Conformal Prediction-Driven Adaptive Sampling for Digital Water Twins

TL;DR

This paper introduces an adaptive sampling framework for digital water twins that leverages conformal prediction and LSTM forecasting to efficiently allocate sensors, significantly improving state estimation accuracy with limited resources.

Contribution

It presents a novel combination of conformal prediction and LSTM for real-time uncertainty estimation in water distribution network digital twins.

Findings

Achieved 33-34% lower demand error compared to uniform sampling.

Maintained 89.4-90.2% empirical coverage with minimal additional computation.

Reduced sensor coverage requirement to 40% while improving accuracy.

Abstract

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.