Looking for advice on a slurry concentration measurement application

[lastone]

I have a problem that I was hoping someone might have some experience with or could offer some tips/advice on. I am looking for a way to measure online a slurry concentration in a vessel.

The operation involves a small vessel, (3.5'-I.D., 4'-Tan/Tan, approximately 330 gal.), in which we mix a solid powder with a solvent. The vessel has a 5 hp mixer which sits on the top flange of the vessel and is run continuously at a constant speed. The solvent is fed into the top of the vessel, and its feedrate is controlled by a flow controller. The solid is educted with N2 into the top of the vessel, and its feedrate is controlled by a Loss-In-Weight-Feeder. The vessel operates at low pressure, (2.5 psig), and is vented into the Unit's Vent Header. The vessel is also jacketed with a coolant and runs around 50-60 Deg F. The process is operated continuously and we set the feedrates of the solid and the solvent so that we maintain approximately a 2% mass concentration of the powder in the solvent, (i.e. 6 LB/HR solid and 300 LB/HR solvent). The powder is not soluble in the solvent. If you were to hold a jar of the mixed slurry you would see the solid quickly settle to the bottom. The mixed slurry is then fed into the suction of a small positive displacement pump, where the fluid is compressed to a high pressure and fed to the next section of the process. The level in the mixing vessel is controlled by adjusting the output of the PD pump using a variable frequency drive. We run the vessel around 50% level. The vessel nozzle that supplies suction to the PD pump is on the side at about the 25% mark. It is a 2" flange and the suction piping rises up about 12" on a 45 Deg angle and then reduces down to 1" piping and tees into the two valves of the PD pump. The PD pump is spared with an identical setup on the opposite side of the vessel. I hope I have described the process enough, so now I'll get to the problem.

We have terrible reliability on the PD pumps. They eventually quit pumping and we are constantly having to switch between pumps. We start one up and prepare the other for maintenance. The new pump will operate OK for a few days, then it will start pumping on only one cylinder, and then that cylinder will begin to quit pumping, so we switch over to the other pump. The maintenance work usually involves removing both valves and 1" manifolds, taking it to the shop and cleaning the springs, valves, and manifold and removing the resultant hard compacted solid powder. Its actually a lot of work and takes considerable time for operations to prepare the equipment for maintenance and it ties up a couple of maintenance techs for a half a day each time. The mechanical engineer and I have everbody yelling at us to fix this problem. We believe that the problem arises from at times there being too high a concentration in the slurry, and the solids simply settle out in the suction valves of the pump. If we line up the pumps so that they are only seeing pure solvent they always pump perfectly. I have measured the Loss-In-Weight-Feeder on multiple occasions and each time it is delivering the desired weight of solids. We have experimented with multiple types of valves in the pumps without a noticeable difference in reliability. The mech eng is convinced that if we can maintain a set concentration of solids the pumps will work much better. I tend to agree with him, because we have observed instances where the pumps have worked for up to two weeks without problems. However, a two-week run has not occurred in a long time and tempers are fairly high. I wish I could set up a sampling program to ensure the concentration, the problem is that due to the nature of the solvent I can't have operators catching any samples. If I could have an online analysis to verify the solids concentrations I think it might help us react to problems much quicker and improve the operation. Does anyone have any tips. It would be much appreciated.

 

[jmw]

Some suggestions and some questions...

Why a positive displacement pump? Have you considered a peristaltic pump for the slurry? or any purpose designed slurry pumps? I assume these have been rejected for some reason.

Concetration measurement can be achived in a number of ways. I would certainly consider a vibrating element density meter.
These are used with a variety of slurries including china clays, chalk slurry (for cement production) etc. for example. These will give you a continuous measurement of the %concentration. You may prefer a gamma ray densitometer.

The two things to look out for are:
[ul][li] entrained air[/li]
[li] solids deposition[/li][/ul]

You are mixing the powder and the solvent together in the tank so it depends on how long air will remian trapped on the powder particles and if this air is carried through.
Solids deposition in the density meter is usually resolved by the installation e.g. flow vertically down which will suit both tube and fork designs of density meters.

