Friday, May 29, 2015

Waterjetting 33c - More on enhanced cutting performance

Frontal photographs taken of waterjets, regardless of pressure, show that the jet spray widens as the jet moves away from the nozzle. Yet, because of the erosion of the outer layers of the jet by the surrounding fluid, the central core of effective jet pressure reduces as that distance grows. The normal way in which this can be seen is in the taper of the cut as a jet penetrates into a target material.

Yet in some cases better results can be obtained if the jet makes a series of passes to cut through that target layer. There is, however, a little problem, and this can be shown by the use of a curve, showing the depth of cut as a function of the number of passes made over the surface.


Figure 1. Depth of cut in two materials, as a function of the traverse speed. (After Hashish).

The graph shows the decline in the cutting ability of the jet with increasing number of passes, and the inability of the jet, at the highest speed, to penetrate through the mild steel plate.

One of the reasons to make passes at a higher speed is to improve the quality of the edge cut, since if the jet makes the pass with the particles of abrasive only cutting on the target one time (rather than the multiple cuts made by a particle at slower speeds, where it bounces down the cut.)

Momber has pointed out the decrease in performance with increased pass number, as well as noting the difference in the amount of energy required to cut through a target as a function of the speed and number of passes.


Figures 2 and 3. Effect of the number of passes on the depth achieved (lhs) and the relative amount of energy required to penetrate material as a function of traverse speed (rhs) (after Momber)

The slight loss in cutting power as the jet cuts deeper in secondary passes comes in part because the jet is constrained by, and thus cuts back into, the walls of the pre-existing cut.

In an earlier post on cutting I pointed out that because of the highly efficient way in which a plain waterjet cuts into material, Chinese investigators have shown that one can achieve a much improved volume removal rate by oscillating the jet perpendicular to the line of travel.

The optimal speed for cutting with an abrasive jet is, however, much slower than that of a plain waterjet, by a couple of orders of magnitude, so that the large scale oscillation that is effective with plain jets will not be similarly so with an AWJ. However the concept remains valid, and has been the subject of significant investigation, particularly in Australia, in the past few years.

The benefits of such oscillation, even over very short angles, can be illustrated with reference to a figure.


Figure 4. Oscillation of a jet perpendicular to the line of travel. The nozzle advances to the grey outline on the subsequent pass. The two lines indicate the range of oscillation. (Motion exaggerated relative to the current discussion)

If the nozzle is oscillated so that the jet moves over the relatively narrow range shown in Figure 4, then after a pass, when the nozzle advances to make the second pass (and it does not have to be in the part to do this) then the jet does not make contact with the target until the back of the previous cut. Thus there is much less energy loss in traversing the jet to the new surface, and cutting performance is improved. If the oscillation is kept small the walls of the cut will still act to confine the cutting ability of the jet, and improve depth-cutting capability.

Shunli Xu looked at oscillating a jet at angles below 10 degrees, while cutting half-inch thick 87% alumina plates. A simple visual correlation showed the relative benefit of oscillation when cutting the plate with a 45 ksi jet, with an AFR of 1.2 lb/min, at a speed of 3.1 inches/min.


Figure 5, Cuts made into a ceramic plate, without (lhs) and with (rhs) a nozzle oscillation of 8 degrees at 10 Hz. (after Shunli Xu)

The study also looked at the effect of changing the oscillation parameters on the surface roughness of the cut achieved, finding that this is controlled by the angle of oscillation, the frequency and the speed of traverse, as well as jet pressure and standoff distance (not shown). The study found that, under optimal conditions, surface roughness could be reduced around 11% relative to linear cutting.


Figure 6. Effect of change in oscillation parameters on the surface quality of cut in a ceramic target (after Shunli Xu).

The study found that the parameters which control the depth of cut gain were a little more complicated to disentangle, given that the density of particles striking an individual area of the target is controlled by both the jet residence time, and the parameters of the jet itself (AFR, pressure, traverse speed).

As a result the optimum value for oscillation angle and frequency varied depending on the jet parameters, but overall it was concluded that an optimal angle of oscillation would lie between 4 and 6 degrees, with higher oscillation frequencies giving better results. An average improvement with oscillation lay on the order of 23% over conventional non-oscillation at the same parameters.

Precision cutting is a task that has a number of complications. In many cases the cuts must follow intricate contours, rather than just making simple linear cuts than separate the material. Increasingly, also, pocket milling has become a valuable ability for this tool. Cut wall quality adequate for final surface finish is increasingly important in this case, and the ability of oscillation to improve that quality and enhance the depth over which a smooth cut was achieved was noted in the work. Similarly the taper of the cut was, on average, reduced 18% with greater improvement at higher oscillation frequencies and angles.

Secondary motions of the nozzle, beyond simple path following, are thus becoming a more important potential tool for the industry, and I will return to this topic again.

Hashish M. “A Modelling study of metal cutting with abrasive waterjets,” Journal of Engineering Materials and Technology, ASME, Vol 106, Jan 1984, pp. 88-100.

