Home
Master`s degree thesis
Contents
1. modelname modelpath uigetfile emx 20SIM Model emx Select 20SIM Model Check if the user pressed cancel on the dialog if isequal modelname 0 isequal modelpath 0 disp Program cancelled return else disp User selected fullfile modelpath modelname end xxsimOpenModel fullfile modelpath modelname e process the model xxsimProcessModel Load mat data file fprintf Please select a SHIPX data press Enter to continue n forinti n pause dataname datapath uigetfile mat Matlab data mat Select SHIPX data Check if the user pressed cancel on the dialog i f isequal dataname 0 isequal datapath 0 disp Program cancelled return else disp User selected fullfile datapath dataname end load fullfile datapath dataname Select condition Define target vessel speed knot vessel velocities 1 852 3 6 foprintf nThe vessel speeds in knots are E PEINTET AG knot 7 Forint ble di vel str2num input Choose target vessel speed in knots s for i 1 length vessel velocities if abs vel knot 1 lt 0 1 velno 1 break end end Define target wave period wp 2 pi vessel freqs fprintf nThe wave periods are Tprinti ig wp Forintt Dl mla period str2num input Choose wave period s for 3 1 length vessel fregs if abs period wp 3 lt 0 1 62 freqno j break end end
2. examination and exclusion from all universities and university colleges in Norway for up to one year according to the Act relating to Norwegian Universities and University Colleges section 4 7 and 4 8 and Examination regulations paragraph 30 am we are aware that all papers assignments may be checked for plagiarism by a software assisted plagiarism check am we are aware that Aalesund University college will handle all cases of suspected cheating according to prevailing guidelines I we are aware of the University Colleges rules and regulation for using sources paragraph 29 Publication agreement ECTS credits 30 Supervisor Karl Henning Halse Agreement on electronic publication of master thesis Author s have copyright to the thesis including the exclusive right to publish the document The Copyright Act 82 All theses fulfilling the requirements will be registered and published in Brage HiA with the approval of the author s Theses with a confidentiality agreement will not be published I we hereby give Aalesund University College the right to free of charge make the thesis available for electronic publication X yes _ no Is there an agreement of confidentiality _lyes Ino A supplementary confidentiality agreement must be filled in If yes Can the thesis be online published when the period of confidentiality is expired yes _Jno Date 05 27 2014 ASSIGNMENT MASTER THESIS 2014 FOR
3. n 2 n 3 i 100 pa r r r r r r I 0 0 2 0 4 0 6 0 8 1 1 2 1 4 Offset m Figure 6 10 Horizontal Offset without Current Force on Load V 1m s 45 Depth m Depth m Horizontal Offset without Current Force on Load v 2m s L L L L _ L Offset m Figure 6 11 Horizontal Offset without Current Force on Load V 2m s Horizontal Offset without Current Force on Load v 3m s L L L L L r r r 2 4 6 8 10 12 O O Offset m Figure 6 12 Horizontal Offset without Current Force on Load V 3m s 46 Horizontal Offset without Current Force on Load v 4m s L L L DNV n 1 A n 2 n 3 j Depth m r F 5 10 15 20 25 Offset m Br O O Figure 6 13 Horizontal Offset without Current Force on Load V 4m s 6 4 5 Discussion e Vertical Offset The FEM vertical offset had a perfect compliance with DNV rules no matter how much FEM parts were set because both DNV rules calculation and FEM have analytical results of the problem Thus for cases where people only interest in vertical offset without current force a single straight line with one translational connector is sufficient enough e Horizontal Offset The result shows that 1 under small horizontal offset condition L lt 1 5 the offset at the bottom end of the cable has good compliance with DNV rules calculation result 2 under l
4. yb axis starboard z axis downward The origin o at the vessel s water plane and the z axis pointing through CoG e Hydrodynamic The hydrodynamic frame H Xp Yn Zn is not fixed to the hull it moves at the average speed of the vessel following its path The xn yn plane is located on mean water free surface and the origin o coincide with o at vessel s stationary status The positive xp axis points forward and it is aligned with the low frequency yaw angle The positive y axis points towards starboard and the positive z axis points downwards In this thesis the hydrodynamic frame moves with the harmonic motion of surge sway and heave that cannot be considered inertial But the wave induced hydrodynamic force is simplified as constant in hydrodynamic frame which means the vessel is assumed to take a straight course low frequency yaw angle is zero so that wave heading change is negligible and so does encounter wave period change due to surge sway 3 1 1 Reference Frame in ShipX In this simulation all vessel data directly comes from ShipX calculation ShipX is possible to compute the hydrodynamic data in the following points CoG centre of gravity CO same as o in the still water plane line on the centreline with the z axis pointing through CG The preferred point is op because it coincides with o at time t0 The ShipX axes are x axis backward y axis starboard z axis upward 3 1 2 Reference Frame in 20 sim In 20
5. zeta disp1l z q w K lambda L Appendix C 20 sim models digital Appendix D Video footage of simulation digital Appendix E 20 sim scripting library 64
6. 2 1 Standard Bond Graph Modelling Readers hereby are presumed to have basic knowledge about bond graph modelling a brief introduction of standardized bond graph modelling are given as follows The language of bond graphs expresses general class physical systems through power interactions The factors of power Effort and Flow have different interpretations in different physical domains In the following table effort and flow variables in some physical domains are listed Table 2 1 Effort and Flow Variables in Different Physical Domains Systems Effort e Flow f Force F Velocity v Mechanical i Torque t Angular velocity w Electrical Voltage V Current 1 Hydraulic Pressure P Volume flow rate dQ dt Temperature T Entropy change rate ds dt Thermal Pressure P Volume change rate dV dt Chemical potential m Mole flow rate dN dt Chemical Enthalpy h Mass flow rate dm at Magnetic Magneto motive force em Magnetic flux f The standard elements in bond graphs include but are not limited to R Elements C Elements I Elements Effort and Flow Sources Transformer Gyrator 1 junctions O junctions Each element is designed to be coded representing specific relations between Effort and Flow with only variation in parameters and vector size 2 2 2 Modified Bond Graph Modelling In conventional bond graph modelling each elements are fixated and constrained by its own funct
7. Coordinates no transformation Help Lancel Figure 6 3 Buoyancy Force Actuator The equation for the 1x1 buoyancy actuator is p e pgV Where p is the water density g is the gravitational acceleration and V is the submerged volume of the load 39 Actuator Properties Actuator Type Force Actuator 6x1 Actuator Name Added Mass and Drag Force Connection Points Connection Con d ConnectionPointl Body Break Connection E dit Name Position IO 0 0 m Orientation Eryant 0 O 0 Degrees Port Properties As Power Port a Body Coordinates In World Coordinates 2 Internal Coordinates no transformation Help Cancel Figure 6 4 Added Mass and Damping Force Actuator The setting inside 3D Mechanics only allows user to define diagonal inertia matrix for the object body while added mass is not always diagonal so the added mass inertia force and coriolis force have to be defined outside the 3D Mechanics The equation for the 6x1 added mass and damping force actuator is p e M i ij p f UMA p f U BG ij p f U p f U with U as the current velocity vector acting on the object The reason transformation jacobian J is not used here is because added mass and damping coefficient for an submerged object is constant with respect to the body fixed coordinate while floating object has a constant added mass and damping coefficient with respect t
