Modeling and Control of Trawl Systems

Karl-Johan Reite, SINTEF Fisheries and Aquaculture

Supervisor: Professor A. J. Sørensen * Advisor: Professor H. Ellingsen * * Norwegian University of Science and Technology

Outline Motivation and objective of thesis Main parts: Mathematical modeling Control concept Control architecture

Summary and concluding remarks

2

Background The world capture fisheries

95 million tons of fish 80 billion dollars

Negative impacts of trawling Damage to the seafloor Pollution (poor energy efficiency) Poor selectivity

3

Goals of trawl control Bottom contact Energy consumption Selection properties

4

Trawl system Trawling vessel

Warp

Trawl door

Bridles Trawl net Cod end

5

Trawl door

To the vessel

To the trawl net

6

Control of trawl system Today Manual control

Vessel speed Vessel heading Warp length

Measurements

Future Automatic control Local actuators Observer

7

Main control objectives

Robustness Integration against existing control systems Integration against existing sensors Complex objectives Industrial constraints

8

Objective of thesis Trawl system model Trawl door control concept Trawl system observer Trawl control system architecture

9

Mathematical modeling Levels of complexity, efficiency and accuracy:

Fast control plant model (MPC optimization) Accurate control plant model (Control concept

optimization, trawl system observer) Process plant model (Evaluate trawl control

system)

10

Angle of attack

Angle of slip

Definition of hydrodynamic angles

αd v

v

βd

11

Trawl door steady-state forces and moments

Wind tunnel experiments Experiment setup 6 force components 90 combinations of

orientation angles

12

Parameterization Parameterization gives:

Smooth model Filtering of measured values

Resulting parameterization:

Extrapolation by interpolating towards simple damping coefficients.

13

Example: Parameterization of lift force

14

Parameterization results

15

Transient effects What forces act on the trawl door during transient

motions? May reduce computational effort Would have a small bearing on trawl system behavior Are estimated using a numerical method

Illustration of 2D unsteady foil from “Marine Hydrodynamics” (1977) by J. N. Newman.

16

Vortex lattice method (VLM)

Ring vortices placed on the mean surface of the foil Strength of vortices calculated to give zero fluid flow through the foil Vortices shed from the trailing edge forms the wake Forces and moments calculated from the strength of the vortices

Γ (i,j)F

Γ (i,1)W

Γ (i,j+1)F

Γ (i,2)W

kvΔt

Foil

Trailing vortices

Bound vortices

Ring vortex

17

Parameterization of the transient results Circulation build-up Step responses of

linear systems Acceleration No cross coupling

Angular velocity No cross coupling

18

Forces and moments from circulation

19

Forces from relative accelerations and angular velocities. (“Added mass” and “angular damping”)

20

Resulting hydrodynamic model

Steady-state hydrodynamic forces and moments Circulation build-up Forces and moments from accelerations (“Added mass“ ) Forces and moments from angular velocities (“Damping“ )

21

Trawl door control concept How can the hydrodynamic forces on the trawl doors be

controlled? Method:

Evaluate former proposed concepts Analyze new possibilities Towing tank experiments Numerical optimization

22

Control concept choice

Towing tank experiments Criteria

Energy efficiency Control performance

Warp control is chosen.

23

Control concept optimization

Optimization routine

Objective function Design

parameters

Objective value

Estimated optimum

Simulations

Simulation results

24

Control concept results

25

Trawl system control architecture How can the trawl system be automatically controlled? Requirements:

Integration against existing controllers and actuators Use of trawl door control concept Ability to include complex objectives Industrial constraints

26

Model predictive control (MPC) Optimization

routine

Objective function

•Dynamic trawl system model

•Constraints

•Performance requirements

•Objective evaluation

Control signal

Objective value

Optimum control signal

Trawl system

Initial states

Measurements

Operator requests

27

Observer

Model corrector

Trawl system

Independent model

Estimated states

Model parameters

What is happening in the trawl system? Position, depth, shape,

bottom contact Catch Velocities

Available measurements Few Low update rate Inaccurate Expensive

Estimated states

Control input

Measure-ments

28

Control system main components

29

Case studies: Model comparison

30

Case studies: Trajectory controller

31

Case studies: MPC, 3000 iterations

32

Conclusions

A mathematical model of the hydrodynamic forces on a trawl door has been developed

A trawl door control concept has been developed A trawl system control structure has been proposed

33

References

Reite, Karl J: Modeling and control of trawl systems. Tapir Akademisk Forlag 2006 (ISBN 82-471-8024-3) 238 s.

34