Waterjetting 30b - An opportunity missed, and a question raised

In the last post I wrote about the benefits of cutting a deep slot around the edges of a tunnel, and made reference to the work done here by Dr. El-Saie back in the mid-1970’s as part of his Doctoral Dissertation.

One of the concerns that we had to address was that the waterjet had to be able to penetrate all the different rock types that it might encounter, and at the same time, since the jet would only cut a short depth on each pass we also had to find a way of cutting the slot wide enough that the nozzle assembly could enter and deepen the slot over consecutive passes around the edge. Given that the available pressures in those days were limited, for us, to 30,000 psi and this pressure was insufficient, by itself, to cut through all the rocks we might encounter, Dr. El-Saie looked at several different ways of enhancing performance. These included adding abrasive to a high pressure jet stream, inducing cavitation into the jet stream and the potential for using the break-up of the jet into droplets to enhance cutting using the impact water hammer effect.

Because of other operating conditions it was not considered practical to try and develop the droplet impact idea for this program, and the work concentrated on examining the potential differences between abrasive waterjet injection and cavitation. To simplify the comparison the same basic nozzle design was used for the tests that were then run, although the shroud fitted to create the secondary (vacuum) chamber was modified to either allow abrasive entrainment, through ports, or to create cavitation. The presence of the ports did, however, allow the strength of the vacuum generated in the chamber to be measured as the jet passed through.

Figure 1. Nozzle designs used by Dr. El-Saie. Note that the upper design has ports leading into the vacuum chamber, so that abrasive can be drawn in by the jet passage. In the lower design there are no ports, and cavitation will be induced in the chamber by the jet passage, with the bubbles then drawn into the exiting jet.

One of the advantages of cavitating the jets is that the cavitation bubble collapse will spread out over a larger area on the target surface, so that the slot generated can be quite a bit larger than the originating jet. This can be shown in two pictures of a block of dolomite exposed to the same cavitating jet, at a pressure of 6,000 psi but one with the jet traversed along the block in a minute, while in the second case the jet is moved at a slower speed, taking five minutes to cross the block, which allows the jet to exploit the cracks generated by the cavitation. A fuller description of the process is given here.

Figure 2. Cavitation damage pattern on a block of dolomite showing the initial width of the jet (red lines), and the zone of damage that is being created around the traverse path.

Figure 3. Cavitation damage on a block of dolomite at a slower traverse speed, showing the width of the damage track that can be created. The slot is about half-an-inch deep.

In the course of the test program different shroud shapes were tested, but in all cases the comparison between an abrasive-laden jet and one containing cavitation bubbles was made with shroud shapes of the same overall dimensions.

The ratio of the exit diameter (discharge) from the shroud (D2) to that of the initial jet orifice (D1) was first changed to one of four different ratios, though the diameter of the initial jet was kept at 0.04 inches (1 mm). Of the different sizes tested the greatest vacuum in the chamber was measured with the smallest of the discharge diameters was tested.

Figure 4. The effect of increasing the throat length of the shroud on the vacuum pilled in the chamber, at different pressures.

If the discharge diameter was increased to 6.35 mm then the jet pressure had to be increased to 12,500 psi to obtain the same levels of vacuum achieved otherwise at 7,500 psi.

Figure 5. Vacuum pressures measured with a larger discharge diameter from the shroud, for different lengths and jet pressures.

Impact force measurements from the jet hitting a target at varying distances from the orifice, with and without the shroud showed relatively little difference in the overall total impact force (not considering abrasive) out to a distance of 6 inches. There was thus no apparent effect due to jet disintegration from the use of the shroud over these distances.

In order to compare the performance of the abrasive-laden jet with that of a cavitating jet, a new nozzle design was developed, and small samples of granite were rotated in front of each nozzle assembly, for 20 seconds. Because the jet had to be brought up to pressure for each test, and shut down afterwards, a steel shutter plate was placed between the nozzle and the target. One of the irritants in doing the tests was that the jets kept cutting through this shutter plate.

Figure 6. Steel shutter plate cut through in 6 seconds during system start-up.

The shroud, made of stainless steel, was also wearing out within a few minutes. Unfortunately we did not recognize that this was demonstrating that abrasive waterjets were an effective method for cutting metal – that commercial development had to wait for the more perspicacious Dr. Hashish to work with Flow Research and bring the technology to the market in 1980.

Part of the reason for our lack of interest was because of a different conclusion that Dr. El-Saie drew from his work, based on the following two curves. The first comparison of different jet results occurred with a jet pressure of 7,000 psi.

Figure 7. Volume of material removed from granite samples, as a function of distance, for four different jet conditions at a jet pressure of 7,000 psi.

Note that the three water jets do not have much significant effect on the granite at this distance and jet pressure (we had to learn some later lessons to make them more productive at this pressure). But even at this pressure the abrasive waterjet was effectively cutting the granite.

But it was the change in the relative position of these curves, as the pressure was then increased to 20,000 psi that caught our attention. (The intermediate plots are not given here).

Figure 8. Volume of material removed from granite samples, as a function of distance, for four different jet conditions at a jet pressure of 20,000 psi.

The water feed was not useful, since the power required to accelerate that volume drew heavily from that available through the jet.

