Back flow in a compressor / pump

Look at the center impeller in the picture. In operation, the impeller is turning clockwise. Each vane of the impeller acts like an airplane wing as it moves through the fluid causing a partial vacuum on the suction (leading) side and positive pressure on the discharge (trailing) side.

So, you have an area of low pressure in the center of the impeller transitioning to high pressure on the outside of the impeller. The impeller in the picture is center suction, right side discharge.

I'm not sure I've done a good job explaining this, but it's a start.

That's very well stated and the illustrations you supplied certainly help in understanding how the fluid moves from the eye (inlet) of the impeller to the tip and then into the diffuser area of the pump casing where the head pressure builds as the velocity energy is converted to head.

So from what you are saying, the suction fluid that get suck into the eye of the impeller "pushes" the fluid at the tip of the impeller through the discharge, meaning that this "push" actually has a higher dP than that between the eye and the discharge?

That doesn't really make sense, because the dP between the discharge and the eye is always going to be higher than the dP between the eye and the suction. So surely, the side with the higher dP will take precedence, meaning that the discharge flow should flow back into the impeller eye?

Like the man said you shouldn't be thinking dP because an impeller (keeping this to centrifugal designs since recips always utilize discharge valves which are in essence check valves) produces TDH or total discharge head based on diameter and speed. The kinetic energy imparted to the pumped fluid/gas by the driver, motor or turbine, because of the HP/KW involved, causes the fluid to leave the tip of the impeller at a much higher velocity then when it entered the eye. This dynamic head becomes converted to pressure based on the specific gravity of the medium.

The rotating impeller basically acts as a check valve as it passes the pump cut-water and the next vane comes up and starts to throw out its volume of fluid. You might say there's no going back as long as the pump/compressor is rotating at design speed. In the event of erosion of a centrifugal pump cut-water you will lose significant pressure output because of the larger gap at the impeller tip to the casing.

It is a mistake to think only in terms of pressure. The impeller imparts Kinetic Energy away from the eye, towards the discharge.

There is a slight reverse flow from discharge to suction side at the plain side (back) of the impeller, but this is almost completely avoided by reducing the clearance to as small as possible. Back pullout pumps are the best because the clearance can be reduced in the shop to the desired clearance.

There are also holes left near the eye to get rid of unwanted pressure.

I think there are two points here:-

1) A pump only provides flow (the system downstream induces pressure)!

2) The reason the pressure does not return back to the suction side is that the compressor discharges through a check valve (one way valve, commonly a plate valve). Theses valves are not perfect so there will be slight leakage backwards.

Hope this helps!

The impeller is increasing the velocity of the water entering from the center. Therefore, inside the impeller itself, the pressure is lower than the pressure at the suction port. As noted by others, the discharge pressure from the pump is determined by what comes after the pump.

If you were considering a positive displacement pump, the pump usually discarges through some sort of check valve (reed valve, for example, in an air compressor).

Centrifugal Compressors: The name defines it. The fluid is ejected at high speed on the perifery of the impeller (Simplifying here) towards the exit aperture of the compressor. A pressure is existing with in the casing. Some of the fluid will flow back to the eye (Suction side) and cannot be avoided 100%. This flow back is minimised by having the clearance between the impeller and the casing reduced as much as possible. This dictates the efficiency of the equipment (not the only parameter for efficiency since the viscosity of the fluid comes in...).Therefore, as you can see, there is a flowback but limited and only from the perifery of the impeller back to the eye, via clearance. the overall result will be a pressure generated at the output. If you shut the output, you will have a static pressure reading wich shows the max pressure developped with the existing backflow .

Other Compressors: Usually positive displacement types etc... have inlet and outlet valves to restrict the backflow...

There are also impellers with vanes on both sides, primary vanes on the front and smaller secondary vanes on the rear to reduce backflow.

Yes.

But I think that the secondary vanes are there to balance the Impeller: The suction generated at the front tends to pull the impeller forward. therefore, the back vanes will be pulling the opposite way and reduce the forward force... I Think this is why but could be corrected.

I believe that you will find the vanes on the back side of an impeller in mostly the open face design on single suction overhung pumps. They are used to reduce the pressure at the stuffing box area and on the rear side of the impeller. The open face design impeller tends to have more unbalanced forces than the shrouded (closed) designs. On double suction sumps the forces cancel and it become moot.

I've just seen a few. You are probably right.

Good question. In positive displacement pumps and compressors when the piston is in the discharge stroke the suction (inlet) valves closes, preventing backflow, and the discharge (outlet) valves open allowing the fluid to leave the compression chamber. Centrifugal pumps and compressors by the design of their impellers impart velocity to the fluid in the direction of the discharge opening and do not have operating suction and discharge valves. This velocity energy is converted to head pressure when the fluid velocity decreases because of the increasing area of the volute of the pump casing. Check out the pump cross section in a pump manufacturer's catalog or a good engineering manual. Only if the pressure on the discharge side exceeded the pump's ability to overcome it would you have back flow. This could happen if the pump selection was incorrect...it's the responsibility of the design engineer to carefully examine the pump curves to assure that the proper pump is selected to match the "conditions of service"...COS.

One thing that we all seemed to have overlooked is that if you were to open both the suction and discharge valves on an idle centrifugal operation where there are existing system pressures then you will get reverse flow and the pump/turbine basically becomes an expansion turbine and is driven in reverse. That is one of the reasons that one generally starts a centrifugal machine with the suction valve open and the discharge valve closed, opening the discharge valve once the machine comes up to speed and has developed sufficient operating pressure to overcome the discharge condition.

"That is one of the reasons that one generally starts a centrifugal machine with the suction valve open and the discharge valve closed"

Wow, that's the FIRST time I ever heard of deadheading a pump to start it.

I've built literally hundreds of machines, with thousands of pumps from fractional HP to 15 HP and we NEVER deadheaded a pump to start it, ever.

I'm not saying you are wrong, just that I've never heard of that practice.

http://www.gouldspumps.com/download_files/3600/Goulds-I3600-10thEd.pdf

Use this link to page 41 of a Goulds API 610 Pump Installation and Start Up Procedure. We're talking a lot more HP here.

Keep in mind that with the discharge valve closed you are pulling very little HP but it is all going into heating up the fluid so one doesn't run there for very long. As soon as you get the pressure up you start opening your discharge valve and crank it open full and let your system control valves and instrumentation take over from there.

Also even cracking the discharge valve of a pump on a high pressure application will cause it to spin backwards if the suction valve is open and you can get a lot of shaft torque applied when you press the start button or open a steam valve. It's not good to try starting a pump or compressor that's rotating in the wrong direction.

Just because that's the way I've always doen it, doesn't make it right!

As you say, it matters more with 100HP pumps than the small one's I used.

I will revise my thinking.

Cheers.

also the pomp has to know when to turn on and shut off (cut in and cut out press.) and when the pump shuts off the check valve holds the static press.

H T H cheers