8. In 3D Mechanics block an actuator puts torque before translational force as a result to connect bond graph model to 3D Mechanics block a transformer STF is added to shift the translational force above torque Force 3D Mechanics ge lt STF lt Eal Bond Graph Moreover to solve rigid body causality problem the causality of two ports in I element shall be both changed to indifferent 3 4 Parameters To input realistic data of rigid body inertia added mass hydrodynamic damping restoring force and environmental forces A MATLAB 20 sim connection is established to transform vessel data from ShipX to 20 sim using Marine control MSS toolbox and the new transcript feature in 20 sim 3 4 1 Calculation in ShipX In ShipX if the reference origin is set on the water plane the equivalent position vector R44 R55 R66 R64 are with respect to the origin on water plane instead of to the centre of gravity In order to generate ShipX data for MSS the vessel hull must be represented by table of offsets Fossen and Perez 2014 Normally 20 offset points on each half section will provide an adequate description of the section shape and assure that correct added mass and damping coefficients are 22 obtained In the pre process of VERES vessel response calculation the following calculation options must be chosen to comply with the MSS data structure e Ordinary strip theory recommended but other methods can be used e Added re
9. Keywords generic modelling simulation vessel crane interaction hydrodynamics 20 sim CONTENTS ASSIGNMEN NEEN 1 PREPAGE EE 3 ABSIR en KEE 4 CONTENTS 222262062 Aa EE 5 LIS TOP FIGURES EEN 7 LIST OF TABLES EE 8 TERMINOLOOGY EE 9 T INTRODUCTION M2272227 2022242700275422 id 10 1 1 BACKGROUND a a e cd o el dea reed bab AA 10 1 2 MOTIVATION AND OBJEOTIVE corra e e ed date ese pd er dado BANG 10 Ta ORGANIZATION OF THESIS EEN 11 A INTE PIO DO OG Vara and 11 1 5 LITERATURE AND PREVIOUS WORK sasaanannnunnnnunnnnnnnnnrunannrnnnunn narar a nanana rara LALA AAAA RADA L Annana a arannana 12 2 MULETEBOD DYNAMICS iii a a a a a 13 2 1 LAGRANIAN MECHANICS ciccici icccccececccececececececcacacacerercncacasesesencncnsnsensacncncasenesencncnsesensnencncnsenenencn 13 2 1 1 Lagrange Equations of the Second bing 13 22 EE re een A Ee 13 2 2 BOND GRAPH MODEIEING aaa haa a ann o o as 14 2 2 1 Standard Bond Graph Modelling aaa ie ee ad 14 2 2 2 Modified Bond Graph Modelling 000n00000anannnaannnnnnannnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnrnnnnnnrnnnnnnnnnne 14 2283 20 8im and SD Mechanics TODA BAK abah i ilagan DA S ia haois ua dee 14 3 VESSEL MODEL cuco 220902600 at a PEKE ed 17 3 1 REFERENCE FRAMES maba E EE A E E 802340 17 3 1 1 Reference Frame in ShIpX ccccccsscccccsscecccnsuesccnsauseccsaceecenscuscessauseessausesenscuesensausesssanaesenas 17 3 1 2 Reference Frame in 20 SIM EEN 17 3 2 MOTION EQUATION oe dera do 18 3 2 1 Equation
10. are the generalized coordinates and q are the generalized velocities The equations can be applied into multi body dynamics modelling where Newton Laws of each physical body can be expressed in its own body coordinates 2 1 2 Kirchhoff Equations In fluid dynamics the Kirchhoff equations named after Gustav Kirchhoff is derived from generalized Lagrangian equations describing the motion of a rigid body in an ideal fluid Kirchhoff 1877 1 e TE 5 a Lo m dav T OT 5 L IK BAT TUX EQ TR dar OT _ e do ee Qu p x fido Fr pido Where w and v are angular and linear velocity vectors at point x respectively I is the moment of inertia tensor m is body s mass fi is a unit normal to the surface of the body at point x p is a pressure at the point x On and F are the hydrodynamic torque and force acting on the body respectively Q and F likewise denote all other torques and forces acting on the body The integration is performed over the fluid exposed portion of the body surface 2 2 Bond Graph Modelling Bond graph is an explicit graphical tool for capturing the common energy structure of systems Complex systems can be described concisely by bond graph in vector form and the notations of causality provides a tool not only for formulation of system equations but also for intuition based discussion of system behaviour such as controllability observability fault diagnosis etc Bond Graph 2014 2
11. at the Offshore Simulator Centre OSC The model consists of several modelling methods from other researchers yet with necessary modification to make a compatible system The general second order differential equation of vessel hydrodynamics is from Fossen and Fjellstad 2011 MATLAB files for ShipX data transformation is from Fossen and Perez 2014 Bond graph modelling technique is from Pedersen 2008 Crane hydraulics and control model is from Chu 2013 Cable model inspired by Johansson 1976 Load model inspired by Halse et al 2014 2 MULTI BODY DYNAMICS 2 1 Lagranian Mechanics Lagrangian mechanics is a re formulation of classical mechanics using the principle of stationary action also called the principle of least action Goldstein 2001 Lagrangian function is the core of Lagrangian mechanics where the kinetic and potential energies of the constituents of the system in a generalized coordinates form the equation defined as L T V Where T is the total kinetic energy and V is the total potential energy of the system Torby 1984 2 1 1 Lagrange Equations of the Second Kind For any system with m degrees of freedom the Lagrange equations include m generalized coordinates and m generalized velocities Euler Lagrange equations also known as the Langrange equations of the second kind Euler 1744 state that for a holonomic system L q q t d OL OL dtXoq dq where j 1 2 m represents the th degree of freedom q
12. curve length do is the unstretched length of ds so that ds d 1 e where e is the local strain R is the tension at the lower end of ds and the resolve components along x y z axes are E E E de ds ds The resolved tension components at the upper end are ER T ds ds ds GE The net tension force components thusly are lt r Va _ d r Vas A ES de ST d ES JELC Let m be the mass per unit length of unstretched cable and X Y Z external force components per unit of unstretched length Also let u v w be the components of the segment velocity The equations of motion can then be written as u _ d r Yu E aa ds with corresponding equations for the y and z directions t represents time For the submerged cable the hydrodynamic forces are included in X Y Z 32 For two dimensional problems the equations may be expressed in terms of local tangential and normal components in the form E 2 p Z At QE 06 Z 2 T Os M Nat at g NG 00 where P Q are tangential and normal components of the external force u v represents the tangential and normal velocities and q is the angle between the tangent and the horizontal plane or some other fixed direction in the cable plane Through the integration of Routh s equation along characteristic lines in x y z t space for some simple examples like straight strings exact analytical solutions can be given However for problems like lifting cable or mooring line
13. emx in MATLAB Load vessel data mat in MATLAB xxsimConnect xxsimOpenModel Select velocity wave Assign parameters into atada heading period and 20 sim amplitude xxsimSetParameters Solana Figure 3 6 Parameters Process from ShipX to 20 sim 3 5 Model Assessment To assess the accuracy and authenticity of the bond graph model an evaluation between 20 sim simulation results and ShipX motion RAO is conducted The default ship s175 from ShipX database is used The time domain 6 DOF motion are compared in amplitude 3 5 1 Results Table 3 1 Vessel and Wave Data en 178 Verte o rt or ba ad ng 2 55 0 9 55 Wave Period s 4 0 5 16 16 1 20 25 30 iesen 24589470 Sampling section t 9900 10000 sec 0 568 Integration method Modified BDF ET rass SE The motion RAO in ShipX only gives one amplitude for one specific wave period and heading meaning there is no transient dissipating vessel motion in its natural period but only steady wave induced motion in wave period However due to the equation 20 sim model based on it allows vessel have transient dissipating motion in its natural period Thus the assessment will be conducted after the vessel in 20 sim reaching the steady state 24 RAO m m RAO m m RAO m m RAO m m RAO m m Period sec 200 100 y Heading deg 0 Sway RAO rad m on I E 0 02 O lt od Sp 200 200 100 100 Period se
14. external force velocity of the submerged object linear velocity potential energy angular velocity motion displacement 1 INTRODUCTION 1 1 Background Modern human civilization is driven by the gigantic energy consumption majorly dependent on fossil fuels discovery and production with a market share of 87 Oil remains the world s leading fuel accounting for 33 1 of global energy consumption where natural gas follows by 24 2 13000 Coal H Renewables Bt a IN Hydroelectricity 12000 H Nuclear energy M Natural gas dd 11000 A Sil 10000 a Ba a9 GA 91 KR 93 94 95 56 97 D I Figure 1 1 World Consumption Millions tonnes oil equivalent BP 2013 g OI OF 03 w 05 06 OF OB 10 11 12 Currently approximately 30 of world oil and gas production comes from offshore and a continuous increase is expected in the future The Petroleum Oil amp Gas sector is Norway s largest industry accounting for 22 of national value creation By 2010 Norway has produced and delivered about 40 of the expected total oil amp gas resources on the Norwegian continental shelf While 35 are reserves yet to be developed 25 are undiscovered resources two thirds of which are expected to be gas and one third oil Norweigian Petroleum Directorate 2010 The easiest barrels have been found and produced so that the way ahead will be demanding in terms of expertise technology and costs Over the last six decades offshore industry has witness
