Waterjetting 31b - short pulse lengths and traverse speed

One of the more surprising things that we learned at the beginning of the research into high pressure waterjet action was how quickly a jet will penetrate to almost full depth of penetration, and how slowly it will cut deeper after that. It is a lesson that often escapes even those who work with the technology today.

A series of tests was carried out in which a jet was exposed for very short periods of time to a fixed block of sandstone. The time that the jet hit the rock varied and the results were used to make the following plot:

Figure 1. Average penetration as a function of exposure time, for a continuous jet impacting a static target. (Polyox is polyethylene oxide) (after Brook and Summers )

The experiment was then repeated using a device that would only allow the jet to hit the rock for much shorter periods of time. When these results were plotted, the following graph was obtained.

Figure 2. The depth of penetration into sandstone as a function of time, for very short intervals. (ibid)

The depths achieved with the longer exposure times were therefore occurring within the first 1/100ths of a second, and the penetration that followed that initial impact time was at a much lower rate.

The reason for this had been suggested by earlier work by Leach and Walker at Sheffield who pointed out that once the jet starts into the hole it has no other path to exit rather than to turn around and come out the way it went in. Since the jet is continuing to flow into the hole, the result is that the pressure in the hole will diminish over time.

Figure 3. The effect of hole depth on the pressure developed at the bottom (after Leach and Walker).

It should be mentioned, however, that Leach and Walker built a special stand to make these measurements and the hole was built out of steel, rather than being eroded by a jet. The reason that this is important is that where the target is weaker then the turbulence generated by the jet:rebound will additionally erode the walls of the hole, particularly at the depth where the jet pressure falls to the threshold pressure of the material. At this point the jet begins to enlarge a cavity at the bottom of the hole. The pressure can then rise in the cavity, as the hole walls are reamed and the pressure bulb can cause spallation of the overlying rock. It is also why one has to be careful in the drilling of holes in glass, since a similar series of steps can also arise with abrasive waterjet cutting, and internal pressures within the drilled hole can cause the glass to fracture.

Rehbinder also built a narrow slot to measure pressure drop with depth of the hole, and showed that the rapid decline in pressure with depth that Leach and Walker found, was related to the relative narrowness of the hole, and that when the holes were wider, relative to the jet, that this decline was not as dramatic.

Figure 4. Changes in hole pressure with depth as a function of hole width. (after Rehbinder)

As I have mentioned in a previous post a logical progression is to then pulse the water so that each slug of water has time to leave the hole before the next one arrives. When this is carried out, in our case by building a small interrupting wheel that spun between the nozzle and the target, the jet will continue to penetrate although at a slower rate than that originally achieved.

Figure 5. The penetration of rock with an interrupted jet. (after Brook and Summers)

There have been several attempts since that time to use pulsed jets as a way of improving breakage, with a lower energy cost. This has centered around some form of water cannon, or similar tool, although the main problem – never really resolved – of maintaining a high firing rate without destroying the seals in the supply lines has led to that approach being shelved.

Other developments first led to a pulsation in the feed line to the nozzle, first described by Gene Nebeker of Scientific Associates at the 3rd ISJCT in Chicago in 1976. Although that work continued for a number of years it was never able to achieve commercial reality at the time. Subsequently Dr. Vijay pioneered the approach that led to the formation of VLN Advanced Technologies Inc. Using an ultrasonic method of pulsation, which produces very short duration pulses at a high rate in the stream, the company has developed a market, particularly in removing coatings from surfaces.

The mechanisms of target failure are different from those achieved with the more conventional, longer pulsed systems, where the length of the individual jet slugs allows more pressurization of cracks within the target. That kinetic energy allows the jets to operate under water, however shorter pulsation lengths (similar in some ways to rain) are attenuated where there is a layer of water on the surface particularly when this is confined, and Brunton and Rochester found that some of the advantages of the technique (including the ability to generate water hammer pressures are diminished when that layer is thicker.

