If it sounds like cavitation, feels like cavitation and is believed to be cavitation, it may or may not be cavitation. The "cavitation" noise we often hear from problematic pumps results from bubbles passing through the pump. But what kind of bubbles? Knowing the differences between cavitation and air entrainment is the first step to preventing equipment damage.
The more classic cause of these noises is reduced suction pressure that falls below the necessary level to suppress transformation of fluid from the liquid phase to vapor phase. Initially, just a few bubbles form because of reduced pressure. This first stage is called incipient cavitation. It can be detected with special instrumentation, but it is not audible to a human ear.
If pressure is reduced further, more bubbles form and grow into larger bubbles, which continue toward the impeller blades. The blades begin to impart energy to the fluid mixture and bubbles, in full accordance with the pressure-volume relationship, causing the bubbles to transform back to liquid. Such transformation is rapid, and the bubbles disappear quickly in an implosion-like manner. While it affects only a small volume of the fluid stream, the inrush of the surrounding liquid into a disappearing vapor bubble is a violent, high-energy, high-pressure event. Some of the bubbles collapse near the walls of the impeller (slightly past the inlets where pressure starts to increase and is already above the vapor pressure), causing the metal to chip away and form cavities. The frequency of the collapse is random but generally is around 15,000 to 20,000 hertz (Hz). The energy is released onto the internal parts, causing pulsations and rotor vibrations. These, in turn, damage the seals and bearings.
To assess a pump's propensity for cavitation, users can compare net positive suction head available (NPSHA) to net positive suction head required (NPSHR). This comparison is not the focus of this article but has been covered extensively in literature. Cavitating pumps are also demonstrated during Pump School sessions, where attendees can observe bubble formation, development and collapse, which cause a drop in flow, noise and other side effects
Air entrainment, on the other hand, may or may not be related to suction pressure level. In cases where there is a relationship to suction pressure, the dissolved air comes out of the solution when suction pressure drops. In accordance with the laws of thermodynamics, this dissolved air grows rapidly until it becomes a large gaseous block—"air locking" the pump inlet. A pump is not a good compressor, and it will stop pumping. As a result, discharge pressure drops and flow stops.
For example, a crack in a pipe, given pressure below atmosphere, allows air to be sucked into the pipe. This air will approach and air lock the impeller inlet. If the volume of air is small, the frothy mixture of liquid and air could continue through the pump, causing noise and vibrations similar to the effects of vapor that results from cavitation. The resultant instabilities and vibrations hammer the internals, causing damage to seals and bearings. In the case of air entrainment, however, the air bubbles do not become liquid, and their energy is substantially less than that of the imploding bubbles of cavitation. For this reason, air entrainment typically does not cause damage to the impellers or casing.
One common cause of air entering a pump's inlet is insufficient liquid level submergence above the suction inlet. Vertical turbine pumps are notorious for these issues. Certain published formulas can help end users to calculate the minimum submergence required to keep air out. Pump size and entry velocity are the main parameters, making a typical minimum required submergence vary from a few feet for small end suction pumps to more than 10 feet for larger vertical pumps.
Suction vortexes are highly unstable and dance around, resulting in radial force. For example, a 10-foot submergence produces roughly 5 pounds per square inch gauge (psig) pressure at the bell entrance. However, because the entire funnel of air vortex is at atmospheric pressure, a local narrow zone (equal to a projected area of the vortex cone) exists and dynamically fluctuates at the bell entrance.
To understand the magnitude of that force, assume a 36-inch bell has a 3-inch vortex core, or roughly 10 percent of the periphery. Also assume that the bell is 2 feet tall, at which point the vortex enters the impeller and is broken up by the blades. The projected area of the vortex core can be approximated as a rectangle that is 3 inches wide and 2 feet tall, or 72 square inches, with vortex core pressure (atmospheric) at that position and 5 psig at the opposite end (surrounding liquid). That equates to 360 pounds of force (5x72=360) that acts on the rotor, assuming some extension of the shaft passes the impeller to the area of the vortex. This force deflects the rotor to take up the clearance in the bushings and impeller tips. As the vortex jumps at a different spot, the force (due to a change in pressure direction across the vortex core) changes direction, and a moment later it changes again. This fluctuation force, typically of low frequency, translates into vibrations and damage associated with it.
The assumed dimensions, as well as the details of the geometry of the rotor, are approximate in this example, and the extent of the penetration of the vortex varies from one design to another. As a result, the calculated radial force is a rough estimation only. It does, however, provide a quantitative description of the instabilities and issues caused by the entrained vortex.
Proper submergence, vortex breakers and sump designs can all improve this phenomenon. The Hydraulic Institute provides formulas, specific recommendations and guidance on the geometry of sumps, required submergence levels and methods to optimize operations. All of these resources can help end users make their pumping systems less sensitive to these issues.