This CFD case study assesses the effectiveness of a weir vessel to mix and separate two mixed process streams. The inlet flow rates are highly variable but the vessel cannot be too buffered (i.e. large) for process reasons, not to mention cost and space requirements.

The concern is that splash from intermittent high surge of inlet flow would go over the weir before beingĀ given the necessary residence time for stratification.

A CFD analysis was performed using the Volume Of Fluid (VOF) method to track the multiple phases and open surfaces. This treats the flow as a series of phase fractions each with distinct properties. The mixture properties simply use a weighted average of the constituents. The interface is kept sharp using an artificial non-diffusive term which compresses the interfaces together over the length-scales of the mesh size. Surface tension is applied at the fluid/gas interface, which becomes important when concerned with small droplets or capillary flow. The mesh was kept fairly coarse since the flow is gravity and pressure dominated.

An initial trial solution was to place a nominal pitched plate over the top of the drainage side, resulting in the flow shown in the video below.

This CFD case study assesses the effectiveness of a weir vessel to mix and separate two mixed process streams. The inlet flow rates are highly variable but the vessel cannot be too buffered (i.e. large) for process reasons, not to mention cost and space requirements.

The concern is that splash from intermittent high surge of inlet flow would go over the weir before beingĀ given the necessary residence time for stratification.

A CFD analysis was performed using the Volume Of Fluid (VOF) method to track the multiple phases and open surfaces. This treats the flow as a series of phase fractions each with distinct properties. The mixture properties simply use a weighted average of the constituents. The interface is kept sharp using an artificial non-diffusive term which compresses the interfaces together over the length-scales of the mesh size. Surface tension is applied at the fluid/gas interface, which becomes important when concerned with small droplets or capillary flow. The mesh was kept fairly coarse since the flow is gravity and pressure dominated.

An initial trial solution was to place a nominal pitched plate over the top of the drainage side, resulting in the flow shown in the video below.

Shown only are the liquid phases, giving an excellent feel for the transient flow features. Notable is how the flow is super-critical around the edges (i.e. is moving faster than the surface wave speed, so with a Froude number greater than unity), coating the vessel inside surface with a smooth film. This rebounds off the pitched plate and over the weir, clearly showing the deficiency in the design although some splash is successfully captured above this plate.

This visualisation alone creates a strong talking point for redesign, and gives insight that would be unavailable with hand calculations and expensive with testing. As shown below, the design was modified with a rebound lip just below the overflow level reaching around the corners where the premature overflow was most apparent.

The overflow is much better contained on the right hand section during the inlet surge, although clearly the flow needs to reduce thereafter given the size of the vessel.

In this example, we were able to apply a combination of broad mechanical engineering knowledge with CFD software modelling to solve the problem, without the need for lengthy testing.

Shown only are the liquid phases, giving an excellent feel for the transient flow features. Notable is how the flow is super-critical around the edges (i.e. is moving faster than the surface wave speed, so with a Froude number greater than unity), coating the vessel inside surface with a smooth film. This rebounds off the pitched plate and over the weir, clearly showing the deficiency in the design although some splash is successfully captured above this plate.

This visualisation alone creates a strong talking point for redesign, and gives insight that would be unavailable with hand calculations and expensive with testing. As shown below, the design was modified with a rebound lip just below the overflow level reaching around the corners where the premature overflow was most apparent.

The overflow is much better contained on the right hand section during the inlet surge, although clearly the flow needs to reduce thereafter given the size of the vessel.

In this example, we were able to apply a combination of broad mechanical engineering knowledge with CFD software modelling to solve the problem, without the need for lengthy testing.