However, if a waterjet penetrates to close to its maximum penetration within a period of around 0.01 seconds, and the jet is cutting a hole that is roughly three times the diameter of the orifice, then it is logical to suggest that after that residence time the nozzle should move further down the sample. If the jet is roughly 0.033 inches in diameter then the nozzle should move roughly 0.1 inches in 0.01 seconds or roughly 10 inches per second, or 50 ft. per minute. Lab studies have shown that shown that speeds in this range are most efficient where plain waterjets are used in cutting. Because abrasive waterjets penetrate material in a different way the best cutting speed for that technology is much slower.

The topic will continue in the next post, since it is often difficult to persuade operators how fast they should be moving tools to get them to be most efficient.

Leach S. J. and Walker G. L. “Some aspects of rock cutting by high speed water jets”. Phil. Trans. R. Soc. 260A, 295-308 (1966).
Nebeker E.B. and Rodriguez S.E. “Percussive water jets for rock cutting,” paper B1, 3rd ISJCT, BHRA, Chicago, May 1976.
Brunton, J.H., Rochester, M.C., "Erosion of Solid Surfaces by the Impact of Liquid Drops," In Erosion-Treatise on Materials Science and Technology, ed Preece, pp. 185 - 248.
Rehbinder, G., "Some Aspects of the Mechanism of Erosion of Rock with a High Speed Water Jet," paper E1, 3rd International Symposium on Jet Cutting Technology, May, 1976, Chicago, IL, pp. E1-1 - E1-20.

Waterjetting 31a - Changing Jet Pressure, Diameter and Exposure

A high-pressure waterjet will penetrate into a material by penetrating into small cracks in the surface and pressurizing those cracks, so that they grow and join together freeing material. This mechanism changes where one moves to add abrasive, but that discussion will come later.

The larger the cracks in the material, then the lower the pressure needed to penetrate into the crack, and to then cause it to grow. Large grained, weakly bonded material, such as for example soil, can, as a result be washed apart by pressures as low as those caused by a heavy rain. As the material becomes more cohesive (think initially of a heavy clay) then the amount of force required to grow the fissures is greater, while the crack lengths are usually smaller. This means that the jet pressure will have to be higher for the same volume of material to be removed.

As one moves from soils to rocks and other materials will increasingly smaller grain size, so the pressure required to cut into the material must be increased. Initially we call the pressure at which the jet starts to dig a hole the initial pressure or threshold pressure of the material.

The way to find out its value is to point the jet at right angles to the jet and begin to raise the jet pressure. When the jet has not enough pressure to penetrate and grow cracks in the target, then it will flow along the surface after impact. However when the jet starts to drill a hole into the target, then the water going into that hole has only one way out – back the way it came, and now the jet comes back along the axis of the jet. (Hitting the operator if the lance is hand-held and this is partly why you need personal protective equipment).

Generally that pressure is not enough to give an economic removal rate and the jet pressure should be raised significantly above the threshold to reach that level. All other things being equal (such as nozzle diameter, standoff distance and traverse speed) then as the jet pressure is raised the depth of the cut will increase in proportion, as will the volume of material removed. This is the case whether the pump providing the water is an intensifier system (usually at higher pressure) or a triplex or similar pump. The main difference in the plot is because of the difference in the diameter of the cutting jets. Berea sandstone is a “standard” rock that has been used in many cutting tests over the decades because of its relatively uniform structure and strength. The uniaxial compressive strength of the sandstone is around 5,000 psi.

Figure 1. The effect of raising jet pressure on the depth of cut achieved in Berea Sandstone with the cuts made at a speed of 12 inches/minute.

This leads into consideration of the second important parameter, that of the flow rate of the jet, which is mainly defined by the diameter of the orifice through which the jet is formed. The flow volume of water is controlled both by the jet pressure (the higher the pressure the faster the water flows out of the nozzle) and by the diameter of the jet. When one is cutting with water alone then it is often better to have higher flow rates at lower pressure rather than the converse. The reason for this is that larger diameter jets hit more flaws on the surface than smaller ones, and the larger the area that is under attack then the greater the likelihood of larger cracks being present and allowing greater volumes of material to be removed. (There are statistical and mathematical justifications for this, but I will forgo going through that math).

