Waterjetting 5b - hose connections and nozzle feed

There is often not a lot of choice, when laying out a project, over the path or distance that water must travel from the time it leaves the initial pump to the point where it reaches the nozzle and is usefully applied. Yet, over this length, if the right choices are made, some considerable improvement in performance might be achieved.

As discussed earlier, the simplest improvement is often just to increase the size of the delivery line, although there have been occasions where it was cheaper for us to use a double line rather than a larger single, in order to lower the friction loss to the nozzle, and keep operating pressure in the range that we needed.

There are other concerns with the layout of the feed line to the nozzle, If it is a hose, then any connections between sections should be whip-checked, so that should a coupling fail (which has been known to happen) then the released sections of hose will not whip around and cause injuries or other damage.

Figure 1. Hose that separated from a coupling, while under pressure.

The major risk comes at the moment of separation, while the pressure in the line is higher, before it drops under the larger area through which the water can now flow. To stop the whipping of the hose ends, the two should be restrained by attaching a cable with two loops that fit over either hose end making the connection.

Figure 2. Hose connection covered with a pressure dissipating sleeve (the blue cover) and with a whip connected to the hose ends on either side of the connection.

Apart from this, and remembering the possible risk about avoiding chaffing points and that, over time, high pressure tubing does fatigue (after many years of operation most of the pipe segments connecting between our ultra-high pressure pump and cutting table failed in a relatively short time period – but they all came at the same time, and were installed together and saw the same loading cycles). (Which emphasizes that, over time, the pressure rating of the tubing should be reduced).

Connections, T-joints and other fittings that are used in the feed line should be sized appropriately. Any time that the diameter of the flow channel changes, then there is a cost in terms of the delivered pressure. This is best checked with the manufacturer to ensure that this is accurately assessed in the flow and pressure calculations.

Moving down the line, this brings us to the end of the feed line, and the entrance to the nozzle. In later posts I will cover different pieces of equipment that can be used, for a variety of different tasks in manipulating the nozzle, but, for now, I would like to consider just the flow from the feed line into the nozzle (without discussing nozzle shape at this time).

Of all the systems I have examined, this is the one point in the assembly of a feed system that was most commonly ignored or badly constructed.

All of us, from time to time, get caught up in traffic flows through road construction. When the lanes are controlled, and traffic feeds are properly directed, it is possible to get through these relatively quickly. But in most cases that is not what happens. There are always drivers who do not ease into the required lanes soon enough, but rather drive rapidly as far as they can and then force their way into the remaining lanes, thereby breaking the steady flow into a process of stop-start-stop-start. The process becomes much less efficient, and instead of the traffic moving at a steady, but slow rate it often can take over half-an-hour or more to get through the restriction.

So it is with water flow down a pipe. If the flow can remain in a laminar mode, with the flow channel slowly constricted to speed the water up to the required velocity, then the resulting flow into and through the nozzle can give jet streams that can throw over 2,000 jet diameters. Instead in most cases the throw is about 125 jet diameters, and I will discuss how to find that distance in a post or two before long.

In this particular post the water has not reached the nozzle yet, and it could still be in a poor condition. The best way to explain why is to use a comparative graph showing jet pressure measurements after the jet has passed through the nozzle, to show how the structure is affected. The work was carried out by the Bureau of Mines (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) and the only difference between the two different plots is that in the upper one a 4-inch straight length of pipe was connected just upstream of the nozzle to allow the flow to stabilize before it entered the nozzle.

Figure 3. The effect of flow conditioning the water, prior to passage into the nozzle.(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)

There is a caveat to this plot, and this is the assumption that the internal diameter of the feed pipe and the entrance diameter to the nozzle are the same size. In virtually every system that I have examined in the field this has not been true, and as I will show, in a later post, the difference that this can make is very large. For the above example the length is 4-inches, but this was for a specific nozzle size, and the more general condition is that the length should be in the range of 100 – 125 pipe diameters.

By the same token, if the nozzle does not make up with the end of the feed line, so that there is a little eddy pocket created, the ensuing jet will be of poor quality.

Which brings us to the final part of this post, because there are many situations where it is not possible, because of space restrictions, to fit that particular length of pipe just before the nozzle. The most glaring example of this that we have had to deal with is where high-pressure waterjets are fed down a borehole, and then used to drill lateral excavations out from the bottom.

But if the borehole is say 10-inches in diameter, and the jet is an inch in diameter, because it is being used for mining out valuable pockets of uranium, then it is not possible to get the required straight section. Again the pioneering work on this was carried out, first in Russia, and then by the U.S. Bureau of Mines, under George Savanick. By placing a shorter length of a flow straightening device within the flow path, just before the nozzle, the flow can be straightened over a much shorter distance, and this will be the topic of the next post. And when George did this he was able to cut cavities that extended more than 30-ft from the borehole (with some unexpected consequences - but we'll cover that later - grin).

Waterjetting 4b - Line losses

High-pressure pumps are, as a general rule, quite efficient at bringing water up to the pressure required for a given task. And yet, time after time, the jet that reaches the target is no longer capable of achieving the work that was promised when the system was designed. More often than not this drop in performance can be traced to the way that the water travels through the delivery system, and out of the nozzle that forms the jet.