The plain jet, without a shroud will cut granite at this pressure, particularly when moved over the surface. (And we later used this system to carve the Millennium Arch_ – as well as the Missouri Stonehenge). But the performance at that time was not that impressive.

Opening the ports on the shroud, without feeding anything into the jet caused, we believed a greater jet breakup and thus some additional droplet impact effects that improved cutting performance over that of the plain, more coherent jet, in part because the jet was spread over a larger contact surface.

Closing the jets induced cavitation in the stream, and this gave the best performance of the four – including abrasive injection. Again this was, in part because of the larger area of damage that the cavitation generated on the target, over the narrower slot of the abrasive-laden and plain jets.

In comparison the abrasive waterjet did rather poorly. In retrospect this is perhaps more of a surprise – but it should be born in mind that there was little attempt at optimizing the feed condition (which later research shows has a dramatic effect on performance) or the chamber geometry. Further the slot cut was much narrower than that created by the cavitating jet.

But it certainly caused us, in that time interval, to look more at cavitation, and to totally miss the implications of the AWJ result.

Most of these illustrations come from the Doctoral Dissertation by Dr. A. A. El-Saie “Investigation of Rock Slotting by High Pressure Waterjet for Use in Tunneling”, Mining Engineering Department, Missouri University of Science and Technology, (Then University of Missouri-Rolla), 1977.

Waterjetting 27a - Cutting materials with internal stress

Safety glass, or toughened glass is typically designed so that, when it fails it will break into small pieces with few of the relatively sharp, and thus dangerous, fragments formed by ordinary glass. It is used in making shower doors, and automobile windows. As such is differs from laminated glass (which I will discuss in a later post in this section). The toughened glass is formed by quickly cooling the glass, after it has been heated. And one way to check if a sheet has been treated this way is to look at it through polarized sunglasses. Tempered glass will show a pattern. The reason for that, and for the rapidity of the breakup of the glass, is that the lines show the internal stresses that the glass treatment deliberately leaves in the material. (There is an interesting variation on this way to tell the difference using an iPhone.)

Figure 1. Broken pieces of tempered glass, showing the small fragments that result. (Floydglass)

Because the treatment puts the outer parts of the glass into compression, while the inner part is in tension once cracks start to appear in the glass, then the glass is designed so that these stresses will cause the cracks to grow, bifurcate and join in patterns that cause the glass to shatter into less dangerous fragments. But this creates a considerable problem if there is a need to reshape the glass after it has been heat-treated.

Note that this treatment is the opposite of the result where glass is annealed, where – by cooling the glass at a slow rate – the internal stresses are much reduced, but as a result, when the glass breaks the fragments can be more damaging.

Figure 2. Sheets of annealed glass, showing how it may break from impact. (ADMglass)

Annealed glass is, as a general rule, relatively easy to cut with an abrasive waterjet system provided that certain simple precautions are taken. However, when it comes to cutting tempered glass, one of the suggestions is to anneal it first, so as to get rid of the internal stresses. Unfortunately, in the process this also removes the benefits of the tempered treatment.

When one tries, without other treatment, to cut tempered glass the results are not pretty. Edgar Hernandez has posted a video of what can be expected to happen.

The problem goes back to the basic way in which waterjets, and abrasive waterjets work in cutting through material. Simplistically waterjet impact will penetrate the cracks that exist in a target surface; the following slug of water then pressurizes the water within the crack, causing it to grow. As cracks get longer it takes less and less pressure, either internally within the crack, or in the surrounding material, for that crack to grow catastrophically to failure of the piece. Where there are relatively few natural cracks in the material – as happens with glass – then abrasive is introduced into the waterjet stream, so that the impact of the small particles will form small cracks when they hit the glass surface. Normally those cracks are relatively small, and when first cutting into or piercing the glass the pressure of the jet is often lowered so that the particle speed is also lower and the crack length that the particles create is also small and localized around the impact point, so that the integrity of the whole piece is not threatened.

Figure 3. Cracks around the impact of single particles of abrasive onto glass.

The problem, from a cutting aspect, with tempered glass is that the internal stresses that are deliberately placed into the glass are designed so that cracks do not have to be very long before the concentrated stress at the crack tip (which increases with crack length) reaches a point where it will continue to grow at an increasing rate to failure of the piece. The longest cut we have made in tempered glass before it shattered was about an inch-and-a-half.

Because the stress in the glass is an inherent part of the nature of that particular type of glass there is no really effective way of cutting the material, after it has been tempered. If a particular shape is required then the glass should be cut to final shape before it is tempered, and care should be taken to ensure that there aren’t any large cracks or chips along the edge of the glass before it is then tempered.

Stress problems aren’t restricted, however, to trying to cut tempered glass. When cutting larger pieces of metal one can also run into problems from stresses that were left in the material after it was initially formed. Perhaps the most common of these is found where a partial cut allows a stressed part to lift slightly above the plane of the rest of the material. If the part is being cut in steps, the raised piece can then move into the path of the cutting nozzle as it moves back over the piece. This can have some unfortunate consequences for the nozzle and focusing tube (there goes bitter experience speaking again).

Other problems that can crop up come from the shifting of the piece in the plane of the part, but where the stress relief moves the edges so that subsequent cuts into the part no longer comply with the blueprint for the cuts, since the material has shifted. This shift can be a relatively small movement – depending on the level of stress that was captured in the material, but it can be enough to take the final part out of tolerance, and thus it never hurts to be sure of the stress condition of the piece before starting to cut.

