Valve for reciprocating flow, >10M cycles: fairy tail?

[TwoZero]

Before I enter the realm of imagination, perhaps it is good to attempt to verify that my expectations are not set too high, or the opposite.

What I'm looking for is a passive check valve which can last at least 10,000,000 operating cycles, with a maximum pressure rating up to 10 bar (145 psi), for filtered water, with a size between DN200 to DN800 (when dimension is too large, parallel placement might be an option as well). When the list of available options didn't approach zero by now, the check valve will have to deal with a variable flow rate up to 1000 m3/h (600 CFM) with minimal pressure loss and leakage, since conversation of energy is essential for operation of the system.

Would there be a suitable valve for this application, or which concept would have the highest chance on success? So far a spring loaded disc-type check valve may be an option, although I've seen no examples yet of these check valves dealing with these high amount of operating cycles. Some check valve designs did pass some certification tests for 1,000,000 cycles without any sign of wear, but no company as of yet has claimed to be able to achieve higher levels. So either I'm asking the wrong companies, or I'm asking something infeasible (read: something which has yet to be engineered ;-) ).

Any thoughts on this matter?

 

[MikeHalloran]

High flow rate and minimal leakage when closed suggest that the valve has to close rather fast, which does not align well with its size.

Beyond filtering, the chemical composition of the water, and any contaminants therein, can have a strong effect on the service life of any moving/sliding/rotating components.

Is it always the exact same water, or is there a continuous supply of new contamination and oxygen as the water flows?

How many truckloads of such valves will you be needing each year for how many decades?



Mike Halloran
Pembroke Pines, FL, USA

 

[georgeverghese]

Are we trying to find a very high integrity check valve to act as a barrier between a firewater system and a connected flammable process system?

 

[TwoZero]

If the shear size would be the main limiting factor, I'm also open to consider multiple smaller valves in parallel to reduce the closing time of the mechanisms. The system itself is a closed-loop system with the exact same water. The reason why a minimization of energy loss is important is that energy is stored in the water being pumped upwards by a (large) reciprocating positive displacement pump, which can be then stored in a reservoir and extracted on a later time by a turbine. In this way, when there is an oversupply of renewable energy from wind turbines or solar panels, the prospected system can act as a battery or buffer, preferable with a minimization of energy losses. I'm aware that there are different pumping techniques which would offer less complexity, but the reciprocating positive displacement pump is one of the main designs features with good reason, which also gives rise to the need of a large check valve which can handle millions of cycles.

If we're going to need yearly deliveries of several truck loads, we would have a serious problem . A valve rated for 1,000,000 cycles would only last months, which is why I'm curious to see if it is feasible to achieve (much) higher operation cycles.

 

[gerhardl]

Hello!

As others have indicated, your optimism regarding number of cycles might be 'a bit' high. Regarding minimum loss you should have a look a 'non slam nozzle check valve' Erhard or similar from other producers. My personal experience is that this is a high-quality valve with long standtime.

http://f2557522.td-fn.net/wp-conten...eet_ERHARD_non_slam_nozzle_check_valve_EN.pdf

 

[TwoZero]

What would be more realistic for operational lifetime of reciprocating valves from an experience point of view?
Is it indeed common to replace check valves with reciprocating flows annually/quarterly after, let's say, 1,000,000 cycles, or are higher reliability levels already achievable (from your experience)?

One of the options I have been considering is this axial check valve from Mokveld (

http://www.mokveld.com/upload/product_document/Mokveld-Brochure_axial_check_valve.pdf),

which looks in principle quite similar to the design of your suggestion, due to its relatively fast closing times and low pressure drop, as well as robustness of the design. Thank you very much for the suggestion, I will also attempt to contact them to see if they could provide a (rough) estimate for the number of operating cycles, so we can calculate how many truck loads of valves we need to order on a yearly basis .

 

[gerhardl]

Hello, TwoZero!

Mokveld is one of the other well-renowned companies producing valves based on this principle.

I would advise you to try to describe each application you have with a minimum. maximum and medium flow, pipeline dimension (note: valve dimension might be suggested different with cone or abrupt reduction (as suitable for the situation) from the valve producer.)

Also try to describe cycles/use per time unit and operational periodes. Also orientation of valve and flow and free pipeline before and after the valve will be relevant. Best and separate support of pipeline with separate support of the valve is a must for minimum wear.

A good supplier or producer will first on this basis be able to advice on life-time.

I will also recommend you to carefully study the factory's (general) data on pressure loss. I see this is not included in the general brochure from Mokveld, but surely available from the factory if you give the correct necessary input.

 

[gerhardl]

Hello again!

What exactly is your definition of reciprocating here?

 

[btrueblood]

A typical triplex pump with stainless steel check valves would operate perfectly fine in continuous duty, or intermittent duty, for more than 10^7 cycles of its internal check valves (I've tested several manufacturer's pumps under much harsher temperatures (180-200 F) than yours for more than 1000 hours at 1800 rpm - that's about 100 million check valve cycles). The outflow pipe could have a simple on/off valve to isolate the tank from the pump when the pump is idled. Or a fancy check valve if you wanted, but it would spend most of the time open, not cycling open/closed, when the triplex (or higher order) pump is running.

