Min flow for old pumps, manufacturer not around

[Patassa]

Hi,

A client is wanting to trip their crude tank pumps at some unspecified minimum flow. I have obtained the data sheets and pump curves for the pumps, they were made by "Pacific Pumps, Inc" and the requisition is dated 1966 so I don't think going back to the manufacture for minimum flow requirements is going to help me much.

How should I go about calculating the min flow at which we should trip the pumps? Is there a good rule of thumb that's conservative enough to pass scrutiny without a manufacture's recommendation? They were built to API 610 standards.

 

[Patassa]

To clarify, I'm looking for thermal protection. The event the client has in mind is and upstream or downstream control valve slamming shut,

 

[BigInch]

You might calculate the heatup rate of the fluid in the pumped fluid at the efficiencies of the low flow levels. The lower flowrates will have poorest efficiencies, causing excess power to be converted to heat and heating up the fluid. When the temperature rise over 10 to 15 minutes, or however long they may want to flow at those low rates is too great, you've hit the low flow limit. I have found this to be around 20% of BEP flow in most cases. I would limit the minimum flow rate calculated in that manner to start-up, or shut-down conditions only. Continuous operation at those levels, in fact any level below 50%, or even higher, has cumulative detremental effects that you will eventually pay for one way or another. Continuous flow rates really should not be much lower than 50-70% of BEP flow. Even that may be too low in some cases.

I hate Windowz 8!!!!

 

[1gibson]

You might be surprised if you check with the manufacturer, a lot of older product lines are still around.

To answer the question for rule of thumb, determine the specific speed (Ns) of the pumps, use 30% of the best efficiency point if Ns is less than 3000 (it should be for a crude oil booster) or 40% if it is greater than 3000.

If you just want thermal, find the pump curve, check the efficiency at 30% of BEP, assume all the lost power goes directly into heating the fluid, see how it looks. This is almost never going to be the limiting factor for determining min flow.

Min flow is not only for thermal concerns, it is also the limit for guaranteed vibration levels during factory testing. Be sure to monitor vibration if you plan on running for extended periods at reduced flow.

But if you just want thermal protection, why not add a thermowell with an RTD into your discharge piping?

 

[Patassa]

Well let me ask you guys another question, if the client wants the pump to trip when a control valve (upstream side) fails closed, what should be my governing flow rate? It's going to pump the line dry pretty quickly so I assumed thermal protection would govern but I don't have much experience with this.

 

[LittleInch]

That was going to be my question which is what are you trying to do / solve?? If the valves either on the inlet (usually to be avoided) or the outlet close (slamming shut) why do you want to find a minimum flow? If these pumps are any serious size, then a simple discharge to inlet re-cycle will heat up very fast if the forward flow through the system is actually zero. If the inlet valve closes first, then you can quickly get into a situation of cavitation, but not normally "pump the line dry".

If your valves fail closed, then just trip the pump. You don't need to know what the min flow is.

It all depends on the size, flow and control systems in place. BI and 1gibson have some good advice, but works on the basis that there is some continuous forward flow through the system. whenever I've tried to get a minimum continuous flow figure from a pump vendor for a decent sized pipeline pump, they often squirm about and then finally come up with a figure not far from the 30% of rated flow mark after a lot of pushing. Smaller more robust pumps could go lower. The same applies to the question how long can the pump operate with a shut in outlet line and no re-circulation - often comes up with a figure of about 1 minute, but again is something the vendors don't like to give guarantees about.

why have you got a control valve upstream your pump - sounds like a bad idea to me.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Patassa]

The way the problem is being defined is "MOV failure upstream of the pump" and "MOV failure downstream of the pump". This equipment is in service and it's beyond our scope to modify their current set up. We are just being instructed to protect the pumps with a "low flow shutoff".

 

[1gibson]

What is the specific speed, or what does the power curve look like? If it is what I am assuming, shutoff will be lowest HP point on the curve.

Now that you've fully described the problem, the solution is to set low amp trip for the motor.

Upstream valve closed -> cavitation -> low hp draw -> low amps -> trip
Downstream valve closed -> reduction in flow -> low hp draw -> low amps -> trip

Run the pump intentionally at shutoff for 30 seconds maximum, record amps, add a small margin and set that as the low amp trip.

 

[Patassa]

1gibson

The power curves are like what you're imagining, more or less linear with the lowest point being at no flow. I mentioned the idea of tripping the motors based on low amperage and my electrical guys tell me it's "hard to do". Right now the prevailing wisdom is saying "cut in orifice plates and have a low flow rate trigger the trip". Which put the calculations back on my plate.

