Overview

In regulated river systems, river operators use water infrastructure to control the flow of water. Irrigators, towns, and other water users submit orders specifying how much water they need. System operators then consider these requests to determine how to operate upstream infrastructure — releasing water from storages, weirs, and managing pumps and channels — so that water arrives where it's needed, when it's needed. They must account for travel time through the system, potential tributary inflows, and transmission losses that may be incurred along the way.

Kalix’s ordering system simulates this process, providing a way for models to represent the operation of infrastructure to meet user demands. The ordering system enables demand nodes (e.g. users and flow requirement nodes) to communicate their requirements to upstream infrastructure nodes (e.g. storages and offtakes) which then respond by making releases and modulating operations accordingly.

Simple Ordering System

Kalix’s simple ordering system follows a heuristic rules-based approach to ordering and operation (c.f. network-optimised approaches available in some platforms). Imperfect anticipation of future inflows, losses, and streamflow routing behaviour can lead to surpluses or shortfalls in water delivery, as occurs in real river systems.

Kalix’s simple ordering system works as follows. Steps 1-2 are done once during the initialisation of the model. Steps 3-5 constitute the ordering phase and are done every timestep. Steps 4-5 constitute the flow phase and are done every timestep.

1. identifying regulated zones,
2. estimating travel times to the nodes within those zones,
3. collecting orders reflecting user demands and other flow requirements,
4. adjusting orders according to expected losses and inflows,
5. directing orders to appropriate infrastructure nodes,
6. simulate operation of infrastructure nodes to deliver orders,
7. simulate water flow through the model network.

This system is similar, but not identical, to the ordering system in other modelling platforms such as IQQM and Source. Some key aspects of the system are:

• User nodes in regulated zones do not send orders by default. They must opt-in to the ordering system by using the regulated = true property.
• The travel time for each node in a regulated zone is an estimate of the streamflow routing lag between the supply (typically a single storage, but caveats are explained below) and the node. The travel time is assumed to be constant throughout the simulation, and is based on the streamflow lag at a typical flow rate typical_regulated_flow = 100.
• Orders can be manipulated using the minimum and maximum order nodes. Orders do not propagate outside the regulated zones.

The simple ordering system is explained in detail below.

Regulated Zones

Areas downstream of storage outlets are designated regulated zones. This storage node is the supply for the zone immediately below it. The zone extends downstream to (a) the next storage that can act as a supply, (b) the end of the system.