Centrifugal Pump Cavitating on tank changeover 3

[beattsmjk]

Hi,

I have searched this topic and found some useful info, but I thought I would write a post to see if anyone can offer any more specific info that might be useful.

I work in a bakery and we have two tanks of liquid yeast (consistency of single cream or slightly thinner). This yeast is stored in chilled stainless cone bottomed tanks, each holding 25,000 litres.

We have both tanks connected to a 'ring main' however we only use one tank at a time, with the other on hold. The contents are agitated at a speed of around 75rpm to prevent settling out of the product.

The yeast flows from the tank in use down to a centrifugal pump approx 8 metres away and approx 1.5m lower(NPSH) distance is ok according to Packo (pump manufacturer), it is then circulated around a 'ring main' and back into the tank through the side of the tank.

As the yeast is required it is drawn from the ring in small quantities at four points (mixers).

The problem we have is on tank changeover. At around 700 litres certain valves close on the tank in use and others open on the other tank which allow for a seamless transition.

However we find that when we reach around 600 litres remaining, the pump starts to cavitate and we lose all pressure in the ring main (normally 3 bar). We then have to manually switch to the new tank and dump the 600 Litres (which is not good obviously - approx £500 each time).

We are scratching our heads as to the cause. like I said, the pump size was calculated for the system and so the NPSH is ok. The pump people have no further ideas.

The yeast is circulating at around 4degrees.

Do you think the agitators could be creating a vortex perhaps and allowing air into the pump? The tanks are fitted with vortex breakers according to the drawings.

We have the pump running at full speed (2900rpm) it is designed for 7m3/h although we only need 4.5m3/h with pressure drops around the system of around 3m3/h this works out about right.

I really look forward to any advice you might have.

Thanks in advance

 

[Artisi]

Just to add a bit more discussion to the problem, a mis-match of flow onto the impeller blade from pre-rotation, operating well left or ever well to the right of BEP can cause uneven operation and considerable noise sometimes mistaken for cavitation.

 

[QualityTime]

Hi Artisi:

Appended is the second article that talks about where the damage to the impeller will be.

 

[QualityTime]

Hi Artisi:

For some reason the article I wanted to download does not seem to open. The following is the actual web link:

http://www.practicalpumping.com/articles/pdfs/Cavitation or Not.pdf

If you do a search elsewhere, you will see that air can cause damage to a pump. If you think about it...you have a mixture of air and water in the pump...air is compressible and water is not...and because of that, the pump pumps easy (because of the air) and then it pumps hard (because it hits water). It is not good for the pump, motor and coupling. If you do have problems with air and you put your ear to the pump discharge you will probably hear that the flowing water in the piping does not sound right.

The damage to the impeller due to cavitation is due to the little air bubbles being compressed. If you have air coming into the suction of the pump (not due to cavitation)...you get the same result in that the bubbles get compressed and that is why you get the same sound as cavitation.

 

[beattsmjk]

Hi,

Just an update, the pump is definately cavitating, not just sounding like it is.

It is getting extremely hot and we lose all pressure on the discharge.

What do people think of running the agitator in reverse so that the blades due to the angle in which they are mounted, in effect push the liquid down? Would this work or might it put greater pressure on the gearbox perhaps?

 

[QualityTime]

As per the article I posted....if it is cavitating the true test is to throttle the pump discharge. The cavitation noise should get quieter and quieter the more you throttle. You will reach a point where there is no cavitation and there is no noise

If the "cavitation" is due to entrained air....no matter how much you throttle the pump the "cavitation" noise will not disappear

In order to back the above practical field test have you done a calculation to determine whether or not it should be cavitating??

 

[Artisi]

If the pump is getting hot, sounds like it is running un-primed due to air or low NPSHa by restriced valve opening.

I would suggest a systematic approach to the problem, look at each likely problem in turn and eliminate them one-by-one,is it entrained air, is it restricted NPSHa due to valve closing sequence, is it operating too far left of BEP etc.

Reversing the direction of the agitator is a fix sometimes used and sometimes successful in the paper industry when it is causing air entrainment problem.

 

[Artisi]

QualityTime, Thans for the link to the second article, seems to me to be a rehash of the earlier article like many others which just restate the same information over and over.

Have attached a link which you may find of interest.

http://www.waterworld.com/index/dis...tion-air-on-centrifugal-pump-performance.html

 

[QualityTime]

Artisi, I beleive the discussion was about damage whether or not air in the pump will do damage to the pump impeller. you seemed to have doubted it and the article I referred you to talked about the potential damage that could occur.

thanks

 

[FMJalink]

Postbeattsmjk,
Probably has been taken into account during the design that most yeasts are seriously non-Newtonian in their behavior. They normally show pseudoplastic behavior, which is quite different from how water behaves. This is especially visible at start of flow and filling of spaces. Another aspect is the carbon dioxide which is normally substantially present in reacting yeast.

You state that: “When the Empty Pipe Sensor (ES1) or the Flow Sensor (FS1) or the Pressure Sensor (PS1)detect a loss of yeast, then the final automatic valve (in this case AV10) is opened to allow yeast from the new tank into the flow plate.”

This includes that you have already a too low suction pressure at the moment the valve is opened. Only by a rapid and full supply of yeast from the newly opened tank will the pump be able to continue normal operation. To my opinion following points could cause your problem:

(1) I do not know whether the primed/flooded part of the pipe below AV2 is without gas. After opening the valve AV2 a substantial gas bubble could be present below the newly entered yeast.

(2) As yeast shows pseudoplastic behavior, it will have a large initial viscosity until it starts to flow substantially. This combined with the required acceleration of the newly connected yeast, could cause a substantial gas bubble in the suction line.

