8 replies to this topic

[onyxgaze]

We have a metering pump which is to pump out liquid from a liquid-liquid washing column into a downstream vessel whcih is at atomospheric pressure. The flow through the pumnp is controlled by a Variable Frequency Drive adjusting speed in resonse to a interface level controller in the column.

The problem is the pump which is at the bottyom of the column sees a static suction head equal to height of column (approx 20 mts) whereas the discharge head is much less.

Due to the suction head, there is flow through the pump even at standstill. We cannot use a control valve downstream of a metering pump since it just wouldnt work.

One option to solve this problem, in our opinion is to have a restirction orifice place at pump suction. Is this a correct solution? If so how does one go ablout designing such an orifice?

Another option would be to change discharge piuping and raise it to the height of the column which would provide a backpressure equal or greater than the suction head.

A third option would be to provide a back pressure NRV whose spring is sized for slightly greater than the suction head.

As per the metering pump vendor, metering action would be incorrect if the discharge head is lesser than the suction head.

What would be the best option to choose from the above.

Thanks in advance.

[pleckner]

The correct solution is to use a backpressure regulator.

If this is a motor driven pump (rather than air driven) make sure you put in a PSV between the pump and the regulator to protect the piping if the regulator won't open. Contrary to popular belief, the PSV installed with the pump is not there to protect the piping or downstream equipment, it is there to only protect the pump. It will not be sized appropriately to protect downstream piping and equipment.

[onyxgaze]

Thanks.

So I would need to connect a line from the Back Pressure Regulator back to pump suction? What pressure should this be designed for?

What would happen if I installed a restriction orifice in pump suction line designed to drop most of the suction head present due to water level in the column?

The pump is a Milton Roy (India) model and is motor driven.

The pump vendor has provided an IRV (Internal Relief Valve) which he says will lift to relieve pressure of process fluid inside the pump before it can damage downstream piping.

Instead of a back pressure regulator, the pump vendor provides NRVs at the discharge of the pump, to be located after the dampner.

[Sandeep01]

Maybe I am missing something, but why do you need to 'pump' if the suction head is more than the dsicharge head required?
Regards
Sandeep

[pleckner]

To Sandeep01: Since this is a metering pump, we must assume that that the flow needs to have some control and cannot just be allowed to flow on system hydraulics. Also, we can assume that there will be a time when the suction level is low and you no longer have the suction head to drive the flow. Of course, "onyxgaze" can confirm these assumptions.

To onyxgaze:

1. Put the back pressure regulator in-line with the pump and set it for anything above the maximum suction pressure expected. This will force the pump to work.

2. The only line that goes back to the pump suction, or preferably back to the wash column is the PSV tail pipe. And I don't agree with the vendor that their internal PSV will necessarily protect the downstrem piping. Have them give you the details on the sizing of that PSV and see if it will do the job. If it does, fine.

3. Putting an orifice in the suction is bad engineering practice. An orifice can cause pumping issues, even for a PD pump. And besides, what are you going to size it for? The orifice is only good for one flow point at one pressure drop. Since you need to control the flow (variable), you will never have the pressure drop you need. If this isn't an issue, then perhaps you should re-examine the issue Sandeep01 brings up.

Hope this helps.

[Sandeep01]

Phil

I don't know how the suction and discharge heads compare at low level, onyxgaze could thow some light on that.

Assuming even under that scenario suction head is more than dischrge head, why not install a flow control valve, with the required flow meter. Keep it simple.

Regards
Sandeep

[pleckner]

Sandeep:

Your option is one that may be possible and that I cannot rule out without seeing all the details. However, since they are using a metering pump, I again have to assume that the flow would vary quite a lot and you may not necessarily have the pressure drop available for good control.

I had a similar type of problem in a previous assignment but in my case, I was meetering from a rather tall atmospheric storage tank into another atmospheric vessel that was at a lower height. However, the level in the storage tank did drop below the destination point over time. To keep the feed under control, we only had to add the back pressure regulator and it worked like a charm. I think that particular device cost me less than $100 and total installed cost was somewhere around $500. The total installed cost of a control loop as you suggest is going to far exceed that.

[onyxgaze]

Thanks for your comments & suggestions Sandeep & Phil.

Normally the flow drawn out from the bottom was done using a centrifugal pump with a control valve which operated to maintain a liquid-liquid interface inside the column close to the top. The flow is expected to vary from low to full loads of the column.

In our case, designed for a plant with much lower capacity, the column has to handle HCl acidfied glycerin water in the tower bottom. The normal flow at 100% load was quite low for which centrifugal pumps were not available.

Hence we had to opt for a metering pump. With a positive displacement pump a downstream control valve would not have been feasible for controlling flow. Hence we had to resort to a VFD control on the metering pump.

So with no control valve in place, my guess is to opt for the back pressure regulator. We simultaed such a device by placing two spring loaded NRV in series (we had to jury rig this design since the site is rate remote) and getting a back pressure regulator is going to take time. Each NRV has approx 1 bar spring lifting pressure and we did see a marked improvement - there was very slight flow at pump discharge in standstill condition and flow picked up once the pump was started.

Will investigate further and post the final results.

Best regards.

Sandeep:

Your option is one that may be possible and that I cannot rule out without seeing all the details. However, since they are using a metering pump, I again have to assume that the flow would vary quite a lot and you may not necessarily have the pressure drop available for good control.

I had a similar type of problem in a previous assignment but in my case, I was meetering from a rather tall atmospheric storage tank into another atmospheric vessel that was at a lower height. However, the level in the storage tank did drop below the destination point over time. To keep the feed under control, we only had to add the back pressure regulator and it worked like a charm. I think that particular device cost me less than $100 and total installed cost was somewhere around $500. The total installed cost of a control loop as you suggest is going to far exceed that.

[pleckner]

onyxgaze:

You reported: "...there was very slight flow at pump discharge in standstill condition..."

You just didn't have enough back pressrue, that's all. With a properly set back pressure regulator, this will not happen.