Backflow prevention to prevent damage to compressors/reciprocating pumps. I've had this same issue with a check valve directly downstream of a compressor. The problem is that with constantly fluctuating flow velocity, typical check valves will chatter and are subject to failure through fatigue or wear of the components leading to sticking.
When we went to a very good manufacturer (Mokveld, axial disk check) for the compressor situation, they looked at our flow conditions and declined to quote and sent us down the street to Hoerbiger, who makes a pipeline size valve very similar to the internal valves on a compressor, and is designed for millions of on/off cycles. (The fact that they didn't just sell me any valve is why I call them very good.) I spoke with Mokveld and off the cuff they said that they would not sell a valve for this service 90% of the time due to risk of failure. I don't think the Hoerbiger valve is suitable for liquid service.
I don't think that either a swing or piston is suitable because the constant motion will eventually cause wear and failure of the hinge-pin or jamming of the piston. I read a report on a valve in the nuclear industry about this exact thing happening in the mid 90's.
I'm looking at a variety of sizes and flow conditions, but most are from 3"-6".
I do have someone who says they can make me a valve that will work, but I'd like to know what everyone's experiences are before I buy these valves.
I've used both reciprocating pumps and compressors. Unclesyd is correct in inferring that you already have a check valve in your pump. It's called the discharge pump valve. If you want to control frequency pulsation with additional check valves, that's OK. It may work, but it depends on the system and other characteristics. A formal system dampner would be much better.
If you're after arresting counter (or back) flow with a check valve in this application, it's a waste of time and money. You already have two (2) check valves doing exactly that: the pump suction and discharge valves.
I've effectively used Hoerbiger "check" valves on compressor discharges and they worked fine for what I wanted to accomplish. They are, as you've noted, nothing more than compressor valves - plain and simple. They will not work with a liquid fluid and I highly recommend you not even think of using them as such.
Your application works out to defining what it is that you are trying to accomplish. If it is pulsation damping, then more check valves on the discharge don't hurt you - but they add cost and maintenance obligations. At 3 -6" sizes, you are facing a formidable cost and maintenance factor(s).
A formal and well-designed capacitance/dampner device would resolve that problem more effectively and positively. I would not have any confidence in someone being able to make a valve that would "work" (doing what?). Your problem seems to be either pulsation or backflow. For pulsation, a dampner will resolve the problem; for backflow, you already have two check valves built into the pump. Which one is it?
It is ridiculously over conservative design in my opinion that assumes that existing measures to control pulsation and to control backflow will both fail. The concern is that because suction and inlet valves are of same type, they will have same failure mechanism, and in case of failure, both will go. Unfortunately, I don't have the authority to remove a valve once it's on the P&ID, I just get to buy it. So my P&ID shows two check valves in series downstream of the "non-slam" type. (Don't even get me started on how often that term is misapplied by our process group.) There is already a pulsation dampener in the system, but a comprehensive analysis shows that there will still be a little bit of pulsation for a moderate length of pipe downstream.
Unfortunately, my very optimistic vendor who was sure there valve would work has now come back and stated that they are "very concerned" about the operation of these valves...just as I anticipated they would.
To top it off, apparently I am not allowed to changed the pressure drop over the valve as that would affect the PSV calcs...
Well if you just have to here is a little information on one type we did use. Hope the other fellows don't get on us too bad.
Give these people a call. We used their valves on a system with Milton-Roy Pumps, relatively low speed, with good success. We have over 40 3-5 stage Bran-Lubbe pumps and the is no check valves other than the pump discharge valves. There is a steam pulsation dampener on each pump discharge line. We do use a conical check valve on a polymer system where we inject an additive to prevent backflow of polymer through a quill.
To reichertc,
The size and required materials of construction can influence the type of check valve selected as much as the requirement for "non-slam" operation. Another term used to describe these are 'silent check' valves. Depending on pressure, size, and material of construction, some vendors you could consider
There was an oxo-alcohol plant with triplex Uraca pumps and Uhde pumps with Rockwell ceramic ball check valves in the discharge piping. Since three and two pumps were manifolded together, the check valves were needed to protect globe valve's stem-disk against operator error when changing over to different pumps. The pumps' internal valves had also been changed to ceramic. Outlet pressure gages could be watched for the 60 to 80 pulses per minute of pump discharge. Check valves were size 3" and 4" of CL1500 design with CrMo bodies but Rockwell retired the technology that made them. Another manufacturer is the Bohler Hochdruck-Technik company with the technology for high pressure valves for urea, ammonia, LDPE services. Two more possible sources are the triplex pump companies, Wepuko Pahnke ,or Union Pump, since pump manufacturers seem to have valve products or valve partners.
If this is not suitable for your installation then a spring assisted check valve such as Noreva or Mokveld non slam valves would prevent disk bounce and avoid fatigue.