I’ll return to this theme, but with a different illustration of stress effects next time.

Waterjetting 23b - From DIAjet to Abrasive Slurry Jetting (ASJ)

When the Direct Injection of Abrasive jet (DIAjet) was first introduced to the general public, back in 1986, there was some initial skepticism as to the overall market potential for the system. Certainly, as the next post will discuss, the ability to transfer higher levels of energy from the pressurized water to the entrained abrasive with no particle fragmentation in the mixing chamber, had a number of advantages, that I will spell out below, but the long-term problem was to develop a method of sustaining a constant abrasive feed in the mix, critical to high precision cutting, while concurrently having a system that can run continuously both day and night. To my own knowledge there have been at least two different designs of ASJ system that have solved this problem, but the market was damaged by the early problems in sustaining continuous flow to the point that customers shied away from this advanced technology, even where it showed some considerable commercial advantage.

The first, most obvious advantage to this new tool was in the smaller cut width that it generated, relative to conventional Abrasive Waterjet Cutting (AWJ). And because DIAjet became known as the original BHRA technology, and there were competitors over the years, let me re-name the technology (as the WJTA did some years ago) as Abrasive Slurry Jetting (ASJ).

Figure 1. Cut size difference between ASJ and AWJ as an illustration. The cuts are in Plexiglas, with the ASJ cut on the left.

The reason for the difference in cut width is, as explained in the last post, that the volume of the cutting jet is cut by 90% when the air carrier for the abrasive is no longer necessary or present in the jet stream. This increase in jet “delicacy” can be illustrated by a small example, and a humorous competition between Don Miller and ourselves back some years ago.

Don was, as the technology evolved, one of the master players in moving the technology toward the micro-cutting market that, to this day, remains remarkably under-exploited.

Because the abrasive particles accelerate to a large extent with the water that is both the cutting and carrier fluid, the cutting ability of the jet is significantly less sensitive to the diameter of the nozzle than is the case with the conventional AWJ. And, because the waterjet is not disrupted within the mixing chamber as a way of helping mix the abrasive with the water, so the jet stream can be kept convergent away from the nozzle, increasing the range, as I will discuss further in a later section.

The high precision cutting, using a finer abrasive since the nozzle diameter is smaller, can create very delicate pieces. We had been asked to use our system to cut jewelry out of silver, since while the ASJ could cut this easily and quickly to the desired shapes (matching necklace and earings) trials with laser cutting had been less successful because of the high conductivity of the metal.

This led on to a demonstration of the precision of the cut that can be achieved. One of the early models cut with an AWJ had been of a dragon, it seemed to be a good idea to match this with a knight on a trusty steed.

Figure 2. Knight vs dragon, in this case the knight was cut with an ASJ, while the dragon was cut with an AWJ system.

Don Miller had put together his precision system in his garage, and was able to control the cutting ability with an on-off switch located upstream of the nozzle, using sliding diamond coated plates to ensure a seal, without the wear problems which are common when trying to valve an abrasive laden flow.

Figure 3. Don Miller’s cutting equipment with precision table. Don Miller).

The need for the rapid on-off design comes where a series of holes must be punched into the target metal in order to effectively create a screen, or similar device. Pratt and Whitney engineers had used a somewhat different concept, and had held the abrasive in a polymer. This is sometimes necessary when using the ASJ system, when the cut-off in flow to the nozzle occurs with abrasive in the feed line from the reservoir to the nozzle itself. If there is any significant delay before flow restarts then the abrasive will settle to the bottom of the containing pipe. When flow then restarts this initial plug of abrasive is picked up by water flow and can block the nozzle when it reaches it. The use of the polymer holds the abrasive in suspension to prevent this happening. (We have seen abrasive held in suspension for over a week, using relatively low concentrations of polymers such as those used to suspend fracking sand for the oil industry).

Figure 4. The diamond-coated valve used by Don Miller to control flow in an ASJ system. (Don Miller)

The risk of abrasive build-up is increased when the shut-off valve is directly behind the nozzle, or where it is mounted vertically, when a polymer is not used, but where the jet is cycling rapidly to drill a precise series of holes in a rapid sequence, then the issue of nozzle blockage doesn’t arise as much.

Figure 5. A grid of 85-micron diameter holes drilled at 2.5 holes/sec at a jet pressure of 10,000 psi (Don Miller).

When we discussed the relative scaling that could be achieved with the technique, Don’s answer to the thickness of the lance that we had given the knight was to put scales on his dragon.

Figure 6. The scales on Don Miller’s dragon. The picture width is around 1 mm.

I had promised to go back and put eyelashes on the horse, but somehow we never managed to find the time.

The delicacy and accuracy of the technique is in marked contrast to manufacturing techniques other than those using abrasive-laden water as a cutting medium. Not only is it possible to cut through metals and other materials without distortion, even with very narrow webs left holding the pieces together, but since there is no heat involved in the cutting process, the precision is retained over the cut and part, after completion. This is illustrated with Don’s construction of a butterfly wing through 150 micron thick stainless steel. (The scale on the illustration shows mm).

