Waterjetting 22d - more on shrouds

The use of a shroud to capture the water and debris from waterjet use, feeding it to an exhaust hose, that will then carry it away from the site, has become more universally applied over the last ten years.

There are, however, different ways in which this new combination (which has been given different names depending on the usage) can be applied, and how the components can be best combined for most effective application.

At the low end of the pressure range, feeding a 2,000 psi waterjet at 2 gpm into the soil at the entrance to a suction hose has created a powerful new tool for deep soil excavation. The technique, known as hydro-excavation, has a variety of different applications – one of the simpler demonstrations was shown by Hydro Spy here on Youtube.

The demonstration lasts some five minutes, and helps show why there is still improvement needed in equipment design, since the vacuum intake is not extracting material at a constant rate, but is only being fed by a hand-held lance that it often not cutting very efficiently, while at the same time the head is either buried in debris or being held too high off the surface to effectively capture the loose material effectively.

Figure 1. Components of a hydro-excavation system, the jet is breaking the soil into pieces that are they removed from the hole with the water, through the suction line. (Hydro Spy )

Because the two actions (jet fragmentation and vacuum removal) are separate they are both working at much less efficiency than if the situation were modified. For example, in the video, the lance is used to pry pieces of soil from the wall and too much time is spent with the lance away from the suction line or with the intake to the line buried and not removing material.

Figure 2. Frame from video showing lance being used to pry soil block from the solid – the suction line is not getting any water or soil at this time.

It is often forgotten that, in soils, the jet penetrates to maximum depth in about one hundredth of a second. Thus, to be effective it has to be moved over the surface relatively fast ( as it is in parts of the video) in order to be most efficient. This is, however, most often best achieved by driving the head mechanically, rather than relying on an operator to move the nozzle as fast as it should be moved. In a simple situation such as this it may be, for example, much more effective to use a dual-jet self-rotating nozzle assembly (which can be obtained from one of several equipment manufacturers) since these designs spin the jets over the surface more rapidly and consistently, so that the material is more effectively broken into relatively small pieces.

However in this section we have been discussing the use of shrouds and intakes to the suction line, and this design becomes of equal importance in ensuring that the system works at its most efficient. If the entry to the suction line in blocked because it has been run up against the bottom of the hole, or into a tight cluster of large pieces of material, then there is no production, until the head is lifted away from that seal. (Or if a short rod is attached to the bottom of the inlet to ensure that there is always a gap between the lip of the line and the bottom of the hole).

On the other hand if the inlet is lifted too far away from the surface, say more than half-an-inch, then the suction force pulling the pieces into the line becomes significantly less effective and production will again suffer. This is made worse where the floor of the opening is very uneven, since this makes it more difficult to maintain the gap at which the suction is most effective.

It becomes more effective – whether removing soil in this way or removing paint from a ship hull at much higher pressures – to integrate the jet action with the design of the shroud/inlet to the suction line. The two cases are otherwise different in that in the softer material the jets are cutting quite deeply (though hopefully no more than about half-an-inch at a time) into the soil, which causes the jet to rebound back up into the shroud body and makes water and debris collection relatively easy.

This is not the case with the removal of paint and coatings, where the layers are often relatively thin, and the jet will rebound, often parallel with the underlying steel that it does not have the power to penetrate. (Nor is this desirable, other than for the jet to penetrate into any corrosion pits in the surface and clean them).

With thin coating removal, since the surface is otherwise relatively smooth, the shroud can be mounted on wheels that allow the operator to set the gap thickness between the shroud and the surface. (The closer the shroud lip to the surface, the higher the force that holds the shroud to that surface, but also the higher the force that the motors must apply to move the shroud against the friction forces that are created). The suction force in this case will hold the shroud against vertical walls and even against the underside of ships hulls, bridge decks, etc. provided that the geometry of the head is optimized to provide that balance of enough suction to hold the head, without it getting too high for the trouble the traversing motors).

There is one other, final thought, in those cases where the jets are cutting into and along paint and other coatings. In some cases the coating can be best removed where the jet is attacking along the surface, rather than almost perpendicular to it, as is quite often the case in head designs. This can give a better and more efficient surface cleaning, but if the jets are at too great an angle to the surface, the operator runs the risk of seeing the jets carry the debris out past the edge of the shroud, making it much more difficult to capture and remove.

One way of getting around this problem is to incline the jet path within the shroud, so that at the distant end of the jet path within the shroud it intersects the path of the next jet around the design, which has sufficient force to stop the jet moving further out. We have successfully demonstrated that this does work in an application, where the jet was cutting relatively shallow grooves in the surface, and with greater penetration the jet will rebound upwards out of the slot, and more easily captured by the overlying shroud.

