How Anglian Water turns demand forecasts into a supply plan

Forecasting tomorrow’s water demand still leaves an operational question: which sources and treatment plants should supply it, and how should water move between them? A source can have water available while a transfer route is full or a treatment asset is offline. The plan has to work across the connected system.[2]

Anglian Water adopted SUEZ’s Aquadvanced Water Supply Production Planning module to coordinate those decisions. SUEZ’s case study reports that its trial generated usable plans in minutes where earlier work took hours or days. It gives no distribution of test times or equivalent workload definition; the timing is a supplier-reported trial result.[1]

First balance the water

The system combines AI forecasting and optimisation with a mass balance: an account of water entering, leaving and remaining within the network. SUEZ describes inputs including historical and current demand, abstractions, storage and infrastructure capacity. Forecast demand supplies a future requirement; the balance constrains how the system can meet it.[1]

Consider a simplified storage tank starting with 1,000 cubic metres. If 150 cubic metres enter during a period and 200 leave, it ends with 950. That illustration ignores other inflows and losses. An algorithm cannot propose an ending volume of 1,100 without identifying another source of water or changing an outflow.

The production module’s documentation describes comparing scenarios against reservoir levels, demand forecasts and operating constraints. It also considers annual abstraction limits and seasonal changes. A plan that looks efficient for one period can consume a resource needed later; modelling several horizons makes that tradeoff visible.[2]

Then add the limits operators actually face

Optimatics’ earlier trial record describes the practical requirements: changing production costs and capacities, asset maintenance schedules, and abstraction permissions that apply over different durations. These constraints restrict the combinations of sources, treatment and transfers available to the planner. They make the difference between a desirable supply total and a schedule the network can carry out.[3]

The trial developed in two phases. Optimatics and SUEZ first modelled part of the trial area using requirements agreed with Anglian Water. The utility constructed test scenarios from historical data. Joint workshops helped refine the model and train staff; Anglian then extended it to the trial’s full geographic scope, adding fixed and variable constraints to represent physical and operational limits.[3][4]

That record anticipated initial rollout in late 2024, followed by wider use as the Strategic Pipeline entered service. The planning programme therefore predates the latest case study. Its expanding scope connects the software to the physical network: additional assets create additional supply routes and constraints for operators to coordinate.[3]

One plan across the teams

In SUEZ’s case study, Anglian supply manager Adrian Hinchliffe describes supply, network, tactical operations and abstraction-licensing teams working from shared instructions. The benefit he identifies is coordinated operation, especially during summer demand.[1]

A rapid recalculation can support that coordination when an asset becomes unavailable or a demand assumption changes. Its value lies in showing what else must change with it: another source’s output, a transfer, or the use of stored water. The forecast provides an expectation. The connected model gives operators a way to turn that expectation into a coherent set of actions.

Report an error or suggest a correction