15. for multi body system modelling in a 3D workspace Rigid bodies are represented by 3D graphics and physical properties connected by joints monitored by sensors and actuated by actuators 3D mechanics toolbox can generate its model into 20 sim code ready for bond graph connection Forces can hence be applied to the joints or on each body directly and 3D mechanics toolbox itself provides options to define springs and dampers to the joint Finally the complete model can be shown by 3D animation toolbox 20 sim Editor on hexape Ele Edt View insert Model Drawing Settings Toos Hep i le OOX H Ht e E E 80 10 We AA mm A Ll e Po a m a gt i I t em l MB du RA je 18 x o Figure 2 1 20 sim Panel 3D Mechanics Editor Tripod 3dn da File Edit View Tools Actions Window Help Ds l lt 1 8 4 RL UU B E 2 WI Model Ta Library gs PALYA Ey Bodies ER Body 1 Joints gt OR Rotation gt 8 Translation CH Weld Joint Free Moving 6 DOF 1 Springs D Dual Quaternion Spring QD Twist Spring Sensors b Position Orientation gt CV Velocity gt A Acceleration J Jacobian Matrix Bxn ba A oh i 2hoir e Body Name Slider2 Show Frame Representation S TL File k F BAON 0 250 5 m Ki SS lt i HL 4 b M Messages 4 Figure 2 2 3D Mechanics Panel j 15 Comparing with direct bond graph modelling Fagereng 2011 which is unintuit
16. load of 50t or 100t have different steady angles 0 035 0 07 rad yet identical amplitude at steady state This is because although the load is heavy enough to cause dramatic leaning of the vessel it is still not heavy enough to significantly change the natural frequency of roll motion and the damping coefficient of the vessel is constant regardless of the leaning angle Therefore the amplitude of the forced oscillation of the vessel will have no observable changes 3 Motions without restoring force surge sway and yaw continue to deviate in the steady state as a result of the coupling effect between different degrees of freedom Counter measures including potential dynamic positioning method is to be expected The deviation is also due to the limited integration accuracy during the simulation 99 7 3 Vessel Crane Cable Load AHC 7 3 1 Model Vessel GlobalParameters inertia Conols Centnpetal Gangral amping Environmental force Cran Control AHC 0 mm MSe 11 mm PI ib Ba Joystick E mm MSe 41 i A boom acfuaforf Joystick1 FID boom base _actustori gt lolas Figure 7 12 Simulation Example Vessel Crane Cable Load AHC With the help of AHC system and possibly other control scheme the crane can not only be manually but also automatically controlled The video in the appendix also showed how the load is being controlled by AHC system 7 3 2 Results Figure 7 13 is the heave moti
17. miiimmmmmrmmmmmm mrmrmm mrmmrmmmmmmmmi 41 6 4 1 DNV Rules Cale ANON tostada ta at ES A pa aan maa leks a 41 6 4 2 ANA BE E ge EE 42 6 4 3 Cable 22172149 EE 42 6 4 4 Ee EE EE EE 43 6 4 5 Tee ee 47 7 SIMULATION EXAMPEE 0 BG AA 49 7 1 NSS PO RAN AA reacia 49 7 1 1 ee EE 49 7 1 2 SS Aa Nana A ee 50 7 1 3 Be ia AA Ee Ee 52 Tie VESSELECRANE E CABLE LOAD kanaba a a haa naaa La aba aka La aaa 52 7 2 1 MOUE PA O E A BENG na Ba 52 7 2 2 EE 54 7 2 3 A O anna eee as ENG akc PG nee eens tea A OILL A tg Eb E 55 Ta MESSEL GRANE4GABLE LOAD E A A a Naa a aani Ciud 56 7 3 1 7 2 10 AORA NABA A BAGS NG BINAGO OT 56 LO O OS A A A A 56 7 3 3 PIS CUSSION AA 58 8 CONCLUSION amp FUTURE WO RK see ENKEN KEREN R ENKEN KNRKNRKNKNENEN ENEE KEEN K ENKER ERKENNEN 59 8 1 CONG ERT EG 59 82 FUTURE WORK add erotica 59 REFERENCES nta GAN UGAT UGE NANA AA O O AA 60 SEN C2279222 Eege ege 61 APPENDIX A USER MANUAL Aen o ed ed NOSE ee 61 APPENDIX B MATLAB SCRIPT Na oe ta EE 62 APPENDIX C 20 SIM MODELS ADICTA A SAA 64 APPENDIX D VIDEO FOOTAGE OF SIMULATION DIGITAL aa 64 APPENDIX E 20 SIM SCRIPTING LIBRARY Ra a AN NAN We a IKAN ieee 64 LIST OF FIGURES FIGURE 1 1 WORLD CONSUMPTION MILLIONS TONNES OIL EQUIVALENT BP 20121 10 FIGURE 21 20SM PANE HE 15 FIGURE 230 MECHANICS GE 15 FIGURE 3 1 BODY FIXED REFERENCE IN SHIPX AND 20 SIM ri 18 FIGURE 3 2 BOND GRAPH OF DGIMPLEVESSEL 000 20 FIGURE 3 3 VESSEL MODEL NZ2DMECHANICR 00 aa 21 FIGURE 3 4 ACT
18. the perturbations are even assumed to be harmonic Then the time derivatives can be further eliminated from the equations which become ordinary differential equations solely about the cable length In this thesis the maximum length of lifting cable can reach thousands of meters which makes deformation along the cable significant and irregular The linearization method does not suit for this purpose 5 3 Cable Model in 3D Mechanics In this model a combination of different types of FEM cable model is used in 3D Mechanics The cable is modelled as rigid bars with free rotational joint and 1 DOF translational joint on two ends respectively The rigid bars are represented by a inertia matrix which has both mass and moment of inertia The free rotational joint has no spring or damping but constrained all translational motions between bars The 1 DOF translational joint has spring and damping property which represent the elasticity and structural damping of the cable respectively The joint connecting the crane tip is free rotational and the joint connecting to the load can be either free rotational or translational depending on the structure of the connecting point The elasticity mass moment of inertia and structural damping of the cable are all realized in this modelling method If the number of discretized parts in FEM is n the number of rigid bars free rotational joint and 1 DOF translational joint shall be 2n n n respectively not including th
19. the project however is based on previous works of others Thorough reference has been made accordingly The thesis project is part of the ongoing research in the Ship Operation Lab at Aalesund University College whose activities support the ongoing development of the activities at the Offshore Simulator Centre OSC The research outline was introduced in one of the previous papers Lifting Operations for Subsea Installations Using Small Construction Vessels and Active Heave Compensation Systems A Simulation Approach from Ship Operation Lab The target is to develop generic modelling method for maritime simulation as a platform for both training and design purpose The thesis project is also an upgrading of my earlier model developed in the summer of 2013 The project is an interdisciplinary task requiring different aspects of knowledge One of the biggest challenge in the research development is to integrate sub systems from various sources into one compatiable system This is also the keypoint for the standardization and expansion of a generic modelling method would like to thank my supervisor Associate Professor Karl Henning Halse for his guidance and support throughout the entire thesis composition The new ideas inspired and new fields of knowledge introduced by him has always been fruitful to me would also like to thank Associate Professor Vilmar Aray for his assistance on 20 sim modelling and project ideas enjoy his teaching very muc
20. 2001 Classical Mechanics 3rd ed Addison Wesley Halse Karl H Vilmar s y Dimitry Ponkratov Yingguang Chu Jiafeng Xu and Eilif Pedersen 2014 Lifting Operations For Subsea Installations Using Small Construction Vessels and Active Heave Compensation System A Simulation Approach OMAE Janssen Sander 2013 A Conversion from SolidWorks to 20 sim through COLLADA Controllab Product BV Johansson Per l 1976 A Finite Element Model for Dynamic Analysis of Mooring Cables Kirchhoff G R 1877 Vorlesungen ueber Mathematische Physik Mechanik Leipzig Lloyd A R J M 1989 Seakeeping Ship Behaviour in rough weather Ellis Horwood Limited Pedersen Eilif 2008 Bond Graph Modeling of Marine Vehicle Dynamics Trondheim Sanfilippo Filippo Hans petter Hildre Vilmar Asoy Hou Xiang Zhang and Eilif Pederson 2013 Flexible Modeling and Simulation Architecture for Haptic Control of Maritime Cranes and Robotic Arms Torby Bruce J 1984 Advanced Dynamics for Engineers CBS College Publishing 60 APPENDIX Appendix A User Manual The manual explains the steps from ShipX calculation to mat file transfer and eventually 20 sim simulation 1 Install MSS toolbox into your 32 bit Matlab see http Awww marinecontrol org download html 2 Run SHIPX Veres get re1 re2 re7 re8 hyd data files see 2008 06 19 MSS vessel models pdf page 18 20 If you do not have re2 file comment on line 131 and line 1
