Waterjetting 14a - an introduction to cavitation

Many chemical plants have heat exchangers built into the circuits through which their various chemical fluids flow. These are long tubes, within a surrounding shell, so that fluid at one temperature flows through the tubes within the unit (the tube side fluid), while fluid at another temperature flows in the shell around the tubes (the shell side fluid), passing heat between the two. It is an effective and economic way, for example, of taking the excess heat remaining in a fluid after processing and pre-heating the feed stock to that process before it reaches the reactor.

Figure 1. Illustration of a heat exchanger (Shell and Tube )

It is a similar construction to the tubes that carry water through a ship’s boiler, generating steam to drive the turbines of the vessel. The tubes and shell will, over time, build up deposits, since the fall in temperature will often cause material to precipitate out of the fluid, and cake the walls. With time the tubes clog, and without cleaning, the unit stops working. So how to remove the deposits? For many years workers would take hammers and long rods and drive them down through the tubes to break out the scale. The problem with this approach, apart from being very slow and messy, was that if the rod bent, then it could be driven through the wall of the tube, which meant that the tube could no longer be used and had to be sealed off, at both ends.

As I noted in a post written a year ago, Naval personnel were some of the first to discover that high pressure waterjets could be used to replace the hammer and rod. Not only was the water not capable of breaking through the tube, but it was much more effective in breaking up, and removing the scale. As a consequence jobs that might have taken 150 man-hours and $2,000 (in 1972) took 10 man-hours and cost $700.

The use of high-pressure water to clean heat exchangers is now ubiquitous, it is a much faster, and cheaper system, and though operating pressures have risen, the volume flow required to immediately carry all the debris out of the tubes has continued to be a requirement. But not all systems work equally well. Back in those early days the Navy held a competition between systems, each capable of delivering water at 10,000 psi at a flow rate of 20 gpm. When they ran competitive trials against similar tubes, they found a significant difference between the performance of five different companies bidding for the work.

Figure 2. Comparative cleaning performance (as a percentage of area cleaned) by competing waterjet systems on Navy boilers (Tursi, T.P. Jr., & Deleece, R.J. Jr, (1975) Development of Very High Pressure Waterjet for Cleaning Naval Boiler Tubes, Naval Ship Engineering Center, Philadelphia Division, Philadelphia, PA., 1975, pp. 18.)

Now there are quite a number of different nozzle designs that have been used by the different companies, and, more recently, rotating nozzle assemblies have also been used to improve efficiency, but back then nozzles were quite simple. So how could one company do that much better than the others?

I was not there for the demonstration, but in talking to operators in that time frame I learned that some of the more skilled folk (and this is back in the day when most of this was done with hand-held equipment, rather than today’s automated systems) would put their (gloved) hand over the edge of feed line, partially blocking the flow of water out of the tube. This filled the tube with water, and the cleaning jets worked better! Why this is, well let me start to explain the mysteries, dangers and potential benefits from what we call Cavitation, since that is what causes the change.

Cavitation occurs when, through one process or another, water is stretched. Because water isn’t very strong in tension, too much of a pull will cause it to form little bubbles of sensibly vacuum within the flow (cavities), as the water is ripped apart. These cavities in the flow aren’t stable and rapidly collapse as soon as the surrounding fluid reaches any significant pressure. The bubbles themselves are individually tiny, but their destructive power can be quite dramatic, as I will explain in a later post.

It is a concern when dealing with high-pressure water systems, because water has to flow relatively rapidly into the high-pressure cylinders of a pump, as the piston pulls back at the beginning of a cycle. If the inlet valves are too small, or the supply pressure is too low, then the water cannot enter the cylinder fast enough to fill the void left as the piston is drawn back. This causes small cavities to form in the water inside the cylinder. One the piston reverses, and the fluid inside the piston is pressured on the pressure stroke, the cavities collapse.

Figure 3. High-pressure pump schematic

No problem, you might think, since one is merely refilling a hole. But that is not the case. Because the bubble does not close symmetrically as it collapses. Rather, at a certain point it will fold over (the term is “involute”) and a tiny jet, known as a Munroe jet, is formed by the convergence of the collapsing walls.

Figure 4. Schematic showing the development of a micro-jet within the collapsing cavity.

