Capture performance data for the design of
Conducting a sea trial is a necessary step when delivering a vessel. This activity not only documents and validates the intended speed and power of the vessel, but if done correctly, it also provides performance data that can be useful for design purposes. The same goes for underwater vehicles in general and thrusters in particular. However, while collecting data may seem easy, extracting useful test data for typical UV can often be quite difficult.
We should start by understanding the system – the Vehicle-Propulsor-Drive system. For âtransitâ UVs whose mission is to move (take readings for example), it is the drag of the vehicle body and appendages which establishes the load on the system. The thruster – an encompassing term for the thruster propeller and nozzle (when used) – provides the thrust needed to overcome drag at a particular speed. The Drive turns the Thruster at the RPM needed to generate the correct amount of thrust. The power generated by the drive is simply what the thruster needs at rpm under the particular test conditions.
Since most transit UVs currently use electric motors as the drive, it is easy for the controller to document RPM and current draw. In most tests, vehicle speed can be captured, which we will see is necessary to determine system performance metrics. This data set is pretty thin if you want to draw real conclusions about performance. But all is not lost…
Consider what performance metrics are needed to fully describe a functioning system. As mentioned, the characteristics and speed of the vehicle are responsible for the drag load on the system. The thruster produces proper thrust at a particular rpm and exhibits torque (i.e. blade rotational drag) to the drive. The electrical input power (from voltage and current) produces a corresponding output shaft torque in the drive. In an ideal setting, each element of the system would be directly measured and documented:
Mechanical: Vehicle drag ï® Thruster thrust ï® Thruster torque ï® Drive torque ï® Drive electrical input
Operational: rpm, speed
Regularly measured items are in bold, so as you can see we have a lot of critical knowledge about UV performance that is not defined. For example, battery budget improvements may come from reduced vehicle drag or greater thruster or drive efficiency. But we need to determine the missing parts before we have any chance of improving system components.
If we cannot directly measure these missing parameters, can we predict or estimate them with enough precision to be useful? Fortunately, in most cases the answer is yes.
Drive motor efficiency
While electrical input power is interesting from a battery usage standpoint, system performance requires the power of the mechanical output shaft. The relationship between the two is engine efficiency, and we want the highest efficiency under operating conditions for the least strain on the batteries.
(Quick sidebar: Pay attention to how torque and power are used to describe drive performance. Torque is a measure specific to a particular rpm. Power is a much better way to communicate power transmission. This is especially true when talking about electrical input power versus mechanical shaft output power.)
Two key points about engine efficiency are that a) engine suppliers usually report engine efficiency only at full load and at particular rpm, and b) propellers almost never get full power from the shaft. a variator. Electric motor suppliers typically do not provide a full efficiency map (like the one shown below) that quantifies the effect of part load and speed. That said, we have the means to estimate the efficiency of the motor at part load so that we can predict the torque developed by the variator – and that should correspond to the torque required by the thruster.
Thus, the operational efficiency of the motor at part load is important information for the design. For more information on electric motors for UV applications, let me refer you to our previous article in the March 2020 issue of Marine Technology Reporter.
Thrust and torque of the thruster
How can we derive or estimate thruster thrust and torque given the limited test data available? For this, we will use the known performance characteristics of the Propulsor as a âdigital dynamometerâ to predict thrust and torque. However, for an accurate prediction, we need accurate test data for rpm, speed, vehicle-thruster interaction coefficients (especially wake fraction) and thrust torque performance coefficients for the thruster. .
In the absence of a direct empirical determination (such as model testing), the interaction coefficients and the propellant performance coefficients must be derived analytically. CFD calculations from an experienced source could be used for this purpose, as could predictions from ‘simplified physics’ calculations such as those found in NavCadÂ® software (for system simulation) and PropElementsÂ® software (for system simulation). analysis of open and channeled propellers).
Let me caution, however, about using popular propeller series (such as the B series or the Kaplan channeled propeller series) for the prediction of propeller performance. In short, there is such a divergence between traditional series and contemporary UV propeller, nozzle and stern geometry that their use cannot be justified without calibration or alignment. For example, blade area ratios are generally smaller than the dataset of these series, and their small diameter is subject to differences in scale effect. Contemporary nozzle styles do not resemble traditional geometries, and in many cases nozzles are just protective casings that do not provide any thrust advantage. (Two examples of contemporary nozzle / shell geometries are shown below.) Stern shapes are often subject to a steep entry angle into the thruster that is not captured by the uncalibrated use of traditional series.
Predicting the performance of the Propulsor component with CFD or a tool like PropElements avoids these drawbacks and can actually expand the calibration corrections for better prediction using a traditional series. With a little care, the vehicle-thruster interaction can be predicted and an appropriate performance model of the thruster can be developed. These will then make it possible to predict the thrust and torque of the thruster at each test condition.
The thruster thrust (as determined above) can then be used together with the thrust vehicle thrust deduction interaction coefficient to determine the vehicle drag at this condition. By performing the tests and running the simulation for a variety of speeds and revs, you can develop a drag curve for the particular design of the body and appendages.
System simulation for design
Even under the best of circumstances, the test simulation may still contain some uncertainty in the various measures of system performance. But that’s okay because improving the design is mostly a qualitative (or comparative) exercise where we want to see if one variation is better than the other. A robust system simulation with appropriate models for the prediction of individual components must be both accurate and faithful to the potential improvements or changes being considered. Capturing test data and deriving performance metrics provides important knowledge for any UV developer interested in improving their vehicle, thruster, or drive design.