Prosjektoppgave:

Dynamisk posisjonering av fartøy ("DP")

Introduction

Dynamic positioning means position control of vessels using thrusters and propellers as actuators to keep the ship at a reference position relative to the seafloor or relative to another vessel or sea platform. DP is an important technology in sea operations.

What this project is about

In this project you will design and simulate a simplified DP system in LabVIEW. The system includes a Kalman Filter for estimation of one of the environmental forces acting on the ship. This estimate is used in the model based position controller.

Practical information about the project

See the homepage of the project.

Software

• LabVIEW with Control Design Toolkit and Simulation Module

Technical information about DP

Information about DP is found in Exercise 8.

Tasks

The simulation time step can be set to 1 second, and you may run the simulator 100 times faster than real time.

Tip: Use the units consequently, for example use N as force unit everywhere in the simulator.

1. Implement a simulator containing the following:
    
   • The ship model of the movements along the surge axis. All parameters should be available at the front panel of the simulator (including ship parameters and wind model).
     Check that the simulator shows a correct response, for example by comparing simulated velocity and manually calculated velocity under static conditions.
     If you want, you can attach a picture of a ship to a numeric indicator (horizontal slide) to animate the ship motion. (Right-click on the arrow on the slide. Select Advanced / Customize in the menu that is opened. Click the Tool button in the toolbar of the new window that is opened. Right-click on the arrow on the slider. Etc.)
      
   • A Kalman Filter (the predictor-corrector version) based on the following:
     • The position is x₁ is measured, and these measurements are available only each 10th second.
     • The wind angle and speed have known values, because they are assumed to be measured (and they are measured continuously).
     • The water current speed u_c is assumed to be constant or slowly varying, but it is not measured, so you must estimate it with the Kalman Filter. Hint: Augment the ship model with a differential equation describing the assumed (modelled) behaviour of the the water current.
     • Use a steady-state Kalman Filter gain, K_s. You can calculate K_s from a linear ship model based in linearization about the present operating point. You can calculate K_s with the Kalman Gain function (K_s is the M-output from this function).
       Use the Observability Matrix-function to check if the system is observable in this operating point. Note: The system is non-observable in the particular "zero-operating point" where the difference between the ship speed and the water current is zero, and therefore you must use a linear model corresponding to a non-zero speed difference when calculating K_s.
       Note: The Kalman Gain function can be used in 4 different ways, denoted "instances" in the LabVIEW Help about this function. You should use the CD Kalman Gain (Deterministic) instance.
       Hint: The calculation of K_s can be implemented outside the Simulation Loop, in a While Loop with a relatively large cycle time.
     • You should implement the Kalman Filter equations in a Formula Node in LabVIEW.
        
     Check that the Kalman Filter produces a correct estimate of the water current (by comparing the estimate with the value that you adjust on the front panel).
      

   • A positional control system for the ship based on feedback linearization, as in Exercise 8, but you are not required to implement feedforward from positional reference. The response-time of the control system is specified as 30 sec (this is used to calculate the numerical values of the PID parameters). The time-step of the control system is 1 sec. Utilize the Kalman-filter estimates in the controller!
     
     Confirm that the response-time of the control system is approximately as specified (there may be a difference by a factor of 2, actually).
     
     What is the value of the steady-state control error?
      

2. Position control error at environmental force impact: Assume that the vessel is excited by a severe wind gust, from calm to hurricane. What is the maximum control error (deviation from the positional reference) during this impact? Are the limits of the thuster force reached during this impact? If not, how far is the force from the limits? Finally, retune your PID controller (by adjusting the specified response-time) so that half the available thruster force is used during a hurricane gust.
    

3. Increasing robustness against measurement failure: Assume that the position measurement can vanish for some time. Assume that the measurement is zero in such periods. How does the DP system behave in such a situation? Is this behaviour acceptable in a practical DP system? Then, make your DP system robust against such a measurements failure! Demonstrate that your solution works.

[Emnets hjemmeside] [Prosjektoppgave-hjemmeside]

Oppdatert 9.2.09 av Finn Haugen. E-post: finn@techteach.no.