21. 36 in veres2vessel m the following calculation options and settings must be chosen to comply with the MSS data structure i Ordinary strip theory recommended but other methods can be used li Added resistance Gerritsma amp Beukelman iii Generate hydrodynamic coefficient files re7 and re8 iv Calculation options choose z coordinates from CO Ob V Vessel velocities must always include the zero velocity it is optional to add more velocities that are needed for manoeuvring vi Wave periods it is recommended to use values in the range 2 0s to 60 0s vii The wave heading must be chosen every 10 deg starting from 0 deg 3 Put SHIPX Veres data files vereas2vessel m into the same folder and set it as the working folder in Matlab 4 Type veres2vessel input in Matlab generate mat file see 2008 06 19 MSS vessel models pdf page 20 21 Add 20SIM script toolbox library and its subfolders into your Matlab toolbox path Run simulation m and follow the commands Modify the parameters block in 20SIM model to modify the rest parameters to run different simulations 8 With 3D animation plot you can directly compare the difference between results from SHIX grey and 20SIM cyan 61 Appendix B Matlab scripts e 20 sim runnine clear 33 Connect to 20 SIM xxsimConnect e Open the model Simple vessel emx fprintf Please select a 20SIM model press Enter to continue n torann pause
22. 5 Define target wave heading hd vessel headings 180 pi fprintf nThe wave headings are fprintf 0 tnd EE TA a head str2num input Choose wave heading s for k 1 length vessel headings if abs head hd k lt 0 1 headno k break end end ForceRAO forceRAOamp zeros 6 1 forceRAOphase zeros 6 1 for 1 1 6 forceRAOamp 1 1 vessel forceRAO amp i freqno headno velno forceRAOphase i l vessel forceRAO phase i freqno headno velno end MotionRAO motionRAOamp zeros 6 1 motionRAOphase zeros 6 1 for 1 1 6 motionRAOamp 1 1 vessel motionRAO amp i freqno headno velno motionRAOphase 1 1 vessel motionRAO phase i freqno headno velno end 55 Curve for Ca and Cd of the object Za rand 7 8 3 CFD matrix for Ca function of period and amplitude Zd rand 7 8 3 CFD matrix for Ca function of period and amplitude 6 p 3 6 Q 3 o aaCla fitting p q Za Curve fitting of Ca aaCd fitting p g Zd Curve fitting of Cd Ca 1 3 Added mass coefficient Cd 1 2 Damping coefficient 55 Set parameters Transfer nested variables xxsimSetParameters g Lpp T EIDEN Mg OML GMT EC Ca La MRE Devo a wd Ma C forceRAOamp forceRAOphase forceRAOw motionRAOamp motionRAOphase motionRAOw vessel main g vessel main Lpp vessel main T vessel main ho vessel main myvessel main GM Ly vessel main GM T vessel main C B C
23. 6 1 CFD EXPERIMENT TO TEST amp VERIFY VELOCITY ACCELERATION DEPENDENT COEFFICIENT 38 FIGURE 6 2 IkOAD MOPELIN 30 MECHANTC Sui didas aiii isis 38 FIGURE 6 3 BUOYANCY FORCE ACTUATOR aliada ANA iaa a 39 FIGURE 6 4 ADDED MASS AND DAMPING FORCE ACTUATOR c cccececececacaciciciciciciccecececececacacacacacass 40 FIGUREO 5 CABLE LOAD MODEL iaa linia 41 FIGURE 6 6 HORIZONTAL OFFSET WITH CURRENT FORCE ON LOAD VaslMS nee rer ne re ee 43 FIGURE 6 7 HORIZONTAL OFFSET WITH CURRENT FORCE ON LOAD MaszMlg 44 FIGURE 6 8 HORIZONTAL OFFSET WITH CURRENT FORCE ON LOAD Mas MlS 44 FIGURE 6 9 HORIZONTAL OFFSET WITH CURRENT FORCE ON LOAD VasdMlS 45 FIGURE 6 10 HORIZONTAL OFFSET WITHOUT CURRENT FORCE ON LOAD VaSIMIS 45 FIGURE 6 11 HORIZONTAL OFFSET WITHOUT CURRENT FORCE ON LOAD VaS2MIS 46 FIGURE 6 12 HORIZONTAL OFFSET WITHOUT CURRENT FORCE ON LOAD V aS2MIS 46 FIGURE 6 13 HORIZONTAL OFFSET WITHOUT CURRENT FORCE ON LOAD VazAMIS 47 FIGURE 7 1 SIMULATION EXAMPLE VESSEL RAN 49 FIGURE 7 2 ANIMATION VESSEL CRANE sada aa Paa balaak 49 FIGURE 7 3 VESSEL MOTION VESSEL CRANE AT TRANSIENT DTATE 50 FIGURE 7 4 VESSEL MOTION VESSEL CRANE ROLL AT STEADY STATE 2 7 7 7 aaa 51 FIGURE 7 5 VESSEL MOTION VESSEL CRANE FROM GHIDA aa 51 FIGURE 7 6 SIMULATION EXAMPLE VESSEL CRANE CABLE LOAp nee rr e earn er nee 52 FIGURE 7 7 ANIMATION VESSEL CRANE CABLE LOAD OO 53 FIGURE 7 8 ANIMATION DYNAMIC BEHAVIOUR OF CABLE Loa
24. CONCLUSION amp FUTURE WORK 8 1 Conclusion This thesis demonstrated a generic modelling approach for marine operation The aim is to propose a standard application of modelling each sub system in the marine operation however the design is The model consists of several modelling methods from other researchers yet with necessary modification to make a compatible system The general second order differential equation of vessel hydrodynamics is from Fossen and Fjellstad 2011 MATLAB files for ShipX data transformation is from Fossen and Perez 2014 Bond graph modelling technique is from Pedersen 2008 Crane hydraulics and control model is from Chu 2013 Cable model inspired by Johansson 1976 Load model inspired by Halse et al 2014 Combined the knowledge of previous researches the modelling method proposed in this thesis enables user to assemble all sub models into a complete operation model The complete model can then be served for virtual dynamic analysis or training purpose In the complete model most of the 3D mechanics modelling are done inside the 3D Mechanics a toolbox of 20 sim but with connectors actuators and sensors the 3D Mechanics model receives information and interacts with bond graph and control scheme outside The bond graph and control scheme expand the function of 3D Mechanics model and give user the freedom to modify the design The realistic physical entities interacts as power flow and information interacts as sign
25. HOGSKOLEN ALES UND Aalesund University College Master s degree thesis IP501909 MSc thesis discipline oriented master A Generic Modelling Approach for Heavy Lifting Marine Operation Jiafeng Xu Number of pages including this page 67 Aalesund 05 27 2014 Mandatory statement Fach student 1s responsible for complying with rules and regulations that relate to examinations and to academic work in general The purpose of the mandatory statement 1s to make students aware of their responsibility and the consequences of cheating Failure to complete the statement does not excuse students from their responsibility Please complete the mandatory statement by placing a mark in each box for statements 1 6 below I we herby declare that my our paper assignment is my our own work and that I we have not used other sources or received other help than is mentioned in the paper assignment i I we herby declare that this paper 1 Has not been used in any other exam at another department university university college Is not referring to the work of others without acknowledgement Is not referring to my our previous work without acknowledgement Has acknowledged all sources of literature in the text and in the list of references Mark each box Is not a copy duplicate or transcript of other work am we are aware that any breach of the above will be considered as cheating and may result in annulment of the
26. STUD TECHN JIAFENG XU A Generic Modelling Approach for Heavy Lifting Marine Operation Background Vessel crane simulation has always been a helpful tool for marine operation training operation planning and product designing Most designing works of crane and vessel require simulation results of their hydrodynamic behaviour while most simulations so far regard crane and vessel isolated without physical interaction with each other the vessel always stays on its weight distribution equilibrium and the crane always experiences a given vessel motion no matter changes on loaded weight and operating posture This issue often lowers the final performance of AHC Active Heave Compensation and ballast water control hence worse positioning of the submerged load Objective Research Study e Lagrangian physics eo Maritime Guidance Navigation and Control theory o Specification rules e Vibration dynamics e 20SIM Modeling and programing skills e CFD Work e Dynamic modelling of vessel crane cable load and other appurtenance eo Implementation of AHC control joystick control e Load hydrodynamics correlation analysis e Integration of all sub models e Simulation test for different vessel crane parameters e Result analysis The thesis should be written as a research report with summary conclusion literature references table of contents etc During preparation of the text the candidate should make efforts to create a well arranged an