These bubbles collapse very rapidly, and are normally very small, but, under very rigorous conditions Al Ellis at USCD was able to capture photographs of bubble collapse showing stages within the above sequence:

Figure 5. Photographs of bubble collapse (Al Ellis UCSD)

The acceleration effect from adjacent wall convergence and collapse has been known and used for years in the development of shaped charges for demolition.

Figure 6. Schematic showing the development of a cutting jet (Munroe jet)

Dr. Baird at MS&T has taken photographs showing the development of this jet from a linear shaped charge, showing that it is not quite as smooth a cutting blade at that scale as it is in the much smaller collapse of the cavitation bubble. The explosive is initiated from the rhs and as the initiation moves along the charge the copper walls of the inner liner collapse together to form the jet. The colors on the liner were used to help monitor the relative displacement of different points on the charge as it reacted.

Figure 7. Formation of a jet during the collapse of a linear shaped charge (Dr. Jason Baird MS&T)

The jet formed in a linear charge is sufficiently powerful that it can, virtually instantaneously, cut through steel bars and armor plate, and is thus useful, for example, in bridge demolition. The small jet from a cavitation bubble collapse is much smaller, though as I will show in a later post, is also extremely powerful.

If cavitation is allowed to continue for long within a high-pressure pump system it will destroy the pump. That is why, where there is concern that it might occur, the high-pressure pump is itself supplied by a lower pressure unit (we typically used one that delivered at 60 psi) that ensures that there is always positive pressure within the pump cylinders. (This was also discussed in an earlier post).

But I will return to explain more about this in the next few posts.

Waterjetting 8a - cleaning with heat

Water is used almost everywhere as a way of cleaning surfaces. Several times a day we typically rub our hands together with water, and usually with some soap, to clean them. Pediatricians and others suggest that children recite a short rhythm, such as a chorus of “Happy Birthday” while doing so to allow the water, soap and mechanical actions to combine and effectively remove dirt. That teaches the child that it takes some 20 seconds for the cleaning action to be effective. The cleaning action is not to sterilize germs, viruses and other obnoxious things on the hands. Rather it is to ensure that they and other dirt particles are physically removed, leaving the hands clean. (This is a different action to the chemical washes that are becoming popular.)

This is not an instantaneous process since the soap and water must reach into all the dirt-collecting parts of the hand, hence the need for the nursery rhythm. The same basic sequence occurs in the cleaning action of a high-pressure waterjet on a surface, although the pressure of the spray means that the water can penetrate faster. But it is why, in using a car wash lance in cleaning a car, it is smart to spray the body of the car with a detergent first, then allow this to work in creating micelle clusters around the dirt particles, so that the mechanical action of the subsequent jet spray will dislodge and remove them. Merely adding detergent to the cleaning water as it goes through the cleaning lance, and strikes the car surface does not give the chemicals in the water time to act before they are gone. Bear in mind that the jet is moving at several hundred feet per second, and that it hits and rebounds from the surface over a path length of perhaps an inch or two. As a result the residence time of the jet on the surface is measured in fractions of a millisecond. This is not enough time for the chemicals to work. (On the other hand it does help keep the sewers under the car wash cleaner than might be otherwise expected.)

With an increase in jet pressure, the speed of the mechanical removal of dirt and other particles from a surface can be fast and effective. The ability of the jet to penetrate into and flush out surface cracks, and joints, means that it becomes a good tool for removing debris from the joints in concrete decks, and, at a little higher pressure, it can also be used to remove deteriorated concrete from surfaces. But I am going to leave that topic until next week.

The other “treatment” that we use when we wash our hands is to heat the water. When used with soap it helps to remove the surface oils on the skin that act as a host to bacteria. Heat is becoming a less common tool than it used to be in high-pressure jet cleaning. At one time steam cleaning, which was followed by hot pressure-washing, had a larger sector of the market. It is a bit more difficult to work with (the handles of the gun get hot, and the operator needs more protection) but for some work it is still the more effective way to go.

Steam, however, loses both heat and mechanical energy very quickly after it leaves the nozzle. It will, for example, lose some 30% of its temperature within a foot of the nozzle. Hot sprays of water can thus be more effective, but when cleaning grease and oils a lower temperature spray will merely move the globs of grease around the surface. Heating the water to around 185 degrees Fahrenheit, or 85 degrees C, will stop that happening and works much more effectively in getting the surface clean.


Figure 1. The effect of water temperature on cleaning different surfaces (A, B and C) of different types of dirt.

