Flowless

Flowless — Calibrated real-time hydraulic digital twins of operating water networks: continuously updated network models driven by live flow and pressure sensor feeds, paired with the non-revenue-water decomposition each twin produces — losses broken down by type into real losses, night use, customer-side leaks and meter errors, attributed to where and why they occur. The asset is not the raw telemetry (which belongs to the utility) but the calibrated model state and the loss-decomposition answers: a network model continuously reconciled against measured reality, plus the simulation records f…

Flowless offers (Alternative, Geospatial) — Est. modest in raw telemetry but dense in derived signal: a district-metered area network yields tens of thousands of flow and pressure readings daily, and each twin additionally generates prediction-versus-observation pairs at the same rate plus simulation runs per scenario tested. Across low-tens of network deployments the reconciliation stream plausibly reaches hundreds of millions of paired model-measurement records, which is where the value sits rather than in raw reading counts.

Training and calibrating hydraulic network models and surrogate models that approximate them at machine-learning speed (the model-error stream is directly usable for this); physics-informed neural networks for pipe-network flow and pressure; leak-detection and non-revenue-water quantification models with cause-type labels; heatwave and climate-stress resilience models for critical infrastructure; synthetic scenario data for network design and maintenance-planning agents; and reinforcement-learning environments for autonomous network operation, which the product explicitly points toward

The data is with 5 years of history.

Coverage spans US, Other; Application Software; alternative, geospatial; equities.