Does anyone in the forum have experience with spring mounted pumps (recip, front-suct, top-disch)? Any background info, modeling tips, etc would be strongly appreciated.

I never designed this pump with spring support. However, I would like you to check if spring support is really necessary in this system. My understanding is typical reciprocating pump system in oil & gas and refinery is small size with low temperature, most of cases are chemical injection and utility system. So, generally spring support is not necessary.
Spring support is not a good idea for pulsating line because it may act as another exciting source of vibration when it coincide with exciting frequency range of pulsation.

Thanks for your response, Sun Wee.

Spring-mounted pumps, as you say, are unusual. My own experience--like yours--is mostly in oil refineries and I have never seen a such a pump in those facilities. I have, however, seen spring-mounted pumps in paper plant chlorine dioxide plants.

My current project is a food processing facility. The pumps are slow speed and for process reasons are mounted directly below a column with a single 5D bend between the tower bottoms nozzle and the pump suction. The project is essentially a copy of an existing plant which also uses the spring-mounted concept. In the new plant the pumps are slightly larger so I need to fnd some way of sizing the springs. We have found no record of how the sizing was done on the existing plant.

I have similar situation as you, which I am dealing with a slurry system, mainly oilsand with high viscosity and abrasion. I use 5D bend with two Victaulic Couplings and slurry pump is fixed. The thermal expansion from equipment bottom can be absorbed by this flexible joint.
If seems that your process is a kind of solid stub or high viscosity fluid. In this case, another suggestion is to use packed type bellows(rubber or metal, it depends on fluid and how much abrasive).
If you cannot use any type of flexible joint in your system, spring mounted pump will be a solution.

It seems that the selection of spring size is not easy. If I will have to deal it,I would like to check;
-consider the total weight of pump assembly and manipulate the coil size to mitigate the flange nozzle load.
-check the natural frequency of piping system and spring assembly and compare to the possible exciting frequency of reciprocating pump. The resonance range must be avoided.
-if system is faced with a vibration problem at field, add some sway brace, snubber, or other anti-vibration devices at field. For the centrifugal pumps deal with slurry and viscous fluids, speed is very low, that means you may need high-frequency tuning.

Let's see if other members have better idea.

I have done similar analysis of pump operating at -196 deg temp. In that case I take complete weight of pump assembly ( as recommended by Mr Sun wee ) at C.G. & connect both suction & discharge nozzle with corresponding piping with connecting node. That pump was hanging from top with hanger spring. I had modeled
Pump supporting rod assembly also. Since pump suction, pump discharge & pump assembly has been considered within a system I believe it will work. Our client also review the stress system with there previous stress package it's working fine. I have not checked frequency. It was a centrifugal pump.

Regards

Habib

Sun Wee & Habib,

Thank you both for your valuable suggestions.

If you want CAESAR II to "size" the springs under the pump, I suggest you use CNodes with anchors at the two flanges and "free" these anchors for hanger sizing. This way, the pump springs will carry all the pump weight and the piping springs will carry all the piping weight.

If you don't do this, you'll have pump too invoved in structural support of your system.

I analysed similar arrangement a while ago for huge axial pump. We found that rubber dumpers were better solution as they ware more robust and less flexible then springs, but still much more flexible then classical anchor.

I modeled rigids, no weight, for the base frame, as per GA drawing and CNoded it to the pump. Six dampers were modeled at their actual locations on the frame. This provided more accurate results with regards to the pump movements.

Good luck!

El Gringo,

I have several experiences on this for a vibration isolation cases where pumps are in spring mounted base sometimes they call it inertia-base pumps. Stress guys I worked with at that time were really highly experienced and since I was a novice that time I need to follow most of the mentoring. So, I have modeled my piping system only up to the pump's nozzle flange and the stiffness would be up to you say 10^5, 10^6 or even lower than this. The stiffer, for me, would be so conservative because in actuality the system reacts in a soft spring in all six degrees of freedom. I tried to acquire spring stiffness from the pumps manufacturer but I was not successful. So for me, if you have to model it properly, you have to model the system including the springs. The base of the spring for your boundary condition should be very rigid I suppose most of the time. Just my opinion.

Ed-Lamigo

the model should be highly detailed, taking into account centroids as well as accounting for ALL weights....

The more detailed the model the better the results....