OpenROV simulation


#1

Dear all!

I’m currently working on a project which consists of simulating the openROV in a specific simulator called V-Rep and to use the results from the simulator to evaluate different control techniques, and finally to test the results on the robot. I was wondering if someone has tried this approach before and simulated the robot in this specific simulator or in another one.

Thank you in advance for your help

Regards,
Carlos


#2

Hi Carlos:
Interesting project.
Is it a software simulator or a “piloting” simulator ? If its a “piloting” simulator, may I know which level of accuracy do you expect from the simulator ?
I’ve played a bit around the maths related to the dynamic behaviour of the ROV, from the naval point of view. I mean, not related to the software itself, but to the “ship”.

Regards


#3

Hi,
It is a software program in which a robot is simulated in a virtual reality environment.
I have another question I would like to ask regarding the propellers used in openROV. Is the dynamical model of the propellers, such as the state space model, available?
Thank you for your help.
Regards,
Carlos


#4

We have basic thrust vs Amperage empirical data for the propellers, but we don’t have a full state-space dynamical model of the propellers.
The data in the link are just from static bollard pull tests of the propellers; but they aren’t really accurate representations of how they work in the OpenROV (since the flow characteristics are very different between a propeller in open water vs impeded flow within the vehicle).

We have more experimentally accurate tests of the thrust dynamics of the OpenROV, and I think @Eric_Stackpole might know where the data is?


#5

Hi:
I’ve performed several tests regarding the ROV’s propellers. There is no technical information available from Graupner, hence I’ve been searching for the most suitable standards and found that those props almost fit with the Gawn Burril standard series.
The Gawn series polynomials (4 quadrants) can be easily found in the net as free downloads.

Hence the polynomial expresions can be used for modeling the propellers performance (Velocity/Thrust/Torque/efficiency/RPM).
As Colin says, the flow characteristics for the ROV are very different from that in Open-Water conditions. But it can be reasonably solved by calculating the wake factor for the ROV.
The suitable propeller is:
R=65: P/D=0.49: AeAo=0.4 NOTE: Correction for Reynolds number required, when needed.

The device design is not “flow-friendly”, hence the flow is quite messy. The incoming flow to the propeller discs is really untidy, and then, the propulsive efficiency very, very much reduced.
About the wake factor, I’ve done several calculations and a number of attempts for modeling it. If you want, I can share my results.
Regarding the dynamic behaviour of the whole ROV under work. The RAOs are the first thing to be calculated. For foating objects the starting point is the sea waves spectrum. But for underwater vehicles, when the wave effect can be neglected, the starting point are the driving, interference and hydrodynamic forces.
Once those forces are duly known, and from the Response Amplitude Operators the motion of a body with 6DOF can be modeled.
Thats not an easy thing, as all factors change when the forces balance changes. Hence a statistical model must be built allowing for the use of only a fixed number of “constant” RAOs.
There are in the market a number of programs that allow for the RAOs computation. From MaxSurf that uses statistical models, to Numeca FineMarine that computes them from the Navier Stokes equations, soving complex and accurate CFD simulations.
As the ROV is already built, my advice is to use a statistical model computed from real “sea trials”. As the speed range is very short, and sea conditions can be completly neglected, from only 6 RAOs (Pitch,Roll,Yaw and linear motions along X,Y,Z) the motions can be duly modeled.

Regards


#6

I ran into a couple of ROV simulators which might be of interest. One which uses keyboard commands and lower down on the link is one which does work with a game controller.
You can check them out at:
http://mate.gearsinc.org/


#7

Hi!
I have tested the ROV horizontal propellers in a tank using a strain gauge and modelled the thrust from the results.
The model is a 2nd order transfer function. I can give you my coefficients and the graphs from the trial+calibration.
Regards