Jim Elsey's tips on addressing cavitation.
by Jim Elsey
Summit Pump, Inc.

My last five articles covered the calculation of net positive suction head available (NPSHa). In the November 2018 issue, I summarized the NPSHa concept and said that firmly grasping the theory and completing the calculation can often be a tricky process. However, with an understanding of the basics and some practice, you can gain confidence and work through most applications. The five examples in the series were selected to cover almost every aspect you will encounter in the real world.

In most real-life NPSH issues, we are not the person that is conducting the initial NPSHa calculation for a system and initially selecting the pump. The more likely scenario is that we are stuck with an existing system problem, and the associated pump is cavitating toward a short and very expensive life ending. The guilty parties are gone or not talking.

Why Cavitation Is a Bad Thing

If there is insufficient NPSHa, the pump will cavitate. Cavitation causes pump damage and a reduction in performance. The pump damage manifests as mechanical seal and bearing damage. In the later stages, it can also destroy an impeller. All damage is expensive.

Most readers know that cavitation (classic) is the formation of vapor bubbles in the liquid. These bubbles form because the pressure on the liquid has dropped below the vapor pressure (NPSH required [NPSHr] exceeds NPSHa). This issue normally occurs near the eye of the impeller since this is the lowest pressure area in the suction system. The bubbles subsequently collapse when they reach an area of higher pressure at about one third to one half the distance along the underside of the impeller vane. The formation of the bubbles does little physical damage. Cavitation will affect the pump hydraulic performance. The collapse of the bubbles potentially creates serious damage to the impeller.

I will have an article explaining how cavitation causes damage in a future issue.

NPSH Margin

To preclude or mitigate cavitation, you must have more NPSHa than the pump requires.

Equation 1
NPSHa ÷ NPSHr = NPSH margin

NPSHr is also equal to NPSH3

How much NPSH margin you need to preclude cavitation varies with each application. The more margin, the better. Guidelines and rules of thumb are as plentiful and reliable as urban myths. I recommend you read American National Standards Institute/Hydraulic Institute (ANSI/HI) specification 9.6.1 to gain a better understanding. The liquid properties and the suction energy level are the differentiating factors.

How to Fix a Cavitating Pump

I am frequently asked this question, and I normally suggest a look at the NPSHa formula and its four components for the solution.

Using each of the four components from the formula, you can map potential solutions to solve the NPSHa problem.

Equation 2
NPSHa = ha – hvpa + hst – hf

ha = the absolute pressure. Absolute pressure as measured in feet of head of the liquid being pumped at the surface of the liquid. This will be barometric pressure if suction is from an open tank, or the absolute pressure existing in a closed tank such as a condenser hotwell or deaerator.
hvpa = the vapor pressure. The head in feet corresponding to the vapor pressure of the liquid at the temperature being pumped.
hst = the static head of the liquid over the pump centerline or impeller eye for a flooded suction in feet (positive value for flooded suction). Not all impeller centerlines correspond to the pump centerline.
hf = the total friction loss in feet of head for the suction side system.

The first factor in the formula is absolute pressure (ha). This factor is always positive. If the suction source is already open to atmosphere, there is little you can do as it is both unlikely and unrealistic to change anything within your control. You cannot change atmospheric pressure or move the pump/system location to a lower elevation in regards to sea level. However, if there is an issue, it will help you understand why the pump is cavitating. If the system is closed and under pressure, there is a possibility you can increase the pressure (consequently the absolute head [ha]) in some manner. My experience with plant owners and operators is that raising the system suction pressure is almost never going to happen due to overriding and/or higher priority constraints.

The second factor in the formula is the vapor pressure (hvpa). The higher the temperature, the higher the vapor pressure and the higher the negative effect. In my experience, I have only witnessed one case where the customer was willing or able to reduce the system temperature, but it is still a question that must be asked. Even a few degrees can have a significant effect.

The third component in the formula is the static head (hst). Sometimes you can convince the system owner to keep the supply tank (flooded situation) or the sump (lift condition) at a higher level. If you are lucky, the few feet the static head is increased can make a big difference. I have been involved in a few cases where the pump was moved to a lower level and in one case, a lower level was created for the pump. These solutions are expensive.

The fourth component in the formula is the friction factor (hf). Of all the factors in the formula, I have had more “luck” convincing the system owner to replace or modify the suction piping in an effort to reduce the friction component. You can increase the pipe size and possibly reduce the number of elbows, tees and other components in the suction system to minimize the friction.