27. Solidworks Parts Agan bla COLLADA Convert to 20 sim Figure 4 2 Steps from Solidworks to 20 sim Tips e The model shall be simplified as much as possible Unnecessary parts that have no effect on dynamic behaviour of the model should not be included eo Mates of parts shall be designed correctly Every mate will result in a joint in 3D Mechanics Only standard mates are allowed e Mass and inertia need to be assigned to each body manually after the conversion e The setup of the coordinate system will need to be adjusted if one would like to use e g the Denavit Hartenberg method D H method in the control scheme e After converting the crane model into 20 sim 3D Mechanics connect the foundation of the crane with the vessel hull on the desired location by using a welded joint e The Solidworks COLLADA exporter can be downloaded at http labs solidworks com Products Product aspx name colladaexport Prerequisites for installation and instructions can be found through the same link as well e NVIDIA PhysX System Software may be required if installation failed in some cases PhysX 9 12 1031 can be downloaded at http www nvidia com object physx 9 12 1031 driver htm Figure 4 3 Crane Model in 3D Mechanics 4 2 Hydraulic System 20 sim provides hydraulic components library for modelling and simulation hydraulic systems In Chu s work the model is designed as close as possible to the Modelica hydraulic library Controllab Produc
28. UATOR IN 3D MECHANICS TOOLBOX ananannnnnnannnnnnnnnnnnnnnnnonrnrnnonrnrnronrnrnrnnrnrnrnnrnrnrnrnnrnrnnnn 22 FIGURE 3 5 USER PANEL FROM MAT FILE TO RUNNING 20 SIM MODEL cc cc ccc sen ne nearer nee e re r ee H 23 FIGURE 3 6 PARAMETERS PROCESS FROM SHIPX TO 20 SImM 24 FIGURE 3 7 MOTION RAO FROM 20 SM GA NANA ARN ANA A NENA 25 FIGURE 3 8 MOTION RAO FROM SHIP Asan NGA da aci n 25 FIGURE 3 9 ERROR BETWEEN 20 SIM AND SHIPX Aa 26 FIGURE 4 1 SOLIDWORKS MODEL OF SIMPLIFIED ROLLS ROYCE RGV 200 27 FIGURE 4 2 STEPS FROM SOLIDWORKS TO 20 SIM 00 aa 28 FIGURE 4 3 CRANE MODEL IN 3D MECHANICS ee MGA ie 28 FIGURE 4 4 SIMPLIFIED MAIN DERRICK HYDRAULIC SKETCH cocooccnccnccnccnccnnnnncnncnncnnnnnnnncnnnnnconnonnnnnnnnnnnnnnnnnnns 29 FIGURE 4 5 BOOM CYLINDER HvDpAULicCeGMopnEt 29 FIGURE 4 6 BOOM CYLINDER SETUP IN 3D MECHANICAL 30 FIGURE 4 7 CRANE TIP VELOCITY CONTROL dias ae e NAG Ge ee etek 31 FIGURE 4 8 DIRECT CYLINDER VELOCITY CONTROL aasanannnnnnnnnnnannnnnnnrnnnnonrnrnnnnrnrnronrnrnrnrrnrnrnnnnrnnnrrnnnrnrnn 31 FIGURE 4 9 CYLINDER VELOCITY INPUT AND PID CONTROL nne 31 FIGURE 4 10 COMPENSATING MOTION VELOCITY ON OPPOSITE DIDECTION 31 FIGURE 5 1 FOUR TYPES OF FEM CGApLEMopnEL 33 FIGURE 5 2 FEM CABLE MODEL IN 3D MECHANICAL 34 FIGURE 5 3 EXTERNAL FORCE ACTING AS ACTUATOR asnnnnnnenannnnnnnnnrnrnnnnnrnnrnrnrnnnrrnrnrnrnrnrrnrnrnrnrnrrrrnrnrennne 35 FIGURE 5 4 CABLE MODEL N 3 i VO aman A a 35 FIGURE 5 5 EXTERNAL FORCE Be EE 36 FIGURE
29. a Cd vessel MRB pas vessel B freqno velno vessel A freqno velno vessel C freqno velno forceRAOamp forceRAOphase vessel forceRAO w 1 freqno motionRAOamp motionRAOphase vessel motionRAO w 1 freqno 6 Run the model XXSimRun 63 DNV checking for cables Matlab script for calc cable responses for crane operations Input variables 9 81 3 Acceleration of gravity m s 2 ho 1025 3 density of water kg m 3 100 Length of cable m 0 05 3 Diameter of cable m 2 1E11 3 Youngs modulus N m 2 c pi D 2 4 3 Cross section area m 2 e tr lt DOA C Brrective area WQ 0 o de dp D HUH O H 8 Mass of cable m 10 2 Mass of cable kg m w Mm g rho g 5A Gif weight of cable in water N m Cd c 1 07 Drag CEL for cable Mass of load M 20000 3 Mass of load kg V 10 W M g rho V g 3 Weight of load N A N 0 Projected area of load m 2 Cd W 1 27 Drag CEL tor load Current velocity U 4 Current velocity m s AO AO Drag force on lifted object Ed load Va 5 Xfh0 Cd WA WU Z7 AO ANO Drag force per unit length of cable q Veo INO CA C D U 27 Weight of Load weight of cable K W w L Drag of Load weight of cable lambda Fd load w L oP e Determine the shape of the cable dz 0 2 Element length ai 207 lt L displ function to calculate static displacement due to current
30. al flow which gives a clear idea to separate the different level of modelling Besides crane a winch a propeller or a drilling tower can also be modelled by following the same idea 8 2 Future Work As a part of the ongoing research there are many things yet to be done So far the modelling method is still based on 20 sim which has many restrictions e g only diagonal matrix of inertia tensor is allowed inside 3D Mechanics only linear spring damping is allowed inside 3D Mechanics the function of real time interaction with Matlab has not been realised In the future development a better platform or direct programming of physical engine may be required to expand the function of the model Also in 3D Mechanics the model is eventually being solved as a big matrix yet only occupying single CPU thread which is time consuming and inefficient Multithreads solution using GPU and parallel computation should be applied in the future The model also requires collision property and event detector to control the different state of marine operation e g crossing splash zone will have a different model which requires splash model and detectors to detect the start and the end of the phase The ideal model shall have the function of Plug and Play which minimize the modelling procedure and maximize the flexibility for different models Thus a standard interface between sub models is also required Differernt parties shall have the same UI to develop th
31. ard to identify the effect in the entire model Moreover the cable model will significantly increase the complexity of the model hence the computation amount As stated in Chapter 6 for most cases the cable can be represented as an straight bar Figure 7 8 Animation Dynamic Behaviour of Cable Load 53 7 2 2 Results Figure 7 9 shows the 6DOF motion of vessel model with cable and 50t of load in 20 sim during the start of the simulation transient state Tn O 001 a AA AAA RA ANA 001 AU TATA 0 02 0 02 m Yaw 0 015 0 01 0 005 O 100 200 300 400 5 time s Figure 7 9 Vessel Motion Vessel Crane Cable Load 50 t at Transient State 0 03 0 035 0 04 0 045 1000 1020 1040 1060 1080 1100 1120 1140 1160 1180 tim O O e iS Figure 7 10 Vessel Motion Vessel Crane Cable Load 50 t Roll at Steady State al A 0 06 0 065 0 07 0 075 0 08 1000 1020 1040 1060 1080 1100 1120 1140 1160 1180 time s Figure 7 11 Vessel Motion Vessel Crane Cable Load 100 t Roll at Steady State Figure 7 10 and 7 11 show the roll motion of vessel model during transient state with the load weight of 50t and 100t respectively 7 2 3 Discussion As the plot shows 1 After putting load and cable into the model the roll motion of the vessel has higher irregularity at the transient state and leans to the load side due to the pulling force 2 The roll motions of the vessel with
32. arge horizontal offset condition lt L 5 1 5 the bottom end of the cable has significant deviation from DNV rules calculation result 3 under small horizontal offset condition nonlinearity shape of the cable trajectory does not affect the accuracy of FEM model The reason for significant deviation under large horizontal offset condition is that by the definition of DNV rules For an axially stiff cable with negligible bending stiffness the offset of a vertical cable with a heavy weight at the end of the cable in an arbitrary current with unidirectional x direction velocity profile is W w z L Sr Ie the formula is only valid under small angle approximation where the vertical depth of the cable equals to the axial length of the cable sin 0 tan0 0 when 0 is small angle 47 yet large horizontal offset cases L gt 1 5 cannot be treated under small angle approximation anymore where the resultant cable length under large horizontal offset is significantly longer than the depth The large horizontal offset condition also contradicts the premise of a vertical cable with a heavy weight at the end of the cable from the description of the formula However the FEM model in 3D Mechanics is built according to a fixed cable length Large horizontal offset will cause the end of the cable to be lifted hence a smaller depth From the small horizontal offset cases the validity ot FEM model has been proved For large horizontal off
