Dear mohamedgomaa,

Pipe wear will not be a problem.

The boundary layer protects the pipe wall against wear.

Bend wear could be a problem.

By applying a boxed bend the wear problem can be solved, although during the first stage of bend wear some stainless steel can be mixed in the silica.

Making a a hole in the outer wall of the bend where a hole is expected that can be prevented and the "wear starts from the beginning as silica on silica.

I do not see any problems with static charges with the silica in the pipeline as silica is not explosive.

Applying a cyclone for a very fine dust is not considered the first choice.

Designing a pneumatic conveying system based on 80 m conveying length (How many meters horizontal, meters vertical and number of bends?) and 7 tph, without considering the particle size, the altitude above sea level, ambient conditions, feeder type indicates that you do not have all the necessary data available.

Whether the system should be a vacuum- or pressure system, depends the number of sources and the number of destinations.

(A vacuum system is certainly dust free)

Dilute- or dense phase is depending on the definition you have for these conveying regimes.

I would design for the lowest conveying velocity, whereby there is no sedimentation/choking.

Also, water condensation is something to pay attention too.

The design you are starting requires the silica particle size (which is in the range of 5 to10 microns, I believe) is a ruling parameter to calculate the lowest allowable gas velocity. (Air or Nitrogen)

You can also ask your local Atlas Copco agent for support on your project ■

Silica flour is very easy convey in dense phase. Done a few of these systems in the past.

You need to keep the velocities low and move it in plug flow so the wear is also low.

Just need to take care in fluidised state it runs like water. ■

Dear Mantoo,

In this case, the decision is already made that it should be a pressure system.

Is a plug conveying sytsen not very sensitive for choking without special controls?

Especially for a length of 80m? ■

Dr Mr Tuinenburg

Silica flour is vey easy to convey in Dense phase. Nothing special is required for plug

control. 80m is very short distance it can go 1000m with no problems. Just form information

plug lengths are natural and no one can control them.

One just need the right vendor for Dense Phase system as not everyone has correct

experience for it.

Have a nice day. ■

Dear Mr. Mantoo,

I am sure that for a dense phase, plug conveying system you need specific experience.

My question is then: what kind of experience?

I once witnessed a 500 m long alumina conveying line (16”), which formed a plug after a horizontal long radius bend at regular time intervals.

There was a pressure gauge at that location and the pressure fluctuated approx. from 1. barg to 2.5 barg and listening with a screwdriver to the noise in the pipe the plug could be heard passing the location very clearly at approx. 1-minute intervals.

There was no heavy pipe shaking, as the pipe supports were very strong.

The installation (a double pressure tank system) was operating initially with a too high SLR and choked each kettle.

After we increased the kettle by-pass air by using the extra control valve from the second, not operating double tank system, we could prevent the choking.

The unit I am referring to, was a pneumatic ship unloader with a filter receiver tank and 2 double tank discharge systems and 2 discharge pipes, one of which was operating for the alumina conveying.

Could it be that this system was operating how you describe it in your reply to this thread?

-Is dense flow and plug flow the same thing?

-Do I understand it correctly that the plugs are formed and broken in a natural frequency (with a natural time constant)?

-Are the bends influencing the formation of plugs, due to friction and velocity loss against the outer wall of the bend? ■

Dr Mr Tuinenburg

First of all it is very important to know if the material will convey in dense phase or not. Silica flour is a good candidate but alumina is not. This is this the place where vendor experience matters the most as not everything is suitable for dense phase conveying. You mentioned Alumina and it is common know fact that alumina cannot be conveyed in dense phase plug flow in normal pipe. Most of the ship loaders are working in lean phase, only because they are conveying at 1.5- 2.5 barg pressures it does not make them dense phase. It is a very common misconception.

But if you look at fly ash or cement unloading then it can be in dense phase but it depends on the vessel configuration. Sometimes you have to put so much air in the vessel to get the material out that the system moves out of dense phase mode in to semi dense phase modes where the screw pumps work. It is also required because if the system is running in true dense phase with few hundred meter of pipe the conveying pressure especially for cement will exceed 2.5 barg vessel pressure limit.

Some grades of alumina can be conveyed in dense phase but they require internal bypass conveying lines. If my memory serves me right the original patent was by Buhler in early seventies now a few companies offer them. These systems tend to convey at 6 barg pressures and I must admit these days are bit uncommon and are limited to relatively short distances.

There are different types of plug flow described in almost all of pneumatic conveying textbooks not much to it.

To answer your questions

- Is dense flow and plug flow the same thing? Dense flow is referred as non suspended flow of material in conveying pipe it can be in plug flow or can be in dune flow.

- Do I understand it correctly that the plugs are formed and broken in a natural frequency (with a natural time constant)? I would consider plug formation random phenomena, but plug length is natural.

- Are the bends influencing the formation of plugs, due to friction and velocity loss against the outer wall of the bend? In dense phase, conveying bends are not that significant. Since the velocities are low and plug wall friction is same in straight section and in bends. Directional change does not causes significant velocity reduction. Verticals have more influence on plug formation and pressure drops.

Have a nice day! ■

Dear Dr. Mantoo,

Thank you for your interesting reply.

I could not agree more.

As you might know, I started to research pneumatic conveying (theory and calculating) in the grain unloading sector in the port of Rotterdam.

These big vacuum unloaders were definitely unloading in dilute phase, no question about that.

When I switched to cement conveying unloaders (vacuum/pressure systems) the same calculation algorithms worked perfectly. I only had to adapt the operation of the vacuum pumps to compressors.

Later, I got involved in cement, barite and bentonite systems on PSV ‘s and riggs.

Those installations were operating at 3.5 to 4 barg.