 

[georgeverghese]

Dont see anything onerous about the duty of this check valve on the pd pump discharge - a standard dual plate check valve will do. An instrumented shutdown valve on the discharge of this pump will help to prevent backflow, since this check valve will leak slightly This SDV should be downstream of the check valve. The manual isolation valve on the pump suction should be fully rated for discharge pressure.

Can see your reasoning for quick closure on this check valve, hence the idea to use a few smaller check valves lined up in parallel to enable quicker closing than a single DN800 check valve. Talk to a rotating machinery engineer about measures on the pump motor to prevent reverse rotation or minimise the effects from it. These check valves should be located on a horizontal piping run to enable maximum possible turndown on flow. Presume this pump is on a VFD drive to save on power demand? You may need an automated startup recirc line on pump discharge to minimise starting torque? If you have one, then you could use this during pump auto stop sequence to further reduce the severity of service on this check valve when the pump is coasting down to stop.

 

[TwoZero]

Thank you all for your responses so far. Unfortunately, I'm unable to change the pump, as it is a reciprocating positive displacement pump (you are quite right gerhardl, that the flow is then indeed pulsating rather than reciprocating, I apologize for the confusion), similar to a pumpjack with sucker rod, but then only single-acting. So this piston-type pump requires at least two valves for proper operation. What I can play around with is pipe dimension (to some extent) and the check valves.

The pipe dimension will be about 600 mm, with a flow rate from 50 m3/hr to 1000 m3/hr and anything in between. The average cycle time for each valve is 15 seconds, or 6000 operations per day, 2,100,000 operations per year. The maximum pressure on the valve never exceeds 10 bar. Arguably, it may be preferable to have several long-lasting smaller valves than one DN600 / DN800 check valve. Whether this means I could use tens or hundreds of 1" valves which may last 100,000,000 cycles, some DN100 valves, or something else, I'm still not sure and will depend largely on the energy loss of the specific solution and lifetime, which is essential in this story.

That the small ball check valves in a triplex pump are actually capable of withstanding these very long duty cycles in harsher environments is actually quite promising. So far, I understand that the shear dimension of the valve is severely limiting the lifetime, so, an interesting point to raise would be if it were beneficial, both from the perspective of energy loss and operational lifetime, to replace the large valves by several smaller valves.

 

[georgeverghese]

Havent seen a simplex piston pump in such heavy duty, high flow service - usually these are constrained to low flow duties. As suggested, a triplex double acting would be a better solution, as a minimum, to smoothen out flow pulsations. Still you'd need a decently sized gas charged discharge pulsation dampener to prevent severe pulsation induced mechanical damage on the piping and the check valve(s). With what you've got planned now, your piping and check valves arent going to last long.

 

[btrueblood]

Triplex pump checks are typically a lot like the caged/guided poppet designs some of the earlier links showed, not a simple ball and spring. Getting the seating area polished "just right" on the mating parts is critical to minimize leakage. But at least your velocities are nice and low, minimizing erosion concerns.

 

[georgeverghese]

Even with a triplex pump, a gas charged suction pulsation dampener is standard practice on the suction line, even if you dont have NPSH issues. But if suction pressure is close to atmospheric, an inline suction flow stabiliser should be used. In either case, the ideal arrangement is to have this stabiliser close to pump suction flange and it should be fully rated for pump discharge pressure.

 

[gustorf]

Sounds to me similar to a Flowserve/Calder DWEER energy recovery System Service.
Noreva has installed some 500 valves (6in) in 2004 that have seen 1 Mio cycles annually.
Feel free to contact Noreva for Details.

 

[gerhardl]

Seems that your dimensions for the valve have been selected after normal (conservative) flow in the pipeline, DN 600 will after my calculations give a speed of 3,5 m/s at 1000m3/h and 0,05 m/s at 50m3/h. The lowest flowspeed could give problems for most checkvalves.

Both regulating valves and checkvalves should for more complicated applications be selected after operating flow conditions suitable for the valve, not pipeline flow.

If you look at the standard allowable flow for a good designed non-slam nozzle check valve, allowable standard speed goes up to 5 m/s, and I know of cases where the factory has given OK for around 7m/s. My experience is with Erhard Non-slam nozzle check-valves, but you may also check others mentioned above (Mokveld, Nereva).

Have a look at my previous advice regarding as exact as possible operating specifications for the flow, contact the factories and check if you can go down to 500mm (should only improve performance at 5,1 and 0,07 m/s) or even down to 400mm (7,9 and 0,1 m/s). I have not checked if any producer can give you 450mm. Downsizing will also of course get the cost of the valves down.

A more complicated solution would be to split the operation to two parallell valves, one small and one larger as done with increasing flow/pressure for overflow valves. The split would have to be done mechanically by opening of the larger checkvalves inlet by increasing flow. An alternative would be to check if a stronger spring (larger operating opening pressure for the largest valve) could solve the split operation. Flow loss to be checked for all operation conditions.