Ship Maneuverability, Control and Safety

Transcription

1 Kazuhiko Hasegawa Ship Maneuverability, Control and Safety Kazuhiko Hasegawa Professor, Department of Naval Architecture & Ocean Engineering Division of Global Architecture Graduate School of Engineering Osaka University, Japan He is working for ship manoeuvrability and its control since He has graduated from Osaka University, Japan in 1974 (BS) and 1976 (MSc). He took his PhD in 1982 (Osaka University). In 1976 he was working as a research assistant in Hiroshima University, Japan and he joined Osaka University in He is professor there since He has also many experiences of short-term and long-term visiting scholars and professors in various universities; University of Twente, the Netherlands, University of Strathclyde, U.K., Pusan National University, South Korea and Ecole Centrale de Nantes, France. He was also invited by many universities and inter-national conferences to give seminars and keynote speeches from U.K., France, Finland, Germany, Belgium, Turkey, Tunisia, India, Bangladesh, Myanmar, Thai-land, Indonesia, Malaysia, Vietnam, China and South Korea, too. He is also contributing in various academic societies in Japan and also internationally. He is a technical committee member of Marine Application (TC7.2) of IFAC since 24. Kazuhiko Hasegawa Osaka University He has engaged in autonomous surface ship control and operation since 198, and has published first paper on automatic collision avoidance in He is using fuzzy logic and control to detect risk of collision as well as waypoint navigation. Since then, he is one of the leading researchers in this subject and many visiting scholars and students have learned in his laboratory. He is also one of the leading researchers in automatic birthing control for surface ships. Many researchers have tried to control a ship to birth, but failed because ship is quite unstable in horizontal plane in low speed and due to its strong nonlinearity and rather strong disturbances. He has also conducted many model ship and full-scale experiments to identify ship dynamics and test the control algorithm from his background in naval architecture and ocean engineering. One of the recent model experiments is to investigate the Korean ferry Sewol, which sank in April, 214. At this moment he is also engaging on signal conflict in AIS (ship-borne Automatic Identification System) communication and ship accident analysis including Costa Concordia. In the seminar he will briefly introduce these research activities to give latest visions in the field of ship automation and accident researches. 4

2 Osaka University Osaka University Towing Tank, Osaka University Free-running Pond, Osaka University 6 EXPERIMENT(1) PRINCIPLE OF ARCHIMEDES REVIEW OF HIGH SCHOOL PHYSICS BUOYANCY 1. Put stones etc. into a milk pack. Put it into water quietly supported by hands 2. Draw the draught (draft) line, when you feel the balanced point with the weight and the buoyancy 3. Estimate the weight you put into the milk pack. 4. Learn about the significant figures 7 8

3 RESULT OF EXPERIMENT(1) EXPERIMENT (2) STABILITY OF A FLOATING BODY What should you check? Describe it in your report - Try to keep photos, sketches etc. as objective as possible 1. Release the hands 2. and observe what may happen. 3. At the equilibrium (balanced) condition, push the milk pack with you hands a little and then release. Observe what may happen. - What did you feel after the experiment? - What did you take care, when you conduct the experiment? - The suggestions to make this experiment more interesting or accurate etc. 9 1 EXPERIMENT (3) RESULT OF EXPERIMENT (3) 1. Replace stones with water, then observe what may happen 1. Why? - Principle of science 11 12

4 DEVICES WE USE IN THE REAL EXPERIMENT ACCURACY, REPRODUCTIVITY AND OBJECTIVITY ARE REQUIRED WiFi antenna for communication between on-land and on-board computers Antenna for high accurate (RTK) GPS Computer for measurement and control PRECISE MEASUREMENT IS ONLY POSSIBLE USING THESE DEVICES GET MORE GENERAL CONCLUSION, COMPARING THE THEORY BEHIND OR WITH COMPUTER SIMULATION Trajectory Rate of turn Gyroscope for measurement of heading angle Motor for taking rudder Motor for main propeller Motor driver to convert command voltage with sufficient current to motors Ballast weights to 13 adjust dragnet line On-land computer to store measured data 14 PRINCIPLE OF SCIENCE IN REALITY Creativity, Originality, Humanity and Sociality are important factors, when you conduct experiment. July 23,

5 IT HAPPENS IN KOREA, TOO Occurred on April 16 th, 214, 34 passengers, mostly high-school students and crew were killed, 172 survived. snafu: a condition of being mess(us slung) 18 What may happen on this ship in the next moment? RESULT (1) 19 2

6 Ship Safety Research Osaka Initiative, University Osaka University RESULT (2) Ship Safety Research Osaka Initiative, University Osaka University Ship Safety Research Osaka Initiative, University Osaka University Structure and Fatigue Failure 23 Structure and Fatigue 24

