Dual Unscented Kalman Filter Architecture for Sensor Fusion in Water Networks Leak Localization

Leakage in water systems causes significant daily losses, degraded service, increased costs, and environmental problems. Efficient leak localization is crucial to minimizing these impacts. Most leak management methods rely solely on pressure sensors due to their reduced cost and ease of installation, potentially missing other sources of valuable data. This article proposes a data-driven hydraulic state estimation methodology based on a dual unscented Kalman filter (UKF) approach, which enables the fusion of pressure, flow, and demand measurements. In this way, the presented method enhances the estimation of both nodal hydraulic heads, critical in localization tasks, and pipe flows, useful for operational purposes. The strategy is evaluated in well-known open source case studies, namely Modena and L-TOWN. In terms of interpolation accuracy, the average performance in both scenarios shows a ~52% error reduction in pressure estimation if comparing with methods only using pressure sensors, and a ~24% error reduction in flow estimation if comparing with a nondual UKF implementation. Moreover, leak localization results in L-TOWN show an average distance error reduction of ~14%, with impressive results in a multileak scenario where a ~68% reduction is reached.