/RaspyGeo

Automated geometry scenario runner (loosely analogous to the Geometry Modification tool) for HEC-RAS.

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

RaspyGeo

Automated geometry scenario runner (loosely analogous to the Geometry Modification tool) for HEC-RAS. Depends on Raspy (raspy-auto on PyPI). Currently works for steady-state simulations. RaspyGeo is available on PyPI as RaspyGeo.

Updates and Added Features

If additional capabilities would be useful to you, open an Issue and I will check into it. Otherwise, I develop updates as I need them.

Usage

Scenario running is fully automated through run. The HEC-RAS project must already be set up with a reference (baseline) geometry and a new (scenario) geometry file, and the current plan must be set up to run the scenario geometry. Then, the user specifies:

  • HEC-RAS project path, which needs to be a full path, not relative
  • Input (baseline) geometry file, which can be a relative path
  • Output (scenario) geometry file, which can be a relative path
  • Scenario data output file, which will contain shear, velocity, and depth for main channel and overbanks
  • Data retrieval locations, which is a list of lists: [[identifier, river, reach, rs]], where identifier is the user's name for the spot and the rest are HEC-RAS location information.
  • Number of flow profiles being modeled (steady state)
  • Scenario specification
  • [Optional] HEC-RAS version (default: "507", for 5.0.7; 6.3.1 would be "631", etc.)

The complicated part is the scenario specification. This is set up as nested dictionaries. The outer dictionary is scenarios, where the key is the name of the scenario (do not use any commas). The inner dictionaries have the keys as River Name,Reach Name (no separating spaces, etc), and the values are the actual geometry specifications. These are simply functions that take a Reach object (see hecgeo.py) and return an updated Reach; the user can specify this however they wish.

The default approach is to design functions that can be passed to Reach.adjust_geometry and to specify datum adjustments with Reach.adjust_datums. Both of these modify any cross-sections either within a given stretch of the reach or for the whole reach. Geometry adjustment functions receive coordinates, roughness locations, and bank stations and return the same. In the provided coordinates, the low point is zero elevation and the left edge is station zero; offsets and datums are stored and added back in before writing.

Coordinates are formatted as [(station, elevation)]. Roughness is, as HEC-RAS handles it, applied from the left point going rightwards, and formatted as [(station, new roughness)]. The bank stations are just [left, right].

Specifying geometry modifiers can be quite complicated. Look at geofun.py for some examples. Two modifier functions are provided by default: set_afp and set_lfc, which are tailored to the developer's use case. LFC specifies a low-flow channel, or trapezoidal channel within the main channel. AFP specifies an LFC which then has a wider flat floodplain, active floodplain (AFP), around it. These are both meant to cut/fill into existing geometry and then daylight. It would be much more straightforward to design, for example, a constant shape that ignores the original channel or just adjusts width based on it, etc. The complications mainly have to do with connecting the new design with the old.

Using set_afp, a full example is included below. This would work similarly with the user's own geometry functions.

from RaspyGeo import set_afp, run, parse
import matplotlib.pyplot as plt


rpath = r"C:\Users\dphilippus\longpathstuff\TestProj\test.prj"
ingeo = "TestProj/test.g02"
outgeo = "TestProj/test.g01"
outpath = "output.txt"


def buildscen(bwidth, tdatum):
    return lambda geo: geo.adjust_datums(0, tdatum).adjust_geometry(
        set_afp(bwidth, 1, 2, lambda x: 0.5*x, 3, 0.05, 0.05, 0.035, 0.035))


def prepscens(bwidth, tdatum):
    return {
        "RiverOne,Lower": buildscen(bwidth, tdatum),
        "RiverOne,Upper": buildscen(bwidth*0.75, tdatum*2)
        }


scens = {"Width %d Datum %d" % (w, d): prepscens(w, d) for w in range(1, 6)
         for d in [1, 2, 3]}

# [[identifier, river, reach, rs]]
locations = [
    ['R1U', 'RiverOne', 'Upper', '2200'],
    ['R1L', 'RiverOne', 'Lower', '401'],
    ['TH', 'Thingy', 'OnlyOne', '100']
    ]

nprof = 10


def runner():
    run(rpath, ingeo, outgeo, outpath, locations, nprof, scens)

Bugs

Note that HEC-RAS geometry files can have various optional components that I may not have tested for, and these may produce odd crashes or similar. If you are running into such a bug, please attach the relevant geometry input file to the Issue.

Summary

The idea is to be able to automatically run through (and retrieve results for) many iterations of some geometry scenario in HEC-RAS. This can be partially automated (editing the geometry) with the Geometry Editor, but that built-in tool, while very useful, is also quite limited and still requires a manual edit-run-analyze workflow.

The Geometry Editor is the best approach if you have a handful of scenarios, and RaspyGeo does not aim to compete with it. However, that workflow becomes a major bottleneck when that scales to dozens or hundreds of variations, especially when those require complex modifications to datum, bank stations, etc, followed by exporting and post-processing data. This project is in response to the developer spending far too much time running geometry scenarios for research.

So the plan is, the user specifies:

  • Scenario location (i.e. reach and river stations)
  • Data retrieval location
  • Datum adjustments at both ends; for now, just linearly interpolate the adjustment
  • New geometry
  • Slope-to-daylight

RaspyGeo runs the specified scenarios, then retrieves profile data.