On TV crime shows, the lead detective will frequently make the following statement to the other associates: “You’re not asking the right questions.” I try to embrace this same kung fu/quanfa mindset when I am solving pump problems for other people.
It is my opinion that the majority of pump problems occur on the suction side of the system. Your experience may be different. I think that once the liquid is properly presented to the impeller eye, then from that point of suction side satisfaction, the pump will perform its proper duty.
Mindset
Think of the system as three separate systems. (1) The suction system must deliver the liquid to the pump, whether through gravity, atmospheric pressure or externally applied forces. The pump can’t reach out and pull the liquid into itself because liquids have no tensile strength. (2) Once the liquid is properly delivered to the pump, it will do the simple job of imparting velocity to the liquid via the impeller, and then the casing will convert that velocity to head (simplification). If the pump is sized correctly for the system curve and the liquid properties, it will then deliver the proper head and flow to (3) the discharge system.
Valve Lineup
Because I spent decades troubleshooting problems in power plants and refineries, I suggest an action step that is mandatory in those facilities, which is to perform a formal valve lineup. Are the valves in their proper position? Nothing is assumed and every valve position is physically checked and independently rechecked. How embarrassing to discover a closed suction valve after a failed startup.
Which Chapter of the Physics Book Are We In?
Is the pump system designed as a “flooded” scheme, or is it in a “lift” condition? In a lift situation, the level of the liquid source is below the pump suction (impeller centerline). If the pump is in a lift arrangement, there are additional concerns when compared to a flooded suction scenario.
Pumps in a lift arrangement (usually self-priming) may present unique problems.
- There are suction leaks (look for any air leaks in, not liquid leaking out).
- The required lift is not physically possible due to the high vapor pressure (as liquid temperatures approach 140 F, proper operation can be compromised).
- Vertical distance (static suction head) is excessive (at sea level with cool water, your best possible lift may be 25 feet).
- Suction lines are too long (presenting both added friction and excess priming time due to the added air volume).
Note: Pumps that are not self-priming can also operate in a lift situation if they are initially primed.
When the liquid level of the suction source is above the center of the pump impeller, the scenario is deemed a flooded suction. This is a simple explanation, so refer to the Hydraulic Institute (HI) standard 14.2.5.2 for a more detailed definition. It is important to understand that even with a flooded suction, the suction pressure can be below atmospheric pressure levels during pump operation.
System Design
Suction piping should be designed in accordance with HI 9.6.6 and 9.8 guidelines. Proper pipe diameter, length and orientation are critical to successful pump system operation.
Sometimes the suction source vessel/tank is too small and/or of a poor design. Improper design may cause high-velocity flow profiles, turbulence and consequential air entrainment. Even as little as 4% air entrainment can measurably handicap pump performance. You can quantify the amount of air entrainment with simple instruments.
The effective volume of the suction side source tank should be at least 4x the pump flow rate to mitigate turbulence and air entrainment. The proper volume affords a higher transient time in the suction vessel to allow for air/gas bubbles to dissipate. If the tank has been designed with the proper weir and baffle geometry (torturous path), the overall size could be reduced. Note my engineering designs and opinions will be conservative in nature.
Gauges
Are gauges/pressure transmitters installed? Are the instruments in the proper range? Are they calibrated? Is data corrected to the pump centerline or common datum point?
NPSHa & System Head Calculations
Verify the net positive suction head available (NPSHa) calculations. Determine the net positive suction head (NPSH) margin by comparing this value to the net positive suction head required (NPSHr). If the pump is operating further out (right) on the curve than expected, the NPSHr will be higher. Often overlooked is the fact that a pump with excessive clearances will require a much higher NPSHr than the published number due to the excessive internal recirculation.
Complete a simple system head curve calculation. If the calculations are different than the sold-for hydraulic conditions, the pump could be incorrectly sized for the application.
Something frequently misunderstood or overlooked is that you can have the proper amount of submergence and still not have sufficient NPSHa. The antithesis is also true—you may have adequate NPSHa and yet have insufficient submergence. You must check for both.
Other Checks
Confirm the liquid properties (specific gravity, viscosity, temperature, pH and vapor pressure). Is the information provided a true reflection of the liquid actually being pumped? Did the liquid properties change? Often a performance problem may be caused by an incorrect assumption related to viscosity or specific gravity.
Pumps operating satisfactorily with viscous liquids in traditional warm climates may exhibit major issues when they encounter those rare but real cold spells.
To echo your 8th grade science teacher, check your units. Just last month I solved a major issue at a plant because the NPSHr was documented in meters while the NPSHa was in feet.
Refer to the pump brake horsepower equation where specific gravity is in the numerator and you can surmise that it directly affects the driver’s required power requirements. With the possible exception of the pump shaft’s power and torque limitations, the pump itself does not care about the specific gravity (assuming the pump is already rated in feet or meters of head and not pressure).
