Help me understand the limitations of a jet pump in a planing hull.

Discussion in 'Jet Drives' started by shaka, Dec 29, 2009.

  1. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    Ok Shaka, attached you find two basic sets of diagrammes, describing the wj operating environment. You have to understand these in order to find the means to avoid "hitting the wall head-on".

    The fast pwc opens a different perspective on jet propulsion, since at low to medium speeds the request for maximum acceleration is demanding maximum flow at low inlet pressures, causing cavitation. At the other end of the velocity spectrum, high nozzle velocities are required for speed. With the configuration you have presented, I find a dimensioning flow of abt 0.2 m3/s.

    At top speed (~32 m/s) this flow is collected from a "streampipe" area of 0.00625 m2, which is the dimensioning throat area for high speed. This area is the required duct area just at the rear of the bottom opening (the "lip" or the front edge of Jim's "shoe"). However, the effective impeller inlet area (just at the blade leading edges) is 0.018 m2, ie nearly three times as big. The flow would need a retardation down to some 11 m/s to keep the impeller filled, which is impossible with the configuration at hand.

    Instead, the horizontal vane ("splitter plate") is used to "peel off" a layer of oncoming fluid. If the transverse width of the hull inlet is 0.12 m, the layer will have a nominal thickness of only 52 mm at top speed! The splitter vane was originally introduced in order to assist in deflecting the flow and give a complete filling of the inlet duct at medium speeds. For high speed applications it creates a "dual mode" inlet instead; smart thinking!!

    But in order to block the rear part of the throat area at high speed, its leading edge must "provoke" a flow detachment from its lower side, which comes almost naturally as a result from leading edge cavitation at high speed. The shape of the leading edge is one of the factors that may influence the jet behaviour. Fine tuning of the inlet flow is done with the "blocking plate" inserted from the ramp top. This also generates a detachment from the roof, so that the incoming flow will be a free jet with much the same velocity as the advance velocity.

    If the pitch notation (14/20) has a reasonable resemblance to local pitch, the local inlet tip angle would be something like 35 degrees. With a tip speed of 58 m/s and the inlet flow ~32 m/s, the relative speed is approaching the impeller with an angle of 29 degrees. The blade leading edge will thus see a flow coming in with an angle of attack of 6 degrees, which is reasonable, considering the guesses made here.

    Now take a look at the right diagram in the attachment. There you see the pump characteristic; basic red curve, showing pump head and pump efficiency (blue) as a function of flow. When your craft is moving forwards, some of the energy in the flow into the inlet is lost due to friction and turbulence; the remainder is adding to the pressure delivered by the pump. In the diagram, this is demonstrated by "lifting" the pump curve accordingly (example: +20 m from 26 m/s and 60 % inlet efficiency). Consequently, the pressure forcing high speed water through the exit nozzle is increased.

    The head required to force a flow through the combined resistance from duct and nozzle is shown by the green line; the final operating point is found where delivery and requements meet. Your high speed dilemma is, that there is an intense, and varying aeration of the incoming flow, and all rotodynamic pumps react to gas mixture with a reduced performance, shown by the thin red "hooked" lines.

    The crash incidents may be trigged by multiple mechanisms, either separately or in combination:

    A/ If a manoeuver produces a critical aeration volume (see notes in diagram), the pump is choked and the "operating point" will follow the duct curve down during the deceleration phase. In practice, this has the effect of projecting the full inlet area to the incoming flow, with very little water passing the impeller. The braking effekt of this sequence will cause a 2 g retardation, which is "hitting the wall". Diagonal pumps are worse in this respect than the pure axial flow type, since they have more fluid "locked" in rotation between impeller and stator.

    B/ If the craft is meeting a wave in nose-up position, the free detached streamline from the splitter vane leading edge may be diverged upwards, thereby opening for additional flow into the rear part of the inlet. This means forcing three times the normal flow through the impeller and nozzle, which is impossible; again you stomp on the brake.

    Both the events above are probably made worse by a rapid rpm reduction, when the driver becomes aware that sh-t is going to happen, and happen fast!

    Now contemplate this for a while (myself,I'm getting hungry now.....), so you get a picture of what is physically happening, and get the feeling for the processes involved. Then it's time to do some creative thinking around solutions; there are some interesting possibilities.
     

    Attached Files:

    Last edited: Jan 10, 2010
  2. jim lee
    Joined: Feb 2007
    Posts: 368
    Likes: 20, Points: 18, Legacy Rep: 247
    Location: Anacortes, WA

    jim lee Senior Member

    The curves doc comes up blank on my machine.