Of course, if we assume the feedrates are well controlled then problem might be with the mixer e.g. solids clumping together occassionally, or solids collecting in the bottom of the tank until the level is high enough for a slug to be discharged.

However, it sounds also as if the pump may have difficulties when the speed is regulated down. Perhaps it would be better to run the pump at a constant high flowrate and use a control valve in a feedback loop i.e. where some of the pump output is fed back to the pump inlet or to the tank.
In fact, recirculating some flow back through the tank might be beneficial in other ways but mostly beacuse it will maintain a higher flowrate through the discharge arrangement.


JMW

http://www.ViscoAnalyser.com

 

[lastone]

Thank you JMW. The main reason for the PD pump is due to the need for a high discharge pressure (> 1000 psi). I'm interested in your density meter suggestion. I don't think air entrainment or solids deposition would be a problem with this application. How accurate would the meter be? Would it be able to differentiate between a 2% and 3% mass concentration? What would I need to install this type of meter? Could it fit through a 3/4" bleed valve on the side of the vessel? How much maintenance time is typically required with a meter like this? And, do you have any suggestions on companies/people to talk with about this application? Thanks again.

 

[jmw]

You can measure the density in the tank, in the inlet to the pump or in the discharge from the pump.

Accuracy depends on the density of the solids. The greater the difference between the solids and the solvent density the better the resolution of concentration so there is some data required to determine the accuracy you will obtain. The density meter options can deliver between 1-2kg/m£ density measurement to 0.15-0.3kg/m3 dependent on type.

Since there are no moving parts, maintenance is minimal. In the installation design flow velocity is an important factor in both maintaini ng sufficient flow and or flow orientation to prevent solids deposition and avoiding excess flow that will cause errosion. In tube density meter installations it is usual to arrange the pipework so the density meter can be cleaned with a "bottle brush" as and when necessary.


JMW

http://www.ViscoAnalyser.com

 

[lastone]

Thanks again for the response JMW.

 

[Ashereng]

Just to add to JMW's post regarding density meter.

A coriolis mass meter may be a good option. There are 2 common designs: the U tube and the straight tube.

In a slurry, a straight tube is preferable, and can be had in 316SS. For a vendor, see Micromotion at:

http://www.emersonprocess.com/micromotion/products/index.html

I would suggest you put this on the suction side (since it is a tube, the pressure drop is the same as that of a pipe - very little loss).


A more cheaper solution may be to install a piece of glass or plexiglass (process conditions permitting of course) in the suction to the pump. It will let you "see" what is going into the pump. If the powder does not disolve in solvent, you should be able to shine a light and see cloudiness. If this works, you can buy a light adsorption measurement device (the name escapes me right now - senior moment) and that will give you a correlation of solids%.

 

[cawse001]

I have had a similar experience trying to pump a sulfur slurry into an autoclave using piston diaphragm pumps. The main problem area was the ball type check valves, which were bound up by solids settling out.

In the end we changed pump types to a helical rotor type with a variable speed drive. The pump was capable of delivering 5000 kPag so you may be able to find a unit to deliver 1000 psi.

Cheers,

NB

 

[smiro]