When carrying out rough calculations on relative cutting performance over the years we have assumed that the relationship between the depth of cut and the diameter of the orifice is a power relationship with an exponent of 1.5. When comparing the data for Berea sandstone which we obtained as we changed jet diameters we found the following:

Figure 2. The effect of increasing jet diameter on the depth of cut achieved in Berea Sandstone with the cuts made at a speed of 12 inches/minute.

The exponents are not quite at 1.5, but using that value gives a fairly close initial estimate as to the performance that we can achieve.

Part of the problem in seeking a correlation between the jet cutting performance and the nozzle diameter is that the cutting range of the jet changes quite quickly with a change in nozzle diameter. And while we often use a first rough estimate that the jet throw is 125 – 150 diameters in reality the jet performance changes over that range, as the structure of the jet itself changes.

One way of showing this is to show how the cut depth varies when the target surface is at different distances from the nozzle, a value we often call the stand-off distance. In this case the rock is a sandstone, and similar to that used above, but the tests are made with the jet firing at the rock for different lengths of time, rather than traversing over it.

Figure 3. The effect of increasing exposure time and standoff distance on the depth of hole achieved in Sandstone.

Note that there is a relatively rapid drop in cutting performance as the target is moved away from the nozzle, which had a diameter of around 1 mm (0.04 inches). But the plot also shows that the cutting depth drops away very rapidly with time. After half-a-second the jet has cut roughly half an inch deep when the target is half an inch (12.5 mm) from the nozzle, but after doubling the exposure to a second the jet has only increased the depth of cut to 0.6 inches (15 mm) and with the time of exposure increased to five seconds the depth only increases to around 0.7 inches (17.5 mm).

This will be the topic for the next post, where the effect on the speed of cutting is the subject.

Waterjetting 16b - Optimum Abrasive Feed Rate and Depth

The post that I wrote last week was focused on the misperception that you need to add more abrasive to an abrasive waterjet if you wish to cut through thicker material. This is wrong on a number of counts, but most particularly because a good operator will have tuned the nozzle to achieve the best cutting jet, based on pressure and abrasive feed rate (AFR) regardless of target material. What the operator may change is the operating pressure (which would change the optimum AFR) and the traverse speed since these control the depth and quality of the cut that the jet makes.

But, before leaving the topic, I would like to discuss, in a little more detail, the concept of the optimal amount of abrasive that one should use with a given jet, and what happens as that feed rate is changed. As I mentioned last time, because of differences in the shapes of the mixing chambers of the nozzles supplied by different manufacturers, the specific sizes and optimal flow rates will differ from nozzle to nozzle but the overall conclusions remain the same.

Last time I pointed out that the driving waterjet had to break up within the mixing chamber in order to properly mix with the abrasive and to bring this up to a maximum speed before the mix left the focusing tube. Where the driving jet is too large then this breakup is not complete and the mixing is not efficient. As a result the jet that comes out of the end is more diffuse and the abrasive will not have reached the full velocity possible. However, if the incoming waterjet is made smaller for the same AFR and other mixing chamber geometries, then the cutting performance will decline.

Figure 1. Effect of increase in jet pressure when cutting aluminum with an AFR of 1.7 lb/minute (after Hashish, M., "Abrasive Jets," Section 4, in Fluid Jet Technology- Fundamentals and Applications, Waterjet Technology Association, St. Louis, MO, 1991.)

For a similar reason adding a polymer to the jet fluid should only be carried out with some care for the consequences. Long-chain polymers can give a jet increased cohesion and this can, at high enough concentrations, inhibit jet breakup in the mixing chamber thus reducing the effectiveness of mixing in the chamber.

Figure 2. The effect of changing cutting fluid on AWJ performance (after Dr Hashish ibid)

Polyox, (polyethylene oxide) is an extremely effective polymer for increasing jet performance by cohering the jet and reducing the friction losses between the pump and the nozzle. However, as the graph shows, adding it to some abrasive systems will reduce performance since the more coherent jet makes it more difficult for the abrasive to mix and accelerate to full velocity. At lower concentrations the polymer allows the jet to breakup, but keeps the slugs of water together making energy transfer more efficient. Higher velocity abrasive means that less is required to achieve the same cutting performance as Walters and Saunders showed.

