The Automatic Guidance and Control of Remotely Operate Vehicle for an Underwater Inspection of Mega-Float

The Automatic Guidance and Control of Remotely Operate Vehicle for an Underwater Inspection of Mega-Float

Copyright © IFAC. Control Applications in Marine Systems, Fukuoka, Japan, 1998 THE AUTOMATIC GUIDANCE AND CONTROL OF REMOTELY OPERATE VEHICLE FOR AN...

1MB Sizes 7 Downloads 49 Views

Copyright © IFAC. Control Applications in Marine Systems,

Fukuoka, Japan, 1998

THE AUTOMATIC GUIDANCE AND CONTROL OF REMOTELY OPERATE VEHICLE FOR AN UNDERWATER INSPECTION OF MEGA-FLOAT

Hiroyuki Oda'" ,Shintarou Miyoshi'" , Tomoaki Ishihara'" Kohei Ohtsu** ,Manabu Sasaki"'**

* Akishima Laboratories (Mitsui Zosen) Inc. ** Tokyo University ofMercantile Marine *** Mitsui Engineering & Shipbuilding Co., Ltd.

Abstract : The inspection and maintenance of the huge structure present an extremely hazardous and tiresome task. This has led to the development of many unmanned, tethered submersibles designed to carry TV and sonar equipment. At the time of the underwater external apparent inspections, the operations are greatly influenced and restricted by the various kinds of condition. The concept design of ROV has been arranged to suit the inspection of the vast bottom area of the floating structure. Navigation control system is needed to fulfill the sufficient accuracy by applying an acoustic positioning and guidance and control system, because of patrolling the vast bottom area efficiently without an oversight. This paper describes the background of underwater inspection, guidance and control system with simulation study and full-scale experiments. The year is the year in which the IFAC meeting was held. Copyright © 1998 IFAC. Keywords : Mega Float , Guidance and control , Remotely operate vehicle Underwater inspection , PID control , Dynamic positioning , Route tracking MATLAB , SIMULINK.

positioning and guidance and control system, because of patrolling the vast bottom area efficiently without an oversight. The former secures the reasonable accuracy by contriving the arrangement of transducers and using the distance between the ROV and bottom of Mega-Float measured by an ultrasonic sensor. The navigation system involves the design of automatic speed control, system for dynamic positioning and tracking, as well as autopilot systems for both heading and depth control. Therefore, the navigation control system was developed using mathematical model simulation and multivariate PID control for ROV. This paper describes the background of underwater inspection, guidance and control system with simulation study and full scale experiments using MATLAB tools and SIMULINK at the model floating structure of Mega- Float.

1. Introduction

The inspection and maintenance of the huge structure present an extremely hazardous and tiresome task. This has led to the development of many unmanned, tethered submersibles designed to carry TV and sonar equipment. Since super-large floating structure (Mega-Float) must be guaranteed as an artificial ground with semi-permanent life and safety, an underwater inspection system is to be developed to observe the bottom of Mega-Float for supplying judgment tools to repair and maintain it. An underwater inspection system is realized remotely operate vehicle (ROY), navigation control system with guidance and control, picture processing system for both optical and acoustic images. Navigation control system needs fulfilling the sufficient accuracy by applying an acoustic

197

Vertil"Al Thruster

Horizontal Thrust e,.

2. Underwater Inspection

Game,.. for N.",;,ation Sen$Orof

In the case of a super-large floating structure with the object of the Mega-Float Technical R & D Association is not able to come into a dock for examinations, inspections and repairs. These operations for the underwater structure are mainly carried out under the water environment. Under water inspection is classified into external apparent inspection and non-destructive inspections. For the underwater external apparent inspections, ROV has popularized along with conventional tools. The underwater external apparent inspections are to detect the total condition of the structure as dropping outs, external wounds and the uneven of the structural parts, etc. At the time of the underwater external apparent inspections, the operations are greatly influenced and restricted by the various kinds of condition. The concept design of ROV has been arranged to suit the inspection of the vast bottom area of the floating structure (Yuasa, et aI. , 1997).