33. blished in body fixed coordinate system Der Yo Zb in order to keep the rigid body inertia and added mass as constants The restoring force and damping force are treated with respect to the hydrodynamic frame H Xh Yn Zn Therefore a coordinate transformation is required before those two elements are being put into the equation 3 2 2 Coordinate Transformation To transform hydrodynamic excitation and restoring forces from H Xn Yn Zn to B Xb Yb Zb a coordinate transformation matrix is established The principle rotation matrices one axis rotations can be obtained by setting x y and z axes rotation as 0 0 Rag o coso singl O sinp coso 0 0 R yo cos sino sino cos0 19 1 0 0 Rzy E cost inp 0 sinp cosy The rotation matrix for linear velocity is R Rg Ban Rzy The rotation matrix for angular velocity is T 0 coso sing 1 sind tan0 coso EN O sinp cos0 cosd cos The overall 6DOF kinematic equation between H Xh Yn Zh and B Xp Yb Zb IS R 0 ta eee h b 3x3 And the final equation used in bond graph model is Mag M4 0 jt Crp Cato NO J 14 Bio JE H t J 14 C gt TO j Texc T T 3 3 Bond Graph Model Vessel Inertia Gonolis Cenfnpetal Environmental force Wave fome Se Vessel Figure 3 2 Bond Graph of Simple Vessel 3D Mechanics toolbox does not have the option for a full inertia damping spring tensor property so all elements of ship motio
34. c G Heading deg Period sec SH Heading deg Yaw E 0 5 has z o Q j 200 or 200 100 20 100 Period sec G Heading deg Period sec de Heading deg Figure 3 7 Motion RAO from 20 sim Roll RAO m m Surge 100 b Heading deg Period sec 0 Sway 100 Heading deg 400 Period sec Heading deg RAO rad m 200 RAO rad m 200 200 0 047 0 02 A lt 200 20 100 Period sec 0 0 Heading deg E D 2 O lt Ke 200 100 Heading deg Period sec Figure 3 8 Motion RAO from ShipX 25 Roll E E o S o O O lt lt S 40 200 200 100 100 Period sec D Heading deg Period sec uil Heading deg Sway 0 2 0 05 E E o E 0 O 2 O E ma l PO Y HT lt 200 p 200 100 100 Period sec D Heading deg Period sec a Heading deg Heave Yaw 0 05 E E E oi E O 2 O 100 0 0 i 0 0 i Period sec Heading deg Period sec Heading deg Figure 3 9 Error between 20 sim and ShipX 3 5 2 Discussion The results show good matching between 20 sim and ShipX in most wave periods and headings yet poor accuracy in vessel s natural rolling period 18 sec There is also relatively high error of surge and pitch in extremely long and short period wave The reason is that all hydrodynamic coefficients in 20 sim are from potential theory but ShipX has additional viscous damping coefficient in its calculation The error exagg
35. ct force The object has its inertia matrix constantly through all phases but different damping and restoring functions respectively e g the air resistance is mostly negligible while the object is in air but the hydrodynamic resistance is significant in the water the buoyancy force varies throughout the splashing phase yet stays constant after the object is full submerged In this thesis a general method of expressing the physical properties of the object in deeply submerged phase is given which can be without loss of generality applied in other phases 6 2 Load CFD 6 2 1 Motion Function of the Load When an object is deeply submerged it experiences a concentrated scattered lifting force from the cable an evenly distributed gravity force downwards and buoyancy force upwards under the premise of a constant submerged volume of which It also experiences hydrodynamic forces including the inertia force and the damping force by the definition in the Morison equation Similar to the ship motion equation for a harmonically oscillating object the motion function of an submerged object in its own body fixed coordinate can be written as Mre H t Malo Ny t Crp C Calo O af t Bo g ilil t Ty Tp where M pp rigid body inertia matrix of the object 0 Oscillation frequency of the object OSCillation amplitude of the object M w C added mass matrix for a specific w and Cpp coriolis force matrix of the obj
36. d well written report To ease the evaluation of the thesis it is important to cross reference text tables and figures For evaluation of the work a thorough discussion of results is needed Discussion of research method validation and generalization of results is also appreciated In addition to the thesis a research paper for publication shall be prepared Three weeks after start of the thesis work a pre study have to be delivered The pre study have to include e Research method to be used e Literature and sources to be studied e A list of work tasks to be performed e An A3 sheet illustrating the work to be handed in A template for the A3 sheet is available on the Fronter website under MSc thesis This sheet should also be updated when the Master s thesis is submitted The thesis shall be submitted as two paper versions One electronic version is also requested on a CD or a DVD preferably as a pdf file Supervision at Aalesund University College Karl H Halse Karl H Halse Jiafeng Xu Program coordinator Stud Techn Delivery 30 05 2014 PREFACE This thesis is submitted in partial requirements for a Master s Degree in Ship Design for the auther It contains work done from January to May 2014 at Aalesund University College The supervisor of on the project has been Associate Professor Karl Henning Hals Faculty of Maritime Technology and Operations AMO The thesis has been made solely by the auther some contents used in
37. e heading wave period and amplitude Several functions from 20 sim MATLAB connection toolbox are being used here gt gt xxsimConnect MATLAB connects to 20 sim gt gt xxsimOpenModel filename Open 20 sim model gt gt xxsimProcessModel Process the model gt gt xxsimSetParameters parameternames values Input variables from Matlab to 20 sim gt gt xxsimRun Run the 20 sim model Full script and user mannul can be found in Appendix Command Window gt gt simulation Please select a 20SIM model press Enter to continue User selected C 1Users Jiafeng Documents Master in Ship Design Master Thesis Appendix Model Vessel 5 Please select a SHIPX data press Enter to continue User selected C 1Users Jiafeng Documents Master in Ship Design4Master Thesis Result run5 mat The vessel speeds in knots are 0 Choose target vessel speed in kmots 0 The wave periods are 30 25 20 19 18 17 16 15 5 15 14 5 14 13 5 13 12 5 12 11 5 11 10 5 10 9 5 9 8 5 Choose wave period 10 The wave headings are 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 Je Choose wave heading 60 na lt gt Figure 3 5 User panel from mat file to running 20 sim model 23 The entire process from ShipX to 20 sim can be described as follows MSS toolbox Generate hya rer enerates mat data ShipX calculation re2 re7 re8 data i file files veres2vessel Connect MATLAB to Open 20 sim model 20 sim
38. e joint on crane tip and load lt 1 DOF Translational Joint Free Rotational Joint Figure 5 2 FEM Cable Model in 3D Mechanics Using the same approach as in vessel motion input the external forces such as current force and buoyancy force can be applied as bond graph model on each rigid bar as a constant or the function of motion and velocity The port power should be defined in world coordinates because of the coordinate where current force is defined 34 Actuator Properties Actuator Type Force Actuator 3x1 Actuator Name Connection Points Connection Con 15 ConnectionPointl Body Break Connection Edit Name Position 0 0 2 m Orientation Bryant 0 0 0 Degrees Port Properties As Power Port DEE Body Coordinates In World Coordinates EA SUIT 7 Internal Coordinates no transformation Help Cancel Figure 5 3 External Force Acting as Actuator Outside the 3D Mechanics each actuator is linked with a bond graph effort source where current force and buoyancy force can be defined by simple Morison equation and Archimedes law or complex expressions The current speed is defined in a signal block by an array with length of n or 2n depending on if two bars connected by translational joint shall have the same current speed Fix Tip amp First Free Rotation Joint Translational Joint Element Bar Actuator for current force and buoyancy A Position Sens
39. e recorded and analysed both theoretically and statistically DNV rules checking will be conducted to make sure its correspondence with legitimate requirements 11 1 5 Literature and Previous Work Currently there are many types of marine simulators on the market Kongsberg Maritime provides Offshore Vessel Simulator for education and procedure training of navigators and winch operators of anchor handling vessels along with the MasterLift ML line of crane simulators dynamic positioning simulators liquid cargo handling simulator ships bridge simulators etc NAUTIS provides DNV certified maritime training solutions for the military civilian maritime industry TRANSAS provides navigational simulators GMDSS simulator engine room and cargo handling simulators crane simulators and simulator development tools CSMART provides two full mission bridge simulators six part task bridge simulators with the ability to simulate fixed propeller and azipod simulation Prof Thro l Fossen and Prof Tristan Perez from NTNU have also developed a Marine System Simulator MSS tool box in Matlab Simulink library which includes models for ships underwater vehicles and floating structures The library also contains guidance navigation and control GNC blocks for real time simulation This thesis is part of the ongoing research in the Ship Operation Lab in Aalesund University College whose activities support the ongoing development of the activities
