Conformal Prediction-Driven Adaptive Sampling for Digital Twins of Water Distribution Networks
This work addresses the challenge of resource-efficient monitoring for water distribution systems, though it is incremental as it adapts existing methods to a specific domain.
The paper tackles the problem of accurate state estimation in Water Distribution Network digital twins with limited sensors by proposing an adaptive sampling framework that combines LSTM forecasting and Conformal Prediction to focus sensing on uncertain nodes, resulting in 33-34% lower demand error than uniform sampling at 40% coverage.
Digital Twins (DTs) for Water Distribution Networks (WDNs) require accurate state estimation with limited sensors. Uniform sampling often wastes resources across nodes with different uncertainty. We propose an adaptive framework combining LSTM forecasting and Conformal Prediction (CP) to estimate node-wise uncertainty and focus sensing on the most uncertain points. Marginal CP is used for its low computational cost, suitable for real-time DTs. Experiments on Hanoi, Net3, and CTOWN show 33-34% lower demand error than uniform sampling at 40% coverage and maintain 89.4-90.2% empirical coverage with only 5-10% extra computation.