality, and setting a new standard in annotation services and quality within the industry.
Leveraging AI for Ongoing Advancements in Rotor Blade Anomaly Detection
The specialized AI model, which is developed with the help of Microsoft technology, is tailored to Sulzer Schmid’s unique use case, and enhances accuracy, speed, robustness, and efficiency in their inspection process. With its recent implementation, Florian looks into the future and explains: “The potential of AI for wind Operations and Maintenance (O&M) is immense and clearly reflected in our future development road map. We are focused on further improving and automating our inspection technology, optimizing our Azure-based software platform to deliver the tools our customers need to turn inspection data into actionable insights and harnessing the power of AI within our processes.”
He continues: “We currently have +700 000 annotations in our platform, and with each inspection, we gather more valuable data to further train and refine the AI engine. Azure Machine Learning Studio plays a pivotal role here, as it enhances our AI engine’s speed, efficiency, and accuracy, enabling customers to receive results faster and make informed decisions based on superior data quality.”
Working Towards an Even More Efficient Wind Turbine Blade Inspection Process
In terms of future developments, Sulzer Schmid is pushing the boundaries and exploring avenues to use AI for automated damage classification and to generate AI-driven repair recommendations. Additionally, AI will play a crucial role in identifying blade erosion with unparalleled accuracy, allowing customers to predict power loss and strategically plan repairs for optimal efficiency and timely action.
Sulzer Schmid also want to use AI to further automate and expedite their rigorous image quality check on-site, ensuring image quality meets the highest standards and lays a solid foundation for damage detection, minimizing the need for costly and time-consuming re-inspections.
Moving On With Valuable Support From Microsoft’s AI Specialists
Reflecting on their collaborative journey with Microsoft, Florian shares his experience: “By being part of the Microsoft for Startups Program, we have been given direct access to highly knowledgeable specialists throughout various software development projects. As we started the AI project, they immediately offered dedicated support through their startup mentors. In monthly meetings, we had the opportunity to discuss our progress and burning topics. Additionally, their swift response to ad hoc requests has been very valuable in the process of developing and implementing our new AI-based anomaly detection for rotor blades. We look forward to continuing our journey together and taking the next steps towards an even more efficient inspection process meeting our customer’s needs.”
For more information on our inspection technology and platform use the contact form below.