Flow Head Boundary

Overview

The Flow Head Boundary is essentially a rating curve that describes the relationship between flow (discharge) and water level (stage) at a specific location in a river or channel network. It serves as a boundary condition for hydraulic models, providing information about how flow varies with changes in water level. This information is crucial for simulating water movement and predicting flooding scenarios. The Flow Head Boundary consists of a set of data pairs specifying flows (in cubic meters per second, m³/s) and corresponding water levels (in meters above a reference datum). These data points help define the rating curve.

Theory and Guidance

All end points on the network require a boundary condition. The Flow Head Boundary defines a set of data pairs specifying flows and water levels above datum.

The boundary condition can be interpolated using either linear interpolation or a cubic spline fitted to the data. The choice of interpolation method affects the smoothness of the resulting curve. It’s essential to position the Flow Head Boundary sufficiently far downstream so that inaccuracies in the boundary condition do not significantly impact the results in the area of interest.

The specified maximum and minimum flows and water levels should be greater than those expected during model computations to avoid simulation failures.

The rating curve can be determined from field observations or approximated using normal depth equations. Sensitivity tests help validate its accuracy.

If the normal flow direction is known (e.g., river units upstream), positive flows represent outflows from the system.

There is only one equation as follows:

The discrete form of the equation is:

At intermediate points, either linearization is used:

Or a cubic spline is fitted. You can choose which form of interpolation to apply.

It is important that the maximum and minimum flows and levels specified are greater than those expected to be encountered in the computations, otherwise the simulation engine will fail part-way through the model run.

The rating curve may be determined from field observations. Alternatively, the normal depth equations may be used as an approximation - care must be taken to ensure that an approximate boundary condition is located sufficiently far downstream so that inaccuracies do not affect the results in the area of interest. Sensitivity tests can be used to test this.

Flow-Head boundary conditions can be extracted from previous unsteady model runs using the Tabular CSV tool (Results tab). However, a "looped rating curve" will usually result from an unsteady run and a best fit single valued curve would need to be estimated before importing back into a model.

Note that in many cases a single-valued rating is an approximation and should be positioned sufficiently far downstream such that the approximation in the boundary condition does not affect the accuracy in the area of interest. This can also be checked through sensitivity tests.

Where Flood Modeller cannot determine the normal direction of flow (for example at Junctions), then outflows from the system are expressed as negative flows and inflows are expressed as positive. However, if the normal flow direction can be determined (for example there are River units immediately upstream of the Flow Head Boundary), then the Q in the Flow Head Boundary data should be set to be positive for flow out of the system.

The boundary condition can only be specified at an extreme point in the network.

There is a limit of 50 data pairs for the flow-stage data.

Unsmoothed boundary data may give rise to step changes, spikes or oscillations in flow and/or stage results. Smoothing data prior to computation or use of the spline function is recommended.

If spline is specified as the interpolation method, then a cubic spline is fitted to the data. This will help smooth data where the curvature is high. However, if the data is noisy, or if there are sudden changes in gradient, splines may lead to oscillations and may not be suitable.

Data

Datafile Format

Line 1 : keyword QHBDY

Line 2 : Label

Line 3 : n₁ [,z] [smooth]

Line 4 to Line 3+n₁ - Q_i, h_i

where Label is the Node label at boundary, n₁ is the number of ensuing flow and water level data pairs, z is the elevation above datum (not used), smooth is SPLINE if a cubic spline is to be fitted to the data, or LINEAR to use linear interpolation (If the field is blank then linear interpolation is used).

Q_i : Flow (m³/s) corresponding to water level h_i

h_i : Water level (metres above datum)

QHBDY
UNIT002
         2     0.000    LINEAR
     1.000     0.000
     2.000     1.000