From Noise to Knowledge: A Comparative Study of Acoustic Anomaly Detection Models in Pumped-storage Hydropower Plants

2025-09-26

AI Summary

To address the challenge of acoustic anomaly detection in pumped-storage hydropower plants under high-noise conditions and limited labeled data, this paper proposes a predictive maintenance framework tailored to real-world industrial settings. We introduce a novel acoustic dataset integrating both synthetically induced and naturally occurring anomalies, and design a noise-robust time-frequency feature extraction and preprocessing pipeline. We systematically evaluate three representative models: LSTM-based autoencoders, K-Means clustering, and One-Class SVM (OC-SVM). Experimental results demonstrate that OC-SVM achieves the best trade-off between detection accuracy (ROC AUC: 0.966–0.998) and computational efficiency (shortest training time), while the LSTM autoencoder attains the highest accuracy (0.889–0.997), making it suitable for high-reliability applications. To our knowledge, this is the first work to conduct a comprehensive, empirically grounded evaluation of acoustic anomaly detection models in an operational pumped-storage hydropower plant environment.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

In the context of industrial factories and energy producers, unplanned outages are highly costly and difficult to service. However, existing acoustic-anomaly detection studies largely rely on generic industrial or synthetic datasets, with few focused on hydropower plants due to limited access. This paper presents a comparative analysis of acoustic-based anomaly detection methods, as a way to improve predictive maintenance in hydropower plants. We address key challenges in the acoustic preprocessing under highly noisy conditions before extracting time- and frequency-domain features. Then, we benchmark three machine learning models: LSTM AE, K-Means, and OC-SVM, which are tested on two real-world datasets from the Rodundwerk II pumped-storage plant in Austria, one with induced anomalies and one with real-world conditions. The One-Class SVM achieved the best trade-off of accuracy (ROC AUC 0.966-0.998) and minimal training time, while the LSTM autoencoder delivered strong detection (ROC AUC 0.889-0.997) at the expense of higher computational cost.

Problem

• Addressing acoustic anomaly detection in hydropower plants under noisy conditions
• Comparing machine learning models for predictive maintenance using real-world datasets
• Improving anomaly detection accuracy while managing computational costs effectively

Innovation

• Preprocess acoustic data under noisy conditions
• Extract time and frequency domain features
• Benchmark LSTM AE K-Means and OC-SVM models