Waterjet Technology – Fitting the water jet nozzle to the system

By Dr. David A. Summers, Curators’ Professor at Missouri University of Science & Technology

KMT Waterjet Systems Weekly Waterjet Blog

KMT Waterjet Systems Weekly Waterjet Blog

Buying a high-pressure system requires a significant amount of money, and, as a result, most folk will make a serious attempt at comparing the quality of the different systems that they are considering buying before they make that choice. Most of the expense goes into that part of the system that sits behind the nozzle and which supplies the water and (where needed) the abrasive that form the cutting/cleaning system.

Often, however, while the upstream system is the subject of such scrutiny, the nozzles themselves and the selection of abrasive often escape this level of evaluation. Both of these “parts” of the system are part of the wear cost of operations, and, as a result, the selection of the “best’ nozzle often involves operational cost considerations with less emphasis on comparative evaluations of performance. To explain this most brutally, a company may spend $250,000 on a system but then degrade the performance of that system by over 50% by choosing a nozzle system that saves the company 15% on purchase costs over that of a competitor. (I will show figures in a later post on this topic).

In the next few posts I am going to explain some of the tests that we and others have run to compare nozzle performance and some of the results that we found. I don’t intend to “name names” because the tests that I will talk about are specific to certain specific objectives, and the reason that you are running a system will likely differ from the conditions and the performance parameters that we needed to match for some specific jobs. The evaluations will range over a number of different applications and will cover some quite expensive tests as well as some very simple ones that can be run at little cost in time or money.

But, to begin, the first question relates to how you attach the nozzle to the end of the supply pipe. Here you are, if you have followed the train of thought of the last two posts on conditioning the water as it leaves the supply pipe through a long lead section or through a set of flow conditioning tubes, the water is nicely collimated and (as I will show) could under certain circumstances have a throw distance of perhaps 2,000 jet diameters or so. Yet the average jet has an effective distance of around 125 jet diameters. Why the difference? An illustrative sketch from Bruce Selberg and Clark Barker* simply makes the point.

Comparison between a typical nozzle attachment and one where the flow channel is smoothed

Figure 1. Comparison between a typical nozzle attachment and one where the flow channel is smoothed (Barker and Selberg)

Right up to the point where the small focusing nozzle is attached to the pipe on the left (a) the flow has been conditioned to give a good jet. But then, just as the flow starts to enter into the acceleration cone in the nozzle it hits the little step at the lip of the nozzle where it attaches to the pipe.

As I will mention in a later post, when a jet hits a flat surface and can’t penetrate, then it will flow out laterally along that surface. (This also happens with wind and is why places such as Chicago are referred to as “The Windy City.”) So the outer layer of the jet hits the lip, and where does it go? It runs right into the path of the central flowing jet into the nozzle and mixes right across it. So much for stable flow, that lateral disturbance turns the flow turbulent, so that it is rapidly dissipated once it gets out of the nozzle. Professors Selberg and Barker calculated the theoretical pressure of the jet coming out of the orifices and compared it with pressure values that they measured.

Measured pressure profiles plotted against the theoretical pressure (small crosses) at different distances from a typical conventional nozzle with two orifices

Figure 2. Measured pressure profiles plotted against the theoretical pressure (small crosses) at different distances from a typical conventional nozzle with two orifices.

In comparison, as a way of ensuring that the flow path into the two orifices was smooth, the two authors added a small section made of brass between the end of the pipe and the entrance to the nozzle body ((b) in Figure 1). They inserted two pins to fit into alignment holes drilled into the end of the pipe in the insert and in the nozzle body itself.

Construction of a feed section between the nozzle body and the feed pipe to stabilize the flow

Figure 3. Construction of a feed section between the nozzle body and the feed pipe to stabilize the flow (Barker and Selberg)

When the pressure profiles were taken with one of the new set of nozzles, the difference as a function of distance was quite marked.

Profiles from the nozzle design

Figure 4. Profiles from the nozzle design shown in (b) with a two-part nozzle (Barker and Selberg). Note that the standoff distance has increased for the two sets of profiles over that in Figure 2.

Further, when the depth of cut was measured after the jets were fired into blocks of Berea Sandstone at various distances from the nozzle, the improved performance was clear out to even further distances.

Depths of cut into blocks of Berea sandstone as a function of distance from the nozzle at two flow conditions

Figure 5. Depths of cut into blocks of Berea sandstone as a function of distance from the nozzle at two flow conditions (Barker and Selberg)

The addition of the flow channeling section does make the nozzle a little longer, and the cone angle of the inside of the nozzle was continued out to the diameter of the feed pipe to reduce any steps that might induce turbulence. In addition the inside of both the transition section and the nozzle were polished to a surface finish of better than 6-microinches.