Other Possibilities Outside of the Formula

If you cannot increase the NPSHa, perhaps you can reduce the NPSHr.

Look for different pump or impeller options that require less NPSH. It is not uncommon for a manufacturer to have different impellers for the same pump with different NPSH requirements. Some manufacturers will offer an inducer that works in conjunction with the impeller to reduce the NPSHr. Do not add an inducer without consulting with the manufacturer, because inducers must be matched to the impeller. Sometimes a different pump altogether is required.

Moving to a double suction impeller (two eyes) will have significant effect on the issue since the NPSHr will be reduced by 50 percent.

Reduce the pump speed either by incorporating variable speed or simply using a pump that will complete the service (flow [Q] and head [TH]) at a lower speed. The caveat is that the pump will likely be twice as big (physically) as the initial pump with an associated higher cost.

In many cases, the solution is to add a booster pump on the suction of the initial pump. In power plants and other steam systems, it is not uncommon to have a condensate pump that pumps to a feed booster pump before the liquid gets to the actual feed pump.


Sometimes there is nothing you can do to prevent the pump from cavitating, so your option is to treat the symptom in lieu of the problem. Different materials offer varying ranges of resistance to cavitation damage. Additionally, some materials offer better protection than others during the course of a phenomenon referred to as cavitation induced erosion-corrosion.

Cavitation damage resistance is defined as the reciprocal of the rate of volume loss for a given metal. The material’s mechanical properties that are part of this equation are ultimate tensile strength, yield strength, ultimate elongation, Brinell hardness, modulus of elasticity and strain-energy.

The most important property from this list is the fracture strain energy of the metals. It is for this reason that variations of aluminum bronze and duplex stainless steels offer better resistance than other materials such as regular carbon steel and iron. Note, as a post-original equipment manufacturer (OEM) fix, there are also several coatings that can be applied. When using coatings, I recommend the decisive phrase and advice for the day be “caveat emptor,” from the Latin for “buyer beware.”

With coatings, there are good ones and bad ones and good ones applied poorly.

Proximity to Best Efficiency Point (BEP)

Look at where you are operating on the pump curve (head and flow). If too far to the right, there is a mismatch with the system and the pump. The NPSHr increases exponentially as you move right. Operating too far to the left on the curve can have similar issues. NPSHr actually increases as you approach areas of low and minimum flow rates. This is not published on most pump curves.

Suction Specific Speed (NSS)

Back in the 1970s, new plants or systems were designed with an ever-increasing strict mandate to save money (sometimes over reliability), especially on the initial construction and material costs. As a cost-cutting measure, the NPSHa of systems was reduced (think smaller and lower tanks and pumps at higher levels). The system owners/buyers subsequently placed increasing pressure on the pump manufacturers to design pumps with lower NPSH requirements. The simplest and quickest solution for the pump manufacturers was to increase the size of the impeller eye. The good news was that the NPSHr was reduced, but the bad news was the hydraulic stability of the pump was also markedly reduced if and as the operating point departed from BEP. I’ll have more on this in a later article.

Note: Also not discussed is the “Hydrocarbon Correction Factor,” the subject of a future article.


No matter what, you will be involved in pump applications whether new or existing from some aspect where NPSH will be a factor. At least now you will know why impellers have big eyes, tanks have long legs and pumps hang out in low places.

To read other articles in the 'Common Pumping Mistakes' column, go here.

Tips for Calculating NPSHa

  1. Always calculate the NPSHa when choosing, applying or troubleshooting a pump.
  2. Always work in absolute values.
  3. Keep the units consistent. I recommend working in feet of head if you are working in U.S. customary (USC) units or meters of head if using metric SI units.
  4. Use the NPSHa formula. It is your friend.
  5. Always calculate for the worst condition (most restrictive) in the system.
  6. Suction pressure is not NPSHa.
  7. Do not confuse submergence with NPSHa. You need to calculate for both.
  8. Almost every pump problem is on the suction side.
  9. Vapor pressure is not your friend. Always know the liquid properties.
  10. In a vacuum, there is still some pressure. It is just at a level below atmospheric pressure.
  11. For a given pump, the same flow rate (Q) using a smaller impeller will require more NPSH. Look at using a larger impeller if feasible. Note the total dynamic head (TDH or TH) will be different.
  12. When in doubt, revert back to this series of articles or call your “pump phone a friend.”