Figure 3. Effect of adding polymer in reducing the amount of abrasive required to cut stainless steel (after Walters, C.L., Saunders, D.H., "DIAJET Cutting for Nuclear Decommissioning," Paper J2, 10th International Symposium on Jet Cutting Technology, Amsterdam, Netherlands, October, 1990, pp. 427 - 440.)

At low levels of abrasive feed Dr Hashish has shown that increasing the amount of abrasive in the feed increases cutting performance.

Figure 4. Effect of increase in AFR on depth of cut in mild steel at a feed rate of 6 inches/min (After Dr. Hashish ibid), waterjet diameter 0.01 inches.

However, as the abrasive flow rate continues to increase the cutting performance reaches a plateau and can decline, as Dr. Hashish illustrated. An AFR of 20 gm/sec is equivalent to a feed of 2.6 lb/minute.

Figure 5. The effect of higher AFR on cutting depth at 3 jet pressures on a mild steel target (after Dr. Hashish ibid)

Note that in this case the nozzle geometry was not optimized for operation at the highest jet pressure. More visibly we ran a series of cuts across a granite sample, where the only thing that changed between cuts was that we increased the abrasive feed rate in cuts from the left to the right. It can be seen that beyond a certain AFR the jet starts to cut to a shallower depth.

Figure 6. Successive cuts made into a granite block at increasing AFR from the left to the right.

Interestingly the optimum feed rate doesn’t just depend on the pressure and water flow rate (waterjet orifice size) of the system. Faber and Oweinah have shown that as the feed particle size gets larger, so the optimum AFR reduces.

Figure 7. Optimal Abrasive feed rate as a function of particle size (after Faber, K., Oweinah, H., "Influence of Process Parameters on Blasting Performance with the Abrasive Jet," paper 25, 10th International Symposium on Jet Cutting Technology, Amsterdam, October, 1990, pp. 365 - 384.)

The process of finding an optimal feed rate for a system is thus controlled by the design of the mixing chamber based on the relative position of the abrasive feed tube and the size of the waterjet orifice. This controls how well the abrasive that is fed into the system can mix with the jet and acquire the velocity that it needs for most effective cutting. Then, as the above plot shows, the optimal AFR is also influenced by the size of the particles that are being fed into the system, since as the particles become larger beyond a certain size, so the cutting effectiveness declines.

Part of the reason for this is that, as the AFR increases so there is an increased risk of particle to particle impact breaking the particles down into smaller sizes. (And an earlier post showed that smaller particles cut less effectively – as does figure 7 above). We screened the particles that came from several different designs of AWJ nozzle assemblies capturing them after they left the nozzle but without further impact, so that the size range is indicative of that which a target material would see,

The table is a summary of some of the results and it shows results for a feed that began at 250 microns giving the percentage of the particles that survived at larger than 100 microns.

Figure 8. Percentage of the 250 micron sized feed that survives at above 100 micron for differing jet conditions. (the numbers are averaged from several tests).

It can be seen that when the feed rate rises to 1.5 lb a minute that there is a drop in abrasive size at higher jet pressures, and this is likely to be due to the increased interaction with particles. Since cutting effectiveness is controlled by particle size, count and velocity the only slightly greater amount of particles that survive above 100 microns at 1.5 lb/minute relative to those that survive at 1 lb/minute suggest that spending the money to increase the AFR above the optimal value (in this case around 1 lb/min) is a wasted investment.

It is therefore important to tune the system to ensure that, for each jet pressure and nozzle design that is used, that the AFR has been optimized.

