Erratic pH results in product stream from PD chemical injection pump? 4

[mjpetrag]

I have a very low flow PD pump that puts an basic chemical into a product stream to raise pH. The flow of the chemical is typically 0.3 GPH and doesn't vary much. The PD pump flow is controlled by the stroke of the piston on a 1200 RPM motor. The flow is measured with a small coriolis meter. The chemical pipe is 1" diameter and the distance to the product stream is approximately 300' away. There is a 50' horizontal run before it 90's down into the product piping.

The product stream that the chemical is pumped into is 4" and the flow is approximately 10000 GPH. After the chemical ties into the product stream, it flows through a static mixer and then travels approximately 100' away to a product storage tank.

The results for pH are taken just before the combined stream enters the storage tank.

The problem we're having is that the pH results vary between 0.5 pH to 3 pH from sample to sample even though the flows have remained the same.

We sampled the tank and the pH is relatively constant at 2.5 pH across samples.

I'm wondering if we have a situation where we have a partially full pipe for the chemical injection and even though the chemical flow is reading fairly constant, the material could be building up and we get waves of the chemical in to the product stream causing the variation in pH.

Am I correct in my thinking?



-Mike

 

[bimr]

Centrifugal pumps and most rotary pumps provide smooth, non-pulsating flow, while reciprocating (PD)pumps produce a pulsating flow. A pulsating flow may require special design considerations in the piping system. A pulsation dampener may be needed to be installed in the discharge piping to reduce pressure surges resulting from the pulsating flow.

A pulsation dampener with an injection valve may improve the operation.

Review figure 1 in the link below:

http://www.miltonroy.com/corp/details/0,10294,CLI1_DIV39_ETI3660,00.html

The flow from the pump is not constant so you should expect different pH readings over time if you are sampling the pipeline. The pumped volume averages out over time in the Storage Tank so that the pH readings in the Storage Tanks should be constant.

If you can not live with the pulsed flow from the PD pump in the piping system, then you will have to install a different type of pump.

 

[mjpetrag]

Right. The thing I'm confused about here though is the pulsation is going at 1200 RPM. So the pulses of flow are 20/second. It takes about 4 seconds to get a sample so it should even out. And it goes through a static mixer and has about 100' of pipe to mix also. I would think this is enough to average out but maybe I'm wrong.



-Mike

 

[bimr]

Are you sure about the 20 strokes per second. It sounds too fast. That may be 20 strokes per minute.

In addition, the static mixer is not going to help that much since there is no back mixing.

 

[Latexman]

Velocity in the 4" line is about 4 ft/sec. Velocity in the 1" line is about 0.002 ft/sec. Does the 1" line just tee into the 4" line? If so, you may have several things going on.

A partially full 1" line? Maybe. Is there a good check valve where the 1" line tees into the 4" line. If not the fluid in the 4" line may backflow up the 1" line.

At 0.002 ft/sec there is no "injection" into the main stream. It "drools" in and clings near the walls and gets smeared down the line against the wall.

Do you have a detail drawing of the union of these two pipes? I hate to go on and not know what is really there.

Good luck,
Latexman

 

[mjpetrag]

Yes I will send more details out later in the day.

Bimr, looking at your post closer, I have to check if the drive from the motor to the pump has a gear ratio on it. It does seem like an awfully low flow/stroke if its 20/second.

-Mike

 

[Latexman]

Is there much difference in density or viscosity between these two fluids? I ASSuME they are soluble.

Good luck,
Latexman

 

[Compositepro]

Your lucky such a system works at all. The velocity in the 1" line is so low that gas bubbles cannot be flushed out. Sunshine on the pipe getting blocked by a cloud probably has more short-term impact on the flowrate out of the pipe than does the pump flow rate.

 

[micalbrch]

Can you post the PD metering pump brand and model? That would help to find out whether bimr's assumption about the pulsation effect is right. I do not think so because the pump seems to be very small but.... How long is your system in operation? Just started?

 

[bimr]

This thread seems like a game of 20 questions because of a lack of details.

It is generally not practical to pump 300 feet with such a small pump. The discharge line should be short and direct, typically 10-20 feet.

If piping contains pockets, it is possible that periodic siphoning or other odd flow conditions may occur. Avoid pockets or high point where air can accumulate in the discharge piping. An ideal pipe run is one where the piping gradually slopes up from the pump to the outlet. This will ensure that any air in the discharge side of the pump can be evacuated to the outlet.

Conveying fluids are sometimes desirable to meter a chemical dose into a stream, Conveying fluids help when pumping over long distances and/or with low chemical doses. Not sure if conveying fluids are practical for your application.

 

[mjpetrag]

Thank you all for the replies.

Going back to it, the majority of the piping is actually 1/2" and expands to 3/4" right before it injects into the 4" main. There is a check valve just upstream of the injection point. There is a 50' horizontal run that contains the check valve, and then expands to 3/4". The 3/4" piping goes another 5' and then 90's down into the 4" main. The connection into the main piping is just a 3/4" coupling at the top of the pipe.

We have a parallel unit that is essential a mirror process, fed by a buddy pump right next to the one in question. This unit runs very similar conditions, however the pH results only swing between 2.5-2.7 between samples. The biggest difference is the length of piping to the unit is about half of that of the other.

Walking out the chemical injection line to the problem unit, I noticed there is definite bowing of the line as excepted with such a small pipe diameter.

Both units have been in operation for at least 20 years. This unit has always been a problem. The comments above seem to point to the problem being in the piping design. I'll have to go back and see what we can do to eliminate the excessive length and pockets in the piping.

-Mike

 

[JJPellin]

A reciprocating chemical injection pump of this type requires a certain amount of back pressure in order to provide accurate delivery of chemical. Even if the line you are pumping into has a relatively low pressure, you need a back pressure regulator valve to get accurate results. We see this commonly with the pumps that add sulfuric acid to our cooling water systems for PH control. Consult the literature for the particular pumps you are using and they should recommend a minimum back pressure. When we added back-pressure valves set to hold something over 40 psi to our acid pumps, the accuracy and consistency was very much improved.

Maintaining higher back pressure in the chemical line also reduces the chance of vapor pockets. It prevents any siphoning.

Johnny Pellin

 

[Latexman]

When designing simple injectors, I like the velocity of the branch to be 2-3 times the velocity of the run. That way it does not "drool" into the near periphery nor get blasted across to the far periphery. Sometimes this requires fabricating an injector from tubing or buying a commercial injector or (spray) nozzle. I place the end of the injector where the ID of the run would be. I've gotten good results from this.

Good luck,
Latexman

 

[mjpetrag]

JJPellin...I think that is the most practical way to go right now. I'll look into that. If it's a viable option we'll try that and I'll post results.

The injector is definitely something we're already looking into.

Awesome ideas from everyone here. Thank you again for all the input!

-Mike