Momber A.W., Kovacevic R, Principles of Abrasive Waterjet Machining, Springer Science, p. 209

Shunli Xu Modelling the Cutting Process and Cutting Performance in Abrasive Waterjet Machining, PhD Thesis, Queensland University of Technology, 2005.

Read more!

Tuesday, May 19, 2015

Waterjetting 33b - More on flow inlet conditions

The structure of the jet flow from an orifice makes a tremendous difference to the ability that the jet then has in terms both of its range and its cutting ability. And one of the major factors that control the structure of the jet lies in the flow conditions just upstream of the orifice itself. From time to time, over the decades, we would go out and buy every nozzle that was available for a certain purpose, and run tests between them, trying to find which would, under otherwise similar conditions, provide the best performance. Rarely did the most expensive nozzle give the best result. And the performance of even the best nozzle was also controlled by the flow channel upstream of that nozzle. This, in turn, controlled the condition of the water entering the nozzle. To illustrate the point let me use the example of some tests we made with fan-jet nozzles. In this particular case to objective was to clean large surfaces, but the generalized conclusion also holds true of nozzles of different shapes and jets up to even the highest of pressures used (and we have gone up to 10 million psi).

Simple cleaning nozzles, of the sort that are used in most pressure washers, have historically produced fan-jets that spread in one plane away from the orifice. There are a large variety of these on the market, of varying flow rate and geometry, and it was an initial challenge to find a simple way of relatively ranking the jet quality. Our initial answer, for the first cut, was to take blocks of polystyrene foam and traverse these at a fixed speed under the jet at different pressures and distances from the nozzle. This foam is very easily cut by a jet. So the tests were carried out at 1,000 psi and 2,000 psi, which is the range of pressures of the electrically powered pressure washers found in most hardware stores these days. The difference between two nozzles that were nominally supposed to achieve the same performance was striking:


Figure 1. Comparison of performance between a “better” fan nozzle (top) and a “poor” one (lower sample) when cutting polystyrene packing foam at low pressures.

As you may note at 1,000 psi the poor design was barely able to remove the surface of the polystyrene, rather than cutting deeply into it, as was the case with most of the nozzles tested, and as exemplified in the top cuts.

My point today however, is not the inherent faults in the design of the nozzle shape itself, but rather to highlight the problems that the particular design had, as a result of the way that water was fed into the orifice.

For in this case, unlike many of the conventional nozzles, where the flow is directed directly at the orifice down a channel aligned with the orifice, the nozzle were small discs arrayed along a spray bar, of the type that is used for car and truck washing rigs where a single channel feeds a number of sprays.

The flow in this case is primarily along the distribution manifold, and, as such, perpendicular to the axis of the resulting jets. When the water, therefore, exits from the individual nozzles it retains a component of this lateral velocity, and this tears the jet apart relatively close to the nozzle. The results are evident in the cut made in the lower half of Figure 1.

It is surprisingly easy to remedy this. A short tube inserted behind the nozzle orifice, and protruding up into the manifold channel allows the water some chance to collimate in the direction of flow before it accelerates through the nozzle orifice, and the result, relative to the original cut is quite significantly better.

Not that short lengths of tube are completely effective, but they are a start. One of the more effective means of getting a water jet to move as a cylinder in short jets (such as those seen at Disneyworld and at Detroit Airport is to run the water from the supply pump through a small stabilizing chamber and then pass it into a collimating tube full of drinking straws (or their technical equivalent) which sit just behind the nozzle. Providing the geometries are properly selected you can get the very smooth cylinders of water that are a feature of the jumping streams.

A similar structure lies upstream of the the nozzle at the Gateway Geyser across the river from the Gateway Arch in St. Louis. The fountain shoots a jet of water to the same height (630 ft) as the Gateway Arch on the other side of the river, and to quote Wikipedia:
the Gateway Geyser was designed and constructed by St. Louis–based Hydro Dramatics. It was completed in 1995 at a cost of $4 million. Three 800-horsepower (600 kW) pumps power the fountain, discharging 8,000 U.S. gallons of water per minute (50 L/s) at a speed of 250 feet (76 m) per second. The fountain has an axial thrust of 103,000 pounds-force (460 kN); water is jetted out of the 6-foot (1.8 m)-tall aerated nozzle at a pressure of 550 pounds per square inch (3.8 MPa).
These are more complex flow straighteners than the simpler ones that are used in low pressure cleaning systems, and with considerable effect in controlling the jet flows from the monitors of hydraulic mining equipment. By channeling the water into a multitude (perhaps 200) small diameter channels and then recombining the water at the nozzle the resulting flow is laminar.

Where the water flow is much lower, such as when being used in an ultra-high pressure system, the flow can be stabilized by allowing a long straight run-up of the pipe leading into the nozzle. (Typically the rule of thumb was that the length should be around 125 pipe diameters, however work at the U.S. Bureau of Mines showed that the length of straight section did not need to be this long – a length of around 4-inches proved effective.)


