Most experts agree that the majority of centrifugal pump problems occur on the suction side of the pump. Based solely on my experience, I would state the percentage is at least 80 percent, and in the case of self-priming pumps I am sure the percentage is higher.
I started a list of self priming pump field issues many years ago and noted that almost all of them break down to 10 fundamental areas.
1. Even a self priming pump has to be primed initially.
Even a self-priming pump has to be primed prior to the first operation. No matter the manufacturer, there is a priming chamber (integral or external) or some portion of the volute that will require filling prior to startup. Please read the manual and/or contact the manufacturer for details. There are other methods to prime a pump, which include ancillary pumps, vacuum, vacuum ejectors and/or eductors. This article only addresses liquid self-priming centrifugal pumps.
Sometimes the pump will require manual re-priming after the initial prime. There can be several reasons for re-priming, one of the most common is evaporation of the fluid, and other reasons include leakage, pump movement and other maintenance related matters.
2. The required lift is too high.
At sea level in a perfect world, you can theoretically lift 65-degree water 34 feet with a self-primer. I normally caution users to limit their suction lift to a maximum of 25 feet due to factors such as fluid temperature (think vapor pressure), specific gravity, friction, system leakage, pump inefficiencies and elevation above sea level.
3. The pump is too far from the liquid source.
Place the pump as close as possible to the suction source. Usually 25 to 30 feet is the maximum recommended distance. Prudent system design dictates that the suction pipe length be held to a minimum to promote long pump life. Every section of suction piping equates to a volume of air that must be removed when the pump starts. Best practices say to reduce priming time to a minimum.
Some system designers will add foot valves to mitigate the prime time and strainers to preclude the introduction of solids into the pump. A foot valve is in essence a check valve placed at the beginning (bottom) of the suction line. My experience is that foot valves add undesired friction and will leak or fail closed (or partially closed) at some point. I typically do not recommend foot valves for use on commercial and industrial self-primer applications. For similar reasons I do not recommend suction strainers. If the pump cannot handle solids and a strainer is utilized, monitor the differential pressure across the strainer. Most industrial self-priming pumps are of robust design and can handle passing solids, but check with the manufacturer. Note: A few applications may perform better with a foot valve.
4. There is a leak in the suction line.
I frequently need to point out to end users that the suction line on a self-primer pump in operation is at less than atmospheric pressure and so there will not be a leak of the liquid out of the suction line. There can, however, be a leak of air into the line. It is possible to have a suction line at 20 inches of Hg (vacuum) when the pump is operating. As a tip for field problem solving, I frequently use plastic wrap around the flanges or suspected areas to test for ingress leaks.
Simply as a general guideline, if your pump takes more than four minutes to prime than you should shut the pump down and look for and correct the cause of the problem.
5. There is no air vent.
The air in the suction side of the system being displaced by the liquid has to have somewhere to go, otherwise the pump will air bind. Centrifugal pumps are not compressors. Water is approximately 840 times denser than air. As an example if a pump was rated at a discharge pressure of 210 psig pumping water, the pump could theoretically compress air to approximately one quarter of a pound (0.25 psig) (210 psig divided by 840 is equal to 0.25). If the pump discharge valve and/or the discharge check valve are shut, the generated pressure of 0.25 psig will not be able to overcome the valves.
Within the confines of the article I will simply state that the air must be vented to an area of lower pressure for the pump to properly prime. There are many acceptable methods to accomplish the process, please contact your pump manufacturer or the author.
6. Consider the pipe size and pump geometry.
Most experienced pump users know that as a general rule you should always design the suction line to be one size larger than the pump suction. Self–priming pumps are an exception, and the suction piping should be the same size as the pump suction. The infraction of the rules is encouraged because of the added air volume that bigger suction lines require. More air means more priming time.
The added friction loss from using the same size pipe is just another reason to eliminate the foot valve and suction strainer mentioned earlier.
The suction pipe should rise continuously to the pump and not higher. In the field, I frequently see suction pipes with high points before the pump suction usually due to obstructions. These high points become a place for the air and other non-condensable gases to collect and will bind the pump suction line. Never install piping that is smaller than the pump suction in any pump.
7. Watch the submergence and NPSHA.
I covered net positive suction head available (NPSHA) in last month's article. I strongly recommend calculating the NPSHA for self-primers, as it is a great method to identify potential problem areas. For example, if the fluid is 160 degrees F, the vapor pressure of the fluid alone will likely preclude you from this application. For example, water at 160 F has a vapor pressure that equates to a negative 11 feet.
The sump you are drawing from will likely have operating levels that are constantly changing. At some value of minimum submergence it will be possible for the system to create a vortex and air bind the pump. I covered submergence in the last article, but simply defined, it is the minimum distance from the top of the fluid to the center of the suction line that will prevent a vortex from initiation. Even if you do not completely air bind, the pump performance can be affected.
8. Avoid freeze damage.
This problem occurs more often in areas that have infrequent freezing weather, but can happen anywhere the temperature will drop below freezing for an hour or more. The fluid in the priming chamber of the pump, usually water, will solidify if the ambient temperature drops below freezing for a sufficient period of time. When water freezes it expands and the casing will crack. The casing will require replacement at a high cost. Either drain the fluid out of the pump or supply a heat source when the ambient temperature is predicted to be below freezing.
9. Avoid reverse rotation.
Unlike an ANSI pump, the impeller will stay in place on most self-primers for a period of time (unless it is an ANSI self-primer. Eventually the impeller may come loose and damage the pump. The backward-running impeller generally will create about 50 percent of the rated flow and, depending on the impeller specific speed (NS), will generate about 50 percent of the rated head. Reduced efficiency of the wrong rotation will likely prevent it from priming or operating correctly but in the simplest of suction lift cases.
10. Flex pipe normally has a reduced internal diameter (ID).
Non-collapsible flex piping is commonly used on portable units. Normally the ID of flex pipe and adaptors are smaller than standard pipe. Think of the dimensions as tubing rather than pipe. Determination of the pipe friction for the NPSHa calculations will be incorrect if the reduced ID is neglected.
Avoid older or incorrectly applied flex pipes because sometimes the pipe ID liner will collapse under suction (vacuum) and block the line.
Conclusion
The pump performance must be de-rated for higher elevation changes (less absolute pressure less NPSHa). If the pump is engine driven in lieu of an electric motor, the resulting intermittent torque introduces limitations to the shaft design capabilities.