2D grid orientation

[Tea-See]

In some hydraulic models it is necessary to adjust the orientation of the 2D rectangular grid to match the general orientation of the river, and this can have a significant impact on water levels, stability and flow through constrictions as flow between cells is always perpendicular to cell boundaries. I don’t fully understand the underlying mathematical rationale for this, but I suspect it will be something to do with the fact that the flow path in the 2D zones will effectively be increased by virtue of the fact that the water has to take a staggered route in order to stay parallel to the river channel, and this will introduce a lag.

HEC-RAS 5 has cleverly derived the flexible boundary cell which, in combination with the use of breaklines, makes it possible to influence the 2D flow directions. All very clever.

My question is whether the orientation of the rectangular grid is possible in HEC-RAS 5, and whether it will make any difference to the model output. All the examples I have seen appear to be orientated North-South. Does anybody have any experience of this?

[cameron]

The orientation of the cells can make a difference to the results. I have tested this and found it does impact results, but can take a lot of time to get the cells in the correct orientation. One thing to note that you get one N value per cell face so if a cell covers both channel and overbank, it might using the overbank and not the channel n value.

Each study is different so you would need to test to see how much of a change you can get.

Simple answer:
If the cells are left North-South, the model should run faster as HEC-RAS solver simplifies some things. Once you change the cell shape, HEC-RAS solver cannot simplify things and calculations take longer.

[Anonymous]

Cameron is correct in this. One thing that should stand out to everyone reading his last post is that each cell face has ONE n value. This is not as apparent as it seems. For example; when creating a manning’s value grid via a land use shapefile, you are able to select the raster size output file.

So if you select 5 and your computational mesh is 10×10 you may assume that overlapping landuse values over a cell face would share both values of n. but this is not the case

[Author]

Posts