Waterjetting 6d - Plywood and Pork, and jet effectiveness

In the last two posts I have tried to show that there is a benefit to running an occasional calibration test on equipment, to ensure that it is giving the best performance. This does not mean that the nozzle needs to be tested every day, although some of the cheaper pressure washer nozzles, for example, will wear out in less than an hour. An operator will learn, over time, about how long a nozzle will last, and can, after a while, tell when it is starting to lose performance. But in working on a number of different jobs in succession that sense of the performance may be missed, and it can be handy to have a standard target that a jet can be pointed at that it should be able to cut in a known time.

One simple target is plywood, and, to continue the saga of nozzle comparisons through a slightly different approach, Mike Woodward used plywood sheets to compare different nozzles in one of the earliest comparisons of performance. We since duplicated his test equipment and ran tests with a more modern selection of nozzles, but the basic results and conclusions remain the same.

In its simplest form the idea is to build a holding frame that will hold small squares of plywood at fixed distances from the nozzle. In the frame shown below the plywood pieces are set at one-foot distances apart, with the nozzle held at a fixed point at the end of the test frame. Tests showed that it takes around 2,700 psi to cut through the plywood.


Figure 1. A simple frame to hold plywood samples

The initial tests that Dr. Woodward ran were run on nozzles that were run at 10,000 psi with a nominal flow rate of 10 gpm. The nozzles that were used cost in the range from $10.00 to $250 apiece. (And these costs were reported in 1985 at the 3rd American Waterjet Conference). Tests such as this are simple to run. Plywood pieces are set into the frame, the nozzle is placed at the end of the frame, and the jet run for ten seconds. Over that time, the jet will cut through any of the pieces of plywood that it reaches with enough power to cut through, and generally the jet will punch a hole through several pieces.


Figure 2. The different designs of nozzle that Mike Woodward tested in 1985.

The profiles show that there was only one of the common nozzles at the time that fitted smoothly onto the end of the feed pipe. In the other cases there is a small gap between the nozzle piece and the feed tube, so that turbulence would be generated just as water entered the acceleration section of the nozzle.

The hole size in each plate was then measured, and that width plotted as a function of the distance from the nozzle, so that a profile of the jet cutting path could then be drawn.


Figure 3. Profiles cut into the different pieces of wood, showing the cutting power of the different jets, as a function of distance and the actual amount of water flow as measured.

As an additional part of the testing a rough measure was kept of the effective nozzle life Some other performance parameters for the different nozzles can be put into a table.


Figure 4. Performance of the different nozzles.

Clearly just going out and buying the most expensive nozzle on the block is not necessarily the best idea. But it also depends on the use to which the nozzle is going to be applied. There are two different applications, that of cleaning a surface, and that of cutting into it. The broader path achieved by nozzle 1, for example, which also removed the largest volume of wood per horsepower, makes it a good selection for cleaning, and for reaching further from the nozzle, as would be needed if one were cleaning the pipes of a heat exchanger bundle.

On the other hand the more coherent flow through nozzle 2, which gave a narrower cut might be a more effective tool in a cutting operation. In other cleaning operations where the nozzle is being operated very close to the surface, then nozzle 3, which has a wider path, might be a better choice, though that is lost if the target surface is further away. And though there was not a great deal of difference in performance between nozzles 1 and 5, there is a considerable difference in price.

A smaller, lighter nozzle may be a beneficial trade-off if the nozzle body is fitting on the end of a lance that will be operated manually for several hours at a time.

There is an alternate way of using plywood as a target that I have also used in teaching class. The student is using a manually operated high-pressure cleaning gun at 10,000 psi and is to swing the gun horizontally so that the jet cuts into a piece of plywood that is set almost parallel with the jet path, but with the stream hitting the wood from the side initially further from the operator, but as the swing completes the jet cuts up where the nozzle almost touches it and then sweeps on past.

The result is that, over the distance that the jet can cut into the wood, a groove is carved into the wood.


Figure 5. Horizontal cuts into plywood. There were about half-a-dozen students who had swiped the nozzle so that it just cleared the left edge of this 4-ft wide piece of plywood, and you may note that the cuts extend roughly ¾ of the way along the surface.

Once the students had seen this cut, I would ask them how far away they thought, based on that measurement, that the jet would cut into a person. Typically they said about three feet, and then, as a precaution, I suggested they add a foot or so more.

