New white paper from Mueller Water Products explores new machine learning techniques for leak detection

Entitled Ensembled-based machine learning approach for improved leak detection in water mains, the white paper from Mueller Water Products presents an acoustic leak detection system for distribution water mains using machine learning methods.

The paper analysis the use of a binary classifier to identify the leak and no-leak cases using acoustic signals. A supervised learning methodology has been employed using several detection features extracted from acoustic signals, such as power spectral density and time-series data. The training and validation data sets have been collected from multiple cities across North America over several months.

The proposed solution includes multi-strategy ensembled learning (MEL) using a gradient boosting tree (GBT) classification model, which has performed better in maximising detection rate and minimising false positives as compared with other classification models such as K-nearest neighbours (KNN), artificial neural networks (ANN), and rule-based techniques. Further improvements have been achieved using a multitude of GBT classifiers combined in a parallel ensemble method called the bagging algorithm. The proposed MEL approach demonstrates a significant improvement in performance, resulting in a reduction of false positive reports by an order of magnitude.