Vibro-Acoustic Condition Monitoring in Internal Combustion Engines

Vibro-acoustic condition monitoring combines analysis of mechanical vibration and sound waves emitted by an engine to detect incipient faults, assess component health and inform maintenance decisions. By deploying vibration transducers and acoustic sensors on key engine components such as cylinder heads, valve trains and injectors, it is possible to capture non-stationary signal patterns that reflect combustion dynamics, mechanical impacts and fluid–structure interactions. Advanced signal-processing methods—including time–frequency transforms, ensemble filtering and feature-extraction algorithms—are coupled with machine-learning classifiers or physics-based models to isolate characteristic fault signatures. This approach addresses the growing need for real-time, non-intrusive diagnostics in automotive, marine and power-generation applications, reducing downtime, lowering life-cycle costs and enhancing safety across diverse operating regimes.

Research from Nature Portfolio

Recent studies have demonstrated the power of acoustic-based intelligence for fault identification and the integration of data-driven twins for enhanced interpretability. One investigation developed an acoustic-emission framework to capture high-resolution surface vibrations and segment their spectra into optimised frequency bands, using convolutional neural networks to achieve near-perfect recognition of abnormal ignition advance angles under multiple operating loads. Another effort introduced a digital-twin strategy combining physical modelling with an adaptive sparse attention network. By applying a novel soft-threshold filter, this method dynamically highlights decentralised fault features in transient signals, yielding clearer visualisation of valve anomalies and improved training efficiency without sacrificing diagnostic accuracy.

Vibro-Acoustic Condition Monitoring in Internal Combustion Engines publication trend

The graph below shows the total number of articles in vibro-acoustic condition monitoring in internal combustion engines across all publications each year (not limited to Nature Index journals).

Technical terms

Acoustic emission (AE): High-frequency elastic waves generated by rapid energy release during structural or mechanical events, used here to detect combustion or mechanical faults.

Digital twin: A virtual representation of an engine system that merges real-time sensor data with a computational model to simulate performance and diagnose anomalies.

Variational mode decomposition (VMD): An adaptive signal-processing technique that decomposes a non-stationary signal into band-limited intrinsic mode functions for noise reduction and feature isolation.

Mel-frequency cepstral coefficients (MFCC): Features derived from perceptually motivated filter banks that summarise the spectral envelope of acoustic signals for classification tasks.

Sparse attention network: A neural architecture that selectively weights important signal segments, enhancing interpretability by focusing on decentralised fault information.

References

1. Analysis of Changes in the Opening Pressure of Marine Engine Injectors Based on Vibration Parameters Recorded at a Constant Torque Load. Sensors (2023).
2. Acoustic emission-based intelligent identification of piston aero-engine ignition advance angle anomalies. Scientific Reports (2023).
3. Combination of VMD Mapping MFCC and LSTM: A New Acoustic Fault Diagnosis Method of Diesel Engine. Sensors (2022).
4. A digital twin auxiliary approach based on adaptive sparse attention network for diesel engine fault diagnosis. Scientific Reports (2022).
5. Diesel Engine Fault Diagnosis Method Based on Optimized VMD and Improved CNN. Processes (2022).