Most projects that have a velocity have it from a model, usually HEC-RAS or an equivalent. Using that number well is mostly a matter of knowing what it is.
What a one-dimensional model gives you
A 1D model solves flow along a channel using cross-sections. At each section it produces water surface elevation, discharge, and velocity — reported as an average through the section or through defined subdivisions of it.
That is a genuinely useful result and it is not a local velocity at a structure. See depth-averaged vs local velocity.
What a 1D model does not resolve:
- Velocity distribution across the section, beyond whatever subdivision was defined
- The flow field around a pier or an obstruction
- Secondary currents on a bend, which is the mechanism that attacks the outer bank
- Local turbulence, including the downflow and horseshoe vortex at a structure
- Three-dimensional effects anywhere
What that means practically
The model output is the approach condition. Getting from there to a design velocity at the armor generally means either applying a location multiplier from published guidance, or running a 2D model that resolves the distribution directly.
Skipping that step and using the section average as the design velocity under-sizes the protection, and it does so at exactly the locations where the consequence is highest.
Questions to ask about the model
Was it calibrated? Against what — a gauged record, an observed high-water mark, a historical event? An uncalibrated model is a geometry exercise with an assumed roughness.
How old is the geometry? Cross-sections from a survey predating channel change describe a channel that no longer exists. On a degrading or migrating reach this can matter within a few years.
What roughness was used, and on what basis? Velocity is sensitive to it.
What discharge and return period? And is the hydrology current? Design discharges get revised, and a model built on a superseded estimate is solving the right equations for the wrong flood.
Does the geometry include the structure? A model built before a bridge was there, or one that schematises it crudely, is not representing the contraction.
Where in the section is the reported velocity? Whole-section, or a subdivided panel?
Was the model built for this purpose? A model built for floodplain mapping is optimised for water surface elevations. Velocities may be adequate for that purpose and not for armor sizing.
That last question is the one most worth asking. Reusing an existing model is sensible and cheap; using it for something it was not built for is where errors enter.
When a 2D model earns its cost
- Complex geometry — braided channels, wide floodplains, multiple openings
- Bends, where the cross-channel distribution is the whole question
- Around structures, where the local field matters
- High-consequence sites, where the cost of under-sizing exceeds the cost of the model
- Where a multiplier approach is being stretched beyond the conditions the guidance covers
Documenting the input
Whatever the source, the design velocity should arrive with its pedigree:
- The value, and its units
- Whether it is section-average, depth-averaged, or local
- The model, version, and whether it was calibrated
- The discharge and return period
- Any multiplier applied, and its source
- The location it describes
That record is what lets someone else check the design, and it is what a reviewer will ask for. See assembling the submittal package.
Where the product figures live
Rated conditions per size are on product specifications, and the calculator applies them. This article does not restate them, deliberately: the site maintains those figures in one place with their conditions attached, and copying them across articles is how a number gets separated from what it depends on.
The standing position
FES is a supplier, not the engineer of record. Building, calibrating, and interpreting a hydraulic model is the engineer’s work. What we can do is tell you what the sizing consumes and be clear about where our own data stops — see when not to use rock bags.