Everybody who works with hydraulic systems has heard of the cavitation and its destructive nature, and I have seen many a mechanic look at those pitted surfaces, nod his head and proclaim: "m-m-m, cavitaiton...", all while pointing with the index finger at the ceiling... Clearly, this means that overhaulers do tend to associate the appearance of pitted and grainy wear spots on parts with the word "cavitation", but it doesn't necessarily mean that all of them are familiar with the (extremely interesting, in my opinion) details of this peculiar type of erosion.
Before all, it must be said that cavitation is an extremely complex combination of physical and chemical processes and it involves mechanisms that, despite numerous studies, are still debated and not fully explained up to this day, so this short post is but a brief overview of this phenomenon, "sprinkled" with a couple of examples form the oil-hydraulic industry and my personal moderately biased opinion.
The term "cavitation" (from the Latin word "cavus" which means "hollow") is used to describe the formation of vapor bubbles in fluid when its static pressure rapidly drops below a certain critical value and their consecutive collapse as the pressure is reestablished. In our case the fluid is most of the times some sort of hydraulic oil, and the critical pressure is fairly close to the equilibrium vapor pressure of the oil at the current temperature.
Cavitation process can be divided into three stages: nucleation, bubble growth and bubble collapse.
Nucleation is the very beginning of the liquid to gas phase transition. Most fluids, hydraulic oil included, can withstand pressures lower than the vapor pressure without changing phase, and in order for the vapor bubbles to actively form and grow the so called nucleation sites must be present in the fluid, like, for example, solid and gaseous contaminants, random and turbulence induced inhomogenities, irregularities and sharp edges of the confining surfaces. Lucky for cavitation (if one can say so) hydraulic systems present no shortage of any of the above, however it is still interesting to note that a more contaminated fluid makes better environment for the formation of the bubbles.
As the bubbles undergo the expansion stage, another interesting phenomenon takes place - the reduced pressure creates conditions for the air dissolved in the oil to diffuse into the bubble through the bubble wall, thus making the cavitation bubbles fill with a mixture of oil vapor and air, rather than with pure vapor. Also, when an entrained air bubble serves as a nucleation site, there's already some air inside before the oil vapor starts to "inflate" it, which means that the oil vapor to air ratio in air-bubble nucleated cavities can vary depending on the initial size of the incipient air bubble. Furthermore, to make things even more complicated, these two distinctly different processes, namely evaporation and degasification, take place simultaneously, meaning that you have growing vapor filled bubbles that have small amounts of air coming in though the process of diffusion, and you have the entrained air bubbles forming (on their own nucleation sites) as the dissolved air is drawn from the solution by the pressure drop, and then these bubbles serve as nucleation sites for the vapor bubbles. And, lastly, you have "normal" aeration bubbles, coming in all sizes from the process of direct mixing of oil with air, which, basically, do the same! Like I said before - a remarkably complex process, in which a large number of factors combine to produce the widest assortment of bubbles. In fact there's a well-grounded opinion that this - let's call it "combinatory" - cavity development process should be called "pseudo-cavitation", and be differentiated for the "pure vapor" cavitation, which makes total sense, but due to the obviously non-scientific nature of this article, this complex phenomenon will continue to be referred to as just cavitation.
The most interesting phase of the cavitation process, and the one that's actually responsible for all the damage, is the bubble collapse phase. In a spectacular manner, as the static pressure around our (so far harmless) bubble rises, the bubble deforms into a kidney-like shape, and as it implodes, a tiny droplet - the so-called micro-jet - is generated, piercing the bubble wall at an extremely high velocity, and then finally, as the walls run into each other, an intense shock-wave is released into the surrounding liquid. When this happens on or next to a solid surface, it suffers innumerous pulse loads from these microscopic but very intense implosions, and it is this cycling stress that eventually leads to the surface material fatigue and, with enough time, fragmentation, giving birth to those grainy eroded craters that mechanics get to gaze at in their workshops later on. The erosion rate will depend on the material strength and the geometry of the affected surface.
In our industry, when you say "the static pressure around the bubble rises", the "rises" part can refer to anything from tank pressure to full system pressure, depending on the place of the hydraulic circuit where the cavitation takes place. Consequently, at low levels of re-established static pressure, the implosion of the bubbles containing relatively large proportion of air inside can be cushioned due to the fact that it takes more time for the air to dissolve back into the fluid than to come out of it, however the same bubbles, when collapsed by a higher static pressure (for example when they are carried inside a bore of a piston pump cylinder block from the suction side to the pressure side) can literally ignite the oil-vapor and air mixture due to rapid adiabatic compression, creating a microscopic but very real explosion - the so called micro-dieseling" phenomenon - which, aside from the obvious shock wave damage, can also cause accelerated thermal degradation of the hydraulic oil.
It is very important to note that in hydraulic systems the cluster of cavities (sometimes referred to as the "cavitation cloud") almost always originates in dynamic, often high-velocity flow conditions, with the bubbles collapsing when they are carried by the flow to regions with higher static pressure. This means that cavitation erosion can and in most cases does happen distant from the cavitation origin site.
Let us consider the places where cavitation can occur. Since we already know that cavitation is primarily caused by a pressure drop, the first logical place coming to mind is hydraulic pump inlet. Truly, pressures low enough to cause cavitation can easily develop when suction conditions are compromised. If this is the case, the erosion is likely to be found inside the pump in the region where the positive displacement elements (like piston bores, or vane and gear chambers) transition from the suction side to the pressure side - i.e. the most probable place for the bubble collapse to happen. Like in this example, showing the typical horse-shoe wear pattern on the open loop pump valve plate, located between the kidney ports in the inlet-to-outlet transit area. When cavitation is prolonged and severe, the erosion spots can be found inside the positive displacement chambers and on the confining walls as well. In this cavitation case, for example, deep cavitation craters developed inside the bores of a piston pump barel.