40. e rolling has the smallest damping of them all 3 The deviation of surge sway or yaw generated during the transient state will be kept during the steady state because there is no restoring force for those three degrees of freedom 4 After entering their steady state the two sets of data 20 sim and SHIPX have the identical amplitude and frequency For natural resonance cases mentioned in Chapter 3 where 20 sim model has a deviated performance from ShipX RAO the damping coefficient can be manually tuned accordingly 7 2 Vessel Crane Cable Load 7 2 1 Model Vessel GlobalParameters Inertia Conols Cenitnpetal General amping Environmental force Wave fore Se gt AM IF Crane Control Joystick PID MSe 41 Joystick Hold1 Gab ZOH gt E D as sisa Cfane2R v4 Cable Load Latz Gain 2 Cable Load Figure 7 6 Simulation Example Vessel Crane Cable Load 52 With addition of cable and load model the interaction between the crane and vessel can be seen in the simulation Apart from vessel crane animation the video in the appendix also showed how vessel is leaning to the crane side due to the effect of the heavy load 50 t amp 100 t As for now added mass and damping of load model is set as constant Ca 1 Ca 1 2 Figure 7 7 Animation Vessel Crane Cable Load 100 t The dynamic behaviour of the cable and load is shown in another video because it is h
41. ect Ca w 9 coriolis force matrix of the added mass for a specific w and 7 n t Object motion Ny Object relative motion to the fluid B w damping force matrix for a specific w and T lifting force from the cable Tg gravity force of the object Tp buoyancy force of the object The added mass matrix and the damping force matrix can be transformed into added mass coefficient matrix M w and damping force coefficient B w for a specific object with its unique shape The damping force here is by definition of Morison equation proportional to the square of velocity The 37 calculation for a set of different frequencies and amplitudes can be done using CFD method Halse et al 2014 6 2 2 Velocity 4 Acceleration Dependent Coefficient The motion of the vessel model is assumed to be only periodic with constant added mass and damping matrices for each frequency and amplitude But the load could be lowered lifted dragged and rotated without periodic feature As the simulation is on a time domain basis the hydrodynamic coefficients of the load should be transformed from frequency amplitude dependent into velocity acceleration dependent Mih and Bin 7 Although the function is not strictly correct because the fluid has memory effect i e the historical motion of the fluid will affect the presence But if a good correlation can be shown by CFD calculation the trajectory of the load can then be removed from the
42. ed manul control can also be added as well as auto control The assembled system shall have a better approximation to the reality also ensure the functionality of all sub systems Thus simulation shall be monitored in 3D visualization and digital records Analysis shall also be made comparing simulation results with regulations and experiments Ultimately the model shall have the potential for quick adjustment and compatibility for simulators Different sub system shall be able to be developed by different parties with standardized interface and Plug and Play function 1 3 Organization of Thesis The thesis is mainly divided into three parts Firstly the principle of Lagrangian mechanics and bond graph modelling technique are introduced which are applied in multi body dynamics modelling in 20 sim Secondly detailed modelling method of vessel crane cable load control system and data panel are introduced separately with argument of modifications from previous works and interfaces between sub systems Alternative methods and potential improvements are also discussed in each chapter followed by evaluation of each subsystem s performance Finally simulation examples of different combinations of sub systems are given with arguements supported by simulation results A user manual is also given as reference 1 4 Methodology The main method used in this thesis is virtual simulation and the primary target of this thesis is to build a real
43. ed a tremendous development Its activities has been expanded from conventional fossil fuel production to areas such as renewable energy development telecommunication infrastructure construction fisheries installation and maritime search and rescue The worksite have also been expanded to deeper water thanks to the advancement in technology Fixed platforms were initially used for the offshore development but as the fields have gone deeper floating production facilities have become the main solution for the offshore production Also for safety and cost concerns subsea platform pipeline and ROV Remotely Operated Vehicles are increasingly being applied into offshore operations Higher requirement for precision accuracy and operability are thusly being raised towards OSV Offshore Supply Vessel and its lifting device along with other relevant equipment on board 1 2 Motivation and Objective Marine operations are usually multi system involved activities with interaction and coordination behind Heavy lifting operation is a typical example often used for ROV deployment and subsea installation demolition which normally involves vessel crane cable load and the manual auto control system Because of the complexity and diverse research focus of its nature people tend to isolate the issue by neglecting the insignificant and simplifying the complicated e g the researchers who intend to study the sea keeping feature of the PSV normally regard the dy
44. eir own sub model which in 3D gaming industry often called as MOD modification Many sub models can then be added into a product library let users and customers to choose and test Remote control and local PC experience is also to be expected in the final package There are still so much to be done 59 REFERENCES Norweigian Petroleum Directorate 2010 The petroleum sector Norway s largest industry 2014 Bond Graph http bondgraph org BP 2013 BP Statistical Review of World Energy June 2013 Chu Yingguang 2013 20sim based Simulation of Offshore Hyddraulic Crane Systems with Active Heave Compensation and Anti sway Control Controllab Product BV 2013 Hydraulic library 20 sim Reference 4 3 DNV 2010 DNV RP H103 Modelling and analysis of marine operations April 2010 DNV Euler L 1744 Methodus Inveniendi Lineas Curvas Maximi Minimive Proprietate Gaudentes sive Solutio Problematis Isoperimetrici Latissimo Sensu Accepti Laussanne and Geneva Fackrell Shelagh 2011 Study of the Added Mass of Cylinders and Spheres University of Windsor Fagereng Christian 2011 Mathematical Modeling for Marine Crane Operations Trondheim Fossen Thor l and Ola Erik Fjellstad 2011 Handbook of Marine Craft Hydrodynamics and Motion Control Trondheim John Wiley amp Sons Ltd Fossen Thor l and Tristan Perez 2014 MSS Marine Systems Simulator 2010 16 3 http www marinecontrol org Goldstein H
45. ent under 10 s wave period 60 deg heading and 1 m wave height Figure 7 3 shows the 6DOF motion of vessel model in 20 sim during the start of the simulation transient state HEEE AE itt ANT ANAN NA PA AAA ARMA BA AAA KA DAYA PARAAPAPADA Pin ARRAARARARARRA RAR AAA AAA AAA A VVVVVV VV VV VV VV VU VY 0 50 100 150 200 250 300 time s Figure 7 3 Vessel Motion Vessel Crane at Transient State Figure 7 4 shows the roll motion of the vessel model during steady state 1000 1180 sec which means transient oscillation has died out Figure 7 5 shows the 6DOF motion of the vessel according to ShipX RAO All motions are harmonic signals meaning there is not distinction of transient or steady state 50 A AAA HHHAHAHHHAHAHAHHHH A e N TTT EEN TINTIN ET ONT 2 M UUU UU UUU UN N T ee HH HH H UH UU Y es FELL Pitch MANANAHAN an INIA INUNDA e H A HA lt e o S 3 5 CH CH S P S L N O n Ss D H Co m a O lt xD CH O gt O D 8 to 9 03 D LG oF Q 4 3 D Q Q a D Li O D S lt O x 5 3 A a I T 5 H O gt CH 71 3 Discussion As the plot shows 1 There is deviation of 20 sim model from SHIPX RAO at the start of simulation which is because the model was at transient state the vessel has both natural frequency oscillation and forced frequency 2 The transient part of roll motion died out slower that the other motions becaus
