Dosing Pumps: typical P&ID

[Subystud]

Hello -

I am posting in regards to a typical P&ID for a dosing pump. Please find the attached .jpg describing my proposed P&ID.

The setup is 3/4" sch. 80 PVC plumbing, necks down to 1/2" through a flowmeter, then expands to 3/4". The reduction/expansion may be removed, but presently this has been built.

My questions are:
1. Is this consistent with what you, personally, have experienced/seen/designed?
2. What are the minimum instrumentation requirements when dosing accuracy needs to be 10 Liters per hour, +/- 0.1 Liter per hour(1% tolerance).
3. Threaded connections can be troublesome - do you recommend socket welding the entire assembly?
4. Calibration columns: my frequency of calibration is 1 year. Would it be wise to simply make a union connection such that a calibration column can be attached, as needed, rather than permanently plumbed in?

Your thoughts and considerations are appreciated.

Thanks -

 

[Subystud]

Attached image.

 

[Subystud]

http://i1292.photobucket.com/albums/b571/rotag4life/PampIDDosingPump_zps7c5f5b86.jpg

 

[Subystud]

Finally got the image posted. Sorry about that.

 

[sshep]

Hey subystud,

It generally looks right. What is the BPV for, what type FM, what type of dosing pump?

The following look unusual: you don't show pulsation dampeners yet a PD pump is indicated. A flow meter seems uncommon with PD dosing pump in my experience.

Best wishes,
Sean

 

[bimr]

The flow meter will not work with the pulsating flow. That is why the calibration cylinder is used. Delete the flow meter and put the cylinder on the pump suction.

The pressure relief valve should be in the discharge side and the valve should relieve to the suction tank.

Ball valves are commonly used.

You need to install some vents to atmosphere to allow the system to be primed. As shown, it will be difficult to start the system.

You may size the piping for a flow of 3 times the average flow rate to reduce the head loss caused by the flow pulsations, or you can add pulsation dampeners. If the piping is sized for the average flow rate, then the system will experience high piping head loss.

See the flow schematic.

http://jaecofs.com/metering-pumps-startup-problems.html

 

[Subystud]

Thanks for the comments. The FM is unnecessary (and non-functional...) as the positive displacement is verified with the calibration column. How often would you calibrate?

Regarding sizing of the pipe: I can perform the calcs all-day to figure the pressure drop, but do not necessarily know what a baseline suction flowrate into a Diaphragm pump? Is it based on the maximum flowrate of 5 ft/s through PVC plumbing? Which pressure drop should I be concerned with?

So, if I'm moving fluid at 1000 L/h, size it for 3,000 Lph? Also, I'm pumping chemical. Sulfuric acid is 6x more viscous than water, adding to the pressure drop.

Thanks-

 

[bimr]

Calibrate weekly. The cylinder also allows the operator to visually verify that the pump is working.

Sizing the suction for 5 ft/sec is adequate unless you have extensive distance between day tank and pump.

Correct, sizing the pipe for 3 times the pumping flow rate helps to account for the pulsating effect of the pd pumping.

 

[Subystud]

I'm still unsure of how to size the piping. What is an acceptable pressure drop per length of pipe? Overall?

The pump can overcome about 25ft of head.

Thanks -

 

[bimr]

The piping should be sized for 4-6 ft/sec with a design flow of 3 times the pumping rate. The "3 times" the pumping rate is used to account for the pulsing flow of the PD pump.

The backpressure on the pd pump will be the piping headloss. pd pumps are typically rated for much higher than 25 ft of head.

 

[Subystud]

So, if 1000 Lph is my desired pump output, then I size my plumbing to 3,000 Lph at 5 ft/s.

Thus, my pump velocity is one-third of 5ft/s? Would I not just design to 1.5 ft/s for all plumbing, for example? Or is this accomplishing the same sizing goal?

Thanks for the input.

 

[bimr]

It is accomplishing the same goal. The concept is that the fluid velocity varies with the pulsations from the pd pump. If you size the piping for the normal flow, excessive headloss will occur because of the pulsations.

 

[sshep]

Subystud,

The pump you selected looks pretty good, and the installation manual can give you some advice. You should especially look at the ancillaries like pulsation dampeners, and the multipurpose valve (relief, anti-syphon, etc). It looks like the pump can take a 4-20 milliamp input control signal and other communications which is a great leap forward from the manual speed or stroke controller. Instead of measuring flow, you can infer it directly and remotely via the pump.

With respect to calibration, my own experience with these calibration cylinders is that they go unused in the long term. I guess it depends on the control objective of the injection, and how accurate you need to be, but usually you find these in the field all yellowed or broken, and showing no signs of use. This is my experience, others may have a different experience.

best wishes,
sshep