    Also, when you say "WJ" Do you mean Jacuzzi WJ?

    -jim lee
     
  3. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    The curves should open directly as a one-sided word document, hence the .doc suffix. If problems, mail me directly and I can send in another format. I use "WJ" as short for waterjet in general. And Jim, there are interesting parallells with your drag machinery, did you use splitter vanes at all, or did you use various length on the shoe/lip for tuning?
     
  4. shaka
    Joined: Dec 2009
    Posts: 17
    Likes: 2, Points: 0, Legacy Rep: 39
    Location: Texas Gulf Coast

    shaka Junior Member


    OK Baeckmo, I have read your post, and the more I read it, the more eye opening the problem becomes. Much of this is hard reading, but I am getting there. This quotation posted IS the heart of the conversation, and if we minimize this issue, then high speed riding will not be quite as dangerous. I'm probably at about 70-80% now in my understanding. I can tell you that armed with this knowlege, I am seeing the problem with a completely differet perspective.

    There is one paragraph that I understood the facts and figures, but did not get your point:

    "At top speed (~32 m/s) this flow is collected from a "streampipe" area of 0.00625 m2, ....... However, the effective impeller inlet area (just at the blade leading edges) is 0.018 m2, ie nearly three times as big. The flow would need a retardation down to some 11 m/s to keep the impeller filled, which is impossible with the configuration at hand." Is the pump to big, is there not enough flow???? Could you paraphrase and tell me what these numbers suggest to you?

    The way this information is presented, it does have me thinking about creative solutions and possibilites. So let's talk about a few.

    Concerning the "splitter vane", by varying the level lazer hight, I can see how the "stream pipe" flow is actually split. Part of the flow is deflected by the splitter vane toward the top opening, and another part of the flow continues to the lip of the "shoe" or the bottom opening. Is there any evidence suggesting that a specific percentage of flow should be deflected and the remainder continue to the shoe? For instance, is it best to have a 50/50 split, or a 60/40, etc, etc.? I can see that by extending the leading edge of the splitter, that I would increase the percentage of streampipe flow deflected. You suggested that there would be a point where the flow would be blocked off in the bottom opening either by physically increasing the leading edge length, or by the turbulance created by the shape of the leading edge. I get the impression that by moving the flow to the top part of the intake, that it would "lift the pump curve" because water is not deflected into the impeller as dramaticly as it would when it hits the lip of the shoe.

    Since I have learned that this pump is a series of compromises, I have to ask the down side to shifting all of the flow to the top opening when at speed.

    I once read a post where a pro tuner had reshaped the leading edge on his splitter vane and promply loss speed and rpm. When he rewelded and reshaped the leading edge, he gained more speed and rpm's then what he originally had. His accounts were that the splitter was a bit longer and the leading edge was thicker. Of course, no specs were given.... racing secrets.

    My splitter is shaped like a wing from front to back, and is slightly concave from side to side on the top inlet side. It also has a relatively sharp leading edge. You can see the splitter in the picture where I show the pump inlet at a 3 degree angle of attack.

    Let's beat this topic up for a while, then we can talk about possiblilites of installing a blocking plate closer to the lip of the tunnel inlet. I'll give you some numbers in a later post.

    Appreciate your imput.
     
  5. jonr
    Joined: Sep 2008
    Posts: 721
    Likes: 11, Points: 0, Legacy Rep: 57
    Location: Great Lakes

    jonr Senior Member

    What kind of instrumentation, if any, would pin down the cause? Accelerometer, tachometer, inclinometer, pressure sensors? I assume at some sample rate fast enough to determine what happened first.
     
  6. speedboats
    Joined: Jun 2006
    Posts: 139
    Likes: 12, Points: 0, Legacy Rep: 127
    Location: New Zealand

    speedboats Senior Member

    So if the intake is 1/3 the size required for the leading edge of the impeller, why would really big manufacturers like Yamaha, Bombardier, Polaris etc make such a big mistake? Do they not have hydrodynamisists working for them?

    Do I understand correctly that you want flow to break off the top of the intake and hit the centre of the impeller? Wouldn't it be better to load the whole face of the impeller?

    We too have a similar experience with varying the shape of the spoon area, some things that look like they should work have an epic fail, while a further seemingly undefined reshape (either sanding or 'grounding') and works even better than the start. While there is no science involved, merely trial and error, this area needs more involved development for increased performance...
     