Figure 3. Showing how, by aiming the jet path into that of the next stream around the shroud the energy of the jets can be contained within the shroud envelope and the splashing outside of that envelope is much reduced.

Waterjetting 22c - More on shroud design

Lowering the pressure in a hose connected to a cutting head, by connecting it to a vacuum pump, will pull a certain amount of the water and debris released from a cutting/cleaning event into that hose. However to ensure that all of this material is captured, rather than just a fraction, requires a little more care and effort in the system design.

At the end of the last post on this subject, I began discussion of the use of shrouds to help to contain the ejecta and to direct it towards the suction line.

Figure 1. Schematic section through a shroud of a device, designed to mine high-level radioactive waste.

There are a number of different lessons that we learned as we developed this tool, and this piece will discuss a number of them. During the development and demonstration of the device we had to use a simulant, and a relatively weak cement was chosen, which would allow us to design the tool to operate it where we could see it, and easily interact with it.

Figure 2. Cutting test under way (no shrouds were used in this early test series)

The easiest way to drive the nozzle system was to run the high pressure tubing through a fixture that contained a hollow shaft electrical motor. This saved a lot of space, and allowed the high pressure tubing to feed into a distribution manifold under the motor, which fed the high-pressure water to a set of rotating nozzles.

Figure 3. Test rig without jets to show the design of the test head.

In figure 2 the jets issue from two self-rotating nozzle sets, themselves fed through a rotating feed tube, itself rotating around a central axis, and driven through a belt drive and gearing.

Various different patterns were cut into the simulant, as the different heads were moved over the surface, with the pattern controlled by the different rotation speeds relative to the overall head speed over the surface.

Figure 4. Computer image of the jet paths over the surface, in one combination of parameters for a head similar to that shown in Figure 3.

The design of the head was aimed at producing a set of jet passes (given that each jet was slightly inclined to the surface) which would produce pieces of simulant that were never larger than half-an-inch in size. Yet at the same time the goal was to remove 4 cu. ft. of waste each minute. The larger we could break out the particles, the less cutting we would have to make into the waste itself, saving energy and time, while at the same time increasing the overall volume we could release in that time.

Figure 5. Deeper cut into the simulant.

At the same time if the cut depth was too great, then several new problems would arise, apart from the initially obvious one of producing particles that would be bigger than the suction hose could easily handle. (Though we overcame that hurdle by running the particles through a high-pressure jet pump that effectively cut any oversize particles down to an acceptable size as part of its design).

The suction line needed more than just the water from the cut, to be able to pick up all the debris from the cutting operation. Air had to be drawn in around the sides of the shroud, yet at the same time the walls of the shroud had to come down to restrict the amount of that air and keep the suction strong enough at the surface to remove all loose material. This is done by fitting a rim of bristles (such as form the head of a paint brush) around the edge of the shroud that come down to brush over the outer edge of the cut, stopping a lot of the material from escaping out from the edge, while limiting the amount of air that feeds into the shroud, and in this way holding the suction pressure inside the shroud.

Figure 6. Early test showing a square shroud with bristles around the edge as it cuts into the waste. (Part of a previous pass has been filled with clay as part of the test). The shroud was larger to ensure that all ejecta was captured – as shown.

During the tests we found that the metal rim should, optimally, be no more than half-an inch from the surface of the material, after it had been cut, to pull all the material from the bottom of the crevices. But the edge of the head has also to pass over the surface in successive passes. So that high points left by deep cutting (Figure 5) will catch on the head, and can interfere with the rotation of the head on the next pass.

The aim of the cutting head design was, therefore, to leave a relatively smooth surface (of the sort shown in figure 3) over the waste after each pass, so that the head could be fed automatically down a fixed amount without any risk of it catching on large peaks left by the previous cut. This risk could also be lowered a little by slightly tilting the head backwards as it moves over the surface, since this allows slightly larger points to enter the head, where they are attacked by the jets before the driving mechanism has to pass over them. This tilting also makes it easier for the head to clean right up to the walls of the tank, where otherwise the edge of the shroud would hit the wall and stop the jets from removing that last rind of material from the edge. (Though it could be cleaned by a subsequent pass with the head turned up parallel to the wall and moved over it in that way – though this wouldn’t capture all the material as easily, due to wall curvature.)

Tilting the jets at a high angle so as to cut material at the edge of the shroud was also a possible problem, since it made it easier for the water to escape from the edge of the shroud, and out into the main body of the tank, which was undesirable. But I’ll talk about that in a later piece. Let me just note that, when these factors were all combined no material escaped from the edge of the shroud.

Figure 7. Test late in development, where a head similar to that shown in figure 1 is cutting over waste, without any material being ejected from around the shroud edges.