Elet;tric

Potent.1

Resoonder

uleraJ Tkruat.er

Fig. 1 General arrangement of ROV

3.2

The concept design of the ROV has been arranged to suit the patrol of the vast bottom area of the floating structure. The function of underwater external apparent inspections has been incorporated in the concept to detect the total condition of the structure. In order to achieve the purpose some items as body stabilization, convenience of handling, motion performance and observation ability. The ROV is operated with the computer or the joystick. The information of the direction and depth and also position sensor, ROV carry out the guidance and control by the computer at the time of underwater inspection. The horizontal thrusters which generate surging force and yawing moment are located on each side of the ROV The vertical and lateral thrusters are located in the middle of the ROV Most of the onboard instruments including cameras, responder, ultrasonic profiler and electric potential sensor are mounted frames . The ROV is keeping balance by the moment between gravity and buoyancy, which realizes good static stability in rolling and pitching.

3. ROV System 3. J

General Arrangement

The ROV is assigned not only to observe the vast bottom area of the floating structure but also to carry the several sensors and support divers. Therefore some functions are investigated to design the navigation control system. The guidance system unifies the information of motion and position of the ROV, compares the results with the set route and notifies the guided order to the control system. Also the control system revises the motions with depth, course and speed ofROV The ROV is an open frame structured robot which measures about Im long and weights approximately 50Kgf in air. The specifications of ROV are shown in Table 1 and arrangement is shown in Fig. 1.

4. Controller Design

Table 1 Specification ofROV system Items

Devices constitution

4. J

Ouantity ( Remarks )

Equations ofMotion

The general structure of the equations of motion can be derived from the basic force and moment relationships for a body moving with six degrees of freedom such that,

Scale of external body : 1.0m(L)*O.6m(8)*0.4m(H) Weight of body : 50Kgf Horizontal thruster: 220W*2sets (fore , back and turn) 220W*2sets Vertical thruster : (upward and downward) Lateral thruster : 220W*lset (sideward and course keep) 150W*2sets (navigation) Underwater bulbs: 150W*2sets (observation) Underwater bulbs : TV camera: Iset (navigation) TV camera : 1set (observation) Iset (flux gate type) Azimuth sensor: 1set (bottom observation) Ultrasonic profiler : Ultrasonic gap sensor : I set (height keep) I set (depth keep) Pressure sensor; 1set (position) Responder : 150m*lset Tether cable : (optic and electronic)

m

U

m(wq -vr)

v

m(ur-wq)

W

m(vp-uq)

p

q

(lzz - Iyy)qr (lxx-Izz)pr

r

(lyy - Ixx)qp

+Fb+Ff +Ft+[:]

(1)

Here m is mass, [u, V, w] are velocity components, [p,q,r] are angular velocity component, [Ixx'/.w.lzz] are inertia components, Fb are restoring moment components, Ff are components of hydrodynamic

198

to keep the advanced direction, so that it must crose while canceling the yawing moment by the current. The yawing moment can be generated by differentially driving both of right and left thrusters, but these must mostly be used to generate the forward force while the body advances. The vertical thruster controls only the depth. The lateral thruster is used to the support at the crabbing motion. The power distribution method as a controllable design method using the weighting matrix on thrust allocation which have wide application among different type of robots and ships (Oda, et al., 1990 ; Karasuno, et aI., 1993). The power distribution method using weighting matrix is shown in Fig. 3.

force, Ft are thruster components. Fa and Ma are influence of some disturbances (penier, 1996; Kapsenberg, 1985). 4.2

Power Distribution

The navigation system of guidance and control for ROY is mainly made up of feedback control and thrust allocation. Improved robustness and performance in presence of environmental disturbances can be obtained by applying a closed loop control system of PID. The PID control power T* is represented as follows (Foss en, et aI., 1995).