 

[BigInch]

Or, assuming you have a process flow minimum, trip the pump when it dips significantly under that. Minimum flow doesn't always have to be based on thermal, cavitation and vibration concerns. Common sense often produces much better results.

I hate Windowz 8!!!!

 

[Artisi]

How about loading the data sheet / pump curves.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Ok, so what I think you're really looking for is a sensible figure to set the low flow switch to so that it trips the pump when the flow stops for whatever reason, but isn't so high that it trips the pump in "normal" operation. Therefore I would still go for the 30% of nominal duty limit. Of course this switch needs to be inhibited on pump start for say 30 seconds otherwise your pump will never set off. You might also want to put a timer on this low flow switch of again about 30 seconds or maybe a little less (15?) so as to avoid tripping the pump on a very short period of low flow or some other transient event.

You should be able to justify this in terms of heat up time unless you've got a really high power unit which could heat your liquid up in a very short time.

I would also add a high temp trip on the casing as a backup as flow figures on a shut in line sometimes go a little odd, especially if your orifice plate is very close to pump discharge.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[ScottyUK]

If you want to trip the pump when the inlet and discharge valves are 'not open' then use a limit switch which makes only when the valve is fully open. Add a casing temperature measurement, either via DCS or a trip amplifier, and trip the pump on that too. Direct measurement of the parameters you are concerned with is always going to have greater certainty than trying to imply them from the electrical power input to the pump driver.

 

[DubMac]

To your original question, the old equation for Temperature rise in a pump (don't know why a new one would be any better) is:

Temp Rise, DegF= (TDH,ft./778) * ((1/Efficiency at point considered) - 1)

If a manufacturer of a certain pump says the max permissible temp rise is 20 degrees, then you can back-calculate your minimum flow using the performance curve.

Several Pacific Pumps are still running at refineries all over the country. However, depending on what model you are talking about (HVC, SV,ROV, JTC are some of their old models), tech info for them will be found through either Flowserve or Goulds current offices. Call a Sales Engineer at one of those two. You also need to realize that your pump has been refurbed many times since '66 and could bear little resemblance to original. The OEM repair shops seldom throw away original files; and their upgrade engineers are typically very helpful to end-users.

If interested in the history: Pacific bought Worthington and became Dresser Pump - Dresser Pump merged with Ingersoll-Rand to become Ingersoll Dresser Pump which then joint ventured with BWIP (Byron-Jackson and United) to form Flowserve.....whew!! Because of monopoly concerns, the Justice Dept. made IDP sell several of the Pacific models to Goulds; that is why the model you seek info on could be at either Goulds or Flowserve.

 

[Rinchterninsob]

On Dubmac's point the pump has definitely been refurbished since '66, although not necessarily by the original manufacturer. Typically, there will be an additional nameplate on the casing of the shop that performed the work. That being said, it could have been restored to OEM spec or updated to current API. I realize it's originally an API pump, but API 610 has changed quite a bit since '66. Also, performance characteristics may have been changed during refurb, such as destaging or trim changes. Even if it hasn't been modified, internal erosion of the casing over time can affect pump curves, so your curve/IOM is almost certainly not accurate.

Minimum flow is indeed established for vibration, not temperature, although it is true that the additional energy put into the pump that is not transferred into flow or vibration will turn into heat. In the case of a suction restriction, the concern is cavitation. A discharge restriction yields suction/discharge recirculation. Both cause an increase in vibration.

Typically, vibration is at it's lowest at BEP. Your best bet is to throttle the discharge while monitoring vibration to determine BEP, then throttle down until vibration reaches an unacceptable level, add a buffer (5-10%), and call that your min flow. Otherwise, you can contact a pump manufacturer with a test facility and have them test the pump and produce a curve for you. Best PumpWorks has a test facility in Tyler, TX and Shreveport, LA that has this capability.

If temperature is the concern, an RTD on the discharge with a hard set point is a poor choice, as there are too many variables which can affect it, such as temperature of the product itself or ambient temperature. You would need one on the suction as well in order to detect temperature rise via SCADA.

For equipment protection, vibration monitoring is going to be your best and most versatile bet. Install a couple of transducers, run them to your PLC, set your alarm/ESD at .20 IPS, and it will definitely trip in a dead head/cavitation situation. Additionally, you're then protected from bearing failures, shaft shears, and mechanical imbalances due to buildup in impeller vanes, as well as being able to perform condition monitoring and logging data so that root cause analyses are easier to perform .

Just my .02.