Electric propulsion design process

Discussion in 'Electric Propulsion' started by Will Fraser, May 17, 2019.

  1. gonzo
    Joined: Aug 2002
    Posts: 16,790
    Likes: 1,714, Points: 123, Legacy Rep: 2031
    Location: Milwaukee, WI

    gonzo Senior Member

    A shorted or connected to a high load generator or alternator will need a very high torque at startup. If the input motor can't generate enough torque at those conditions, it will stall. Actually, it will also stall at any RPM if the load is in excess of the available power output for the motor. It is a simple power balance equation. The motor's available power output must equal or be larger than the needed power to turn the generator.
     
  2. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    Pick up a brushed motor and spin it with your fingers. I have one here and I can hardly feel the difference in torque between open and shorted terminals.
    The key variable to consider is speed. Please see my experimental results, the motor turns with ample torque to spare with the generator shorted. This is despite driving it through a step-up gearbox.
    Without the gearbox, the shorted generator puts even less load on the motor. Conversely, with a high enough step-up gear ratio even the static friction of an unloaded generator can cause the motor to stall.

    To other readers, I know this looks like thread drift but much of what is discussed here will become very applicable when selecting the correct propeller for a motor.

    Whereas a loaded generator's torque will increase linearly with rpm, that of a propeller will increase by rpm^3. The overall response is the same though - when you apply a voltage to the stationary motor, it kicks into life with maximum torque which immediately starts decreasing linearly as rpm picks up.
    The torque on the propeller (caused by hydrodynamic drag) starts at zero and ramps up with rpm cubed.
    The motor will continue to accelerate until it reaches an rpm at which the two torque curves intersect. This all happens very quickly.
    If the correct prop was selected, this intersection should occur at the motor's rpm for maximum efficiency for best performance on limited power.
    To achieve maximum boat speed with no restriction on available power, the prop should be selected such that the torque curves intersect at the motor's rpm for maximum power, i.e. half its no-load speed.

    Boat speed also comes into play, and the torque curve looks different at different speeds, so it is not as simple as writing out the two equations and solving for rpm. Different sized props can result in the same motor rpm, but due to differences in efficiency of the props themselves, boat speed will not be the same.
    I will try and put together a family of these curves to illustrate.
     
  3. gonzo
    Joined: Aug 2002
    Posts: 16,790
    Likes: 1,714, Points: 123, Legacy Rep: 2031
    Location: Milwaukee, WI

    gonzo Senior Member

    You are incorrect that a generator's torque will increase linearly or that of a propeller to the 3rd power. They both have non-linear responses. You also need to study propeller theory. The resistance (not torque) is generated by more than hydrodynamic drag. If drag was the only force applied to the propeller, boats would not move. Seems like you have a lot of ideas, but little technical background. If there were no restrictions on available power, the limiting factor would be either the structural integrity of the boat, or loss of stability.
     
  4. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    What I meant with "no restriction on available power" was power input available to the motor. At a fixed voltage this translates into current not being limited. An example would be a large battery bank that can supply whatever current the motor draws at a fixed voltage. The motor itself has a finite maximum power output at that voltage and the prop should be sized to load the motor to that exact rpm if you want maximum power from the motor.

    A rotating propeller blade experiences a force that can be resolved into two components: for the sake of explanation, call them drag and thrust. Consider a small segment somewhere along the length of a slender aircraft prop blade (it is just easier to visualise than a typical boat propeller). The drag component on that segment is tangential to the arc of rotation, i.e. it is perpendicular to the prop shaft and lies in the plane of the prop disc. Multiply the drag force by the distance to the shaft and you get the torque required to rotate that portion of the blade at that rpm. Add up the torques for all the segments along the blade and you get the total torque required to turn the blade.
    The second force component is thrust. It lies perpendicular to the prop disc and it pushes the boat forward. Since it lies parallel to the prop shaft it does not contribute to the torque on the shaft.
    As far as the motor is concerned, it does not exist. It only seeks equilibrium with whatever torque load there is on the shaft.