The nozzles themselves were specially constructed for us using electro-formed nickel on flame-polished mandrels and were thus quite expensive. Our particular purpose, however, was in the development of a mining machine that, with the nozzles that we used, was able to peel off a slab of coal to the height of the seam and to a depth of 3 ft at a rate of advance of at least 10 ft/minute. (A later design in Germany went over 6 times as fast, when operated underground).

The advance rate was achievable because the jets were cutting a slot consistently about 2 ft ahead of the machine, and with two jets the coal between them was washed out without having to be mined. But that is a subject for a different post a t some time in the future.

Before I leave the subject, however, some folk might comment that their nozzles sit in holders that are then threaded onto the end of the pipe – thus they should be in alignment, and they are tightened until the holder is tight on the pipe. There are two caveats with this: the first is that this does not necessarily mean that the entry into the nozzle smoothly butts up against the end of the pipe, and in alignment with it. (Hence our use of pins.) In field visits, we have measured for other operators the relative distances involved and found that there can be a gap between the end of the nozzle body and the end of the pipe, both contained within the holder. Even though the two diameters are the same, the presence of the larger chamber before the entry into the nozzle will again create turbulence and a poor jet.

The fix in both cases is a small transition piece, which is simple to design and insert to fill that gap and smooth the passage though it does bring with it the second caveat. You need to make sure that the number of threads of engagement of the holder on the pipe remains enough so that the holder won’t blow off if the nozzle blocks. (One time one of ours did, but it was in a remote location, so thankfully no-one was hurt, although there was some damage as a result).

In the next post I will start to discuss the different ways that we have used, after the nozzle is in place, to make sure that the jets were doing what they were designed to and producing a jet of the quality needed.

* The information that I used in this article can be found, in more detail, in the paper: Barker, C.R. and Selberg, B.P., “Water Jet Nozzle Performance Tests”, paper A1, 4th International Symposium on Jet Cutting Technology, Canterbury, UK, April, 1978.

Waterjet Pumps – Pump Pressure is not Cutting Pressure

By Dr. David A. Summers, Curators’ Professor at Missouri University of Science & Technology

KMT Waterjet Systems Weekly Waterjet Blog

KMT Waterjet Systems Weekly Waterjet Blog

When I began this series, I pointed out that whenever a waterjet is going to be used both the target material and the waterjet delivery system have to be considered if the work is to be done well.

In the last four posts I have tried to emphasize the role of cracks and flaws in the way in which water penetrates into and removes material. It is easier to see this with large-scale operations, such as in the removal of large volumes of soil, but it equally holds true in the abrasive cutting of glass. Now in this next series of KMT Waterjet Blogs, the focus is going to swing back to the ways in which high-pressure waterjets are developed, particularly in the different choices of equipment that can be used.

Because this KMT Waterjet Blog Series is meant to help folk understand how systems work and through that how to improve production and quality, it will tend to shy away from putting a lot of formulae into the presentations. There is a reason that I, an academic, don’t like having students learn equations by rote. It is that it becomes quite possible to misremember them. If you are used to looking them up (particularly true in today’s computer world where formulae can easily be used to generate tables) then you are less likely to mis-remember the exact relationships and to make a possible critical mistake.

But, as I showed here, when the tables of jet flow, horsepower and thrust were generated, there are a few critical equations that need to be born in mind. And the one that underlies the economics of many operations is tied up in the size of the power that is available to do the work. The basic power equation itself is relatively straightforward:

Relationship between hydraulic horsepower, pressure and flow

Figure 1. Relationship between hydraulic horsepower, pressure and flow.

But the calculation gives different values, depending on where the calculation is made in a circuit. To demonstrate this, let us use a very simple drawing of a flow circuit.

Components of a simple flow circuit

Figure 2. The components of a simple flow circuit. Water is drawn from the water tank, through the pump and delivered down a hose to a high-pressure lance, where the water is fed, through a nozzle and aimed at the target, where it does the work.

In the course of this small set of posts the different components that make up this circuit are going to be discussed in turn. But at the end of the first set, I mentioned that in an early comparison of the relative cleaning performance of 10,000 psi waterjets of nominally equal power, and flow (10 gpm IIRC) there was a dramatic difference in the cleaning efficiency, as the Navy reported at the time.

Relative cleaning efficiency in areal percentage cleaned of five competing systems in cleaning heat exchanger tubes in Navy boilers

Figure 3. Relative cleaning efficiency in areal percentage cleaned of five competing systems in cleaning heat exchanger tubes in Navy boilers. (Tursi, T.P. Jr., & Deleece, R.J. Jr, (1975) Development of Very High Pressure Waterjet for Cleaning Naval Boiler Tubes, Naval Ship Engineering Center, Philadelphia Division, Philadelphia, PA., 1975, pp. 18.)

Why such a difference? Consider how the power changes from the time that it first enters the pump motor, and then is converted into power along the line to the target. The numbers that I am going to use may seem extreme, but they actually mirror an early experimental set-up in our laboratory, before we learned better.

