Water Science & Technology, Journal Year: 2024, Volume and Issue: 89(11), P. 3147 - 3162
Published: June 1, 2024
Real-time and model-predictive control promises to make urban drainage systems (UDS) adaptive, coordinated, dynamically optimal. Though early implementations are promising, existing algorithms have drawbacks in computational expense, trust, system-level coordination, labor cost. Linear feedback has distinct advantages interpretation, coordination. However, current methods for building linear controllers require calibrated software models. Here we present an automated method generating tunable that only system response data. The controller design consists of three main steps: (1) estimating the network connectivity using tools causal inference, (2) identifying a linear, time-invariant (LTI) dynamical which approximates network, (3) designing tuning based on LTI approximation. flooding safety, erosion prevention, water treatment performance evaluated across 190 storms separated sewer model. Strong results suggest knowledge required effective, safe, UDS is surprisingly basic. This allows near-turnkey synthesis solely from sensor data or reduction process-based
Language: Английский