Figure 7. Detail of a butterfly wing cut by Don Miller, using his ASJ system (Don Miller)

Waterjetting 18d - Abrasive considerations

It would be best if, before I ended this short session on abrasive, I mentioned some of the practical constraints that sometimes limit the options for choosing abrasive types. To give a simple example, we were, at one time, demonstrating the ability of a waterjet drill to penetrate limestone. In the demonstration that morning we had used garnet as the abrasive and had made a steady penetration down to about 70-ft but the contracting office on the project did not seem overly impressed. So, after lunch, I suggested that we switch to an aluminum oxide abrasive, since we knew it was more aggressive.

Unfortunately for the afternoon program we were using a DIAjet type of system, where the abrasive is added to the water under pressure just downstream of the nozzle, and upstream of the delivery nozzle. While that worked well with the garnet abrasive (which passed without significant damage through the swivel on the end of the drill) that was not the case with the aluminum oxide. This is a much sharper abrasive and less prone to damage in mixing. As a result once we had the rig back in operation we were immediately struck by the black color of the water coming out of the hole – as the aluminum oxide stripped the inner lining from the hose carrying it to the nozzle. We then watched as, in real time, the pressure gage on the driving pump slowly slid back from the 10 ksi initial pressure to about 2 ksi as the abrasive ate out the orifices of the nozzle. Needless to say, having pretty much destroyed the downstream equipment in about five minutes, the afternoon demonstration was a bit of a disaster.

I also remember the first time that we used steel shot to try and cut through some rock, without giving too much thought to encasing the cutting operation. Those small spheres retained a lot more energy than most particles, and we were dodging the equivalent of shotgun pellets which ricocheted around the lab as we raced to shut the system down.

Both abrasives are, in their place, very effective tools in cutting materials that might be more difficult or uneconomic to cut by other means, but the peculiarities of their nature require that special precautions be used when they are used to make sure that there are not unintended consequences.

Sometimes the choices are simply practical. When we were cutting the walls of the Omnimax theater under the Gateway Arch in St. Louis, where we had to cut straight down (within half-an-inch either way over 15-ft of cut depth) through dolomite and chert it took less than a day to realize that the cost of using garnet to achieve the 12 – 15-inch individual cut depths was going to drive us out of economic reality within a week. Changing to a blasting sand (which we bought by the ton) did not change the cutting performance by much, but had a remarkable effect on overall costs.

Figure 1. Effect of abrasive type, size and feed rate on the depth of cut and optimal cutting condition when cutting rock. (after Yazici*)

Abrasive type and abrasive size both effect the depth of cut, and thus the economics of a cutting operation. Yet it is not possible to draw absolute rules since the different abrasives have different relative cutting efficiencies in different materials. For example, in the above plot boiler slag was relatively ineffective in cutting rock. On the other hand, with the right type of slag and steel Faber and Oweinah** have reported that slag can cut steel more than three times as efficiently as garnet. (This is partly because the slag shatters on impact and the fragments go on to scour the uplifted edges of the cavities generated by the initial impact of the particle.)

And while the British Welding Institute use smaller particles to cut softer materials, they have found it critical to use larger particles to get viable performance as the target material gets harder. In cutting steel I had mentioned in an earlier post, that garnet becomes less effective at a particle size below 100 micron. Yet in cutting aluminum (which is softer) the particles can be smaller and yet still effective.

Figure 2. The effect of particle size when cutting aluminum using corundum particles (after Faber and Oweinah ibid)

Yet, as discussed at the beginning, the cost of the abrasive must not only be set off against the potential for improving the cutting rate, one has to also look and see if there is an increase in the operating cost of the system when a harder, and thus often more effective cutting abrasive is used. Zaring et al showed this with a plot that they published at the 6th American Waterjet Conference***.

Figure 3. Relative benefits and costs of changing abrasive type (after Zaring et al***)

All things are, however, relative, and in some small cutting operations we have found it more economic to sacrifice the nozzle over the cutting time required in order to achieve a cut that could not be effectively achieved any other way.

As with many things in the waterjet business, while there are general rules that can be laid down to guide operations, when it comes to specific cases then it is often worth running a small series of tests on the projected target material, using different abrasives, at varying size ranges and feed rates, before calculating (usually using a normalized cost in dollars or gms per area of cut) the most effective abrasive for a given operation.

*Yazici, Sina, Abrasive Jet Cutting and Drilling of Rock, Ph.D. Dissertation Mining Engineering, Univ. of Missouri- Rolla, Rolla, MO, 1989, 203 pp.
**Faber, K., Oweinah, H., "Influence of Process Parameters on Blasting Performance with the Abrasive Jet," paper 25, 10th International Symp Jet Cutting Technology, Amsterdam, Oct, 1990, pp. 365 - 384.
***Zaring, K., Erichsen, G., Burnham, C., "Procedure Optimization and Hardware Improvements in Abrasive Waterjet Cutting Systems," 6th American Water Jet Conf, Houston, TX, Aug, 1991, pp. 237 - 248.

Waterjetting 16c - Optimal AFR and cutting curves

The discussion on surface quality which forms this month’s topic has, to date, focused on linear cutting since this has been the simplest way of explaining some of the factors that go into choosing an optimum abrasive feed rate (AFR) for a system. Along the way, however, I have pointed out that the internal design of an abrasive nozzle has a considerable impact on the relative performance of different systems.