  7. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    No, it's not a mistake, the full throat area is needed to feed the impeller at low to medium speeds. This is where the variable throat area would be of great value, as used from time to time in the russian high-speed hydrofoil vessels with wj propulsion. Using the horizontal vane to actually split the flow at high speed; one part into the inlet, one part blowing by, is clever thinking. The demands for acceleration and top speed of the pwc's are extreme, requiring odd solutions, and when you squeeze that extra performance from them, odd things happen.
     
  8. shaka
    Joined: Dec 2009
    Posts: 17
    Likes: 2, Points: 0, Legacy Rep: 39
    Location: Texas Gulf Coast

    shaka Junior Member

    Speedboats,

    What you are describeing is what is happening in my inlet according to my lazer. None of the stream pipe flow ever seems to get any higher than 1 cm from the top of the bottom half of the impeller. Betweem using this lazer and Baeckmo's comments, my eyes have been opened. I can literally see what the stream pipe water is doing by using the lazer. I may never be able to solve this problem, but at least I feel that I understand it much better and am not chaseing some crazy off the wall ideas.

    If there is no formula, only trial and error for how much flow must go to the top vs the bottom inlet, then perhaps extending the leading edge of the splitter vane by a few mm would not be to far fetched. There have been cases where some have lowered the front mounting plate of the intake grate (which has the splitter vane on it) by 25-40 thousandths, and reported an increase in speed, with this "bucking" resolved. There is a greater percentage of flow into the top inlet by lowering the front part of the intake grate. The problem is that the "buck" returned at a higher speed when other changes to the hull or engine were made.


    jonr,

    at this point the problem is just guess work, with Baeckmo giving us some sound advice on what more than likely the problem is. Perhaps with a bit more data, we can demonstrate that intense aeration is our culprit, and find a clear path to minimize it.

    As for the instrumentation, there is a point that I can tap into the roof of the intake that is not intrusive. I will be using a pressure gauge as suggested by Speedboats and will slowly try to get a relationship between speed, rpm, and pressure. At this point, I do not know what I am looking for.

    I would like to use electronic data collecting equipment, but would have to ask some of my co-riders to throw in with me for the cost. Getting money from them falls under the heading of "Fat Chance". So I will do the best I can.

    Texas gulf coast weather is not cooperative this time of year.......
     
  9. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    One of the main issues in establishing the operating point with WJ's is finding, within "field testing tolerances", what is the dimensioning flow!

    To do that, you have to perform a bollard pull test, including measuring the inlet static pressure one inlet radius upstream the impeller inlet.

    Using a tension scale, measure the pull and static inlet pressure at increasing rpms, starting from, say 2000 rpm, in 300 rpm steps. When plotted, you will find that first the pull is increasing as a function of rpm^2 up to a certain rpm. Then the increase will deviate to a lower rate, until finally levelling out at a maximum, when fully cavitating.

    The point, where the deviation from the parabolic trend sets in, marks the cavitation onset, and those data may be used to A/: find the flow at that point, consequently the system characteristics, and B/: with the pressure, flow and rpm the cavitation performance, in terms of the so called specific cavitation speed is found.

    Now, checking this pressure against rpm and constant speed will give a picture of what is happening in your inlet. I would start with this and see what comes up before I decided on a more extensive testing programme.

    And, yes I think the vane may be changed slightly to improve the situation, more on that later.
    For the pressure tap, place it ~80 mm upstream the impeller blade leading edges ~at a two o'clock position. Hole dia 3mm, sharp edges, no chamfer and no burrs. To this, we also need a picture (dimensioned sketch?) of the metering nozzle.
     
  10. speedboats
    Joined: Jun 2006
    Posts: 139
    Likes: 12, Points: 0, Legacy Rep: 127
    Location: New Zealand

    speedboats Senior Member

    Why the two o'clock position for the presure sensor? We've always used a 12 o'clock position to determine what is happening along the roof, are you expecting no pressure on the roof and perhaps better readings at 2?

    Also static pull. Would this determine real world situation of the water entering the inlet at speed? During a static test the impeller would be required to suck the water from under the hull, while when underway the water is more or less driven into it...

    Not trying to pick holes, but trying to get a sense of the direction you are seeing it from...
     
  11. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    I don't want the pressure pick-up in the "shadow" of the shaft. Although the static pressure is fairly constant over the area, in spite of the velocity gradients, there may be secondary (mostly transverse) flows in the shaft wake, that upset the readings. The data from the B-P test reveals the "true" characteristics of the system losses as well as the impeller design flow.

