Computational fluid dynamics (CFD) modelling at Featherstone gauging station

Location: River South Tyne, Northumberland                                 Client: Environment Agency

Featherstone gauging station on the River South Tyne is a two-stage weir with sloping faces and a central dividing pillar. There is gravel build-up in close proximity to the weir, and concern that the gravel is producing low confidence in the rating curve and inaccurately predicting flows.

Challenge

Normally, under modular weir flow conditions, a theoretical rating curve for a weir structure that is calibrated with spot gaugings would provide a reliable stage-flow relationship. However, the gravel build-up at Featherstone means that confidence in the rating is low and there are concerns that:

  • Infill upstream of the crest will reduce depths and increase approach velocities
  • Infill downstream of the crest will cause non-modular flow to occur at lower flows
  • The impact of these factors is not uniform across the channel, and the local stage-flow relationship varies along the width of the weir

These points can potentially lead to inaccurate flow derivation for flood forecasting and warning downstream.

To investigate the need for gravel removal, we were commissioned by the Environment Agency to use 3-Dimensional (3D) Computational Fluid Dynamics (CFD) models to produce and assess the impacts of the rating curve for three scenarios: existing gravel build up in the channel, dredged, and aggraded riverbeds.

Solution

We used a drone survey of Featherstone reach and coupled with in-channel GPS points, we produced a digital elevation model (DEM) of the river channel, at 10cm resolution with a vertical accuracy of ±30mm. We transferred the DEM survey data into a computational fluid dynamics (CFD) model. This 3D model solves the full three-dimensional form of the Navier-Stokes flow equations with sophisticated turbulence modelling and a relatively fine grid, and with less dependence on user-defined parameters than in 1D or 2D models. We also built a 1D HEC-RAS model to generate boundary condition for the 3D model.

We then ran several steady-state models in the 3D CFD model, using a range of upstream flow rates to produce the rating curve for the existing gravel build up (baseline scenario). We also did this for dredged and aggraded riverbed scenarios, to give a comparison of the impacts on the rating curve.

Figure 1: Survey and Delineation of Important Features

Survey and delineation of important features_CFD modelling at Featherstone gauging stationSurvey and delineation of important features_CFD modelling at Featherstone gauging stationSurvey and delineation of important features_CFD modelling at Featherstone gauging station

Benefit

Our modelling work demonstrated that the impact of the gravel build-up on the rating is minimal. Our CFD modelling also showed that, should the channel be dredged, the gravel is likely to be re-deposited even at relatively modest flows. This confirmed that dredging and repeated removal of gravel is not a sustainable option. Removal of gravel also leads to significant environmental disturbance to in-situ wildlife and potential spawning sites, and there is also believed to be eel passage benefits associated with the gravel.

We recommended continued monitoring of the bed evolution – with use of the model to adapt the rating if necessary – rather than any intervention to remove the gravels.

A detailed CFD modelling study is a cost-effective solution, and the use of a drone survey also has reduced costs associated with time-savings. The drone survey minimised the time surveyors were in or close to the river, and the use of CFD modelling to check high flow rating avoided the need for manual gaugings at potentially dangerous flows.

 Want to know more?

Email Andy Collier for more information about the Featherstone gauging station project, or visit our Computational Fluid Dynamics webpage for more information.



1 Comment

  • Mark Franklin

    A great and innovative approach, so much information about structures can be learnt in his way, an applied to improve high and low flow estimates. Good stuff

Leave a Reply