Everyone's heard of Cavitation, but what is it? Cavitation occurs when pump intake pressure falls so low that the water is pulled apart by suction to form cavities. This is most likely to happen during low-speed acceleration, when there is little forward motion to force water into the pump intake. It can also happen when the engine and impeller are mismatched.
Cavitation causes two related problems: a drop in pump efficiency, and erosion of impeller and/or pump surfaces. Pump efficiency falls during Cavitation because the volume of cavities formed is subtracted from the normal water flow. This drop in water flow translates to a thrust loss.
Cavitation erosion is caused by the implosive collapse of the cavities against interior pump surfaces. The extremely high pressures generated in cavity collapse soon find any defects in metal. At first, Cavitation damage looks like light sandblasting. As the process advances, the metal assumes a porous, spongy appearance. Continued long enough, Cavitation can destroy metal parts.
As an impeller slices through the water being fed to it by the intake grate, there is a pressure difference across each blade. Pressure on the downstream side is high, caused by the fact that the fast-moving blade is accelerating the water on that side. Pressure on the upstream side is much less, because the only pressure acting on that side comes from three sources:
- The depth of the water at that point, generating about 1/2 a psi per foot of depth.
- Atmospheric pressure pushing down on the water's surface, equal to about 14.7 psi.
- The so-called "dynamic pressure" - that caused by the speed of the boat, causing water to ram into the fast-moving intake grate. Dynamic pressure increases as the square of velocity.
Now imagine that you nail the throttle from a standing start, so item (3) equals zero - no dynamic pressure. You have a powerful engine and it spins the impeller so fast that the water flow can't keep up with the demand of the pump. The water on the low-pressure faces of the impeller is physically pulled apart - it cavitates, forming voids in the water which stream back across the impeller face.
What is in those cavities, and why do they form? The first answer is water vapor. The second is that these cavities form because very little is holding water together in the first place. Familiar solids like wood, steel, and bone are held together by strong electrical bond forces, but water stays together mainly because of gravity. If we apply a low-enough pressure to water, it readily forms cavities filled with water vapor.
Water in lakes, rivers and the ocean also contains quantities of dissolved gases (the existence of fish is good proof of this). The tendency of these gases to form bubbles when pressure is removed from the liquid makes Cavitation occur more easily. This is the same action we see when uncapping a bottle of beer or soda. The previously dissolved carbon dioxide in the liquid is invisible so long as there's enough pressure in the bottle to keep it dissolved. When we pop the cap, the pressure is released and the gas comes out of solution to form bubbles. The same happens in the water flowing into your pump if the intake pressure drops low enough. This bubble formation gets the Cavitation process started, and the formation of water vapor continues it.
Strangely, minor Cavitation can actually improve pump performance slightly. This is because the Cavitation bubble just behind the leading edge of the impeller gives the water flow a smooth curve rather than a sharp edge to flow over. But if the Cavitation becomes more general, whole areas of the blades will be covered with cavitated regions, and the result will be a drop in pump delivery. This is because, for full delivery, water flow has to follow both faces of each impeller blade, entirely filling the spaces between.
As a cavitation bubble streams back from its point of creation, it eventually reaches regions of higher pressure, which cause it to collapse. There is a great deal of interest in precisely what happens during such collapse, because extremely high pressures are produced as the water rushes radically inward from all directions and then stops suddenly when the cavity disappears. If this process occurs in the free stream of water, it makes a noise but is otherwise harmless. Many of the familiar sounds made by home plumbing -squeals and roars - are caused by cavitation; fast-moving water cannot follow the sharp turns in elbows and fittings, and cavitation is the result. This can generate a stream of partly cavitated flow. The sounds you hear result from the collapse of the cavities. When pumps are run under conditions of too-low intake pressure, roaring or screeching sounds can result as cavitation occurs.
It is worth quoting a descriptive passage from an old text, "Fluid Mechanics" (R.C. Binder, Prentice-Hall, 1958):
"As (cutoff - the drop in pump efficiency after peak) is approached, the sound emitted by the pump changes. At first it sounds as if sand were passing through the pump (with clear water entering). Then the sound or noise may change (as the discharge is increased) to give the impression of rocks passing through the pump, or a machine-gun barrage. If the pump is operated for any length of time at these conditions, the impeller may be badly eroded and pitted."
If the cavitated region is in direct contact with an impeller blade or water-tunnel surface, noise isn't the only result. The extremely high extinction pressure, as the cavities collapse, is now exerted against the metal itself, and the results are familiar to all. There is actual erosion of the metal, as a constant succession of high-pressure cavitation events hammers it. Over time, the metal can be deeply penetrated by the process, becoming spongy and eroded, looking for all the world like another familiar disaster - detonation damage on an engine's pistons or cylinder heads. As with detonation, a mild case resembles the effect of sandblasting. Because cavitation is speed-dependent, it will be seen most often near the tips of the impeller vanes.