    Unless you have someone at the manufacturers backdoor, whispering data into your ear, the only way to find out is to test, following a logic route.
     
  12. shaka
    Joined: Dec 2009
    Posts: 17
    Likes: 2, Points: 0, Legacy Rep: 39
    Location: Texas Gulf Coast

    shaka Junior Member


    Just wanted to keep you guys in the loop about what has been done so far.

    I've contacted Jack at Marine Performance per Jim Lee, and yes Jim he remembers you. He is in the process of setting me up with a data collection system. This system was exclusively designed for boating buffs. They are sure expensive, but I will have 3 points that I can measure accurately at the same time with many other points available as long as I am willing to buy more probes. It will take a while to set my boat up with this system.

    Weather is also a factor here.

    I will start with the pump inlet approx. 80mm away from the leading edge of the impeller blades, the roof of the tunnel, and I was thinking the venturi. I really do not know exactly where I should mount this one to get the most accurate measurement. The simplest place would be at the venturi inlet. The outlet is out of the question since the steering nozzle surrounds it. It may not be completely accurate at the inlet, but it will be measureable and (hopefully) relative. Perhaps I should ask Jack.

    About this quotation, it is mentioned that only a 52 mm layer of high velocity water is coming into the inlet. I have managed to manipulate the intake grate slightly to actually reduce the flow into the bottom side of the intake by about 5%. With a little coaxing and a lot of welding on the shoe and intake, I may be able to reduce it by a total of 10% (5-6 mm). I will use my standard intake and shoe to get some data, then install the modified version to record any changes.

    The part where I am having a problem is a leading edge that would "provoke" a flow detachment. I get the impression that we are trying to block any incoming water into the bottom portion of the intake since it will be turbulent and has to make a radical change in direction in order to get to the impeller. The angle on this splitter vane (front wing) is not aggressive at all. I'm sure the oncoming water will detach at some point, but to far away from the leading edge to do us any good. When it detaches, what we have is turbulent water entering the bottom inlet. I believe that this would be true because I have an extra intake that was painted. The paint was peeled off about half way from the leading to trailing edge.... presumeably from the negative pressure created at high speeds on the bottom of the vane.

    Is there a web sight that may show the best leading edge design that would work for this application? I was thinking that this leading edge would not be a factor at low speeds since it would be so small, but at higher speeds it would actually force water away from the lower inlet.
     
    1 person likes this.
  13. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    Sounds you are steaming on here. I'll be gone for a few days now, but will get in touch regarding inlets when back.
     
  14. jim lee
    Joined: Feb 2007
    Posts: 368
    Likes: 20, Points: 18, Legacy Rep: 247
    Location: Anacortes, WA

    jim lee Senior Member

    I'm glad you got in touch with Jack. Good luck on your project!

    -jim lee
     

  15. Doc Nozzle
    Joined: Apr 2009
    Posts: 14
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: USA

    Doc Nozzle Thrust Whisperer

    I tried REAL hard not to join in on this thread... but... can't... fi... fight it any... longer!!! The contribution (or maybe distraction) I would like to throw in here is this:

    Can we look at a PWC (personal water craft) as an approx 1/3rd scale model of it's bigger cousin the 18-20' jetboat? And do some of the scaling laws of hydrodynamics apply?

    Ultra prepped PWC's with 150hp engines and numerous mods to their inlet geometries but running axial flow hardware are hitting high 80 to low 90mph speeds and are dealing with the "bucking" phenomena (and other handling issues.)

    Ultra prepped mixed flow top fuel jetboats (see example in photo) are running 1000+hp engines, numerous mods to their inlet, ride geometry et cetera and are hitting 160-190mph. And they have discovered/dealt with the "bucking" phenomena (also known as "the back of my drag boat blows up if my engine stops") for a couple of decades. (Their "coping mechanism" is the combination of a pop-off valve and a "jet-a-way" - one to bleed off overpressure in the intake and the other to try and decouple the motor from the impeller.)

    I'm a bit rusty on my scaling laws but isn't this the right equation for scaling speed: Vmodel/Vboat=(scale factor)^.5

    If so then pardon my highly unscientific comparison but:
    ((Top speed PWC)/(Top speed Jetboat))^2=(90/175)^2=0.26 which is in the same general neighborhood as the scale of the hulls.
     

    Attached Files:

Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.