T* == Kp . e(t) + Kd . e(t) + Ki·1 e(t)dt

Ad ---. cJJ*Ad+cJ2*Ro ---. Th r

(2)

Here T* means the required control power of surging (A d) , yawing (Ro) , heaving (Vt) and swaying (Tr) . Also, Kp, Kd, Ki are control gains and e(t) is deviation from desired and value. The feedback control flow of ROY is illustrated in Fig. 2.

Ro Vt Tr

Fig. 3

X ---. ---. ---.

c2J*Ad+c22*Ro ---. Th I c3J*VI

~

Th v

c4J*Tr

~

Th s

Power distribution with weighting matrix.

(Controller)

The coefficients of [cl l,cl 2,c21,c22,c3I,c4I] are factor of weighting matrix on thrust allocation. Also, Th J is the thrust of right -side horizontal thruster and Th - I is the thrust of left-side one. Th- v and Th - s are the thrusts of vertical and lateral thrusters.

Settina: (Depth .head)

(loop )

4.3

Simulation techniques are presented as a possible aid in the development of guidance and control system of ROV To solve the problem of guidance and control of ROY, we use MATLAB and SIMULINK to build a reasonably simple yet effective model of response to thruster control (Alberto, et aI. , 1996 ; Bishop, 1997). The block diagram of simulation using MATLAB and SIMULIMK is shown in Fig. 4. The hydrodynamic characteristics of ROY made the database of the resemblance forms. The thrust performance that was obtained with tank test about

'----~I"utom.t lc Mode Horizontal Thruster )( 2

Later.1 t hruster

Fig. 2

I

1

----C:~.Ck

Simulation Study

M R.. " ,..

Block diagram of control system.

When the combined forces of advanced speed and the current is oblique, a moment acts to the body. In the case, the longitudinal axis of the body is desirable

~------------------~I d~ X

... Hph

·[email protected]

Depth

Mux

f - -- f- - '- 1

U \

.. leading

Mux ROV Model Block

--- -- - - - --_. ------------- -------::=:-----==. Cont.ro l Blo c k

~

' - - -<:KG~1

PlO Depth PlO

Contrnllerl

Fig. 4

Basic block diagram of control simulation. (set depth : 2m , set head: -120deg)

199

_ Sur"4

----- ~ S··

I De

Ih

>--__--=:.::.B.:....=..:=:::P-=----'

_ I J . V , Vt.I ,

1

the thrust characteristic, using a first set value. The adjustment of the confirmation and also control gain of a fundamental dynamic characteristic adjusted it while doing a simulation. The results that advance speed simulated with O.4m1s be shown in Fig. S and Fig. 6, to confirm the performance regarding the heading and depth control. These figures show the situation that is converging at the setting direction and also the setting depth with about lOseconds.

The combined performance of guidance and control system was evaluated in simulation method. Typical result is shown in Fig. 7 and Fig. 8. Figure 7 shows both set route and trajectory of ROV and suggest the performance of route tracking performance of this control system. The ROV starts from (Om,Om) and pass (20m,Om), (20m,20m), (-lOm,20m), (-lOm, -lOm) and then return to the start point. (x, y) means x coordinate and y coordinate.

I+..•.. •. •. ..T.i..:..·.: :r.··•••:'······::·· ·r·····:·::·:

~!- . ~

· 120(u.,) .,

50

t.

..... .

o

Fig. S

!(m)

I f

d

. .. ... . .

B 10 n 71me ( s~r::c:"ld)

" '"

Simulation result of heading control performance with horizontal thrusters. (set head: -120 deg) .1.