    Here are some test results for generator torque vs rpm. I would call that a linear response.
    The load was a shorted 12V brushed motor driven through a 3.4 step-up gearbox. Speed was varied with the ESC.

    Generator torque curve.JPG
     
  5. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    Here are some propeller torque curves generated from propeller theory.
    A 12" x 6" and 12" x 10", each at static (think bollard pull) and at different boat speeds.

    At near-static conditions the prop torque curves are proportional to rpm^2, but at speed rpm^3 fits the curves much better. The exact response is irrelevant.
    Also shown is the brushless motor's torque vs rpm at 13V as reported on earlier, adjusted to reflect the shaft torque and rpm of a 2.7:1 reduction gearbox.
    This is just to illustrate that, like the generator, the load torque rises with rpm while the motor torque (at constant voltage) decreases with rpm. The point where they intersect will be the actual operating torque and rpm.

    The closely spaced data points on the shaft torque curve corresponds to the motor's max efficiency rpm as tested earlier. They clearly fall between the 12x6 prop's curves for static conditions and up to 4kts.
    This prop and gear ratio combination will therefore automatically load the motor to its maximum efficiency rpm at 4kts and less.
    These curves are not useful on their own since they do not contain any info on prop efficiency or thrust.

    The static (0.1kts) curves will cut both axes at the origin, but if you extrapolate the curves for 4kts and 6kts downwards they cut the rpm axis at a few hundred rpm each. The rpm axis represents zero torque, so those points of intersection indicate the free-wheeling rpm of each prop at that particular boat speed. Any rpm lower than that and the prop would act as a turbine driving the motor like a generator. For this to happen you would of course need some alternative propulsion to maintain boat speed e.g. sails.

    12in prop torque curves.JPG
     
  6. gonzo
    Joined: Aug 2002
    Posts: 16,790
    Likes: 1,714, Points: 123, Legacy Rep: 2031
    Location: Milwaukee, WI

    gonzo Senior Member

    Stating that "current is not limited" is unrealistic. All electrical motors have a current limit before they go up in flames. You need to read some basics on electrical components.
     
  7. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    Gonzo your comments have become a nuisance, I have patiently explained my approach and have been able to back up everything with actual experiments.
    This includes disproving the majority of your opinions.
    You misquote me and then proceed to try and lecture me in my own field of expertise.

    I have never had to do this before, but I need to ask you to please refrain from making any further comments on this thread.
     
  8. gonzo
    Joined: Aug 2002
    Posts: 16,790
    Likes: 1,714, Points: 123, Legacy Rep: 2031
    Location: Milwaukee, WI

    gonzo Senior Member

    Sorry, but I have done a fair amount of work on electrical testing and research. Your post do not reflect you being an expert in the field. I have never misquoted anything, but asked questions and pointed out mistakes. Statements like "current is not limited" are only valid in a theoretical application that is not possible in the physical world. In a public forum, specially one with a large percentage of engineers and other members in technical fields you will find that we demand proof to statements and claims.
     
  9. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    Testing continues, this time with the batteries in series for 26V.

    The pinion in the generator gearbox eventually failed under the increased rpm.
    This old starter motor for gas rc planes turned out to be a very convenient replacement. It ran at 5520 rpm with 13.3V, KV = 415, just about double that of the brushless.
    So with the brushless doing 5500 rpm on 26V, this motor was generating just a little more than its rated 12V without the need for any gearbox.
    I also discarded the pwm drill trigger as a load regulator and connected the bulbs one at a time.

    The bulbs (now 5 of them) provided just enough load to get the motor up to 89%efficiency.
    For higher loads I connected the starter (now generator) clips directly onto a long piece of 2mm stainless steel cable. To adjust the generator load I now simply adjusted the distance between the clips, and therefore also the resistance.

