How to assess drawdown and pumping rate sensitivity to excavation retention embedment depth

Help & Troubleshooting
1

Purpose:

The objective of the modelling is to assess the relative difference in modelled groundwater drawdown and pumping rate between two excavation dewatering scenarios that differ only in the embedment depth of excavation retention.

  • Scenario 1: Excavation retention embedded 2 m below the excavation base.

  • Scenario 2: Identical model setup, except excavation retention embedded 4 m below the excavation base.

The expected conceptual outcome is that deepening the excavation retention (Scenario 2) should:

  • reduce groundwater inflow to the excavation, and

  • reduce the magnitude and spatial extent of drawdown immediately outside the excavation.

Problem:

Despite the conceptual expectation that deeper excavation retention should reduce inflow and drawdown, model results indicate that Scenario 2 produces greater pumping rates and drawdown compared with Scenario 1.

Expectation is that this behaviour arises because deepening the excavation retention increases the saturated thickness of Layer 1, which in turn increases its transmissivity and therefore increases discharge (Q) in accordance with Darcy’s Law. This effect occurs even with no leaky barrier present.

Base Case Model – Scenario 1:

The base case represents a small-scale, short-duration excavation dewatering scenario, with the following characteristics:

  • Single transient timestep: 30 days

  • Layering and aquifer settings:

    • Two-layer model

      • Layer 1: Unconfined

      • Layer 2: Confined

    • Homogeneous and isotropic hydraulic properties assigned to both layers.

  • Model simplification assumptions:

    • No infiltration or recharge applied.

    • Far-field boundary represented by a head-specified (HSEG) outer boundary at a radius of 1 km.

    • Specific yield assigned to Layer 1 is equal to the storativity assigned to Layer 2.

  • Initial groundwater level conditions:

    • Hydrostatic groundwater level of 8 m elevation in both layers.

  • Excavation geometry and target groundwater level:

    • Excavation base and target groundwater level at 6 m elevation. Implemented using:

      • an internal head-specified line boundary, and

      • a spatially variable area sink (SVAS) head-dependent flux,

      • both applied with identical geometry and coordinates.

  • Model discretisation:

    • Nested SVAS polygons used to discretise the model domain.

    • Finer spacing applied within and immediately surrounding the excavation to better resolve gradients.

  • Excavation retention representation:

    • Retention installed as a leaky barrier boundary condition, offset 0.1 m outside the excavation footprint.

    • Barrier conductance set to 1 x 10-3 day-1.

    • Leaky barrier applied to Layer 1 only, representing partial cut-off of the shallow unconfined unit.

  • Layer elevations:

    • Layer 1 bottom / Layer 2 top elevation: 4 m.

Scenario 2:

Scenario 2 is identical to Scenario 1, except for the depth of excavation retention:

  • Layer 1 bottom / Layer 2 top elevation: 2 m, representing excavation retention embedded an additional 2 m below the excavation base.
2

Kevin - - It’s a bit difficult to discern your setup, but yes deepening the wall around an excavation pit should reduce the flow into the bottom of the pit. For the model you described (it would help to have the pit dimensions) you will probably not see a large reduction in inflow by changing the wall depth from 2 m to 4 m, but you should see some. You will also see more if vertical anisotropy in the layers around and beneath the pit is large.

It is not clear what you are doing with the internal head-specified line boundary”. You may want to try creating a three-layer model; use your second layer to extend your wall depth the additional 2 m for Scenario 2; create a separate pit domain over which you specify a head dependent flux with reference head of 6 m; run the model, and then use Analysis > Vertical Leakage over Polygon Areas to report the flux (and change in flux as you alter the model).

You might also want to not use a leaky barrier element for the retention wall; create a thin domain between the outer pit domain line and the surrounding aquifer (instead of a leaky barrier element); and make sure to use refined zones of SVAS points outside and inside the wall (SVAS points by polygon; make sure your order of applying them is correct).

Hope this helps.

1 Like
3

Hi Kevin. Can you email me the files?

4

Many thanks Charles, I found your response helpful. Apologies for delayed reply as I’ve been on-site most of this week. I did try implement what you’ve suggested, but I could not make your suggested approach work either.

I’ve sent through the model files to Jason today who may be able to provide some additional insights by inspecting the .anaq model itself. Thank you

5

Kevin - - I’m sure Jason will be able to help and probably has already. (Sorry I’m late getting back here.) I have a simple model set up along the lines of what I suggested to you in my post. If you would like a copy please let me know and I will try to get the files to you. [Not sure if we can do file transfers through the Clubhouse site, or whether we will need to exchange emails.]