What distance from impeller for IVR calculations?

Discussion in 'Jet Drives' started by taffyinoxon, Jul 29, 2014.

  1. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    Hi,

    Having waded through many papers I have still yet to find an answer, so hope someone on here can help.

    For waterjet calculations on IVR etc., is there a conventional distance from the impeller that the 'impeller plane' should be; eg. for a 150mm diameter impeller, the impeller inlet/plane being 150mm forward of the blades leading edge?
     
  2. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    There are two critical areas in the inlet of a jet unit:

    A/ The impeller inlet area is the transverse area that is "swept" by the leading edge of the blades. This area is critical for the layout of the pitch angle and the rest of the blades (ie the overall pump design).

    B/ The inlet throat area is the duct area at the "lip" (or corresponding) section in the hull opening. This area should be considered when discussing inlet velocity ratios, since it is a function of the forward velocity of the vessel, the shape of the boundary layer and the pump flow required.

    The (eventual) velocity change from throat area to impeller inlet area will result in a change in static pressure at along the streamlines. A velocity reduction will lead to a reduced cavitation risk, but simultaneousley it will increase the risk for boundary layer separation from the "roof" of the ramp section of the inlet. I'd say that this part of the waterjets are where you find most of the misunderstandings and mistakes that cause operating troubles and bad efficiency.

    Everything between the throat area and the final jet exit (nozzle) area can be regarded "a black box" and treated separately; this is the pump section, where the energy is added, but the conditions along the boat bottom and down to the throat is dictating the operating situation for the pump. You often find expressions like "ram effect" and the like, but that is pure nonsense; the throat area must match the flow and speed. Too small a throat will starv the pump, too big a throat will act as an efficient drag anchor. Note that the throat is operating as a hydraulic "sink" at low to medium speeds and acceleration, ie fluid is ingested from the sides of the opening.

    Finally, a Word of warning: There is no consensus as to the definition of IVR, not even within the ITTC propulsion committees. The only definition that has a physical meaning that is relevant to the design of jet units is Va/Vthroat, where Va is hull speed of advance and Vthroat the velocity in the throat area. Unfortunately, many use the inverse of this ratio, which renders the high flow - zero speed cases difficult to treat. For instance, the inlet losses are varying with IVR, how do you define the zero speed IVR and the corresponding loss in that case?
     
    Last edited: Jul 29, 2014
    1 person likes this.
  3. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    Thanks for your reply. However, I'm still wishing to know where to take the Vpump measurements from. Obviously it has to be a little distance upstream, but how far upstream, as I'm surprised how far away from the pump it is before my scalar plots resemble that in other published work. Surely there must be some kind of convention to set distance from the impeller?
     
  4. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    Have you included inlet, inlet pipe, and boundary layer losses? There is actually a lot less flow into the impeller than you think. Rather than start the work with velocity ratios and try to get the volume flow rate (Q), start with mass flow (m dot) and exit velocity and then work backwards to find the necessary inlet area.
     
  5. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    This is the problem with many of those velocity-related coefficients; you have to know exactly what you are talking about. And since there is no strict consensus on definition, you often can't compare the results from different sources.

    In the litterature will find IVR based on the pump inlet velocity (V1) as well as the throat velocity (Vthroat), the variation through the duct either neglected or compensated for.

    As Jehardiman has said, the incoming flow shows high velocity gradients, the steepness of which are depending on a number of variables. An example: Depending on boat size and design speed, one and the same jet may operate with completely different boundary layer thickness in the inlet. The result is interpreted as a variation in IVR.

    So, again:

    For the pump design it is the mean impeller inlet velocity (V1) that counts. Note that when there is a velocity gradient in the incoming flow, there is a variation in angle of attack for the leading edges, just as the ship wake causes a varying flow field into an open propeller.

    For the operating point of the pump, it is the throat velocity that counts. You could regard the hull opening ("intake", "throat" etc) as a system variable that together with the exit nozzle influences the operational situation of the pump, just as the total piping will determine the working point for any pump.

    With a fixed throat area and constant pump flow, the outside "supply area" will vary with boat speed, hence my comment on the hydraulic sink as a model for the flow coming into the throat area. The "source area" for this flow is, of course infinite. For calculation purposes, you may use a finite area, based on knowledge of the whereabouts of the limiting streamlines of the inflow into the throat area. Generally, there is little influence from the hydraulic sink outside a range of about three to four throat diameters. But please note that this is a rough approximation that may be completely wrong in certain circumstanses.

    And again:

    The relevant areas are those mentioned in my previous comment, nothing else.
     
  6. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    Yes, I am aware of the variations in flow according to boat velocity, RPM etc. and am running a series of simulations to highlight this. And yes, Q, m(dot) are being recorded, including each particular phase.