Waterjetting 22b - Steep seams and shrouds

It was not until we started to move the auger that I described last time, that I realized how heavy and cumbersome it remained. It is true that we could lighten it considerably (and we did in the UK version), but it was still largely a platform mounted device that was relatively easy to move on the surface, but which would be much more difficult to move around in the confined spaces underground.

In this regard it is worth comparing two photographs.

Figure 1. Coal being transported in the pipes at the Hansa hydraulic mine in Germany.

Notice in this first one how the coal is confined, there is no dust, and the tunnel is relatively clean and clear. Contrast this with the typical mining operation where the coal is carried from the working face to at least the main haulage drifts, and often all the way out of the mine using conveyor belts. (The method of transport that would also likely be used in a typical underground auger section).

Figure 2. Conventional belt conveyor carrying coal underground. (Famur )

Note that in the second photo the belt occupies most of the space in the roadway making passage more difficult, and that the coal is openly exposed. The problem that this occasionally generates is that the coal transfers from one belt to another as it moves through the mine by falling off the end of one belt onto the next. This puts dust into the air, and if not properly maintained coal dust can accumulate under the belt and around the support rollers and drives. If not cleaned this can create heat through friction, and can lead to disastrous fires.

Figure 3. Studying a belt fire underground (Office of Mine Safety and Health Research )

Confining the coal, and using the water that mined it as a transport fluid – or at least part of such – has many advantages.

The Russians were aware of this as they developed some of their different mining machines. One of these was a small monitor that could be pushed up a seam, from a lower drift (without the need of a higher one), by adding segments as the unit was jacked forward.

Figure 4. Russian GVD monitor

The monitor could be advanced up the seam, cutting a channel about 3-ft wide, and to the height of the coal. This relatively narrow channel confined the water and the coal produced, so that both ran back down to the drift, where they could be either enclosed in a pipe, or run into an open flume, that would carry the coal away. Once the drive had reached the end of the section, then the two hydraulic cylinders that sat under the monitor could be turned, so that, from a protected section down-slope, the monitor could successively ream out strips on either side of the entry, until the roof collapsed, or the operation holed through to the previous panel.

Figure 5. Schematic showing the sequence of extraction in a) a coal seam with a relatively strong roof b) a seam with a weak roof, where a pillar of coal is left between successive lifts, typically around 30 ft, so as to provide additional support to the roof as the coal is mined out.

Production from these machines averaged about 76 tons/hour from seams that were in the range from 2 to 4 ft thick, this was more than double the production output achieved by more conventional means, with significantly less manpower.

However while there are some conditions where gravity helps to bring the coal and water back together, there are many cases where this is not possible due to other constraints. This is where it becomes necessary to use a shroud to confine the jets, debris and water so that they can be extracted together, often using a vacuum to assist in the process.

Figure 6. Rendition of the combination of three cutting jets rotating around a vacuum tube to slice into material and remove it.

The particular device shown in Figure 6 was developed as a way of remotely slicing into high-level radioactive waste for the Pacific Northwest National Laboratory. The waste is held in underground storage tanks, where the levels are too high for human access. As a result the waste had to be broken into pieces and removed remotely.

The initial problem was that the tanks, though holding perhaps half-a-million gallons of waste, had only small (18-inches or less) ports through which they could be accessed. Thus a relatively small tool was required, yet the mining rate had to exceed 4-cu. ft/minute. The situation required an excavation system on the end of a robotically controlled arm. But the arm would have to be more than 60-ft long (in the end a cousin of the arm used on the Space Shuttle was used). Any mechanical force applied at the end of such an arm would have tremendous leverage on the holding fixture, and a relatively low overall force would have to be found, yet one capable of cutting material perhaps as strong as a weak cement.

The answer came in the form of the device shown in Figure 6. By placing three cutting jets to rotate around a central tube, connected to a vacuum line, one could cut into the material and break out relatively small pieces, which could be aspirated away with the cutting water. (The requirements were that there would be no water left in the tank for any significant amount of time). Because the bottom of the tank was some 40 or 50 ft below ground level a small, high-pressure (10,000 psi) jet pump was developed by Michael Mann, capable of drawing particles of up to an inch in size into the line, and then projecting them up with sufficient energy to carry them out of the tank.

Figure 7. Basic system conceived to mine high-level radioactive waste and remove it from a storage tank.

While this description of the system makes the tool seem to be relatively simple to build and operate there are a number of features to the design that are fairly critical in order for it to work well, and I will cover some of those next time.

Waterjetting 8d - Choosing angles

How times change! I was reading a column in the British Farmer’s Weekly, and came upon this, where the author is discussing the need for a generator.:

It will also be vital to keep the fuel flowing into the tractors, and power the pressure washer, and light the security lights, and all the other essentials of an average arable farm.