1

l , •• {uc:

;.\~

---:---------,---::::::===:=--:----,

J

1.5 ·

!l

"

l._______ ____ _______

~,,::----:::-~

: 1· - - - - - - : : - - - - - -

-<>.5

T l ft :uc)

Fig. 8 Fig. 6

Simulation result of depth control performance with vertical thrusters. (set depth : 2m)

Simulation result of route tracking ( upper : time series of depth, lower : time series of vertical thrust)

Figure 8 shows time series of depth and vertical thrust. Upper one shows time series of depth and lower one shows the vertical thrust. From the results, it is clear that the depth control is able to carry out well accuracy within 0.2m. Also, it understands that the thrust fluctuation of a vertical thruster is small except for the start of control.

@

15

Set Route

5. Full Scale Experiments

5.1 Position Sensor

~

--

The performance of the guidance and control system depends on the accuracy of the positioning system. The results of acoustic positioning test in the sea indicated the error on the average about O.3m. The tests were carried out by using the floating structure with the width about 60m and the length 300m. The six sets hydrophones were located about 4m below the bottom. The ROV with a responder cruised in the plane of about 1m below the bottom. The method of measurement was SBL with 40KHz. The depth information were measured with both ultrasonic gap sensor and pressure sensor. The results in the sea indicated the error on the average about O.OSm.

-5 End

® - 15

-25 -10

Fig. 7

10

20

Simulation result of route tracking ( set route and path in horizontal plane )

200

5.2 Result afFull Scale Experiments The full scale experiments of guidance and control were carried out by using the model floating structure of Mega-Float with 300m*60m as shown in Fig. 9. ~ c:

30CJ(m)

1

• Hydroptlone 2



"

• Hvdrophone 1



::J

• ,

.. I

RO V

Fig. 9

Model floating structure of Mega-Float.

Fig. 10

The ROV with a responder cruised in the plane of about lm below the bottom. The control system was composed with host processor and slave processor of DSP(Digital Signal Processor) driven by MAlLAB and REALTIME WORKSHOP as shown in Fig. 10. The main computer manages the hardware interfaces, such as AID and D/A converter, DIO and RS232C. These systems are wired as network with high-speed serial line.

Structure of the controller system.

The total systems of full scale experiment are shown in Fig. 11 . The horizontal position (x,y) are measured with six hydro-phones and the depth(z) is obtained by pressure sensor and ultrasonic gap sensor. The heading angle is measured with azimuth sensor. These measured signals, thruster commands and observed data are transmitted through optical tether cable between ROV and guidance and control system, etc.

Optical and Ultrasonic Data

1

Position (x.y.z)

Image Processor

Acoustic Positioning

I Joystick I I

Serial Signal

Optical Tether Cable Signals

ROV

CD
z (Depth.Height) X.y (Horizontal)

® Triuer of Responder @ @

Hydrophone (6sets)

Fig. II

Schematic view of full scale experiment.

201

Head.HeightDepth Thruster Command

The results of full scale experiment in route tracking are shown in Fig. 12 and Fig. 13 for current running at about O.lrn1s. The ROV starts from (80m,40m) and pass (lOOm,40m), (lOOm,20m), (80m,20m) then returns to (80m,30m) point last. The ROV cruised in the plane of O.5m below the bottom. (x, y) means x coordinate and y coordinate of Mega-Float.

ISO.O '/ : - -- - -""-~r~ 1. . ---"'~~~~_-.;,5.L',...;S'_'H_"~d_--""'---I

120.0·

'

--

---\ ' - ---

He .....

"t-+-------->-----<'-------i -- -- ------ ---

-lO.'

-- ----~- I

·60,0

- 150.0

., Tn. (wo)

Fig. 15

The ROV cruised in the plane of 2m below the bottom. Fig. 14 shows the horizontal path and Fig. 15 shows the time series of heading. The set head is estimated during in the route tracking based on own position and next desired position. Figure 16 shows time series of horizontal, vertical and lateral thrusts. The direction is controlled by only the difference in horizontal thrusters. At the time of the control starts, especially as for the exception, be not admitted the fluctuation of the large control command. The sample of current data shows in Fig. 17.

10

10 -

- - .