The water flows that are used in a broad range of operations are quite low. Ten gallons a minute (gpm) and flows below that volume are mainly used in cutting operations and higher-pressure cleaning. Further, there are few occasions where hand-held operations will use flows much above 20 gpm, because of the thrust levels involved. And low flow rates mean that there is little pressure loss between the pump and the nozzle, right? UM! Well not exactly.

The pressure losses due to overcoming friction in the feed lines (whether hose or tubing) from the pump to the nozzle can make a significant difference in the operation of the system, as I mentioned in one of the early posts of this series. In that post I pointed out that a well-known research team (not us) spent two weeks running a system with 45,000 psi water pressure going into a feed line, but with only around 10,000 psi being usefully available when the flow reached the far end. (And I will freely confess later in this piece to having made a similar mistake myself). So the question naturally arises as to how these losses can be avoided.

In a word – diameter! The smaller the diameter of the feed line through which the water must flow, then the higher the pressure that is required to drive the water through that line, regardless of the nozzle size at the delivery end. The diameter of concern is, further, the inner diameter of the hose or tubing, not the outer diameter (though the combination is important in ensuring that the line can contain the pressure that the water is carrying through the line).

There are concerns over the condition of the line, the fittings that join the different parts together and other factors that I will cover in the posts following this one, but this will deal just with the simple pressure drop that occurs along a tube at different flow volumes. There are formulae that can be used, but a reasonable estimate of the loss can be obtained, either with the design tables that most manufacturers supply with their product, or through a simple nomogram that I will place at the end of this piece.

To begin with consider the basic equations that govern the pressure drop:

Figure 1. The equation relating pressure drop to flow volume and pipe diameter.

Note that in the above equation the pressure drop is related to the fifth power of the diameter of the tube – such is the power that even a small change in flow channel diameter will have on the pressure drop in the line.

When flow begins through a channel it is initially going to occur with the flow being laminar, in other words the water moves in layers. (There is an interesting video of this here and a video of one of the designs used, for example, to give the “solid” jet slugs that you might see jumping around the hedges at one of the Amusement Parks.

Figure 2. The difference between laminar and turbulent flow. (Equipment explained)

As water speed increases, however, the flow will transition from laminar flow into turbulent flow, where the roughness of the flow channel wall becomes more important. The roughness, resulting friction factor and the flow volume all then combine to allow the calculation of the pressure required to overcome the friction in the pipe. This holds true whether the flow is at the one or two gpm used in cutting at high pressure, or the relatively low pressure, high volume flows used in fighting fires.

But (outside of us academics) few actually calculate the numbers. There really is no need, since most of the manufacturers provide the information in their catalogs. There are two ways of presenting the information. The older convention was just to provide a graph, from which one could read off the pressure drop, as a function of the pipe internal diameter, and for a given pipe length.

Figure 3. Pressure drop along a tube, as a function of flow rate and tube internal diameter. Note that the scales are logarithmic.

Charts such as this are a little difficult to read, and being on a log plot small mistakes in reading the value can give significantly wrong estimates so that a more spread-out method is often more helpful. The one that I prefer to use is a nomogram, where it is possible to do comparisons between different options on a single figure with a slightly expanded scale.

Consider, for example, this nomogram from the Parker Catalog which shows the relationship between the volume flowing down through a line, the inner diameter through which it is flowing, and the resulting velocity of the flow.

Figure 4. A nomogram to determine the best pipe diameter, based on the allowable velocity of the flow. (Parker)

While this is not generally a concern in feed lines to nozzles (because of the high levels of filtration of the water) in lines that carry away spent water and debris the velocity can be of concern, and also in abrasive slurry systems, where flow rates above 40 ft/sec can lead to erosion of the line.

The more useful nomogram, however, is one that I have adapted from the U.S. Bureau of Mines (a Government agencies that is now, sadly, defunct).

Figure 5. Nomogram to calculate pressure loss along a 10-ft length of tubing.

Knowing the flow rate through the line, and setting a straight-edge (usually a ruler) to mark the level, the ruler is then positioned so that it also crosses the inner diameter of the tubing. In the example above that would align the ruler along the line shown, that runs from 20 gpm to 0.1875 inch pipe diameter (3/16ths of an inch). The point at which the line crosses the pressure drop gives the friction loss in the line. In this case that reads at 3,600 psi per 10 ft of pipe.

The example was taken from a field trial where we were drilling holes into the side of a rock pillar. We had no problem drilling the first ten feet, but when we added a second length of 10-ft tubing to allow us to drill holes 20-ft deep the drill did not work. It was not until late in the afternoon that we realized that by adding that second length of pipe we had dropped the cutting pressure coming out of the nozzle so that while the gage pressure was 10,000 psi, the initial jet pressure had been only 6,400 psi and when the second pipe length was added, the pressure fell to 2,800 psi. This was below the pressure at which it was possible to effectively cut the rock. And so we learned!