(3) I assume you have a kind of non-return valve installed, as at the moment you have both tanks in parallel on the pump, the yeast from the full tank will flow to the nearly empty tank. At that moment the pressure in the suction line will stay that of the nearly empty tank. As far as I know, no return valves should not be used in a hygienic application, due to difficulty in cleaning.

(4) In order to cope with the gas should the pump be self venting.

Looking at your system, it appears that the Empty Pipe Sensor (ES1), the Flow Sensor (FS1) and the Pressure Sensor (PS1) should be located in the legs between the valves AV2 – AV10 and also a set between the valves AV1 –AV9. This would allow switch over at the moment the pipe below the valves AV9 and AV10 is still full off liquid.

 

[QualityTime]

I think what is happening is that when the TANK IS FULL and the mixer is running there IS ENOUGH depth and residence time to allow the air that is introduced by the mixer to rise to the surface. When the LEVEL IS TOO LOW there is NOT ENOUGH TIME FOR THE AIR TO RISE TO THE SURFACE BEYOND THE ZONE OF INFLUENCE before it gets sucked into the pump.

I have run across this type of problem before in a cooling tower installation and it is documented on this site. The cooling tower manufacturer's rep set the working water level as per the levels they have always been doing it. The rep advised that they have always set the working water levels to this level. We were getting air being introduced into the cooling tower pumps and it was causing a lot of problems. I could see the effect of the air on magmeters that I installed on the cooling water pump discharge. I then reviewed the working water levels DIRECTLY with the cooling tower manufacturer and found that THEY HAD historical conflicting published data showing what the water levels should be. The manufacturer looked into it and advised I was right. At the end of the day, the manufacturer had the rep raise the working water levels and the air problem disappeared....and we were only talking about raising the water level about 1.5 - 2 inches!!

Getting back to the problem…some potential solutions:

1 ...if you can live with shutting off the mixer when the level reaches a certain point you that is the easiest solution.

2. If you need mixing I would talk to your mixer manufacturer and look at a blade design that does not entrain air. The other thing is that mixers are typically designed to work well when the tank is FULL. When the tank level drops the mixer performance IS VERY DIFFERENT so you should look at this fact very carefully with your mixer supplier.

3. You might want to look at some form of non violent jet mixing using a recirculation pump.

4. As far as the valve sequencing issue you should be able to solve that by reprogramming the PLC

 

[beattsmjk]

Many thanks for the excellent comments, in particularthe last two.

FMJalink, I appreciate your thoughts very much and you have some valid points there that I will definately look into.

In terms of the changeover, I switched off the agitators at 1500 Litres (they normally switch off automatically at around 100 Litres). I then watched the process over a period of 5 hours and am please to say the tank changeover took place without any problem.

There was a drop of pressure to zero when (as mentioned in FMJalinks post) the old tanks outlet valve closed and the new one opened. There is no check valve because the system waits for one valve to close before allowing the other to open, therefore for a second there is no flow.

I will look into adjusting this, perhaps by repositioning the sensors as you mentioned.

Many many thanks chaps, it is much appreciated. I will keep you posted on progress.

 

[QualityTime]

I really want to stress Point (2) I made in my last post:


THE MIXER MIXING ACTION IN A TANK IS COMPLETELY DIFFERENT WHEN THE TANK IS FULL COMPARED TO WHEN IT IS PARTIALLY FULL. It could be that when the tank is full there is no air being introduced into the process and when it is partially full the action of the mixer blades is quite different and as a result you might be causing air to get into the process. THE DEPTH OF SUBMERGENCE OF A MIXER BLADE IS REALLY IMPORTANT. I WOULD TALK TO YOUR MIXER MANUFACTURER FOR POTENTIAL SOLUTIONS. Maybe there should be two mixers..each placed at stratgic levels and they turn on or off when the tank reaches a certain level...just a suggestion

 

[beattsmjk]

Thanks Qualitytime,

I completely agree with you. When the mixer is fully into the liquid, there are no bubbles at the surface, yet when the blades are submerged by say 500mm, there is a dense layer of foam on top of the yeast.

I will definately look into this more

Many thanks again!

 

[QualityTime]

...where do I send my bill.....

 

[QualityTime]

Hi Beattsmjk:

One other thing...and I don't want to sound belabor the point I made earlier in this post....BUT IF IT IS THE MIXER INTRODUCING AIR INTO THE SOLUTION AND EVEN THOUGH IT SOUNDS LIKE CAVITATION IS OCCURRING IT IS NOT CAVITATION BY DEFINITION. In this case the potential solution is to fix the problem with the mixer so you don't get the foaming.

Cavitation, by definition, is because you have a lack of NPSHa and the solution to that is to, for example, reduce the friction losses in the suction piping by, for example increasing the suction pipe size.

As I have said earlier in this post, THE TWO PROBLEMS PRODUCE THE SAME SIMILAR CAVITATION SOUNDS BUT THE SOLUTION IS ENTIRELY DIFFERENT.

 

[Artisi]

It would seem that the first response from Compositepro and my first response were pretty well on the mark, the agitator beating air into the liquor.

One of life's lessons (free of charge), if an agitator is involved in a process, it is the number 1 suspect when it comes to inlet problems.

 

[QualityTime]

So....what has happended since?

 

[beattsmjk]

Hi Qualitytime,

we have had flawless changeovers since we stopped the agitator at 1500 litres. I have now written this into the program and all seems well.

Many thanks for your help.

 

[Artisi]

Well, what a surprise!

 

[QualityTime]

Hi...you are welcome!