It is an indication of how far the use of pressurized water has come, that it is now seen, at the lower end of its application, as a vital farming tool. Which is a good introduction to talk a little further about the use of cleaning streams, and how to interact with differing target materials.

There was an initial first step, when someone would send the lab a mystery block of material and asked – how do I cut it? Generally the samples were small, but we would find a flat surface on the material, and carefully point a jet nozzle perpendicular to this surface. (In the early stages this was hand-held). When a jet strikes a surface, but can’t penetrate it, then it will flow out laterally around the impact point, under the driving force of the following water.

The test began with the jet at low pressure, and this was slowly raised, until the point was reached when the pressure was high enough to just start cutting into the material. At this point the jet had made a small hole in the target, and so the water flowing into that hole had to get out of the way of the water following. The sides of the hole stop it flowing laterally, and so it now shoots back along the original jet path. This spray can hit the lance operator if the nozzle is hand-held, but it is a fairly graphic way of determining the threshold pressure at which the material starts to cut. (and I’ll get into what happens as the pressure continues to go up in a future series of posts).

But for the purpose of cleaning, the jet has to move over the surface, once it has made that initial hole, at pressure. But, in many materials, if the jet comes vertically down onto the target, then only the material directly under the jet will be removed. And so the jet has to be played on every square inch of the surface in order to ensure that it is cleaned, or that the coating/layer is removed. In some sandstones, for example, two jet paths could be laid down, almost touching one another, and yet the rib of material between them would remain standing.


Figure 1. Adjacent jet passes in sandstone, the cuts are about an inch deep, but note that even though the narrowest rib is about 1/8th of an inch wide, it is only when the cuts touch that the intervening material is removed.

Yet that rib of material was, in that case, so weak that it was easy to break it off with a finger. (This turns out to be a weakness in making delicate sculptures out of rock). To use the full pressure of the water can be a waste of energy, if the material is very thick, since it all must be eroded with such a direct attack.

Yet the minimum amount of material that needs to be removed is that that attaches the layer to the underlying material (the substrate concrete, steel etc) and this can be quite thin. Thus, in attacking a softer material, particularly one that can be cut with a fan jet, a shallow angle directed at the edge of the substrate can be more effective.


Figure 2. Round v fan cleaning from Hughes (2nd US Waterjet Conference)

Because there is a balance between cutting down through the material to be removed, and cutting along the edge to grow the separation crack between the materials, some practice is needed to find, for a given condition, what that angle would be.


Figure 3. Choice of angle from Hughes (2nd Waterjet Conference)

The more brittle the material, then the greater the angle to the surface, since rather than just erode the material, the jet may also shatter the layer into fragments that extend beyond the cut path. But otherwise using an angled jet to the surface can be more effective. Hughes (from whose paper at the 2nd Waterjet Conference I took these illustrations) has a simple test for orifice choice.


Figure 4. How target response influences nozzle selection. (Hughes 2nd Waterjet Conference)

Some of the more advanced cutting heads use a series of nozzles that spin within an outer protective cover, as they remove anything from layers of damaged concrete to thin layers of paint from ship hulls. Increasingly these are connected to vacuum systems that will draw away the spent water and debris from within the contained space, without it entering the work space, and creating problems for the worker.

In order to reduce any collateral damage to the surroundings these jets are often made very small (thousandths of an inch in diameter) so that their range is short, and they are inclined outward to cut to the edges of the confining shield.

We have had some success in turning those angles the other way, so that they cut into the shield, rather than away from the center, and also so that each jet is directed towards the path of the next jet around the circumference. The intent in this case is to allow the use of a slightly larger jet, with a greater cutting range. In this case the individual cleaning/cutting path is a little larger, but because the jet at then end of the cut moves into the range of the adjacent jet, then any remaining energy that it and the dislodged debris still have, will not be enough to get through this second jet.


Figure 5. Inclined jet and shroud design.

The action of each jet then becomes not only to cut into and remove material, but also to contain the spent material from the other jets dispersed around the cutting arm, and to hold the debris in the center of the confinement for the very short time needed for it to be caught up in the vacuum line.

In all cases the choice of pressure, nozzle size, and operational factors such as angle of attack, come down to the target materials, those that have to be removed, and those that need to be left undamaged. And it is why it is useful, at the start of any new job, to take the time to do a little testing first, to make sure that the right choices of nozzle and angle have been made to get the job done quickly and efficiently.

Incidentally the idea behind the test of effective pressure, that the jet flows laterally when it hits something it can’t cut, can help, for example in easing the meat from the bone when a jet cuts a deer leg.


Figure 6. Cut across a deer leg, note how the jet has cleaned off the meat from the bone, undercutting the flesh.