Figure 2. The improvement in jet performance with a straight inlet section (after Kovsec et al*)

A similar improvement can also be seen when the flow conditions are correct when working with higher pressure jets.

As a general rule, however, such care is not taken in the construction and lead-in to the nozzles, and the jet will begin to taper and reduce in effective diameter from the time that it leaves the nozzle.

I’ll talk more about that, next time.

*Kovscek, P.D., Taylor, C.D. and Thimons, E.D., Techniques to Increase Water Pressure for Improved Water-Jet-Assisted Cutting, US Bureau of Mines RI 9201, Report of Investigations, 1988, pp 10.

Read more!

Tuesday, May 12, 2015

Waterjetting 33a - Waterjet structure and its effect

When a waterjet first comes out of an orifice the flow (providing the upstream conditions are properly aligned) will form a cylindrical stream, with the jet pressure across that stream relatively constant. Within about an inch, depending on the flow conditions, ripples start to appear on that smooth cylinder (Rayleigh waves) and these grow and gradually disrupt the jet as it travels further from the nozzle.

Looking at the jet under ordinary light, this makes the jet appear to grow larger, and potentially more powerful as it moves away from the nozzle. However, when the jet is back-lit, or when a pressure profile is taken of the jet at different distances from the nozzle, a different picture emerges.


Figure 1. Pressure profiles across a 6,000 psi jet at 6-inch intervals from the nozzle

What this shows is that the initial even pressure distribution across the stream gradually transforms into a curve very similar to that described as “Gaussian” in mathematical literature. These pressure profiles are generally carried out only at lower jet pressures, because of the way that we have to protect the pressure transducer from the direct jet impact.


Figure 2. Instrumentation for measuring pressure profile.

At higher jet pressures the erosion from the jet on the protective steel cap very rapidly wears the entry hole into the pressure transducer channel, and makes the readings less reliable as the channel shape begins to change. For this reason we have relied on either physical damage to the target, or photographs of the jet, to see what the jet structure had transitioned into, with backlit photographs giving the better set of information.

This damage, interestingly, comes more from droplet impact caused by the breakup of the outside of the jet, which can be easier shown through a front-lit photo.


Figure 3. Microsecond exposure of an ultra-high pressure jet, the orifice is at the 10-inch mark on the rule.

At the nozzle the jet emerges as a smooth cylinder, but small ripples develop on the edge of the jet as it flows. There has been a considerable amount of study of this, and the development of the waves relates to the surface tension in the liquid, and the relative densities of the two fluids (in the case the water in the jet, and the air into which it is injected). At very low pressures (such as water from a tap) surface tension effects dominate and the jet stream is pulled into droplets over a relatively short distance.


Figure 4. Breakup of a low pressure jet into droplets. (Taken from an MIT lecture on fluid jets)

Incidentally, for my male readers, this is why you should stand within 6-inches of the wall of a urinal. (Read to the bottom of the post).

As the jet velocity increases small surface waves develop on the outside of the jet. They look similar to these stationary capillary waves, except that they grow in magnitude as they move away from the nozzle, and were first discussed by Rayleigh, for whom they are named.


Figure 5. Wave generation on the surface of a jet (the grid has a 1-mm spacing for scale). (From an MIT lecture on fluid jets ).

At the jet velocity grows higher – as can be seen by looking closely at the jet in Figure 3, these waves grow large enough to be pulled from the surface of the jet, and, being slowed by the surrounding air, appear to be pulled backwards as the jet flows.

The wave deceleration and break up into a fine mist reduces the size of the central core of the jet, and also the drag reduces the velocity of the outer layers of the jet, giving the pressure profile that is shown in Figure 1.

When cutting with a plain jet (i.e. without polymers or abrasive) a slight deceleration over the edge of the jet is helpful in eroding and removing the material in the target. When a jet impacts close to the nozzle, and the pressure across the target is relatively uniform (as it is 15-cm or 6-inches from the nozzle in Figure 1) then there is no pressure difference between the water that penetrates into the cracks on the target within that central zone. Because of this, while the material may compress, there isn’t enough difference in the forces on the material to cause it to be removed, and the central stub of material will therefore remain in place.

However, in the zone where the pressure differential has developed, on the sides of the jet, there will be a large difference in the pressures in the fluid within the cracks on the target as one moves away from the jet center. This will cause significant material removal. I wrote about this in an earlier post and used the same illustration of what we called a butterfly – which is the erosion pattern that a 10,000 psi jet makes on an aluminum target, where the pressure profile of the jet is still relatively even, close to the nozzle.


Figure 6. Erosion pattern around the point of impact of a waterjet on an aluminum target. The erosion occurs on the outside of the impacting jet, while the central core, being under an even jet pressure, is not removed.

Having digressed a little from the initial topic to explain what happens with the water on the outside of the jet as it is peeled away from the central core, I will return to the topic in the next post, because it is the tapering in of the cut, as the central higher-pressure cutting jet moves away from the nozzle, that is the subject of this short sequence, and so we will return to that topic, next time.

Read more!