This problem is not so simple and according my opinion there should be taken more measures in the same time.
1. Vessel should be equipped with buffles (min. 3 pieces) mounted at 102o angle distance.
2. Buffles should be fastened to vessel inside with distance road and between must be a gap 1 cm or more. Any settling or pile up of the solids must be avoided – check at raw material dosing.
3. Mixer should have more mixing elements (pitched blade turbine type or similar) on the shaft and at the bottom should have (well-balanced) rotating disk with sharp teeth for better slurry dispersion.
4. This disk as lowest element must be relative close to the vessel bottom. Please confirm at the representative of well-known mixer manufacturer (Chemineer, Ekato or similar).
5. Positive displacement pump is not well suitable for feed of the slurry. Single rotor screw pump is much better even for much thicker suspensions.
6. For pressure increase feed can be done also in two steps with two pumps.
7. Measurement should be done with coriolis mass flow meter type U-type or straight type. Vendor are Endress&Hausser which I prefer or Micromotion from Emerson. It is important to install mass flow meter in vertical position with slurry bottom inlet.
8. Density can be taken from mass flow meter.
9. For density inline control I can recommend to install continuously agitator current measurement at the control room. Accuracy of this measurement is should be as high as possible about 0.01 A or more. This is not real viscometer but his reaction is very quick and accurate. This measurement can be connected with alarm and can automatically stopped feed of raw materials.
10. Preparation vessel should be filled with 75-85 % of the volume and equipped with level control to increase residence time to improve dispersion of solids, to prevent any clumps or much worse pile up of the solids.
11. For better dispersion it also usable use of separate tank for preparation which can be discharged continuously into separate tank as feed tank. This tank must be equipped also similar as preparation tank (buffles, agitator).

 

[smiro]

Please for feed back info for any improvement after any fom above mentioned measure has been taken. Thank you.
P.S English is not my native language.

 

[jmw]

Smiro,
density can indeed be measured by many mass meters and with good accuracy but if the flow rate varies significantly we can be caught between two problems with twin tube versions:

Firstly, if the flow drops we can have solids drop out and this may not be equal in both tubes. Indeed, as solids build up in one tube the flow will further split with proportionately more flow through the less blocked tube. This leads to a further reduction in the partially blocked tube until it is completely blocked.

Secondly, with slurries we need to be careful about errosion which will occur just after each bend. Massmeters, and many density meters, have thin walled tubes and any errosion may lead to a loss of accuracy but also the risk of tbe failure in extreme cases. However, since we do not need to measure flow I will allow we do not need the high flow velocities necessary to generate the pase angle measurements for mass flow and that as a consequence, that errosion may not be a problem since we would size the meter for low velocity.
However, faced with a choice, a single straight tube device is preferable, especially if arranged with the flow vertical down which will help prevent solids formation. Single straight tubes are also easier to clean if there is a settlement problem since they can be cleaned with a bottle brush.
Your point seven does pick up on this but only for a single straight tube design.
However, if all we are interested in is density there is no need to specifically choose a sigle straight tube mas meter mass meter over a single straight tube dedicated density meter, or an insertion style density meter which will be by far the most economical approach.

JMW

http://www.ViscoAnalyser.com

 

[lastone]

Thank you all for the postings. I certainly like the thought of a mass meter as ASHERENG describes, however we just don't have enough room to put one in the suction of the pump. The suction piping is very short and any re-piping or addition of length would only add to the settling problems we currently experience. As far as the discharge is concerned, I don't think a mass meter would work well here either. Putting the mass meter into the discharge of a PD pump causes inaccurate readings and I would need to install pulsation dampeners. (I know this because I have another area of the plant with a mass meter near the discharge of a PD pump and it doesn't work well.)

To respond to SMIRO, the vessel is equipped with baffles and blades on the mixer shaft as you describe. I believe the material in the vessel mixes well. It just doesn't stay mixed as it moves through the piping.

CAWSE001, we have had the same problems with ball check valves as you describe. We have switched to cone-type check valves and it appears to work a little better, but not much. Do you really think I can pump a slurry at this small feedrate with a helical rotor pump? I am pumping 500 LB/HR through a 1" line to a downstream vessel operating at 1650 psig/11,375 kPa?

At this time I think I am leaning toward installing one of the vibrating element density meters as JMW is describing, and making some modifications to the suction piping to increase the velocity and flow distribution to the pump. It looks as if I can use an existing 3/4" bleed on the vessel to install the meter.

 

[Ashereng]

lastone,

Too bad about the lack of suction piping.

Let me know how the vib element density meter works out.