    20191011_092755-1.jpg

    I have superimposed the results over the 13V tests. The primary (left) y-axis scale was adjusted to fit the new rpm range, all the remaining variables are still on the secondary axis.

    The two current curves are co-linear. For a given motor, a certain amount of current will result in the same amount of torque, regardless of voltage or rpm.
    Once you know the slope of that line (a constant expressed as Nm/A) you can calculate the torque for any load if you know the current draw. This could be very useful if you want to evaluate propeller torque during sea-trials where measuring torque directly is not practical.

    Keda dyno graph2.JPG

    When efficiency is plotted against power instead of torque, it becomes obvious why voltage needs to be selected to match the desired power.
    I have plotted both Power-in and Power-out curves, the horisontal distance between each pair of curves representing the power losses.

    keda eff vs power.JPG

    For solar applications the Power-in will be known. So as an example, if you have a 100W panel, the 13V (red) curve shows that you can get around 85% efficiency from the motor, i.e. 85W on the prop shaft. At 26V (pink curve) you will be lucky to get 50%. This assumes that the ESC is at "full throttle" in both instances and that prop selection was used to achieve the desired rpm.

    There is an another significant penalty when voltage is too high. At 13V the motor consumes 100W at 2600rpm. At 26V it does so at over 5400rpm. This rpm can only be achieved with a much smaller propeller operating reduced efficiency. This assumes identical gear reduction ratios, if any.

    I also tested the ESC (still connected to 26V) at partial throttle . I picked a point on the 13V curve and adjusted the throttle setting and generator load until I got the same combination of torque and rpm. The motor-esc combined efficiency was 77%, down from 87% at 13V full throttle. By choosing the same rpm and torque it can be assumed that a large, slow prop can still be used as with 13V. It will just have to be a fraction smaller due to the lower shaft power.
    If you do not have the option of feeding the ESC with an ideal low voltage, this is still a much better option than running a small prop at WOT.
     
    Dejay likes this.
  10. Dejay
    Joined: Mar 2018
    Posts: 721
    Likes: 138, Points: 43
    Location: Europe

    Dejay Senior Newbie

    Thanks for sharing your tests and your explanations Will !
    This is all very instructional to learn the basic relationships between these curves.

    Looking forward to see how you are going to select the propeller. Matching a suitable motor, propeller with the available voltage and resistance/speed curve of the boat seems a bit tricky.
     
  11. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    Video of the dyno as discussed above.
    Also shown are the bevel gears from a small angle grinder. These will be used to reduce speed and increase torque on the propeller shaft. For a given motor operating at its maximum efficiency - or any other specific rpm - a reduction in prop rpm offers the opportunity of swinging a larger diameter prop. This generally means better efficiency on the prop.
    The increased prop efficiency would have to compensate for any losses in the gear drive, so the reduction ratio would have to be large enough to ensure a net gain in efficiency.

     
    Dejay likes this.

  12. Will Fraser
    Joined: Feb 2014
    Posts: 170
    Likes: 37, Points: 28, Legacy Rep: 11
    Location: South Africa

    Will Fraser Senior Member

    I finally got around to making the video on propeller performance.
    It is still just a high level explanation without detailed equations.
    The focus is on performance graphs to explain the differences and similarities between model aircraft props and marine propellers.
    In the process I also demonstrated how variables such as pitch to diameter ratio, blade area ratio and blade count affect efficiency.

    Data for the model props were taken from wind tunnel tests done at UIUC: UIUC PDB - Vol 1 https://m-selig.ae.illinois.edu/props/volume-1/propDB-volume-1.html
    It happened to include the prop I had used on my solar kayak. This made for a convenient case study since I had first hand test data for performance in water.
    Marine prop performance were based on the Wageningen B-series.
    Reynolds number is also addressed and I show how it affects performance for model aircraft as well as marine propellers.

    In a sequel to this I will look at the equations and show how to use them for different applications.

     
    BlueBell likes this.
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.