IU

Fig. 12

!O

lOO

11 0

Result of full scale experiment. ( time series of heading)

120

Result of full scale experiment. ( route tracking : set route and path )

--'E ~j

. Horuoftlal ( riJht l

}

i '0

Fig. 13

Result of full scale experiment. (time series of depth : set=O.5m)

- 3.0

Next results of full scale experiment are shown in Fig. 14 and Fig. 15. This experiment started from (90m,50m) and reached (90m,40m) point straight line target.

~, L---------------~~



Fig. 16

Result of full scale experiment.

0 . .30

0 . 25

~

..§.

i:i

~

0 . 010

0 . 15

0.10

Q

0 . 05 50.0

0 . 00

i

0 : 00

I.."" Fig. 14

.

8 :00

t2 : 00

16 :00

20 : 00

Time

( time series of all thrusts)

Fig. 17

l "

<1 :00

Sample of current data. ( velocity )

The main flow was longitudinal direction of the float and it was about O.Im1s. From the results of some full-scale experiments, it can be seen that the horizontal performance of this control system is within Im deviation, vertical one is within O.2m and heading one is within 5degree deviations.

i

Result of full scale experiment. ( path of moving between two points)

202

Oda, H., Masuda, K. and Karasuno, K. (1990). A portable automatic control system for ocean research operation of a ship with a controllable pitch propeller, a rudder and a bow thruster , 9-th SCSS. Perrier, M. and Canudas-de-wit, C. (1996). Experimental comparison of PID vs. PID plus non linear controller for subsea robots , Underwater Robot ,Kluwer Academic Publishers. Yuasa, H., Iyama, T. , Harada, H., Hirai, Y , Yarnada, M. and Okada, M. (1997). Underwater inspection ROV system for the bottom appearance ofMega-Float, OMAE'97.

6. Conclusions This paper described a guidance and control system of ROV for the undeIWater inspection at the model floating structure of Mega-Float. The ROV system is assigned not only to observe the bottom of the float but also to carry the sensor of electric potential, or support divers who work to investigation in advance, transport tools and keep safety. The navigation system involves the design of automatic speed control systems, system for dynamic positioning and tracking, as well as autopilot systems for both heading and depth control. Therefore, the navigation control system was developed using mathematical simulation and PlO control. Also, the power distribution method as a controllable design method using the weighting matrix on thrust allocation which have wide application among some vehicles. This paper describes the background of undeIWater inspection of the bottom of the float, guidance and control system with simulation study and full-scale experiments using MATLAB, SIMULINK and REALTIME WORKSHOP. From results of simulation and full-scale experiment, it can be concluded that the guidance and control system of the ROV system is very useful tool as below, . heading control : under 5degree . . depth control : under O.2m . track keeping : under Im . dynamic positioning: under O.2m These results suggest that the horizontal performance and vertical one is allowable limit. Acknowledgment This research and development were carried out under the Mega-Float Technical R & D Association. Authors would like to acknowledge Hajime Yuasa of Akishima Laboratories (Mitsui Zosen) Inc., Hidetoshi Harada of Mitsubishi Heavy Industries Co., Ltd. , and Tadahiro Iyama of Ishikawajima Harima Heavy Industries Co., Ltd. References Albert, c., Roberto, S. and Francesco, V. (1996). Using MATLAB : SIMULlNK and control system toolbox, A practical approach , Prentice Hall. Bishop, R.H. (1997). Modern control systems analysis & design : Using MATLAB & SIMULlNK , Addison-Wesley. Fossen, T.I. and Fjellstad, O.E. (1995). Robust A adaptive control of underwater vehicles comparative study , CAMS ' 95. Kapsenberg, G.K. (1985). A step towards the introduction of simulation techniques in the world of remotely operated underwater vehicles, ISP, No.368, Vo1.32 . Karasuno, K., Matsushirna, H., Oda, H. and Igarashi, K. (1993). Advanced portable control system under simple operation lO-th SCSS. I

203