If, for example, the internal geometry is such that there is not an optimal transition of energy between the high-pressure waterjet stream and the abrasive particles, then trying to draw conclusions over the influence of some of the operational parameters, such as pressure, can lead to false conclusions. The optimal AFR changes with the relative sizes of the waterjet orifice, the location of the abrasive feed line, the length of the mixing chamber and the geometry of the focusing tube. These parameters are generally held fixed since most folk buy only one cutting head design, and tend to stick with it once purchased. However, as I pointed out at the beginning of this blog, there is a considerable difference between the performance of different abrasive cutting heads.

Figure 1. Comparison of the relative cutting performance of twelve different abrasive nozzle designs, when operated otherwise at the same pressures, water flow rates and AFR.

The best design, for the particular waterjet and AFR parameters that were tested in generating Figure 1, was 24% more effective than the average performance of the nozzle designs tested. This is indicative that the design was more efficient in accelerating the abrasive to a higher velocity than the competing designs. Those designs were tested at a number of pressures and AFR values to ensure that the conclusions held within the range of test – and they did. But as the pressures and AFR values change so there is a change in the optimal design with consequences on the optimal AFR as it relates to the operating pressure of the system.

Without an awareness of these inter-related parameters it is possible to draw erroneous conclusions about the best choice of cutting parameters for a given operation. The situation becomes even more complex where the paths being cut are no longer straight but involve complex contour cutting, and where there are requirements for zero taper and high surface quality on the cut faces of the part being generated.

One solution to the problem is to accept the limitations of the system, and cut the part at a constant speed, slow enough that the jet cuts through the piece on first contact with the abrasive stream over the length of the cut. (In other words after the abrasive bounces away from the initial contact plane along the cut it does not meet any more material before it exits from the bottom of the cut). At a pressure of 40,000 psi the cutting speed to achieve these requirements over an half-inch thick titanium target lies at around 0.3 inches per minute.

Figure 2. Change in cut face taper angle with traverse speed at a cutting pressure of 40,000 psi.

However as the pressure of the jet is increased the cutting speed to sustain that quality cut goes up significantly, so that there is a significant benefit to the increased pressure. But the optimization to achieve this is geared to ensuring that the optimal abrasive feed rate has been selected, for a given nozzle design and waterjet pressures. Without a short series of tests to ensure that the system is being run at this optimal condition it is not possible to accurately state how a system can best be used.

I have described, in an earlier post, how such a simple test can be run. It should be stressed, however, that the selection of an optimal AFR for a nozzle is based on the nozzle geometry and the operating pressure of the system. That selection will provide the best cutting jet and this jet will have different capabilities in different target materials. Composite materials will cut at a different optimal speed depending on the material type and thickness, and these values will differ when metals, or ceramic materials are being cut. But, as a general rule, the selection of the best cutting conditions are first established by knowing the thickness and type of material to be cut. This should then produce, based on tested performance tables, recommendations for the cutting speeds at different pressures, where the cutting pressure in turn defines the optimal abrasive feed rate. Based on an assessment of the different categories of cost of an individual operation one can then decide which set of conditions would provide the most economical and acceptable answer to providing the quality of cut required.

In some cases it may be that the cutting head can be tilted so that, particularly with straight cuts, the part being isolated will have a perpendicular edge, while the scrap piece will have a tapered edge at twice the normal angle. For example under the conditions illustrated in figure 2 tilting the nozzle by only one degree will allow cutting at 4 ipm rather than 0.3 ipm, a 12-fold gain in performance, depending on the assurance of the quality of the surface being sustained.

As mentioned earlier this option becomes more difficult as the part being cut acquires contours. At higher pressures the angle of the cutting face curve is reduced, but in thicker parts there is often a slight displacement backwards (a rooster tail as it is sometimes called) from the top edge of the cut to the bottom. When the nozzle comes to cutting around a curve that backward projection at the bottom of the cut can pull the cut edge away from vertical unless the cutting head is adjusted to ensure that this difference is minimized to the levels acceptable to the customer. Most commonly this is achieved by slowing the head speed according to the radius of the curve, with sharper turns being made at slower speeds. Some adjustment in the angle of the head can also be made, but this requires a more advanced method of control and programming in developing the cutting path for the head.

Note: Because of the season this site will be Dark next week, so let me take the opportunity of wishes the readers of the waterjetting series all the Compliments of the Season, and with hopes that you have a Prosperous and Happy New Year.

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 16a - Abrasive Feed

I was talking with someone the other day who mentioned that it was necessary to increase the amount of abrasive feed to a waterjet whenever the material to be cut was thicker. This is actually a myth, or put another way, wrong! It is the equivalent of saying that you should use a duller knife if you want to cut thinner material.

Today’s topic therefore is about optimizing the abrasive feed rate into an abrasive waterjet cutting system, but you should remember that different manufacturers have different nozzle assembly designs. Thus the graphs and tables that I will use to illustrate the discussion relate to one particular nozzle design that was used at that time. Not all the nozzle designs were the same, but the results illustrate the points that I am going to make. But it underlies the recommendation that you should each run some standard tests with your system so that you have a baseline of performance and data to tell you where your system works best.