But, as with many tools, heated water needs to be applied with a little bit of background knowledge. I mentioned that just pointing a large jet of water at, for the sake of discussion, a boulder covered with an oil spill would, at lower water temperatures, just move the oil around the surface. At higher temperatures the oil would break into smaller fragments that are removed from the surface, but they need to be captured, otherwise the treatment is just spreading the problem over a larger area. This is why it becomes more effective to use smaller, higher pressure systems that have lower contained jet energy, and which can be used with a vacuum collection system to pick up the displaced water, oil and debris.


Figure 2. Using hot, pressurized water streams in cleaning up after the Exxon Valdez oil spill (NOAA )

With the streams used in the picture shown in Figure 2, the energy in the jet will move the oil, but without containment it was being washed down to the water, where it was collected using booms. This is not particularly effective, since in the process the jets also washed the silt out of the beach, and drove some of the oil down into the underling beach structure, so that it continued to emerge in later years contributing to an ongoing problem.

What is needed is to provide enough energy to drive the oil away from the surface, and yet not enough to move it great distances or to disrupt the surrounding material. This can be achieved by using a higher-pressure, but lower flow rate jet. Because some of the water will turn to steam as it leaves the nozzle, Short (PhD U Michigan, 1963) showed that the droplet size will fall from 250 microns to 50 microns when the water is heated above 100 degC.

Obviously that also will reduce the distance that the jet is effective, and so a balance needs to be achieved between the heat put into the water, and the size of the orifice(s) if the jets are to remove the contamination, but in such a way that it can be captured. And here again there is a benefit from having a suction tool associated with the cleaning spray. Because of the problems that oil and grease can cause, it will require special care in designing the capture systems downstream. Incidentally it is generally better if the water is heated downstream of the pump, since there are higher risks of cavitation in the inlet ports if the water is too hot.

And sometimes the two can be combined in ingenious ways. For example Bury (2nd BHRA ISJCT, Cambridge, 1974) added a steam shroud around a conventional waterjet at 5,000 psi as a way of cleaning hardened plastic from the insides of a chemical plant pipe.


Figure 3. Wrapping a conventional waterjet in a steam shroud (Bury et al 2nd BHRA ISJCT, Cambridge, 1974)

Without the steam assist the plastic was not removable, even at higher jet pressures, but with the steam to soften the plastic the pipe was successfully cleaned.


Figure 4. High-pressure water fails to remove hardened plastic, (lhs) but with a steam shroud a lower-pressure jet effectively cleans the pipe (rhs). (Bury et al 2nd BHRA ISJCT, Cambridge, 1974).

Waterjetting 1d - Not quite that simple!

When I first began the research on the applications of high-pressure water that was be one of the major parts of my professional life I must confess to a certain naïve innocence in regard to other folk’s work. One assumed that other folk had made similar mistakes to mine, and then corrected them, so that when different systems were compared that the early, obvious, mistakes had not been made.

One of the first times I found that this wasn’t the case was when we were asked to go and demonstrate that high-pressure waterjets could economically cut granite, in quarries located in the heart of the Granite industry, in Elberton, Georgia. We were working with Georgia Institute of Technology (Georgia Tech) at the time and were asked if we could, at very short notice, go down to a couple of quarries and run a demonstration.

Back during my graduate studies I had found that Russian claims were true that said that it was possible, with a 10,000 psi jet pressure to cut through a rock with a compressive strength of 30,000 psi. (I'll tell you how later)

Figure 1. 9-inch thick block of granite drilled through by a 10,000 psi waterjet at Leeds University. It took over 30 minutes. (Summers, D.A., Disintegration of Rock by High Pressure Jets, Ph.D. Thesis, Mining Engineering, University of Leeds, U.K., 1968.)

Knowing this, and having a suitable pump at Rolla, our group ran some tests at the RMERC to get the angles right between the two jets that we were to use, and then, about a week later, we went down to Elberton and set up a system in the quarry.

Figure 2. Starting to cut a 1-inch wide slot in granite, pressure 14,000 psi, 90 rpm, linear cutting speed around 9 ft/min, areal cutting rate around 20 sq. ft./hour.( Raether, R.J., Robison, R.G., Summers, D.A., "Use of High Pressure Water Jets for Cutting Granite," 2nd US Water Jet Conference, Rolla, MO., April, 1983, pp. 203 - 209.)

