INDUSTRIAL HEALTH MONITORING

09:00 – 09:20

Registration and coffee

09:20 – 09:30

Wärtsilä Dr Tero Frondelius

Welcome speech

09:30 – 09:50

Nome Tommi Helander

High-Frequency Measurements in Condition Monitoring: From Vibration to Acoustic Emission. This presentation provides an overview of high-frequency measurement techniques used in modern condition monitoring. It introduces key methods such as vibration analysis, airborne and structure-borne sound and ultrasound—each offering unique insights into the health of mechanical systems. The focus is on how high-frequency data can reveal early-stage faults, such as bearing failures, frictional anomalies, or tool wear, often invisible to traditional low-frequency monitoring. The session will also highlight practical applications, sensor technologies, and signal processing strategies that enable predictive maintenance and real-time diagnostics in demanding industrial environments.

09:50 – 10:10

Wärtsilä Rami Annila

Wärtsilä SOUND: Laying the Foundation for Predictive Engine Health Monitoring. The project aims to revolutionize engine diagnostics by developing a system that listens to engine acoustics and detects deviations from normal sound patterns—before they escalate into critical failures. By capturing high-resolution sound data and synchronizing it with engine cycles and power levels, the system creates a digital acoustic fingerprint of engine behaviour. This foundational capability not only enhances failure prevention but also generates structured data streams that feed into advanced models for predictive maintenance. This presentation introduces the vision, architecture, and early outcomes of the SOUND DIGITAL TWIN initiative, setting the stage for deeper insights into Wärtsilä’s intelligent sound monitoring ecosystem.

10:10 – 10:30

Wärtsilä Pasi Halla-Aho

Counterweight Measurements Device. A highly flexible, custom-developed smart measurement device was created to meet Wärtsilä’s specific needs for monitoring vibrations inside medium-speed internal combustion engines. Designed for harsh environments, it operates reliably in oil, high temperatures, and intense vibration. The device performs real-time signal processing and analysis onboard, transmitting results directly to monitoring platforms. Its low power consumption enables extended operation, making it ideal for long-term testing, product development, and troubleshooting.

10:30 – 11:00

Coffee break and Nome product introductions/demos

11:00 – 11:20

Wärtsilä Tero Kujamäki

Precision in Motion: Acoustic and Vibration Monitoring for Smart Machining of Critical Engine Components. This presentation explores how real-time vibration and sound monitoring is transforming the machining of critical engine components—such as connecting rods—into a data-driven, quality-assured process. By integrating tri-axial accelerometers and sound sensors into the final machining stages, Wärtsilä is able to detect anomalies, optimize cutting tool performance, and ensure surface quality in high-load areas. These insights feed into a digital feedback loop that supports predictive quality control, reduces reliance on manual inspection, and ensures that only reliable components enter the engine.

11:20 – 11:40

Wärtsilä Dr Jalal Torabi

Ultrasonic inspection with CIVA for engine components. At Wärtsilä, we have an advanced robotized immersion ultrasonic testing facility for some engine components, and simulation of this ultrasonic inspection can significantly improve our understanding of results from real field measurements. So, the applications of CIVA software in the simulation of ultrasonic inspection of engine components are the main topic of this talk, where I start by explaining different features in CIVA and then show some use cases. Some comparative results will be presented for some engine components, like the W32 connecting rod and big end bearing housing, to see the accuracy and capabilities of the simulation.

11:40 – 12:00

University of Oulu Saana Bergman

Bayesian approach to defect characterization in ultrasonic non-destructive testing. Bayesian methods are combined with an efficient approximate model for ultrasound measurement with the aim of developing a probabilistic method for defect characterization. This novel method carries the potential to significantly improve defect sizing, especially in defect cases such as non-metallic inclusions or unfavorably oriented cracks, where the methods currently in use lead to unconservative size estimates. The information provided by this method in the form of probability distributions for the size and position of the defect can be used directly in fatigue design and service life estimation of Wärtsilä components.

12:00 – 13:00

Lunch break

13:00 – 13:20

Wärtsilä Aamos Vaara

AI improved immersion ultrasonic defect detection. We propose a method for automating defect detection in ultrasonic testing data by leveraging Wärtsilä’s extensive archive of scan data. The approach utilizes 3D semantic segmentation models, drawing inspiration from recent advances in medical imaging, and is trained on millions of simulated examples to enable precise, automated defect identification. By combining segmentation probability maps with critical size thresholds derived from finite element modelling (FEM), the system also enables automatic assessment of defect criticality, significantly reducing the need for human intervention.

13:20 – 13:40

TAU Repekka Kovanen

Smart Diagnostics for Hidden Damage: Acoustic Emission Monitoring of Fretting. A practical and scalable method for detecting fretting-induced damage in large mechanical components using acoustic emission (AE) technology. The study demonstrates how AE can identify wear and fatigue mechanisms in running conditions in real time—without disassembling components. By correlating AE signal parameters with damage types such as adhesive and abrasive wear and cracking, the method enables early detection of critical issues. The approach is especially suited for large, flat-on-flat contacts under harsh conditions, offering a non-invasive, cost-effective tool for predictive maintenance and condition monitoring in demanding industrial environments.

13:40 – 14:30

Coffee break and Nome product introductions/demos

14:30 – 14:50

Nome Juha Hautala

Machine Heartbeat is a highly adaptable monitoring solution developed to meet the unique needs of industrial processes and machinery. Inspired by how doctors interpret heartbeats, it captures and analyzes sensor data—such as sound, temperature, and force—to detect anomalies and assess machine health. The system is modular, supports various sensor types, and can be easily expanded or retrained with machine learning algorithms. Its flexibility, ease of installation, and real-time insights make it ideal for predictive maintenance and process optimization.

14:50 – 15:10

Wärtsilä Toni Hakkarainen

From Sound to Insight: Intelligent Acoustic Monitoring for Engine Health in Harsh Environments. This presentation showcases a flexible, custom-developed acoustic monitoring solution tailored for Wärtsilä’s engine platforms. By combining real-time sound capture, synchronized signal processing, and onboard data analysis, the system enables early detection of anomalies in medium-speed internal combustion engines. Leveraging structured data streams—including high-resolution .wav recordings and angle-segment-based .txt monitoring—the system provides deep insight into engine behaviour across varying power levels and fuel types. Integrated with cloud-based infrastructure, this approach supports predictive maintenance, reduces downtime, and enhances product development through intelligent, scalable diagnostics.

15:10 – 15:20

Closing words and wrap-up. End of the event.

Limited seats available

We have limited room for participants. Last registration date is
15th of September 2025.

Register here