Search

• An LSTM-GAN generative model is employed for enriching the leak detection dataset. • The model captures the time-series features of leak signals in WDNs. • The performance of LSTM GAN is evaluated through a series of validation approaches. • The method enriches the leak signals and enhances the water leak detection accuracy. Water distribution networks (WDNs) experience significant water loss due to leaks, necessitating advanced water leak detection methods. However, machine learning-based acoustic method heavily relies on signal information and is limited by data scarcity and the limited diversity of available data. To address this challenge and enhance water leak detection in WDNs, this study proposes an LSTM-GAN approach. Acoustic signals are collected from WDNs to train the LSTM-GAN model, which generates synthetic leak signals to enhance the dataset. The validity of the generative method is evaluated through t-SNE and acoustic characteristics analysis. LSTM-based water leak detection models are established and compared using the original and the generated datasets to confirm the efficacy of generated samples in improving water leak detection performances. The capability of LSTM-GAN has been evaluated through different perspectives, including sensitivity analysis and model comparison. The results validate the quality and consistency of the generated acoustic signals under leak conditions. Besides, the optimal number of generated samples should be determined according to the requirements and characteristics of the leak detection task. Furthermore, the comparison between the proposed method and other acoustic generative methods demonstrates the superiority of LSTM-GAN-generated signals in enhancing the performance of leak detection models. The proposed generative method offers an innovative approach to facilitate machine learning-based leak detection models with limited data, thereby enhancing robustness. [Display omitted] [ABSTRACT FROM AUTHOR]

● Essential factors are discovered for the deterioration modeling of water distribution pipelines. ● A new machine learning approach is proposed to predict the failure risk of water pipelines. ● The proposed method can accommodate limited, high-dimensional, and highly censored data. ● The advantages and drawbacks of the proposed method and other state-of-the-art machine learning paradigms are discussed. Population growth and urbanization worldwide entail the need for continuous renewal plans for urban water distribution networks. Hence, understanding the long-term performance and predicting the service life of water pipelines are essential for facilitating early replacement, avoiding economic losses, and ensuring safe transportation of drinking water from treatment plants to consumers. However, developing a suitable model that can be used for cases where data are insufficient or incomplete remains challenging. Herein, a new advanced meta-learning paradigm based on deep neural networks is introduced. The developed model is used to predict the risk index of pipe failure. The effects of different factors that are considered essential for the deterioration modeling of water pipelines are first examined. The factors include seasonal climatic variation, chlorine content, traffic conditions, pipe material, and the spatial characteristics of water pipes. The results suggest that these factors contribute to estimating the likelihood of failure in water distribution pipelines. The presence of chlorine residual and the number of traffic lanes are the most critical factors, followed by road type, spatial characteristics, month index, traffic type, precipitation, temperature, number of breaks, and pipe depth. The proposed approach can accommodate limited, high-dimensional, and partially observed data and can be applied to any water distribution system. [Display omitted] [ABSTRACT FROM AUTHOR]