The trials demonstrated that high-pressure water could cut granite at commercial rates, we cut a slot some 11 ft long and about 2-ft deep, and, after a couple of days of work, we went home. Georgia Tech then went to one of our competitors who set up to run a similar test. We had been done in 2 days, it took them two weeks to cut a slot about 2 ft long and 6-ft deep. They were running a jet system at 45,000 psi, roughly 3 times the pressure of our system. Why did they do so badly?

Well it turned out that they connected their pumps to the nozzle through a very narrow length of high-pressure tubing, and we calculated (as later did they) that of the 45,000 psi being supplied at the pump, some 35,000 psi had been lost in overcoming friction between the pump and the nozzle, As a result they were trying to cut the granite with jets at a pressure of 10,000 psi effective pressure, and it was much slower than our system which retained most of the 14,000 psi from the pump to the nozzle. (Hilaris, J.A., Bortz, S.A., "Quarrying Granite and Marble using High Pressure Water Jet," paper D3, 5th International Symposium on Jet Cutting Technology, Hanover, FRG, June, 1980, pp. 229 - 236.)

Now you may note that I said something about mistakes – it turns out that we had made an identical mistake a few years earlier and had added a second 10-ft length of narrow diameter tubing to the nozzle, and suddenly a system that had cut adequately with 10-ft of tubing did not work with 20-ft. The reason was the pressure loss in the tubing was too great at the longer length, and the pressure fell below that required to cut into the rock. (But at the shorter length we were drilling the hard sandstone at 12-ft/minute).

It is a very simple mistake, and many folk have made it over the years. The system has to be designed from one end to the other to ensure that all the parts are properly sized for the systems that are to be used. (And I will refer to other cases such as that above as we go through this series.)

It is not just the diameter of the feed lines that is important. In 1972 it took, on average, 150 man-hours and about $2,000 for the U.S. Navy to clean a single ship boiler using chemicals and mechanical scrubbing and cleaning. An enterprising company showed the Navy that it was possible to use waterjet lances to clean the tubes. In the demonstration they cleaned a boiler in 10 hours, and it cost around $700. This being Government work, the Navy then arranged a competition to find the most effective contractor. Based on the performance of the system that had been used in the first demonstration they asked 5 companies to compete in cleaning boilers. The operating equipment was designated as having to operate at 20 gpm, at a pressure of 10,000 psi. The results were not even close, even with systems nominally the same.

Figure 3. Relative cleaning efficiency in areal percentage cleaned, of five competing systems in cleaning heat exchanger tubes in Navy boilers. (Tursi, T.P. Jr., & Deleece, R.J. Jr, (1975) Development of Very High Pressure Waterjet for Cleaning Naval Boiler Tubes, Naval Ship Engineering Center, Philadelphia Division, Philadelphia, PA., 1975, pp. 18.)

One of the differences between the competing systems, you won’t be surprised to hear, was that some had smaller feed hoses than others.

There are many different reasons that the various systems performed as they did. One of the aims of this series is to ensure that, should you be asked to engage in such a competition, you will know enough to follow the path of company A, rather than company E.

As systems have become more sophisticated the different factors that control the performance of the jets have increased in number. As a simple example, when abrasive particles are mixed with high-pressure water in streams of abrasive-laden waterjets at pressures that can run up to 90,000 psi in pressure, for high precision cutting of material, the factors controlling performance now include not only the delivery system for the water, but also that for the abrasive, the type of abrasive and the configuration of the nozzle through which that final cutting jet is created.

Again, when we were asked to compare the performance of these different systems we set up nominally identical test conditions under which to determine which nozzle system would perform better. If I were honest I would tell you that before the tests began I expected that the variation in performance of the systems would vary by perhaps 10% between the best and the worst. We were quite surprised by the result.

Figure 4. Comparative performance between 12 nominally similar abrasive waterjet cutting nozzles in cutting through steel at a standard speed, pump pressure, and abrasive concentration.

I use these last two figures to show that all the details of a high-pressure waterjet system are important, when it comes to optimizing performance. One of the reasons to write this series is to ensure that folk that use these systems in the future do not make the mistakes that we made, as we learned how to tune the systems from getting poor performance to the commercially viable rates that are achieved today.

Unfortunately much of the early research and tests that are the basis for this knowledge were performed before the Internet existed. As a result I will have to use references to books and papers (as above) rather than using the electronic references that are the more common habit now.

This concludes the basic introduction to the series, which will now focus on more specific subjects.