When you hit the throttle from a standing start, your engine begins to spin the pump. It pulls water out of the intake tunnel, and something has to push more in or the pump is going to cavitate. That something is the pressure of the atmosphere, plus the water pressure at the face of the intake grate. Because the boat is hardly moving yet, there is no dynamic pressure.
One point that's easy to miss is this: There is really no such thing as suction. When fluids move, it is always because there is more pressure on one side than on the other, causing the fluid to move toward the lower-pressure region.
"Wait a minute!" you may object. "What about drinking with a straw - that's suction."
It is not. What is really happening is that by "sucking," you reduce the pressure above the liquid in the straw, and the greater pressure below it pushes it up. It is really atmospheric pressure pushing down on the surface of the drink that pushes it up the straw when you suck. You can check this easily with 30 feet of clear tubing. Put one end into the drink, then climb up a ladder to a point 10 feet above the surface and try to drink. You can do it, but it's hard work. Try 20 feet - maybe hard suckers can still get a drink at this level. Now try 25 feet - nothing. You can't pull the fluid that high. Even if you connect a vacuum pump to the top of the hose, you won't be able to lift the liquid 30 feet. When you start the pump, the fluid rises quickly, but it slows down and stops at about 28 feet or so.
Why is there a limit on how high you can "suck" water? The answer is that this is as far as the pressure of the atmosphere can push it up. In theory, the atmosphere's 14.7 psi ought to be able to lift water 32 feet, but two factors prevent this: One is the presence of dissolved gases in the water, and the other is the evaporation of the water itself to form vapor. (Some may ask how water is lifted from deep wells; this is done by placing the pump at the bottom of the well. It pushes the water up.)
When you open the throttle from zero speed, a contest results between how fast the pump blasts the water out the back and how fast the little pressure at the intake can push more water in. The forces present to push more water into the pump are as already noted above:
- The pressure of the water at the depth of the intake grate.
- Atmospheric pressure (pressing down on the surface of the lake, stream or ocean).
If this much pressure is not enough to push water in as fast as the pump is throwing it out, cavitation will occur somewhere. The usual location is just behind the leading edges of the impeller vanes, on the suction side, but there are other possibilities as well. In water-jet installations on big boats (in sizes up to thousands of horsepower), the intake tunnels are very carefully designed as smooth S-bends, bringing water in from below the hull, turning it slightly upward to raise it to the height of the pump, then turning it horizontal again to enter the impeller squarely. There is no intake grate and there are absolutely no sharp edges anywhere.
Personal watercraft pump intakes would look just like this in a perfect world, but real-world water near the shoreline is full of things such as swimmers, waterlogged sticks and lengths of polypropylene rope, none of which we want to enter the impeller. Another difference is that large jet-drive craft are not usually beached. Although a perfectly smooth, open intake would be best for performance, our world contains liability lawsuits, which are far worse than polypropylene rope. Because intake grates contain sharp direction-changes for the waterflow, they are rich sources of both drag and cavitation. All the little mismatches where the pump meets the hull are likewise prime sources of this. Smooth is the word on the intake side.
I was forced to learn this from the coolant-water pumps on racing motorcycles, which often have a 90-degree change of direction right at the entry to the eye of the pump impeller. I learned by experience that careful smoothing of this pump-entry region would often earn me a five-degree operating temperature reduction. What was happening was that water, flowing at low suction pressure across sharp edges or changes of direction, was cavitating and partly filling the impeller with emptiness instead of solid waterflow. Smoothing the entry resulted in less cavitation and more coolant flow. The same situation exists on a much larger scale in the intake arrangements of a personal watercraft.
As the boat gets under way, another source of intake pressure develops: the speed with which the water approaches the grate. This dynamic pressure is why cavitation is much less likely at higher boat speeds.
Because the grate does not squarely face the water, it has to be equipped with one or more vanes whose purpose it is to scoop water up into the intake tunnel. These turning vanes are in effect little wings. Each of them has a high-pressure side (the side facing the incoming water) and a low-pressure side (the opposite). And like wings, these vanes can stall. The flow over the high-pressure side has no choice but to follow the vane surface, but the flow trying to follow around the low-pressure side may "feel" so much centrifugal force as it makes the turn that it can no longer remain attached to the surface. If this happens, the grate vane itself generates a sheet of cavitated flow, which can then cavitate the pump itself.
Impeller and pump specialists point out that leading edges of things like the anti-swirl vanes behind the impeller are rounded for a good reason. The angle at which the water approaches them changes with impeller and boat speed, and round edges are easier for the waterflow to curve around. Knife-edging seems like it ought to be an improvement, but when the water approaches such an edge at an angle, the flow will cavitate behind it on the low-pressure side, reducing efficiency.
Because there can be such low pressure on the pump's intake side, any leakage here can induce cavitation and thrust loss. This is why racers seal the hull-to-pump gaps with silicone, and smooth away all roughness.
Cavitation is an ever-present possibility in pumps and marine propulsion systems. Knowing something about it offers some defense against the problems it can cause. pwi