Due to the lower specific gravity of hydrocarbons, you must pay stricter attention to the pump running clearances. Clearances will need to be increased for low-gravity liquids. When the specific gravity approaches 0.70 or below, the lubricity factor is approaching zero—but more importantly, the Lomakin (centering) effect is diminished. As specific gravity decreases, the potential for gasket material fidelity also decreases.
Pumps are not “plug and play” items. There is a mandatory set of tasks that must be performed prior to startup. A common misconception for smaller (less than 200 horsepower) pumps is the belief that the pump was aligned at the factory and so there is no need to realign. The pump was aligned to some degree in the factory, but the moment it was picked up for shipment, the alignment was negated. Any remaining alignment fidelity was destroyed during the pump installation process.
Pumps are customarily shipped without oil in the bearings. They are shipped with the mechanical seal in a sleeping position or not installed at all. The direction of rotation for the motor has to be determined in the field and is best done without the coupling connected. For certain types of pumps, the rotation check must be completed with the pump uncoupled, or damage will occur.
Note: For horizontal split case pumps, it may be possible for the pump shaft to turn in the correct direction, but the impeller is reversed.
Pipe Strain
If the piping is not properly supported, the resulting stress will transmit to the pump as hot bearings and vibration issues. A pump is the most expensive pipe support you can purchase (sarcasm), so don’t use it that way.
Location
Where is the pump? This seems like a silly question, but location can make a difference. What is the elevation above sea level? Altitude can affect NPSHa, vapor pressure, potential lift, gauge readings and driver performance. High ambient temperature, salt air, humid environments and dust can all lead to premature failures.
Forgot to Mention
Wisdom comes from (bad) experiences. Experience is why I ask a lot of questions. Sometimes facts are omitted that make all the difference. I apologize in advance if a few comments denote a curmudgeon tone.
On more than one occasion, the client has failed to mention that the pump was operating on a variable frequency drive. Another instance was at a remote site where the diesel generator was mistakenly set to 50 cycles for a 60-cycle application. More than once it was not mentioned (or was considered unimportant) that there was a significant suction lift and the liquid temperature was approaching 140 F.
Several times I have discovered that the pump in question was operating in parallel with another pump, but this critical tidbit of data was omitted in the original report.
Sometimes the pump has just been overhauled, repaired or remanufactured and is back fresh from the shop. It is important to know if something impeller- or clearance-related was changed during that evolution. Why was the pump in the shop in the first place?
How old is the pump? Acquire information on the duty cycle and determine how long it has been operating. Check the impeller clearance and ring clearance. If the clearances are now 3x larger than new, the pump will require work or adjustments to reestablish them.
Field Kit
Years ago, I traveled worldwide to solve pump problems in person. I carried a field kit that consisted of some basic journeyman tools but also included:
- Gauges of various ranges
- Portable flow meter
- Handheld vibration detector
- Camera
- Stethoscope
- Infrared temperature detector
- Voltmeter
- Amp meter
- Strobe tachometer
- Dial indicators
- Micrometer set
- Flashlight
- No. 2 pencil with eraser
- Cameron Hydraulic Data Book
Safety Note: You must have formal training, protective equipment and permission from the owner to perform voltage and amp readings.
Not all facilities will allow you to turn wrenches, so ask permission first. Many of my millwright brothers and sisters were overjoyed to see an engineer get dirty, even if it was for “instructional purposes.”
Problem Is Not on the Suction Side
What if the issue is not on the suction side? If you believe this to be the case, investigate the following:
- Pump speed
- Impeller diameter
- System curve “as is” data versus the original data
- Valve lineup
- Liquid viscosity
- Driver limitations
Keep These Steps in Mind
The person with the pump problem is already familiar with the situation, but the “pump-phone-a-friend” is hearing about it (not seeing it) for the first time. When you call for advice, keep the following in mind. (Proceed in the order you think best tells the story you are trying to relay.)
- Describe the pump: Share the size, speed, number of stages, sold-for condition of head and flow, casing orientation and shaft orientation.
- Describe the system: Include pipe sizes, static head, tank sizes, geometry and elevations. Is the suction source open to atmosphere? Is it flooded or lift?
- Describe the liquid: Tell them the liquid name, temperature, viscosity, specific gravity, suspended solids, pH and whether it is Newtonian or not.
- Describe the problem and symptoms as best you can in your own words.
- Describe the pump performance relative to the published curve or with real data of differential pressure (head) and flow.
Sometimes the answer is that you have the wrong pump for the application. For example, the pump was not designed to pump a large amount of suspended solids. The materials of construction are wrong for the application. The motor wasn’t sized for the high viscosity. Perhaps it should have been a positive displacement pump. The pump wasn’t sized properly for the system. The pump was asked to cover an impossible range or myriad of conditions.
No pump is one-size-fits-all.
References
- Troubleshooting Centrifugal Pump Field Problems, Robert Perez
- Pump Users Handbook: Life Extension, Heinz Bloch and Allan Budris
- Centrifugal Pump Sourcebook, John W. Dufour and William E. Nelson
- Centrifugal Pump Clinic, Igor J. Karassik