This design was chosen to keep the pipeline diameters small (6 to 8”)

The developed pneumatic conveying algorithm was improved over many years of gathering information from the build and operated installations.

The algorithm gave (and still give) results to calculate a Zenz diagram in the sedimentation region, which can also be regarded as dune conveying.

If dune conveying is considered dense phase, my conclusion is that the written software also calculates dense phase to a certain level.

When the airflow becomes too low and the pressure too high, the software results in choking and a plug flow calculation must be applied.

As far as I know, some researchers consider plug flow as plugs in a pipeline, obeying the flow formulas is silos (Janssen formula)

Then, controlling the plug length becomes the main issue to prevent arches.

That must be where internal bypass conveying lines are used.

Assuming that you mean that the plug length is natural, why is the formation of plugs not a natural phenomena?

Bends are tricky things.

The entering velocity might be low, but then the velocity reduction does not to be much to create a plug in the bend.

This is even more the case where the gravity causes an extra velocity reduction. F.i. in a bend going from horizontal to vertical.

That is why those bends should not be long radius bends as the gravity component works for a much longer time on the velocity reduction than in a short radius bend.

(In powder coal injection systems for blast furnaces, horizontal to vertical bends are generally T-bends)

Must say, I enjoy this conversation. ■

Maybe I should have clarified bit more when I said dune flow I meant dune plug flow. Where a solid plugs is moving on top of the sliding bed of material.

I think you are referring to dune flow in lean phase which is observed in the acceleration region. They are both different. The sliding dune plug flow can

only exist in horizontal lines and it has to move to full plug or fully suspended flow in the vertical.

No algorithm calculates dense phase plug flow. I would exclude fly ash and cement from this statement as they easily go through unsteady state zone.

So lean phase algorithm may have some validity in unsteady state zone for these material.

Plug formation can be controlled by air velocity, rate of material flow into the conveying pipe and conveying route geometry at the beginning.

So I will not consider it natural.

Bends in dense phase are relatively insignificant. There is research available in comparing pressure drops in bends in lean & dense phase and

between horizontal and vertical orientations.

There is no significant velocity loss in bends with regards to R/D ratio. In dense phase as the material is already sliding on the pipe wall.

In vertical up it is not the velocity loss it is the weight of the material that causes extra pressure losses.

Blast Furness injection systems are bit different as the end pressure is 2 -2.5 barg and not ambient like normal conveying systems. Again

they are running in lean phase with very high phase densities and the conveying pressures could be anywhere from 3-14 barg. At these

elevated pressures the standard norms of conveying are changed and fully suspended flow can be achieved at very low velocities due to

high compressed gas densities. T Bends are used there if in case a line blocks they can rod it or put compressed gas in it to clear blockage.

Bend Pressure drop is least important in coal injection system, there are too many other things to worry about. These type systems are not

everyone’s cup of tea only a handful companies in the world offer them. ■

Dear Dr. Mantoo,

Here, the importance of exact definitions is shown as very important in a discussion.

Before you know a conversation starts to diverge, because of an un-noticed confusion.

Maybe I was in my mind referring to dune flow and plug dune flow with the small top of the dune touching the upper side of the horizontal pipe wall, which has only a short lifetime.

It is obvious that there no dunes against a vertical wall.

In a vertical pipe, the sedimentation against the pipe wall, where the actual suspension velocity is higher than the boundary air velocity, the material falls back forming a smaller conveying cross-section with higher air velocities. The dune is smeared along the pipe wall.

I believe we are on the same page.

I tried to make one in Excel by applying the silo formula as described by Janssen whereby the plug length and the pressure difference over the plug are considered the driving force and the resistance force.

The condition is then that the material flow properties (internal friction, wall friction etc.) were used to check the occurrence of arch forming.

The calculation results were indicating the observed phenomena, but was not considered a reliable outcome, because of the sensitivity to the many (not exactly known) flow variables.

One conclusion was that very long plugs made the system choke. (As in high but small diameter silos)

Another problem was the expansion of the air pockets between the plugs.

That is exactly what I experienced with the cement unloaders I designed and operated.

Good to hear that you can confirm those conclusions

If the plug formation is controlled, I would say the plug is not formed randomly.

Matter of interpretation?

The pressure drop in a bend is insignificant.

However, the velocity loss in a bend is not insignificant.

In a bend full segregation exists, making the air resistance minimal.

The real bend loss becomes manifest in the acceleration zone after the bend.

And the velocity loss in a bend is not only depending on the friction along the outer bend wall, but also of the time that the gravity component is pushing the material against the outer bend wall.

Depending on the bend orientation, the gravity component can be negative or positive and can become significant in long radius bends.

The lean phase bends can therefore cause significant variations in related pressure drop, as in dense phase bends, where the friction is supposed to be low and there is no acceleration zone (constant plug velocity), the bend pressure drop can be assumed fairly constant.

The pressure loss of lifting the material up is equal for dense phase as for lean phase as the required energy is mass times height, which is equal for the same mass flow.

I was involved in a project, where the conveying pressure was approx. 10 barg and the back pressure in the blast furnace approx. 4 to 5 barg.

The pressure ratio of such an installation is then approx. 1.83 to 2.2, which is very low and are typical for suspended flow as you mention.

What I heard from the experienced engineer was that for the horizontal to vertical bends a long radius bend was chosen to get rid of the frequent blockages. (Which phenomena is supported by the theory-based calculations)

Indeed, there are many things to consider, as those system are used to convey the powder coal from a long distance away from the blast furnace and then split through a distributor to many injector lines of equal length to guarantee equal division over the burners, whereby the distributor pressure is the back pressure for the supply line and the conveying pressure for each injector line.

And the use of nitrogen to prevent explosions.

I do not think we have helped Mr. mohamedgomaa answering his original question. ???? ■