This research addresses the challenge of monitoring large-scale machine halls, particularly in the context of iron making processes. We propose an acoustic sound-based condition monitoring (ASCM) system to detect potential faults and damages in machinery. The study focuses on selecting suitable audio features, integrating physical insights regarding the fault, and determining optimal window lengths for feature extraction. Our fault detection method utilizes outlier detection with a Gaussian mixture model trained on features extracted only from normal operating conditions. We compare conventional audio features with physically motivated features and conduct a window length analysis. The results demonstrate the effectiveness of our approach and the impact of incorporating physically motivated features for fault detection performance.

Monitoring the health of machinery in industrial environments is critical to prevent costly downtime and production disruptions. Acoustic measurements offer a promising alternative to traditional methods like vibration analysis due to their simpler instrumentation. However, accurately detecting fault sounds amidst high background noise remains a significant challenge. Machine learning approaches, for example, require extensive datasets encompassing normal and faulty operation to learn the machine's behavior. In this paper, we propose a different approach by focusing on knocking sounds, which are typical indicators of faults in industrial machinery. We describe these fault conditions using an appropriate signal model and use a General Likelihood Ratio Test as a detector. As demonstrated in this paper, by accurately describing the fault pattern based on a small amount of fault data,very low false positive rates can be achieved, significantly reducing the effort required to collect extensive data sets for faulty machine operation.

Maintenance and condition monitoring of machinery is an essential part of almost every industrial branch. An auspicious approach for machine surveillance is to use an acoustic sound based condition monitoring (ASCM) system. With such a system complex equipping and wiring single machines with sensors gets obsolete. ASCM has seen intensive research for monitoring single machines. In this work we consider aspects for the application of ASCM for monitoring entire machine halls. In contrast to the application of ASCM for single machines, the sound scene offers larger variations for normal operational conditions, making the detection of faulty states more challenging. We present measurements of a machine hall and address these aspects. We then investigate a statistical signal modelling technique to describe the sound scene of the hall and analyse the feasibility for the detection of different fault scenarios. We show that model based detection methods indicate a good detectability of fault states.

Acoustic sound-based condition monitoring (ASCM) systems, which typically utilize machine learning algorithms on established audio features, have demonstrated effectiveness under controlled conditions. However, their application in real-world industrial environments presents significant challenges due to complex and variable soundscapes with high noise and limited fault data. The presence of random interfering sounds and variability in operating conditions can lead to lower performance and high false-positive rates. To overcome these limitations, we propose a fault detection method that leverages the underlying physical characteristics of the sound signals. By investigating the components of the acoustic signal, we found that fault-related sounds can be modeled as exponentially decaying oscillations. This insight allows for the development of a physically based signal model, setting our approach apart from purely data-driven methods. Using this model, we developed a robust detection method based on a Generalized Likelihood Ratio Test (GLRT). The effectiveness of this approach was validated using both synthetic and real-world data from a steel industry facility. Our results demonstrate that the proposed model-based approach provides superior performance compared to standard audio features, particularly in high-noise conditions. On real-world data, the GLRT-based approach outperformed all audio features, as clearly shown by the Receiver Operating Characteristic (ROC) analysis. Specifically, the Partial Area Under the Curve (pAUC) of the GLRT is more than twice that of the best-performing audio feature, demonstrating good detection at significantly lower-false-positive rates compared to audio features. Furthermore, simulations showed that our method maintains robust detection down to a Signal-to-Noise Ratio (SNR) of −13 $\mathrm{dB}$, significantly outperforming audio feature-based detection, which was limited to approximately −10 $\mathrm{dB}$. The physically informed nature of our model not only provides a more reliable and robust solution but also enables the method to be generalized to other industrial scenarios with similar fault properties, offering broader applicability for reliable acoustic condition monitoring.

Acoustic condition monitoring is a cost-effective approach to maintain proper machine operation. Defective machines often generate knocking sounds, e.g., due to loose metal parts, which can be modelled as an approximately periodic sequence of acoustic pulses. Based on a signal model for these sequential pulses, we derive a z-transform-based maximum a posteriori ratio test (MAPRT) detector that reliably detects such knocking noises. The proposed detector is tested with measurement data showing that it is more robust than the commonly used generalized likelihood ratio test (GLRT) pitch detector. Furthermore, the proposed MAPRT detector can be implemented efficiently using the chirp z-transform.