A water flow of 10 gallons a minute (gpm) at a pressure of 10,000 pounds per square inch (psi) pressure will contain – using the above equation;

10,000 x 10/1714 = 58.34 horsepower (hp)

But that is the power in the water. Pumps are not 100% efficient, and so there has to be some additional power put into the pump to allow for the relative efficiency of the pump itself. For the sake of illustration let us say that the pump converts the energy at 90% efficiency. Thus the power that is supplied to the drive shaft of the pump will need to be:

58.34/0.9 = 64.8 hp

But that is still not the power that we have to supply, since that power – usually – comes from an electric power cord that feeds into a motor, which then, in turn, drives the pump shaft. That motor itself is also not 100% efficient. Let us, for the sake of discussion, say that it is 92.6% efficient. Then the electrical power supplied will be:

64.8/0.926 = 70 hp

Now, as the calculation progresses, remember that this is the power that is being paid for. And so, in the first part of the flow, the power is transformed from electric power to water power, but at the pump.

The change in power from that input to the motor to that coming out of the pump

Figure 4. The change in power from that input to the motor to that coming out of the pump.

The water coming out of the pump then flows through either a length of pipe, or high-pressure tubing until it comes to the tool that holds the nozzle. There are a number of different factors that change the flow conditions to the point that it leaves the nozzle. The most critical, and often overlooked, is the size of the hose/tubing that carries the water. Particularly as pumps get larger and more powerful and the flow rates increase, it is important to ensure that the passage for the water is large enough so that it does not require too much pressure to overcome the friction acting against that flow. I have, myself, put an additional 10-ft length of tubing on a drilling lance, and seen the cutting pressure coming out the end fall from that which drilled a rock at 12 ft/minute to where it could not drill at all. (The pressure drop was around 200 psi per foot). I mentioned in that earlier post that a competitor, running at a pump pressure of 45,000 psi was losing 35,000 psi of that pressure just to overcome friction in pushing the water down through a tube that was too narrow. As a result the water coming out of the nozzle had barely enough pressure (10,000 psi) to cut into the rock.

At the same time, very few people pay a lot of attention to how their nozzle fits on the end of the feed line, or how well it is made. Think of this – you have just spent $200,000 on a system, and yet, because the nozzle is a disposable part, you look around for the cheapest source you can get. You don’t size it for a good fluid fit, nor do you check how well it is machined. And yet the entire performance of your system is controlled by that small item. The difference between a very good nozzle and a standard nozzle can give as much as a factor of 10 improvement on performance – but who checks. The one you use saved you $15 relative to what you would have paid if you had bought the competing product, what a bargain – right?

There are different ways in which pumps operate and produce the high-pressure flow. With a fixed size of orifice in the nozzle and with a given pressure drop along the feed line, the pressure at the nozzle will be correspondingly reduced. So that if, for example, we use a 0.063 inch diameter nozzle then the chart you developed after generating the table will show that this will carry a flow of 9.84 gpm at 10,000 psi. But let us suppose that the hose loses 20 psi per foot of length, and that the hose is 200 ft long, then the pressure drop along the hose will be 20 x 200 = 4,000 psi.

Thus the pressure of the water coming out of the hose will be only 6,000 psi. And at an orifice of 0.063 inches, the flow through the orifice will now only be 7.62 gpm. (The way in which the pressure is controlled is assumed to be through bypassing extra flow back to the reservoir through a bleed-off circuit).

Now the pump is still putting out 10 gpm at 10,000 psi, but now the flow out of the nozzle is only 7.62 gpm at 6,000 psi. The power in this jet is (7.62 x 6,000/1714) only 26.7 hp. This is only 38% of the energy going into the pump.

The power losses to the nozzle

Figure 5. The power losses to the nozzle.

Unfortunately this is not the end of the losses. Particularly in cleaning operations there is a tendency for the operator to hold the nozzle at a comfortable distance from the target, so that the effect can be seen. But, as I will show in later posts, the jet pressure can fall rapidly as stand-off distance increases, particularly with a poor nozzle. A good range for a normal nozzle in a cleaning operation is about 125 nozzle diameters. So that at a diameter of 0.063 inches this range is less than 8 inches. Many people hold the nozzle at least a foot from the target.

If the nozzle is held about that far from the target the pressure will have fallen by perhaps 65%. The water thus reaches the target at around 2,000 psi. The flow rate is 7.62 gpm, and the actual horsepower of the water doing the work is 8.89 hp. This is 12.7% of the power that is being paid for through the meter. And the unfortunate problem is that no-one can tell, just by looking at the jet, what the pressure and flow rates are. So that often these losses go undetected, and folk merely complain about how the target material is more resistant today, not recognizing that they are throwing away 87% of the power that they are paying for.

Power losses from the power cord to the target

Figure 6. Power losses from the power cord to the target.

One of the objects of this series is to help reduce these losses, by avoiding those mistakes that those of us who started in the industry some 40-odd years ago made all the time.