Parallel centrifugal pump operation using motor and steam turbine 2

[CaracasEC]

Dear Experts,
How reliable is the parallel operation of centrifugal pumps using steam turbine and motor as drivers?
Thanks....

 

[rmw]

yes, if and only if you are able to govern your steam turbine at the same speed as the motor driven pump is turning (assuming same pump/same impeller diameter). Remember that the ST will probably have a different droop characteristic than the motor so you have to make sure to take that into account.

rmw

 

[CaracasEC]

Hi RMW, thanks and noted. Yes, the pump in question have the same performance curves.

 

[Artisi]

Question: how reliable
Answer: reliable.
however, not really sure what you are asking.

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.)

 

[CaracasEC]

Hi Artisi, The pumps in question is the reliable operation of a centrifugal pump driven by a motor and steam turbine working in parallel.
Normal arrangement will be pumps in parallel operation using motor as driver.

 

[pennpiper]

CaracasEC
We need a clarification here.
You wrote:"How reliable is the parallel operation of centrifugal pumps using steam turbine and motor as drivers?"

Did you really mean they would both be operating at the same time?
or
Did you just mean that they are connected in "parallel"?

Installing Pumps in parallel in pairs is very common in large process plants such as Refineries and Chemical Plants. This is done to insure continued operation in case of some unexpected failure. It is NOT normal to run both "Paired" pumps at the same time for long durations.
There is an operating plan (for most plants) for having one pump on "Run" mode and the other on "Stand-by" mode to keep the bearings and seals operational but the start-up and shut-down is done on the fly and is a short time event.

I feel you may need to get more information about your situation.






prognosis: Lead or Lag

 

[Artisi]

I don't see any difference between electric motor, diesel engine, steam turbine, pedal power, horse, slaves, windmill, hydraulic or air motor or perpetual motion as the driving unit to operate a centrifugal pump, so long as the driving unit has the capability and reliability to fulfil its task.

Still not clear on your question, are you asking about compatibility of having two different drive arrangements? Again, don't see any problem with an electric drive and a steam turbine drive providing the steam turbine has the same or very similar characteristics as the electric motor under any load variation.



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.)

 

[JJPellin]

We had a catastrophic pump failure and fire caused by a pair of pumps running in parallel with one motor driver and one turbine driver. This occurred in 2007. The problem in this failure was not associated with the differences between the two drivers. It was associated with the process conditions and controls.

As already noted, running any pumps in parallel (actually running both of them for normal operation at the same time) requires some additional analysis. The pumps need to be hydraulically matched. If they are the same model, the need to have the same impeller diameter, same speed ,etc.

If the system was not designed to run both pumps, you have another set of challenges. I will refer to the typical arrangements in our plant. We normally have two pumps in parallel sharing a common suction and discharge. Each pump has a check valve. There is normally only one flow meter measuring the combined flow from both pumps. There is usually only one control system to control the flow. There may be one spill-back system set up to control the minimum flow. We usually don’t have run status from both pumps into our DCS system.

In the event with the fire, we found that the total flow was more than two times the manufacturer’s recommended Minimum Continuous Stable Flow (MCSF). But, since there was only one flow meter, it was possible to determine how much each individual pump was contributing to that total. We believe that the condition of the two pumps was not exactly the same. One pump was running most of that flow, and the other one was running below MCSF. Vibration levels were very high. But, the operators had become used to high vibration on these pumps and did not take action. The weaker pump had a thrust bearing failure, seal failure and fire.

Following this event, we set up requirements for parallel operation of process pumps. First, we analyze all of the issues discussed above. Then we calculate a minimum flow that is based on the work of Val Labonoff. It is based on specific speed and suction specific speed. This results in a value for minimum flow that is higher than the MCSF. This minimum flow is more conservative and allows that there may be some difference in the condition of the two pumps. We set up a policy that requires operations to get approval before they run any pumps in parallel. And, they have to set up alarms in DCS that will alarm if they drop below our required minimum flow limit.

Since we have implemented this program, we have not had any more failures associated with parallel operation.

