Modeling Delaware Sand Filter

[UHydro]

I am also trying to route a storm hydrograph through a bank of Delaware sand filters, as described in thread789-250097.

I'm not sure I understand the proposed solution. It seems to me the challenge is equalizing flow between the forebay (into which runoff enters) and the filter bay (which is loaded via wier/orifice flow from the forebay. Once the water surface elevation in the filter bay has risen to - or above- the orifice inlet you can no longer drain the forebay via orifice flow. It seems to me you've now got potentially bi-directional culvert flow. Standard software - like Hydrocad- does not seem to have the capability to stop orifice discharge from the forebay into the filer bay - even if the wse in the filter bay > forebay elevation.

I'd be very interested in a functional solution to this, short of CFD model.

 

[psmart]

If you use a tailwater-sensitive routing procedure (i.e. Dynamic Storage-Indication), HydroCAD evaluates tailwater conditions on the orifice or culvert and will reduce or stop orifice discharge from the forebay into the filer bay as the WSE in the filter bay rises to the WSE of the forebay.

Note this is not bidirectional (reversing) flow. That's a more complex situation. It's just standard tailwater where the water continues to flow in the same direction but at a reduced rate.

For details see

http://www.hydrocad.net/tailwater.htm

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[TerryScan]

If you read the previously cited thread, the key points I explained for modeling were:
* one ignores the "forebay" storage ("sedimentation chamber" in the previous thread), since it is typically full of water up to the weirs.
* the limiting flow through the system is perforated outlet pipes - the chambers WILL constantly equalize.

Making a more detailed or complicated model does not always yield better results. Sometimes it is better to ignore negligible components and simplify the model.

One must also look at the sizing guidelines for these filters. There isn't much storage available in the filers for attenuating flow (again, the forebay does not have AVAILABLE storage below the weirs). In looking at the previous thread, one would need many standard Delaware sand filters to handle over an acre of drainage area. The cause of the oscillations in the original thread may have been caused by overloading the capacity of the sand filters.

 

[UHydro]

Yes thanks - I appreciate the filter representation you describe. A little more detail -

I'm trying to model a bank of filters - two sets of 4 filters in series. The runoff is equalized among the 8 forebays (in both series and parallel) through one 6" orifice between each pair of the adjacent forebays. With runoff entering all 8 forebays through the surface grates, flow is fairly equalized to begin with, and rapidly redistributes through the 6" orifices between forebays.

We want to modify the existing bank of filters to introduce runoff from a perforated drain in the subbase of a pervious pavement, rather than through overland flow from the impervious pavement (that will be replaced with pervious). For logistic/constructability reasons, we only want to introduce the subbase drainage into the two forebays on one side of the filter bank, letting that flow equalize over both the adjoining sand filter and ALSO to the next "downstream" forebay. The concern is that the "upstream" pairs of forebays will have a higher head than the down stream pairs - until they all equalize, or until the downstream wse reaches the overflow elevation. We want to route design storms through the system to determine the storm that results in surcharge or overflow of the upstream forebay-filter vaults - or determine that the downstream overflow can prevent surcharge.

I appreciate your perspective on rapid equalization of the head within each forebay-filter pair (since rectangular orifice area >> 6" orifices. So each forebay-filter combination can be modeled as the filter footprint (for exfiltration) with twice the area (filter and forebbay) providing storage as the wse rises above the elevation of the rectangular orifices between forebay and filter (kind of a composite filter-forebay for purposes of accounting for storage within each filter-forebay pair).

You're right about the limited attenuation; not really the purpose of the SF , especially with a 15" overflow. We just need to satisfy ourselves we're not going to create standing water over the forebay grates (and the filters).

Thinking if I had two 6" orifices between adjacent "composite" filter-forebay pairs (one flowing in each direction) that might allow Hydrocad to equalize the flow without the need to reverse flow direction.

Hope that makes sense.
Thanks for your feedback.

 

[psmart]

I fear you're making the model much too complex...

Unless the flow actually changes direction, you don't need the pair of orifices "one flowing in each direction". Furthermore, a reverse outlet will not yield a stable routing unless the ponds have significant storage, which your system does not.

I recommend a single outlet in the direction of the predominant flow. Better yet, since you know that the two volumes will equalize fairly quickly, just model them as a single pond. In fact, your storage is probably negligible in comparison to the inflow volume, so a zero-storage pond is recommended for a stable routing.

If you're primary concern is "standing water over the forebay grates", just do a steady-state head loss calculation through the system. Attempting to route a hydrograph (and that's what HydroCAD is for) is overkill.

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[TerryScan]

If equalization between units is your main concern, I would do a manual check of the inter-chamber orifice capacity. If the orifice capacity is a problem, create more inter-connection with additional holes or an inlet manifold connecting the chambers.

I have designed a similar system, but it was not a retrofit. We were using pervious pavers. Because the flow to the filters had already been filtered by the pervious pavement system, we modified the sand filter to do away with a forebay, and used both chambers in each unit as a sand filter to maximize throughput per unit.