There are several reasons why different designs produce different cutting results, and I will point out some of them in what follows. To begin, however, consider again the basic elements of the waterjet mixing chamber.

Figure 1. Section through the mixing chamber of a conventional abrasive injection system.

A small high pressure stream of water enters the chamber through the upper nozzle, passes through the chamber, creating a vacuum that pulls abrasive into the chamber, and mixes with that abrasive before the resulting AWJ exits through the focusing tube.

One of the first things to understand is that, in the cutting jet that issues from the tube the actual cutting comes from the abrasive particles. From work that has been carried out at a number of places we know that the higher the velocity of the particles, the greater the damage that they will do on the target.

Figure 2. Relative mass loss when steel balls hit a mild steel plate (after Hutchings, I.M., Winter, R.E., Field, J.E., "Solid Particle Erosion of Metals;The Removal of Surface Material by Spherical Projectiles," Proceedings of the Royal Society, London, Vol. A348, 1976, pp. 379 - 392.)

Discussion in earlier posts pointed out that there are different ways in which waterjets attack ductile and brittle materials. However the relationship between an increase in impact velocity and damage occurs whether the targets are brittle or ductile. In figure 2 the target was a ductile steel, in more brittle material it is the coalescence of cracks that removes material, and higher velocities create larger cracks, as shown in figure 3.

Figure 3. Effect of impact velocity on crack length, (after Evans, A.G., "Impact Damage Mechanics: Solid Projectiles," in Erosion, Treatise on Materials Science and Technology, Vol. 16, ed C. Preece, Academic Press, 1979.).

And the same form holds true if steel balls are fired into sandstone.

Figure 4. Relative amount of Berea sandstone removed by the impact of steel balls of varying size (after Ripkin, J.F., Wetzel, J.M., A Study of the Fragmentation of Rock byImpingement with Water and Solid Impactors, Final Report on U.S. Bureau of Mines, Contract HO 210021, February, 1972.).

The above graphs show that it is more effective to have the abrasive moving faster in terms of the damage done by individual particles. Which brings up the first consideration in the design and use of an AWJ mixing chamber.

In order to get the particles moving as fast as possible they have to get their energy from the water jet entering the chamber. But the jet enters the chamber as a solid stream that then breaks into droplets (as shown in earlier pictures) as it passes down the chamber. If the jet remains in a solid stream all the way down, and out of the focusing tube, then the abrasive will find it difficult to penetrate into the center of the jet stream and pick up all the needed jet energy.

This can be illustrated by looking at the velocity and distribution of particles coming out of a nozzle with two different sizes of waterjet orifice but the same size of focusing tube diameter.

Figure 5. Relative particle distribution across a 40,000 psi jet with a focusing tube diameter of 2.3 mm (0.09 inches), and an AFR of 1 lb/min for two waterjet orifice sizes (after Mazurkiewicz, M., Olko, P., Jordan, R., "Abrasive Particle Distribution in a High Pressure Hydroabrasive Jet," International Water Jet Symposium, Beijing, China, September, 1987, pp. 4-1 - 4-10.)

The smaller waterjet breaks up fully within the chamber entraining and accelerating the abrasive particles and providing the desired cutting stream. The larger sized jet does not completely breakup, and fewer particles can mix into the center of the jet giving a more diffuse and less efficient cutting stream. In this case changing from a 0.005 inch waterjet orifice to a 0.013 inch diameter orifice (at roughly 7 times more power, because of the higher flow rate) produces a poorer result.

It is therefore important to ensure that there is an efficient energy transfer between the water and the particles. But the jet energy can only be diffused to a certain number of particles before it significantly begins to reduce in the amount of energy that it imparts to each particle. In other words if you put too much abrasive into the jet stream, then the amount of energy each particle gets is reduced, as it the overall cutting efficiency.

In an example I have used in class I noted that if I pick up a small child and run down the corridor, then I can carry the child at about my normal running speed, on the other hand if I pick up a couple of football players and try the same run I will be barely able to stagger. So the optimum carrying capacity of any jet can be determined for a given water flow rate, which is itself based on the waterjet orifice diameter and the pressure at which the water is supplied.

I will return to this topic next time, but you can see, in the concluding figure, that when a lower abrasive feed rate is fed to the nozzle, that the percentage of the abrasive moving in the higher velocity range rises to over 60% compared with only 20% of the particles when the abrasive feed is too high. And that means that the cutting performance will be less with the higher abrasive feed rates. (The numbers are a little high to reinforce the point).

Figure 7. Particle velocity distribution on leaving the focusing tube (after Isobe, T., Yoshida, H., Nishi, K., "Distribution of Abrasive Particles in Abrasive Water Jet and Acceleration Mechanism," paper E2, 9th International Symposium on Jet Cutting Technology, Sendai, Japan, Oct, 1988, pp. 217 - 238.)

Waterjetting 15d – More thoughts on cut surface quality.

When a high-pressure stream of water hits a surface, the arrival of subsequent lengths of the waterjet stream forces the initial water away from the initial impact point, into and along any weakness planes in the target material. As a result there is some preferential cutting of the material, especially where there are defined weakness planes in the material. One illustration of this is where a jet that contains cavitation bubbles impacts on a rock surface (figure 1) and as the water enters the narrow eroded channels where preceding lengths of water have preferentially eroded out the weaker rock the pressure in the channel increases, collapsing the remaining cavitation bubbles and further exacerbating the damage within that narrow channel, causing it (them) to grow preferentially relative to the surrounding rock.