Then I took them over to a metal frame where we had hung a piece of pork. We carefully measured off the “safe” distance from the end of the nozzle to the pork.

“Now assume that is you”, I would say, “swing the jet as fast as you can, so that it barely has time to hit “your arm”, and we’ll just check that distance is correct.”


Figure 6. Piece of pork that has been traversed by a 10,000 psi jet several times, with a typical standoff distance from the nozzle of more than four feet.

Invariably we got the result shown in Figure 6. The jet would cut into the meat to a typical depth of around two inches and groove the underlying bone. It was a salutary way of getting their attention about the safe use of the tool, and I noticed that the staff also got a bit more cautious after we ran this class every year.

Waterjetting 6b - The Triangle comparison test

This post is being written in Missouri, and while the old saying about “I’m from Missouri, you’re going to have to show me,” has a different origin than most folk recognize*, it is a saying that has served well over the years. We did some work once for the Navy, who were concerned that shooting high-pressure waterjets at pieces of explosive might set them off, as we worked to remove the explosive from the casing. We ran tests under a wide range of conditions, and said, in effect, “see it didn’t go off – it’s bound to be safe!” “No,” they replied, “ we need to know what pressure causes it go off at, and then we can calculate the safety factor.” And so we built different devices that fired waterjets at pressure of up to 10 million psi, and at that pressure (and usually a fair bit below it) all the different explosives reacted. And it turned out that one of the pressures that had been tested earlier was not that far below the sensitivity pressure of one of the explosives.

That is, perhaps a little clumsily, a lead in to explain why just getting simple answers, such as “yes I can clean this,” or “yes I can cut that” doesn’t often give the best answer. One can throw a piece of steel, for example, on a cutting table, and cut out a desired shape at a variety of pressures, abrasive feed rates (AFR) and cutting speeds. If the first attempt worked then this might well be the set of cutting conditions that become part of the lore of the shop. After a while it becomes “but we’ve always done it that way,” and the fact that it could be done a lot faster, with a cleaner cut, less abrasive use and at a lower cost is something that rarely gets revisited.

So how does one go about a simple set of tests to find those answers? For many years we worked on cutting steel. Our tests were therefore designed around cutting steel samples, because that gave us the most relevant information, but if your business mainly cuts aluminum, or titanium or some other material then the test design can be modified for that reason.

The test that we use is called a “triangle” test because that is what we use. And because we did a lot of them we bought several strips of 0.25-inch thick, 4-inch wide, ASTM A108 steel so that we would have a consistent target. (Both quarter and three-eighths thick pieces have been used, depending on what was available). The dimensions aren’t that important, though the basic shape that we then cut the strips into has some advantage, as I’ll explain. (It later turned out that we could have used samples only 3-inches wide, but customs die hard, and with higher pressures the original size continues to work).

>br>Figure 1. Basic Triangle Shape

The choice to make the sample 6-inches long is also somewhat arbitrary. We preferred to make a cutting run of about 3 minutes, so that the system was relatively stable, and we had a good distance over which to make measurements, but if you have some scrap pieces that can give several triangular samples of roughly the same shape, then use those.

The sample is then placed in a holder, clamped to a strut in the cutting table, and set so that the 6-inch length is uppermost, and the triangle is pointing downwards.


Figure 2. The holder for the sample triangle.

The nozzle is placed so that it will cut, from the sharp end of the triangle, along the center of the 0.25-inch thickness towards the 4-inch end of the piece. The piece is set with the top of the sample at the level of the water in the cutting table. The piece is then cut – at the pressure, AFR, and at a speed of 1.25 inches per minute, with the cut stopped before it reaches the far end of the piece, though the test should run for at least a minute after the jet has stopped cutting all the way through the sample.

The piece is then removed from the cutting table, and, for a simple comparison the point at which the jet stopped cutting all the way through the triangle is noted.


Figure 3. Showing the point at which the jet stopped cutting through various samples, as a function of the age of the nozzle – all other cutting conditions were the same. (A softer nozzle material was being tested, which is why the lifetime was so short). The view of the samples is from the underside (A in Fig 1.)