    I think part of the issue is where so many studies make use of an 'actuator disc' to represent the impeller and those that do simulate using a 'correct geometry' impeller gloss over/ignore where to locate the plane to obtain the velocity gradient. I only asked as I'm not entirely comfortable with "guesstimating" it's location
     
  7. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    Ok then, for any kind of simulation (what kind are you aiming at btw?), if you study the inlet separately, then you can use the impeller inlet area (as defined here) as a sink, or outlet, fi in a CFD model. You specify the mass flow (NOT a constant velocity) through that "inlet disc" and see what happens when the mass flow and the inlet geometry remain constant and the hull speed varies.

    Substitution of the complete, finite length impeller with an actuator disc is incorrect when it comes to the study of the inlet up to the impeller eye, hence my recommendation to use the meridional inlet area at the vane leading edge. I'd say this is where some of the confusion comes in. If you really want to model the flow all along to the exit (...there is no rational reason, as I see it, only a waste of labour and computer time), you put a source area with the design impeller outlet velocity at the impeller exit location.

    Now, if the impeller is not operating at its design flow (oooh yes that happens.....), then there is a prerotation set up in front of the impeller, adding or subtracting to/from the eventual rotation set up by an open shaft, generally of the free vortex form. Question is, how deep you want to dig, and what resources are at your disposal.

    Any CFD (or else) study should include the setup and mapping of a reasonably correct boundary layer in front of the inlet ramp; the BL flow is a considerable share of the ingested mass. In order to validate your simulation you must have a benchmark design available, which can give comparative values for losses and velocity gradients, but that is of course standard procedure.

    Good luck and keep us informed (not only of success, but we learn more from the s**t that went wrong!).
     
  8. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    Per "Hardiman's Unified Propulsor Theory" the answer is 11.18 diameters.
     
  9. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    My main focus is on analysing cavitation and noise levels through a compact unit, but want compare the single-phase models with the skeleton work done by a previous person to act as the benchmark before my multi-phase modelling begins in earnest. Though how far I can take it will depend on how much access to the cluster that I can shake out of the powers that be vs acceptable simulation runtime
     
  10. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    Since cavitation may occur at either side of the lip, at the impeller inlet (in most cases leading edge tip), at the impeller trailing edges and at the stator inlet, these locations are the relevant sections to study. In particular the asymmetric distribution of velocity and static pressure in the impeller inlet has a prime influence on the cavitation situation downstream.

    Cavitation and boundary layer detachment at, or upstream the lip is of course dictating the events in the impeller inlet, so you are back to the two critical areas I noted earlier.

    With that in mind, it should be clear that any simulation where the impeller is represented by a single actuator disc is unreliable when it comes to impeller cavitation. Better to use the sink/source model in preliminary stages and then maybe refine into a rotating mesh for the impeller.
     
  11. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    But I'm not using an actuator disc. I'm using an impeller, and hence my reasons for trying to find at what point ahead of the impeller should the plane for Vpump be located
     
  12. baeckmo
    Joined: Jun 2009
    Posts: 1,654
    Likes: 670, Points: 113, Legacy Rep: 1165
    Location: Sweden

    baeckmo Hydrodynamics

    Define "Vpump"! Since the meridional area is changing along the flow, "Vpump" can be anything - anywhere.

    The standard point of reference for cavitation/NPSH calculation is impeller inlet area as I have outlined above. There is no reason to use any loosely defined area "ahead of the inlet".
     
    1 person likes this.
  13. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    Sighhhh...Are you going to try to time step an impeller? For cavitation? With multi-phase modeling? Remember...CFD == Colourful Fancy Drawings. Too many known unknowns to even believe anything that would come from that.
     
    1 person likes this.
  14. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    Baeckmo, yes I know. All I want is a point/plane that is acceptable to take both the mean value and gradient of a vector or scalar from. In this instance we are talking about velocity of fluid entering the impeller to obtain IVR values. I gather that you view such info as meaningless, but I am just trying to follow statements in papers regarding IVR to see whether they are agreeable or not.

    Jehardiman, All I'm looking at is cavitation number rather than bubble modelling. Having not yet analysed results, I may well find that a reduced number of phases sufficient to use. Not forgetting noise is also being modelled too. It may well be colourful fancy drawings, but as long as the parties receiving the final report are happy with it then it's job done
     
    Last edited: Aug 2, 2014

  15. taffyinoxon
    Joined: Jul 2014
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Oxford, UK

    taffyinoxon Junior Member

    And the impeller inlet area that you mention is agreeable with me, but it is not the point that appears to be used in other published work, hence my confusion with it
     
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.