Flow pattern through a Line Blender (pipe mixer) at Post Mixing

[ News ] [ Services ] [ Products ] [ Mixing Forum ] [ About Us ]

[ Up ]

An Line Blender is a mixer in a pipe. This Line Blender is made of a T-pipe section with 10" diameter. It has a shaft with 2 5"-radial impellers extending through the T into the pipe. The impellers are 4 bladed turbines without a disk or RP4. This particular Line Blender has a baffle plate that forces the fluid to the upper impeller, through an orifice, and down to the bottom impeller and back to the pipe. The overall flow is from left to right. More details.

A detailed study of 4 other styles of In-liners or Pipe Mixers has just been completed. Which design do you think is the best? Which one will give you the most mixing for the least amount of power. The results will amaze you.

Here is an animation of the vector plot. Horizontal slices are taken from the top of the T to the bottom of the pipe. Each slice is shown for 1 second. To repeat the animation, right-click and select Play.

Velocity [m/s]	Animation of the velocity profiles	Comments
		The first picture is shown until the video is loaded unto your computer. The white ring represents the top cross-section of the T. The cross represents the diameter of the impellers. The ring within the cross represents the location of the shaft. The pipe section appears first as a narrow slice of its top. The vertical parallel lines to the right is the upper baffle plate. The slice goes through each impeller when you see the large amount of green and yellow arrows coming off the blades in CW rotation. The slice is midway through the pipe when the outer pipe lines are tangent to the circle of the T. You will notice no vectors between the parallel lines on either side of the T. This is because the slice is going through the horizontal baffle plate. The flow is all going through the central orifice, which is the smaller inner white circle. The outer parallel pipe lines become smaller now because each slice is getting closer to the bottom of the pipe. The vertical parallel lines to the right are now gone and a set of vertical parallel lines are now to the left of the circle T, representing the bottom vertical plate. The last slice is at the bottom pipe wall and therefore shows no velocity vectors.

A side view of the Line Blender is shown here as a CFD mesh.
The flow through the pipe is from left to right and top to bottom.	The animation starts at the top of the T. Horizontal slices move from top - down. The upper impeller appears first. Then comes the dividing plate with an orifice. Then comes the bottom impeller. It ends at the bottom of the pipe.

A close up of the impeller region.

If you want to see these pictures better click on the appropriate thumbnail below.

Details of the animation:

Physically, this is a nominal 10" pipe with "tee", with flat blade impellers. The fluid is nominally water, and the flow rate corresponds to 1100 gal/min, and impeller speed of 1750 rpm. The pressure drop shown on the vertical pressure plot below is in Pascals, which corresponds to approx 8 psi in English units.

The segmented region is 20" long. This means that the volume in this region is 25.7 Liters = 6.78 gallons.

Just a couple of words about the solution itself - it looks like there is a "stable" oscillation some where - i.e. I ran about 100 iterations which took just under 7 hours, but the solution reached its stable oscillating mode in about 20 to 40 iterations depending on how you call it. That is, that the nonlinear residuals are solved pretty well - and the step to step velocity was changing. I need to investigate this a bit further - but usually this is like a vortex moving around - which is highly probable. The Froude Number is 11!

Discussion of the results:

Assuming that Nq = 0.6, the pumping capacity of each impeller is 0.0358 m³/s or 567 gallons/min. Compared to the flow through the pipe (1100 gallons/min), the flow from the impellers are about 1/2 of the flow from the pipe. The animation and the vertical plot shows an overwhelmed impeller system, which is to be expected in this case.

In the vertical velocity vector image shown below, you can see quite well the up stream and down stream flow implications. The up stream effects are not too large - but this is a function of the flow parameters - speeds, fluid viscosities, etc. The down stream shows a significant recirculation loop at the top. That is not surprising.

Notice that there are 6 significant regions of less active zones and recirculation in the Line Blender:

The bottom of the upstream pipe before the forward vertical baffle.
The top of the T
The top near the rear vertical baffle
The bottom near the forward vertical baffle
The bottom under the rear vertical baffle
The top of the downstream pipe.

Backmixing and short-circuiting will be a serious consequence of this arrangement.

The torque and power number when taken about the impeller axis - is as follows: impeller speed = 1750 RPM, total torque = 4.865980 Nm, total power = 891.737368 Watts and total power number = 3.267261. A single, fully baffled 4-bladed paddle (no disk) would have a power number = 3.4. Obviously, the power number is less in this unbaffled situation.

The segmented region is 20" long. This means that the volume in this region is 25.7 Liters = 6.78 gallons. For this region P/V = 892 Watts / 25.7 Liters = 34.7 kW/m³ = 173.5 Hp/1000 Gallons.

At 34.7 kW/m3 you would think there was really enough energy there to mix just about anything! What we see is a totally swamped impeller. The impellers have no significant influence on the flow pattern. If the flow was uniform through the orifice, the velocity would be 4.5 m/s. The tip speed of the impeller is 11.6 m/s.

Click here for more CFD and CFM examples.

[ Home ] [ Up ] [ Inliner ] [ Line Blender ] [ Process Intensifier ]

Send mail to webmaster@postmixing.com with questions or comments about this web site.
Copyright © 2003,2004,2005 Post Mixing Optimization and Solutions
Last modified: Feb 6, 2005