AcquaTerra

Know which water network segments need attention next.

AcquaTerra turns pipe network records, maintenance history, failure events, repairs, and external spatiotemporal signals into segment-level risk forecasts for predictive maintenance and capital planning.

Network History Context Risk Action

Plan a backtest See validation model

Train on the past. Hold back the last year. Score the forecast against what actually happened.

Network Model

A traceable network model for utility reality.

AcquaTerra links pipe inventory, valves, service areas, materials, install dates, work orders, repairs, inspections, and failure events into a segment-time graph that can be scored and audited.

Assets: pipe segments, valves, hydrants, zones, material, age, diameter.
History: breaks, leaks, repairs, inspections, replacements, complaints.
Lineage: source, time, quality state, confidence, update cadence.

Forecast

Risk forecasts at pipe-segment level.

Each segment receives a forecast window, risk score, confidence, driver evidence, and recommended action class. The output is built for maintenance and capital teams, not for passive map browsing.

Segment: stable network identity and geometry.
Window: forecast period and refresh cadence.
Drivers: material, age, pressure, weather, soils, roadwork, history.
Action: inspect, monitor, repair, renew, or defer.

Validation

Backtest before trust.

Start with historical data and hold back the latest year. Train on the earlier period, rank segments for the hidden year, then compare the forecast against the breaks, leaks, and repairs that actually happened.

01 Train
Use older network history, repairs, failures, and context.
02 Hold back
Keep the most recent year hidden from model training.
03 Rank
Score every eligible segment for future failure risk.
04 Compare
Measure lift against the real failures and repairs.

Signals

Signals that move failure risk.

Failure risk is not only pipe age. The surrounding world changes the priority of a segment, and those changes can be modeled through space and time.

Network condition

material, diameter, age, pressure, repair frequency, inspections.

Ground and weather

soils, flooding, drought, freeze, heat, slope, land movement.

Street activity

traffic load, road class, pavement cuts, nearby construction.

Operational history

complaints, service interruptions, field notes, prior actions.

Operations

From risk score to maintenance plan.

AcquaTerra packages risk into ranked work, review queues, field checks, renewal candidates, budget evidence, and performance measurement for the systems utilities already use.

Ranked segments

Driver evidence

Field checks

Renewal candidates

Budget scenarios

Outcome tracking

AcquaTerra

Turn water network history into predictive maintenance.

Bring the network, event history, and a holdout year. We will show whether the model would have found the risk before the year unfolded.

Plan a backtest