Figure 1. Looking down into a channel cut by a cavitating jet that traversed from left to right, at a speed of 0.4 inches/minute. Note the preferential attack into weakness planes within the rock.

As the weakness planes grow and join, so individually larger pieces of rock can be broken free from the target and the path, and pressure profiles of the water in the cutting zone change quite significantly. For this cavitation to have a significant impact on the erosion pattern, however, the traverse speed over the surface must be controlled, and be relatively low. At more effective speeds the cutting process does not allow for the development of this fracture mechanism. Rather, with plain jets, the process concentrates just on crack growth around individual grains. Optimum cutting speeds are much higher, depending on the intended result.

The efficiency of waterjet cutting has, historically, been assessed in terms of how much energy is required to remove unit volume of material. This we call the specific energy of the cutting process, and a common unit is joules/cubic centimeter (j/cc). When using a waterjet to cut into material, in part because of the interference between different segments of the jet stream, pre and post impact, the most efficient cutting speeds are quite high.

Figure 2. The change in cutting efficiency with traverse speed of a high-pressure waterjet cutting stream

The downside to using higher cutting speeds (apart from the simple inertial problems in driving systems at higher speeds in other than straight lines) is that the depths of cut achieved become smaller on individual passes, as the jet has less cutting time on each path increment.

Figure 3. Change in cut depth as a function of traverse speed, for varying different rock types.

In linear cutting systems it is sometimes possible to align secondary or a higher multiple array of nozzles along the cut, so that thicker materials can be cut with a sequence of jet cuts along the same path. Alternately a single nozzle can make multiple passes along the cut path and sequentially deepen the slot.

Unfortunately while this is an effective way of solving some problems, it becomes less efficient as the slot gets deeper.

Figure 4. The change in cutting efficiency with increase in the number of cutting passes.

At higher pass numbers with the target surface at a growing distance from the nozzle, and with the edges of the cut starting to interfere with the free passage of the jet to the bottom of the cut, less energy is arriving at the bottom of the slot and thus the effectiveness falls.

While there are differences between abrasive waterjet cutting (where the optimal cutting speed is much lower than that for a plain high-pressure water jet) the form that the cutting jet takes through the target material is of similar shape in both circumstances.

Figure 5. An abrasive waterjet cut through 1-inch thick glass

As the jet cuts through the piece, so the cutting edge curves backwards from the top of the cut to the bottom. The rate of this curvature is, inter alia, a function of how fast the nozzle is moving over the surface. Dr. Ohlsson showed this effect in cutting through 0.4-inch thick aluminum and mild steel plates, back as part of his doctorate at Lulea in 1995.

Figure 6. Change in the cutting edge profiles and cut groove patterns in metals as a function of cutting speed (L. Ohlssson PhD Lulea, 1995)

The growth of the striations in the cut surface, as the depth of cut increases is one of the larger concerns with cut surface quality, since customers are often concerned that these be minimized, and further if they become large enough they can make it difficult to separate the pieces, particularly if the parts are cut with a complex geometry.

Early in the understanding of the way in which waterjets work, it was thought that the jet would incrementally cut strips from the material in front of the previous cut, inducing steps into the cutting plane that worked their way down the material.

Figure 7. Early concept of cutting front development (L. Ohlssson PhD Lulea, 1995)

However, as higher speed cameras recorded the development of the cutting front, this concept has been rethought. Henning, for example at the 18th ISJCT, used a camera taking 520 frames per second to establish the development of the cut profile as the jet cut through clear plastic. In figure 8 the profiles are shown as they developed at 35 frames/sec to allow them to be distinguished.

Figure 8. Cutting front development as an abrasive jet cuts from right to left (Henning 18th ISJCT)

As Ohlsson had shown this profile develops as the abrasive laden jet impacts then bounces, then impacts and cuts further into the material, as it moves down the cut.

Figure 9. Frames showing a sequence as an abrasive waterjet cuts through 2-inches of glass. ((L. Ohlssson PhD Lulea, 1995)

In his work Henning correlated the change in the “bounce angle” with the jet properties, while Ohlsson also correlated with the traverse speed.

Figure 10. Change in the “bounce” angle as an abrasive jet moves down the cut (Henning 18th ISJCT)

Two things should be remembered in this analysis, since they also explain causes of the increased roughness of the cut each time the jet bounces. The first is that the jet is not only laden with any initial abrasive, but as it cuts into the material, and removes it so that cut material is entrained in the jet, so that there is some abrasive cutting, even with a plain waterjet once the initial cut has been made. The second point is that when the jet bounces it is not constrained to bounce just in the plane of the cut, but can and does take up some deflection into the sides of the cut. Thus, with each bounce and reflection, the cut becomes rougher as that side cutting becomes more pronounced.

However the number of bounces can be slowed by slowing the speed at which the nozzle moves over the surface.

Figure 11. Change in the angle along the cutting edge as the speed of cutting and the jet pressure are changed (H. Louis, Waterjet Conference, Ishinomaki, 1999)

Henning uses a different term, but nevertheless it is clear that increasing the jet pressure and changing the diameter of the jet stream also controls the edge profile, and as discussed, with a smaller number of bounces so the edge quality improves.

