Featured Case · 01
Water Production Measurement & Soft Sensor

The Challenge
Discharge readings did not always reflect real-time water production: storage changes, pump operation, sensor timing and operating transitions all distorted them. Building a reliable production label was part of the modelling problem.
A soft sensor estimates a process value that is difficult to measure continuously from other available operating signals.
My Role
- Mapped evaporation, condensation, storage and discharge into a measurement model the business could read.
- Rebuilt minute-level production labels from anonymised discharge and storage changes, and checked them against independent metering.
- Built features from upstream process signals only, excluding label sources and downstream outcomes to prevent target leakage.
- Validated in time order and on a later operating regime to test generalisation beyond the training window.
Method Flow
- 01Reconstruct the measurement chain
- 02Build a reliable minute-level label
- 03Model upstream process history
- 04Validate forward across operating regimes
- Measurement reliability before model complexity.
- Multi-million-row source dataset with more than 200,000 stable operating minutes.
- Time-ordered evaluation instead of random train/test splitting.
Offline Validation
Measured forward, qualified carefully
Offline validation results; the model had not yet entered production deployment.
Measurement assumptions still need field confirmation, and steady-state rules must be refit on training data before a final benchmark.