An abrasive jet cuts into material in a couple of different ways - the initial smooth section where the primary contact occurs between the jet and the piece, and the rougher lower section where the particles have hit and bounced once on the target, and now widen and roughen the cut. Since some work requires the quality of the first depth, we take the steel samples, and mill one side of the sample, along the lower edge of the cut until the mill reaches the depth of the cut, and then we cut off that flap of material, so that the cut can be exposed. Note that the depth is measured to the top of the section where the depth varies.


Figure 4. Typical example of a steel triangle that has been cut and then sectioned to show the quality of the cut.

I mentioned, in an earlier article, that we had compared different designs from competing manufacturers. Under exactly the same pressure, water flow and abrasive feed rates, the difference between the cutting results differed more greatly than had been expected.


Figure 5. Sectioned views of six samples cut by different nozzle designs, but at the same pressure, water flow, AFR and cutting speed.

There was sufficient difference that we went and bought second, and third copies of different nozzles and tested them to make sure that the results were valid, and they were confirmed with those additional tests. Over the years as other manufacturers produced new designs, these were tested and added into the table – this was the result after the initial number had doubled. (The blue are results from the first nozzle series tests shown above).


Figure 6. Comparative depths of cut using the same pressure and AFR but twelve different commercially available nozzle designs.

There were a number of reasons for the different results, and I will explain some of those reasons as this series continues, but I will close with a simple example from one of the early comparisons that we made. We ran what is known as a factorial test. In other words the pressure was set at one of three levels, and the AFR was set at one of three levels. If each test ran at one of the combination of pressures and AFR values, and each combination was run once then the nine results can be shown in a table.


Figure 7. Depths of cut resulting from cutting at jet pressures of 30,000 to 50,000 psi and AFR of 0.6, 1.0 and 1.5 lb/min.

The results show that there is no benefit from increasing the AFR above 1 lb/minute (and later testing showed that the best AFR for that particular combination of abrasive type, and water orifice and nozzle diameters was 0.8 lb/minute).

Now most of my cutting audience will already know that value, and may well be using it, but remember that these tests were carried out over fifteen years ago, and at that time the ability to save 20% or more of the abrasive cost with no loss in cutting ability was a significant result. Bear also in mind, that it only took 9 tests (cutting time of around 30 minutes) to find that out.

__________________________________________

* The reason that the “I’m from Missouri, you’ll have to show me,” story got started was that a number of miners migrated to Colorado from Missouri. When they reached the Rockies they found that, though the ways of mining were the same, the words that were used were different. (Each mining district has its own slang). Thus they asked to be shown what the Colorado miners meant, before they could understand what the words related to.

Waterjetting 3d - High-pressure pump flow and pressure

When I first began experimenting with a waterjet system back in 1965 I used a pump that could barely produce 10,000 psi. This limited the range of materials that we could cut (this was before the days when abrasive particles were added to the jet stream) and so it was with some anticipation that we received a new pump, after my move to Missouri in 1968. The new, 60-hp pump came with a high-pressure end that delivered 3.3 gpm at 30,000 psi. which meant that a 0.027 inch diameter orifice in the nozzle was needed to achieve full operating pressure.

However I could also obtain (and this is now a feature of a number of pumps from different suppliers) a second high-pressure end for the pump. By unbolting the first, and attaching the second, I could alter the plunger and cylinder diameters so that, for the same drive and motor rpm, the pump would now deliver some 10 gpm at a flow rate of 10 gpm. This flow, at the lower pressure, could be used to feed four nozzles, each with a 0.029 inch diameter.

Figure 1. Delivery options from the same drive train with two different high-pressure ends.

The pressure range that this provided covers much of the range that was then available for high-pressure pumping units using the conventional multi-piston connection through a crankshaft to a single drive motor. Above that pressure it was necessary to use an intensifier system, which I will cover in later posts.