Figure 12. The effect of changing jet pressure and jet diameter on the gradient of the cutting edge profile (Henning 18th ISJCT)

Waterjetting 13c - On Milling and bas reliefs

In the last two posts I have been discussing how, either with the use of masks, or with an orbiting nozzle tool, it is possible to mill the material from a confined space within a surface, thereby creating a pocket.

There are a number of advantages to the latter technique, albeit it does require a special tool, rather than using masks that can be made from material already available at a shop.

Figure 1. Using a steel plate to provide a mask, while cutting a square pocket in glass (Courtesy of Dr. Cutler)

Figure 2. Detail of the corner of the pocket (Dr. Cutler)

With the oscillating tool, which can go deeper into the part to keep the distance from the nozzle to the work surface short, the corners can’t be as sharp as they are with the mask, since the outer radius of the focusing tube provides a limiting bound, once it moves into the cut. However, for shallower pockets where the nozzle can be further away, then the limiting corner radius sensibly becomes the orbiting radius of the nozzle.

Figure 3. A milled pocket in glass made using the Wobbler.

Note that the floor is relatively even in both cases, though in the masked case the view is taken only after the first pass over the glass. With the orbiting head it is possible to slightly tilt the head (it only requires a couple of degrees – depending on the other operational parameters) to ensure that the walls are being cut to as tight a tolerance to spec as desired (give or take a thou).

Dr. Hashish has noted, from some of the early work that he carried out, that it is possible to mill materials so that very thin skins (around 0.02 inches) can be left at the bottom of the pocket. As I will note in more detail next time, it is also possible to mill using abrasive waterjets in such a way as to leave intervening walls between adjacent pockets that are only that thick. If you have never had to do this in a conventional machine shop, you should know that as the wall of the pocket gets this thin, particularly at significant milling tool depth, the heat from the milling process, and the forces on the metal under the cutter are such that the wall will likely have some permanent deformation after the milling is over. Such is not the case where an abrasive waterjet system, of either variety, is used to cut the pocket.

Depths of cut uniformity can be held to a thousandth of an inch, though this requires some careful selection of both the abrasive size, and feed rate as a function of the other operational parameters of the system. As I mentioned last time, and Dr. Hashish demonstrated, as increasing precision is required in creating the floor of the pocket, so the abrasive being used must become finer and more precisely sieved to keep the wear pattern consistent.

. Figure 4. The effect of change in abrasive size on the smoothness of the pocket floor (Hashish M.: An investigation of milling with abrasive-waterjets, Trans. ASME, Journal of Engineering for Industry, Vol. 111, No. 2, 1989, pp. 158–166.)

There is an interesting niche market waiting to be developed in sculpting, I believe, based on putting some of these factors together. It was Professor Borkowski of the Unconventional HydroJetting Technology Center at Koszalin University of Technology* who first demonstrated that, by controlling the jet feed rate over the target, that the depth of cut into the material (and thus the depth to the floor of the pocket) could be controlled.

If now a photograph is scanned, so that the color of individual pixels along the photograph can be identified, then this color can be translated into a required depth By then setting the speed of the nozzle over that point on the target surface to give the required depth, then the jet will profile, from the color changes along the scanned path, the depth of cut on the milling path over the target. The details of the process are specified in the paper cited above, and the result has been the transfer of a 2-D image from a photograph to a 3-D bas relief cut into metal or other material surface. The depth control was well achievable using the rotational frequency of a stepping motor to drive the motion of the nozzle.

Figure 5. Outline of the process turning pictures into bas-relief (Dr. Borkowski).

The initial pictures that were obtained with the very first experiments were somewhat simple, though more than adequate to validate the concept. Where a smoother surface was required secondary passes could be made either in a parallel or orthogonal direction.

Figure 6. Early ball shape cut into metal to demonstrate speed control effect (Dr. Borkowski)

The next trial was with a ladies photograph:

Figure 7. Early trial of the technique to validate the effectiveness of the computer control (Dr. Borkowski)

More recently, as the process has been refined, much more detailed profiles have been demonstrated, as was seen, for example at the 2010 BHRA meeting in Graz.

Figure 8. Lizard bas-relief as shown at the 2010 waterjet meeting.

The concept of changing depth of cut, and thus being able to transfer photographs from the screen or paper onto metals or rock was an interesting academic challenge, that MS&T chose to address in a slightly different way.

Consider that the depth can be achieved by changing the speed of the nozzle on a single pass, so that the depth is controlled, or one can control the depth when only plain waterjets are used, by rapidly switching the jet on or off, as it makes sequential passes over the projected picture area.

The first image on steel led the subject in the first photo to mutter something along the lines of putting them on tombstones to remember those who had passed, so the next tests used photographs of my Grandfather and Dr. Clark, who founded the RMERC.

Figure 9. Images of my Grandfather and Dr. Clark transferred to basalt. (Dr. Zhao)

The technology advanced to the point that it was used to generate the plaque presented to me on my retirement from active academia.

Figure 10. My retirement plaque

Which seems to be a good point to close until next time.

*This University was kind enough to give me an honorary diploma.