Highlights

• The open source model ABIMO allows for a simple calculation of the urban water balance.

• The deviation from the annual natural water balance can be used to detect hotspots for WSUD.

• Transferability of ABIMO is currently tested between the German cities of Berlin and Cologne.

With 120 Mio. m3 per year of lost water globally, leakages in drinking water distribution networks (WDN) still pose a major challenge to water utilities, furthermore, resulting in a multitude of cascading effects such as operational disruptions, environmental hazard, property damage, and sanitary issues. In the last decades there has been a growing focus on leakage detection within the scientific community leading to the development of numerous computer-based solutions for leakage detection. Despite these developments, practical approaches employed by water utilities in their leak management routines still primarily rely on in-situ acoustic devices in combination with periodic water audits, altogether falling short of ensuring continuous system monitoring and leaving much further potential for leakage reduction. Conclusively, further dissemination and widespread implementation of automatic leakage detection technology in the near future will be paramount to contain water losses and foster robust and climate-resilient water supply systems.

Currently available computer-based technologies for leakage detection can be categorized either as data-driven or model-based, primarily depending on their requirement of a hydraulic model. Algorithms for leakage detection based on hydraulic models may accurately detect the occurrence and location of leakages, yet they are highly sensitive to model inputs and, thus, are required to be well calibrated. On the other hand, data-driven models operating on the premise of anomaly detection merely require data without any anomaly, i.e., leakage, for their calibration. However, these data-driven models cannot compete with the localisation accuracy of model-based leakage detection, as they do not incorporate geophysical information about the underlying WDN. Altogether, while yielding great improvement over in-situ technology, the requirements of automatic leakage detection technology still hamper its practical implementation. While both model-based and data-driven approaches have different requirements, their combination may ultimately enable mitigation of high technical requirements and, thus, enhance its practical applicability, thereby potentially facilitating a more efficacious, robust, and widespread implementation of leakage detection technology in water distribution networks.

In this work, we explore the trade-off between model-based and data-driven leakage detection on the basis of two award-winning state-of-the-art leakage detection algorithms developed by our consortium in previous research, i.e., the data-driven LILA and the model-based Dual Model. Through the integration of both algorithms into a unified application, we aim to mitigate technical barriers and bolster detection robustness. To validate our approach, we quantitatively evaluate its performance regarding false alarms, time-to-detection, and localisation accuracy against the individual algorithms while considering different levels of confidence and availability regarding the input data, i.e., hydraulic model, water demand estimation, and pressure data.

Short-term fecal pollution events are a major challenge for managing microbial safety at recreational waters. Long turn-over times of current laboratory methods for analyzing fecal indicator bacteria (FIB) delay water quality assessments. Data-driven models have been shown to be valuable approaches to enable fast water quality assessments. However, a major barrier towards the wider use of such models is the prevalent data scarcity at existing bathing waters, which questions the representativeness and thus usefulness of such datasets for model training. The present study explores the ability of five data-driven modelling approaches to predict short-term fecal pollution episodes at recreational bathing locations under data scarce situations and imbalanced datasets. The study explicitly focuses on the potential benefits of adopting an innovative modeling and risk-based assessment approach, based on state/cluster-based Bayesian updating of FIB distributions in relation to different hydrological states. The models are benchmarked against commonly applied supervised learning approaches, particularly linear regression, and random forests, as well as to a zero-model which closely resembles the current way of classifying bathing water quality in the European Union. For model-based clustering we apply a non-parametric Bayesian approach based on a Dirichlet Process Mixture Model. The study tests and demonstrates the proposed approaches at three river bathing locations in Germany, known to be influenced by short-term pollution events. At each river two modelling experiments (“longest dry period”, “sequential model training”) are performed to explore how the different modelling approaches react and adapt to scarce and uninformative training data, i.e., datasets that do not include event pollution information in terms of elevated FIB concentrations. We demonstrate that it is especially the proposed Bayesian approaches that are able to raise correct warnings in such situations (> 90 % true positive rate). The zero-model and random forest are shown to be unable to predict contamination episodes if pollution episodes are not present in the training data. Our research shows that the investigated Bayesian approaches reduce the risk of missed pollution events, thereby improving bathing water safety management. Additionally, the approaches provide a transparent solution for setting minimum data quality requirements under various conditions. The proposed approaches open the way for developing data-driven models for bathing water quality prediction against the reality that data scarcity is common problem at existing and prospective bathing waters.