7 Ship Safety Research Osaka Contents Initiative, University Osaka University Ship Safety Research Osaka Contents Initiative, University Osaka University Capsizing Dynamic Stability 2 26 Ship Safety Research Osaka Contents Initiative, University Osaka University Ship Safety Research Osaka Contents Initiative, University Osaka University Collision Agrounding 27 28

8 Support Vessel and its Simulator Offshore Support Vessels hyp:// EMAS, Singapore Simulator for Harbour Manoeuvring Simulator for River Barge Taiwan Ocean University, Taiwan hyp://telema`n.france2.fr/?page=chronique&id_ar`cle=47182

9 Ship Mo`on and Interac`on nearby a Lock River Transporta`on hyp:// Huanpujian, Shanghai, China Ship Mo`on in Current Ship Mo`on in Current Upper Mekong River, Myanmar Upper Mekong River, Myanmar

10 Where did it happen? INVESTIGATION OF THE ACCIDENT OF THE KOREAN FERRY SEWOL IN SOME FACTS OF THE INCIDENT MODEL SHIP MODEL SHIP (SCALE 1/7) Departure was delayed about 3 hours, because of deep fog. The officer ordered the course instead of rudder angle, when it changes the course at the place. It leaned about 4 deg. probably due to turning, then the engine has stopped and drifted by current. L (O.A.) 1.9 [m] L (P.P.) 1.76 [m] Bm.29 [m] dm 7.67 [m] Cb.46 The model is too small to install high accuracy instruments. 39 AIS record of Sewol 4

11 MEASUREMENT SYSTEM (2) - DIRECT ROLL ANGLE MEASUREMENT - Ship Safety Research Osaka Initiative, University Osaka University Camera 41 Sewol model experiment, Osaka University, Ship Safety Research Osaka Contents Initiative, University Osaka University TOP-VIEW, SIDE-VIEW AND ON-BOARD CAMERAS Sewol model experiment, Osaka University, 21 Ship Manoeuvrability Second A. Yucel Odabasi Memorial Colloquium, Nov , Istanbul, Turkey 43 44

12 Ship Safety Research Osaka Initiative, University Osaka University GM=.9 M, Δ=3 DEG Roll angle [deg] `me [sec] Ship trajectory 3 Rate of Turn [deg/s] [m/l] 2 1 Sewol model experiment, Osaka University, Time [sec] [m/l] 1 4. GM=.6 M, Δ=3 DEG Roll angle 1 6. GM=.3 M, Δ=3 DEG Roll angle [deg] -1-2 [deg] Time [sec] Ship trajectory Time [sec] Ship trajectory 3 Rate of Turn 4 3 r [m/l] 1 [m/l] [deg/s] 1 2 [deg/s] Time [sec] [m/l] Time [sec] [m/l]

13 8. GM=.2 M, Δ=3 DEG Roll angle Comparison btw Exp. and Sim. Roll angle (GM=.2, δ=3) (GM=.9, δ=3) 1 1 [deg] Time [sec] Ship trajectory [deg] Time [sec] Roll angle (GM=.3, δ=3) Maximum Heeling Angle =-19.2 deg. [deg/s] Rate of Turn [m/l] [deg] Numerical Simulation by KRISO p.86 of the Investigation Report, 21 Maximum Heeling Angle exceeds 2 deg. Free-running Experiments by Osaka University Time [sec] [m/l] Time [sec] Relation between GM and maximum heeling angle in initial turning Maximum Heeling Angle [deg] Maximum Heeling Angle =-19.2 deg. δ=1 δ=3 Estimated MOTIVATION GM [m] Numerical Simulation by KRISO p.86 of the Investigation Report, 21 2

14 3 4 FIRST RECOGNISED TIME AND DISTANCE from the court documents First recognised distance (km) (mile) CONCEPT OF MARINE ITS First recognised time (min) K. Ohnaka: Reliability of Watch Systems in Collision Avoidance, Journal of Navigation, Japan, Vol.62, pp.89-96,

15 AUTOMATIC COLLISION AVOIDANCE SYSTEM 1987 AUTOMATIC COLLISION AVOIDANCE AND MARINE TRAFFIC SIMULATION Original Course Of Own Ship Own Ship 9 6

16 SHIP AUTO-NAVIGATION FUZZY EXPERT SYSTEM (SAFES) MARINE TRAFFIC SIMULATION IN 198S AND BEFORE Own Ship Virtual Ship 2 Virtual Ship 2 Own Ship Virtual Ship 1 Original Course Of Own Ship Rudder Angle Collision Risk 63 [34] Y. Fujii, T. Makijima and K. Hara: Marine Traffic Engineering (in Japanese), p.13, Kaibundo Shuppan, Kobe, [3] A. Nagasawa: Marine Traffic Simulation Including Collision Avoidance, Navigation, Bulletin of JIN, 8, pp.28-34, June,