As you can imagine, a number of circumstances can compromise pump inlet conditions, the most obvious being:
and also, unlikely, but still possible and therefore worth mentioning:
Another factor, that can lead to hydraulic pump cavitation damage, but which is often overlooked, is the reaction time of certain open loop variable displacement pumps, typically when applied in closed center load sensing systems with frequent cycling between stand-by and full flow operation. Some pump models are capable of changing the displacement from zero to maximum so quickly, that the stationary column of oil in the suction line can't keep up with the pump's suddenly increased suction rate, thus creating momentary cavitation enabling conditions even in cases where the suction line is correctly sized. In these cases, if the system requirements allow it, the problem can be solved or mitigated by installing orifices in servo-piston lines thus effectively increasing the pump's reaction time.
The next place for the cavtiation to develop would be the actuators, typically in systems where over-running load conditions can create suction and, consecutively, generate low pressure in one of the lines. The most common example would be hydraulic motors and rotary actuators, although cavitation can happen in cylinders as well. Correctly designed systems address this problem through application of some sort of over-center control and anti-cavitation arrangement, typically in the form of anti-cavitation check valves connected to a slightly pressurized (e.g. return) line to make sure the "starving" part of the actuator gets adequately filled with oil. But sometimes these valves fail, get misadjusted, or the load conditions are too fast for the anti-cavitation system to keep up. In these cases, if cavitation damage is discovered in the actuator, the anti-cavitation and load control measures must be double-checked. Here's another "good example of a bad example" - a gerotor hydraulic motor, severely damaged by cavitation. In this particular case the fault lay in bad design - the load inertia wasn't taken into consideration and no anti-cavitation system was used, resulting, as you can see, in severe erosion damage to the rotor.
Another cavitation-generating place to consider would be any place in a hydraulic system where high pressure oil needs to be throttled, producing narrow "oil-in-oil" jet. High velocity flow through restricted passages, like orifices, pressure control poppets, or spool metering notches, can easily create low pressure regions where cavitation can develop, and due to the fact that the originating cavitation cloud is also fast moving, the implosion area can cause erosion literally somewhere "around the corner". While hydraulic pump and actuator cavitation can be addressed by a number of solutions that effectively normalize the static pressure or even "supercharge" it in the problematic area, it is much harder and often impossible to do so in "throttling" cavitation cases, although it also must be said that more often than not the damage is likely to be negligible, since "normal" systems are built to avoid continuous high pressure throttling scenarios. But, as is always the case with hydraulics, it depends... And if unacceptable cavitation damage is indeed discovered downstream of a restriction point, the solution usually must be tailor-made for that particular case. Off the top of my head I can name a couple of possible fixes for such situations, for instance: changing the geometry of the problematic area so that the collapse phase happens farther from the confining walls, or maybe replacing a single restriction point with two connected in series.
Have a look at this very interesting "oil-in-oil jet" cavitation example. In this pump model the high speed stream originating from a decompression orifice of the valve plate is directed inside the pump case at an angle towards the end plate, having resulted in a relatively harmless, but nevertheless spectacular erosion spot. Here's another example - an eroded overcenter valve spool. The cavitation damage is hardly noticeable at first glance, but it is actually extremely severe, with the erosion cavity of almost 4mm deep (!). If you zoom into the picture you can see that cavitation ate out a big chunk of material inside one of the spool notches.
For some reason there's a strongly established opinion that whenever cavitation occurs you should hear a loud rattling noise comparable to something like pump "running with marbles inside". Throwing in my two cents, I can tell you that indeed, in case of a severe pump cavitation, yes, you will hear this distinctive loud noise, and if nothing is done immediately the pump will fail catastrophically, in fact in many of such cases when you arrive to diagnose the problem it's "already too late"... However, I've also seen many components with signs of cavitation damage, that came from hydraulic systems that didn't produce any humanly noticeable abnormal noise. And these systems, by the way, also appeared to have normal service history with acceptable component life spans.
Which brings me now to the revelation of my point of view on all this cavitation thing. Simply put - cavitation is extremely complex and generally bad, but - some degree of cavitation will happen in many, if not all, hydraulic systems, and we just have to learn how to live with it. Even the fact that you discovered signs of cavitation erosion during the last scheduled overhaul doesn't necessarily mean that you have to run and "save the patria" by taking all possible and impossible measures to eliminate it no matter what. I believe that you should use common sense and consider costs when evaluating the presence of this type of damage, and only consider taking measures if the normally expected life span of a component was significantly shortened. Which is why it is also important to make sure that you have the correct information about how many operational hours led to the damage.
Obviously, if you hear loud and clear that a pump is cavitating, or when you find a motor partially destroyed by erosion after a short operation period, immediate investigation and corrective actions are necessary. But if you design your hydraulic system right and run it clean (the word "clean" here should be regarded as a collective term meaning everything that's good - i.e. contaminant-free and correctly chosen oil, right temperature and operating conditions, adequate preventive maintenance in place - you know - clean), chances are that you will never have to worry about cavitation, and even if you eventually encounter minor cavitation erosion spots, most likely it will happen during a scheduled overhaul and to a part that was going to be replaced anyway, thus not having impacted the system's expected life.
I used to say - "cavitation must be avoided at all costs", now I say - common sense and numbers, gentlemen, it's all about common sense and numbers...
P.S. A short Add-On to this article from 12 December, 2017.