However there were a couple of snags in using this system to explore the cutting capabilities of waterjet streams in a variety of targets. The first of these was when the larger flow system was attached to the unit. In order to compare “apples with apples” at different pressures some of the tests were carried out with the same nozzle orifice. But the pump drive motor was a fixed speed unit which produced the same 10 gpm volume flow out of the delivery manifold regardless of delivery pressure (within the design limits). Because the single small nozzle would only handle a quarter of this flow, at that pressure (see table from Waterjetting 1c) the rest of the water leaving the manifold needed an alternate path.

Figure 2. Positive displacement pump with a bypass circuit.

This was provided through a bypass circuit (Figure 2) so that, as the water left the high-pressure manifold it passed through a “T” connection, with the perpendicular channel to the main flow carrying the water back to the original water tank. A flow control valve on this secondary circuit would control the orifice size the water had to pass through to get back to the water tank, thereby adjusting the flow down the main line to the nozzle, and concurrently controlling the pressure at which the water was driven.

Thus, when a small nozzle was attached to the cutting lance most of the flow would pass through the bypass channel. While this “works” when the pump is being used as a research tool, it is a very inefficient way of operating the pump. Bear in mind that the pump is being run at full pressure and flow delivery, but only 25% of the flow is being sent to the cutting system. This means that you are wasting 75% of the power of the system. There are a couple of other disadvantages that I will discuss later in more detail, but the first is that the passage through the valve will heat the water a little. Keep recirculating the water over time and the overall temperature will rise to levels that can be of concern (it melted a couple of fittings on one occasion). The other is that if you are using a chemical treatment in the water then the recirculation can quite rapidly affect the results, usually negatively.

It would be better if the power of the pump were fully used in delivering the water flow rate required for the cutting conditions under which the pump was being used. With a fixed size of pistons and cylinders this can be achieved, to an extent, by changing the rotation speed of the drive shaft. This can, in turn, be controlled through use of a suitable gearbox between the drive motor and the main shaft of the pump. As the speed of the motor increases, so the flow rate also rises. For a fixed nozzle size this means that the pressure will also rise. And the circuit must therefore contain a safety valve (or two) that will open at a designated pressure to stop the forces on the pump components from rising too high.

Figure 3. Output flows from a triplex (3-piston) pump in gpm, for varying piston size and pump rotation speed. Note that the maximum operating pressure declines as flow increases, to maintain a safe operating force on the crankshaft.

The most efficient way of removing different target materials varies with the nature of that material. But it should not be a surprise that neither a flow rate of 10 gpm at 10,000 psi, nor a flow rate of 3.3 gpm at 30,000 psi gave the most efficient cutting for most of the rock that we cut in those early experiments.

To illustrate this with a simple example: consider the case where the pump was used configured to produce 3.3 gpm at pressures up to 30,000 psi. At a nozzle diameter of 0.025 inches the pump registered a pressure of 30,000 psi for full flow through the nozzle. At a nozzle diameter of 0.03 inches the pump registered a pressure of 20,000 psi at full flow, and at a nozzle diameter of 0.04 inches the pressure of the pump was 8,000 psi. (The numbers don’t quite match the table because of water compression above 15,000 psi). Each of these jets was then used to cut a slot across a block of rock, cutting at the same traverse speed (the relative speed of the nozzle over the surface), and at the same distance between the nozzle and the rock. The depth of the cut was then averaged over the cut length.

Figure 4. Depth of cut into sandstone, as a function of nozzle diameter and jet pressure.

If the success of the jet cut is measured by the depth of the cut achieved, then the plot shows that the optimal cutting condition would likely be achieved with a nozzle diameter of around 0.032 inches, with a jet pressure of around 15,000 psi.

This cut is not made at the highest jet pressure achievable, nor is it at the largest diameter of the flow tested. Rather it is at some point in between, and it is this understanding, and the ability to manipulate the pressures and flow rates of the waterjets produced from the pump that makes it more practical to optimize pump performance through the proper selection of gearing, than it was when I got that early pump.

This does not hold true just for using a plain waterjet to cut into rock, but it has ramifications in other ways of using both plain and abrasive-laden waterjets, and so we will return to the topic as this series continues.