Johnny Pellin

 

[1gibson]

The steeper the pump curves, and the smaller the speed difference, the more reliable they will be. If your pump curves are very flat, then a 10 rpm difference could cause the slower pump to run below MCSF (worst case at shutoff, producing no flow at all) while the faster pump runs off the right end of the curve.

Create or obtain the pump curves for max and min expected speed of each driver, put them on two pages (page 1: maxrpmA minrpmB, page 2: minrpmA maxrpmB.) Draw vertical lines at slow minimum flow, and fast maximum flow (confirm your NPSH is adequate at max flow.)

If you can draw a horizontal line and hit both curves in allowable operating region, you should* be ok. If you can't hit both curves in allowable operating region, you will have problems. If a horizontal line from your fast end of curve flow is above your slow shutoff head, you will have MAJOR problems.

Short version: If fast pump runout head is higher than slow pump min flow head, you are in trouble. If fast pump runout head is higher than slow pump shutoff head, you will destroy your equipment.

*Verify the parallel conditions with your system curve, all of my comments are based on the potential for safe operation.

 

[Artisi]

The question was, "How reliable is the parallel operation of centrifugal pumps using steam turbine and motor as drivers". Without anything to the contrary, I think we should assume at this point that the OP has reviewed the hydraulic performance of the pump system and is only looking for advise on using different driving methods for each pump.

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.)

 

[CaracasEC]

@ Penniper the discussion is about two pumps in parallel operation. Pump A driver is steam turbine while Pump B & C are motor driven. Normal operation will be Pump A + Pump B while Pump C will be stand-by. See attached p& id.@1gibson&artisi, hydraulic performance, performance curve, datasheet for the pumps have already reviewed and found no deviations including pump performance curve in parallel operation.@ JJPellin, as per attached p&id, the pumps share a common suction and discharge pipes having its own check valve. The pump are API bb2 type, they are not yet on operation as we are just on a precommissioning stage of the project (Ethylene Cracker Plant Project). The fire experience you have on which the root cause is one pump running below mcsf for parallel pump operation will be very important to look into since total flow doesnt always mean that it was Q/2. As the pump were running, erosion on the wear rings and suction filter condition were not the same forcing one pump operating point to the left of the BEP or to the right of the BEP. In addition to the speed monitoring of the steam turbine compare to the motor speed.

 

[ccfowler]

CaracasEC,

It is important to know the duty of the subject pumps. There is nothing unusual about pumps operating in parallel, but the risk and potential for severity of damage varies greatly with the power levels and pumped materials involved. Artisi is absolutely correct in stating that the specific drivers or mix of drivers is involved so long as their operation is controlled suitably for the application.

It is common in steam power plants for boiler feed pumps to operate in parallel. Depending on reliability requirements for the plant, there is likely to be at least one extra pump in addition to the number required for maximum power production. Typically, multiple boiler feed pumps operating in parallel are of identical design, but it is not unusual for some to be driven by electric motors and others to be driven by steam turbines. Commonly, the power level of each boiler feed pump is several thousand horsepower. At these power levels, destructive conditions can develop in seconds, so these pumps are well instrumented to assure safe operation. The number of pumps allows operation of the steam cycle over a wider range while keeping the individual pumps operating within their safe range by operating fewer pumps when the steam generator(s) are operating at reduced power levels.

Where power levels are much lower, operating conditions can be much less critical. For example, several pumps of different designs may be supplying water to a header system where the total flow requirement may vary greatly but the header pressure is kept within a very narrow range. So long as all of the pumps involved can operate properly within this pressure range, they can work very nicely together. Pumps are started and stopped as needed to maintain the needed flow rare. Problems can develop if the header pressure is allowed to get too high or too low. The pumps with the steeper head vs. flow curves will be the least troublesome, but the pumps that have flatter curves can get into trouble much more easily. If the header pressure rises too high, flow can stop in these pumps turning them into mechanical water heaters. If the header pressure drops too low, they can run out on their curve and overload their drivers. For an application such as this, the pumps are likely to have one or two stages. Instrumentation for a system such as this is likely to be minimal. Here too, drivers can be of different types so long as they can develop sufficient power to keep their pumps operating at proper speeds.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[stanier]

Suggest each pump has its own flow control. That way you can balance the system sT can be speed controlled and/or flow control valve and electric motors flow control valve.

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/