Estimating the Immeasurable: Soil Properties...Estimating the Immeasurable: Soil Properties CEDA...

Estimating the Immeasurable: Soil PropertiesDredging Tools for the Future

CEDA Dredging Days5/6 November 2009J. Braaksma, J. Osnabrugge and C. de Keizer

Estimating the Immeasurable: Soil PropertiesCEDA Dredging Days: 5/6 November 2009

2

Contents

• Introduction• Control problem for dredging equipment• Model based adaptive control• Examples

• Overflow loss estimator for hopper dredgers• Trail force estimator• Grain size estimator• Anchor position estimator

• Conclusions


3

Introduction

Dredging: relocation of dredged material

Environment part of process

Soil characteristics have large influence on the system behaviour


4

Control problem for dredging equipment• Typically we use feedback control for set-point tracking

Process+

-Controller

Set-point Controlaction

Output


5

Control problem for dredging equipment• Typically we use feedback control for set-point tracking• Process is characterised by dredge equipment,

environment and their interaction

Process

+

-Controller


OutputDredge equipment

Environment

Interaction


6

Control problem for dredging equipment• Typically we use feedback control for set-point tracking• Process is characterised by dredge equipment,

environment and their interaction• Environment may vary significantly

Process

+

-Controller



Environment

Interaction


7

Control problem for dredging equipment• Fixed controllers will mostly not work under varying

dredging conditions• Varying process behaviour requires special control

solutions• Adaptive control


8

Model-based adaptive control

• Models are available that describe the dredging processes• Use models for predicting soil properties

Model

Measuredsoil properties

Predictedsystem behaviour

Inverse Model

Measuredsystem behaviour

Predictedsoil properties


9

Model-based adaptive control

Predictedsoil properties

Process

+

-Controller



Environment

Interaction

Inverse modelof process

Adaptationlaw


10

Examples

The paper describes the following examples:• Overflow loss estimator• Trail force estimator (for DP/DT and Trail Speed Controller)• Grain size estimator• Anchor position estimator


11

Overflow loss estimator for hopper dredgers• Overflow losses reduce dredging production• Reduction is required in sensitive ecosystems, causes

turbidity• Measuring means a technical challenge and requires

expensive equipment• Therefore an estimator is developed that requires only

sensors already available on every modern hopper dredger


12

Overflow loss estimator for hopper dredgers

0 20 400

2

4

6

8

Time [min]

Flo

w-r

ate

[m3 /s

]

0 20 401000

1100

1200

1300

1400

1500

1600

Time [min]

Den

sity

[kg/

m3 ]

Qi

Qo

ρi

ρo


13

Possible applications for the overflow estimator• Active control to reduce the losses in case of environmental

regulation• Active control to increase the losses in case of agitating

dredging• Warning system to operators for prevention of excess

overflow loss


14

Trail force estimator

• Trail force during dredging determined by drag head ground interaction

• Force sensors in pipe hinges are expensive and fragile• Solution estimate force based on differential pressure drag

head and accurate model of the ship dynamics

-200

0

200

400

600

800

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Differential Presure [mbar] Force [kN] Speed [cm/ sec]


15

Grain size estimator

• The grain size is an important soil property for the dredging process for both the TSHD and CSD.

• Behaviour of dredge pumps, pipes and hopper (sedimentation speed) are significantly influenced by the grain size

• The estimator has been successfully implemented onboard several CSD for optimizing the discharge process

• Other applications of the estimator are: • Optimizing the loading process of a TSHD • Useful for the development of advanced wear models of

pumps, pipes and valves (e.g. life time prediction)


16

Grain size estimator

• Experimental data from the CSD Rubens with a discharge pipeline of 7720 m

• Measured inputs are the discharge pressure, mixture flow and density

• The estimated mean grain diameter (309 µm) nicely converges to the measured value (285 µm) resulting in an relative error of 8.4%


17

Anchor position estimator

• Correct positioning of the anchors of a CSD is important for the swing process (unfavorable anchor positions -> lower swing speeds -> lower production)

• The anchor position estimator gives an advise to the operator when to reposition the anchors

• Dragging of the anchors is detected and reported to the operator

• The estimated anchor positions is also used to further improve the automation onboard the CSD (e.g. swing speed controller)


18

Anchor position estimator

The anchor position estimator is implemented and in use onboard several CSD


19

Conclusions

• Soil properties determine large part of the process behaviour

• Sensors that directly measure soil characteristics are either not available or expensive/fragile

• Solution: use estimation techniques where sensors/measurements are not feasible


20

Thank you for your attention

Estimating the Immeasurable: Soil Properties...Estimating the Immeasurable: Soil Properties CEDA...

Documents

Transcript of Estimating the Immeasurable: Soil Properties...Estimating the Immeasurable: Soil Properties CEDA...