17 Ship Safety Research Osaka Contents Initiative, University Osaka University MARINE TRAFFIC SIMULATION TOKYO BAY, 199 Automatic Ship Collision Avoidance System in Tokyo Bay Marine Traffic Simulation 6 66 WITH AIS COMMUNICATION SIMULATION TOKYO BAY, 26 Ship Safety Research Osaka Contents Initiative, University Osaka University 67 Intelligent Ship Handling Simulator 68

18 Input layer Hidden layers Output layer Io An Am NEURAL NETWORKS v r Wno Wmn AUTOMATIC BERTHING/DEBERTHING ψ x y δ d1 d2 Wlm Ol δord Mean squared error (MSE) is used as evaluation function and Lavenberg- Marquardt algorithm in training. u nord 69 7 Start Goal 71 72

19 Initial positisons trajectory ACCIDENT OF COSTA CONCORDIA, ITALY, st Error Change in voyage plan 2 nd Error Incomplete route planning RESEARCH FOR ACCIDENT ANALYSIS 3 rd Error Wrong use of INS Chart Alarm 4 th Error Faulty route monitoring th Error Lack of crew challenge 6 th Error Fault in conning order 7 WMTC 76 21

20 First Error (Change in Voyage Plan) Second Error (Route Planning) Fact W1: Captain s Mentor in Giglio Island. Fact W2: Hotel Manager asks the Captain to sail past Giglio Island. Fact O2: Do not allow change in voyage plan. Fact O4: Do not allow without prior approval Organization Decision C3: Informal Procedure (O2, O4, C1, C2) = C3 Decision S2: Conducts route planning on large scale charts because of limited time and informal procedure. (Limited Time, C3) = S2 Decision C1: Change in voyage plan. (W1, W2) = C1 Decision C2: Do not want to take prior approval Decision C3: Informal Procedure (O2, O4, C1, C2) = C3 Fact CA1: Captain agrees whatever SOOW decides regarding voyage plan WMTC WMTC Third Error (INS Chart Alarm) Fourth Error (Route Monitoring) Fact S4: Chart alarm was set to go on if the radar distance is 2m or less. The alarm was not set for crossing 1m bathymetric line. Decision C3: Informal Procedure (O2, O4, C1, C2) = C3 Fact S2: Route planning conducted on large scale charts. (Limited Time, C3) = S2 Decision J4: No danger observed Fact J2: No crew challenge (J4, C3 = J2) Bathymetric Chart WMTC WMTC 21 8

21 Fifth & Sixth Error (Faults in Crew Challenge and Conning Orders) Guests in the Bridge No Danger No Danger Decision C: Captain s orders were not correctly transferred/ executed to helmsman. PREDICATE AND PROPOSITIONAL LOGIC Structure of a Logic Premise 1, Premise 2, Premise 3 Premise n, Predicate Variables Conclusion Example of a Logic Ship has speed. Engine stopped. Bow thruster shutdown Ship is uncontrollable Helmsman Fact S3: At the final stage of the approach the Captain took over command form SOOW. During this period no alarm/danger was observed and valuable time was lost. Predicate Logic Propositional Logic WMTC Logic Model: Agent Based Perception-Action Method AGENT BASED PERCEPTION-ACTION MODEL Schematic Representation of Crew Perception-Action THE FOUR DISCONTINUITY In Control Engineering WMTC

22 THE FOUR DISCONTINUITY BRUCE MAZLISH 1. The discontinuity between nature and religion - Nicolaus Copernicus: On the Revolutions of the Heavenly Spheres, The discontinuity between human and monkey - Charles Darwin,:On the Origin of Species, The discontinuity between me and others - Carl Jung: Die Beziehungen zwischen dem Ich und dem Unbewuβten, The discontinuity between human and robot

23 FUTURE ROBOTS SHOULD BE

24 THIS YEAR AT HAWAII, DEC. 216 CONCLUSIONS Ship Osaka Safety Contents University Research SUGGESTIONS We are seeking for safer sea including hard, soft and these integration system 9 96

25 IT WAS ONE SMALL STEP, BUT TOYOTA FOR EXAMPLE Mitsubishi (187- ), TOYOTA (189- ), Panasonic (1918- ), Honda (1946- ) and so on were founded and now they are world-leading companies, but did you know what is the origin of TOYOTA, for example? It has started from a company manufacturing human-powered textile loom machine in 189. Neil Armstrong, TOYOTA WHAT IS LOOMING MACHINE? TOYOTA HARD TIME IS THE CHANCE! Toyota Automatic Looming Machine 99 1