46. eo Vertical Offset Table 6 3 Vertical Offset with Lifted Load m 0 00387 0 00387 0 00387 0 00387 e Horizontal Offset The horizontal assessment was conducted with variables of FEM parts number current speed and linearity of the shape e without current force on load the cable will have higher nonlinearity of its shape Horizontal Offset with Current Force on Load v 1m s L L L L L L DNV n 1 A n 2 n 3 j Depth m 100 amp i l 8 0 0 5 1 1 5 2 2 5 3 3 5 Offset m Figure 6 6 Horizontal Offset with Current Force on Load V 1m s 43 Depth m Depth m Horizontal Offset with Current Force on Load v 2m s L L L L L _ L DNV n 1 n 2 n 3 j r r r r 28 2 4 6 8 10 12 14 Offset m Br O O Figure 6 7 Horizontal Offset with Current Force on Load V 2m s Horizontal Offset with Current Force on Load v 3m s L L L L _ L 100 L E F E B 0 5 10 15 20 25 30 Offset m Figure 6 8 Horizontal Offset with Current Force on Load V 3m s 44 Depth m Depth m Horizontal Offset with Current Force on Load v 4m s L L L L L L L L 1 00 E r r r r r r r r r 0 9 10 15 20 25 30 35 40 45 50 Offset m Figure 6 9 Horizontal Offset with Current Force on Load V 4m s Horizontal Offset without Current Force on Load v 1m s L L L E L L t DNV n 1 _
47. erates itself when high velocity occurs in natural rolling frequency Also the motion RAO values at where high surge and pitch errors exist e g surge in beam sea pitch in 4 sec period wave are extremely low hence higher potential for large relative error mathematically speaking One way to improve the model quality is to add linear viscous damping coefficient for natural rolling frequency and to tune the motion RAO in 20 sim closer to that in ShipX The reason not using non linear viscous damping is that non linear viscous damping will cause the vessel yaw drastically because of the coupling effect between degrees of freedom A linear approximation is more favourable for having a stable course The user can tune the damping coefficient for each wave period and heading to have a better approximation towards ShipX motion RAO 26 4 CRANE MODEL The following chapter is based on Yingguang Chu s crane model in 20 sim including kinematics system hydraulics system and control system Chu 2013 4 1 Dynamic System The 3D Mechanics toolbox in 20 sim provides 3D animation environment for rigid body dynamics modelling where the crane can be modelled as well as the vessel In Chu s work the 3D modelling of crane was firstly done in Solidworks and then transferred into 3D Mechanics by using a converter called COLLADAto20 sim COLLADAto20 sim is developed by the Controllab group using C and COLLADA an inter application exchange file f
48. h The gratitude also goes to Yingguang Chu for his crane model and assistance on 3D Mechanics modelling The project is far from over and am looking forward to our cooperation in the future Finally the thesis is also to commemorate my amazing two years in Norway Aalesund University College May 2014 Jiafeng Xu ABSTRACT This thesis project introduced a generic modelling approach for heavy lifting marine operation based on 20 sim simulation and Matlab control The model is a multi body dynamic system which can be divided into vessel crane cable load and control system Physical entities are modelled either in bond graph or directly using 3D Mechanics toolbox and connected by interactive power port All control scheme is modelled as signal flow separated from physical entities The vessel is modelled as 6 DOF bond graph using parameters from SHIPX data then connected to the crane model inside 3D Mechanics unit Crane model is controlled by outside manual auto control scheme Cable and load are modelled inside 3D Mechanics with hydrodynamic behaviour represented by actuators The performance of each system is evaluated respectively by regulations and analysis The simulation examples of different combinations of sub systems are given at the end The project is aiming at developing a generic modelling method serves as a multi user training and design platform Standardization and potential for upgrading are to be expected in the future
49. ion But thanks to software like 20 sim and SimulationX bond graph elements are more flexible and in some extent can be regarded as a coding language For example several elements can be modelled in a single block with extra power ports and equations New types of element can also be invented and named by user s preference In this thesis three translational degrees of freedom and three rotational degrees of freedom of the vessel and for other objects are being treated homogeneously as six elements in motion vectors Standard elements such as R element C element and I element are being recoded to cope the expanded expression of degree of freedom Other Transformers and user defined elements will be explained in the following chapters Names of each element are simply to reveal the functionality of which in real physical system for reader s convenience 2 2 3 20 sim and 3D Mechanics Toolbox 20 sim is a commercial modelling and simulation software developed by Controllab Products B V lt enables user to enter model graphically similar to drawing an engineering scheme and to simulate and analyse the behaviour of multi domain dynamic systems and create control systems It also provides tools that allow users to create models using equations block diagrams physical components and bond graphs C code are also compatible in this software to use on hardware for rapid prototyping and HIL simulation 3D mechanics toolbox is a packaged module in 20 sim
50. istic real time simulation model for marine operation As for economic and safety reasons a full size experiment of marine operation is practically impossible for the industry A virtual environment of maritime operation shall thusly be developed in the computer which allows users to interact with Physical objects are being represented as components with different functionality in the system The system can accept input from the user e g wave spectrum vessel inertia matrix hydrodynamic damping parameters crane model CFD data of the load and produce output to the user e g vessel motion during the operation crane hydraulics performance load motion in the water power output mechanical behaviours The model is built based on studies about realistic feature of the vessel and hydrodynamic statistics and coded according to general physical principles with simplification for faster performance The parameters the model chooses depend on field measurement but are also tuned for better approximation to the reality Each sub system is modelled separately whose accuracy and applicability are tested before being integrated into bigger system During the simulation variables are controlled to test the influence of each Although virtual simulation has the flexibility and compatibility to adjust and to replace input with minimum cost and maximum fidelity it always stands on certain assumptions that loses reality in some extent Thusly simulation results ar
51. ive but with less code and consumes less computer memory 3D Mechanics Toolbox generates enormous lines of code and computer resources However the biggest advantage of 3D Mechanics is its user friendly interface and modelling accuracy For efficiency and feasibility purpose 3D Mechanics Toolbox is used in this thesis 3 VESSEL MODEL The following chapter is based on Fossen s equation of vessel hydrodynamics Fossen and Fjellstad 2011 and Pederson s bond graph model Pedersen 2008 3 1 Reference Frames Following the guiding spirit of Lagrangian Mechanics it is convenient to use different reference frames in each particular case In this thesis the Earth surface is assumed to be flat in the range of where the vessel moves Only three right handed orthogonal reference frames are used in this thesis e NED The North east down frame E Xe Ye Ze is fixed to the Earth The positive x axis points towards the North the positive ye axis towards the East and the positive z axis towards the centre of the Earth The origin On is located on mean water free surface and coincides with body fixed coordinate system B x Yo Zp at time t0 This frame is considered inertial because the force acting on the vessel due to the Earth rotation is negligible compared to the hydrodynamic forces acting on the vessel e Body fixed The body fixed coordinate system B Xp Yo Zp is fixed to the hull The directions of the positive axes are xb axis forward
52. n 53 FIGURE 7 9 VESSEL MOTION VESSEL CRANE CABLE LOAD 50 T AT TRANSIENT STATE 54 FIGURE 7 10 VESSEL MOTION VESSEL CRANE CABLE LOAD 50 T ROLL AT STEADY STATE 54 FIGURE 7 11 VESSEL MOTION VESSEL CRANE CABLE LOAD 100 T ROLL AT STEADY STATE 55 FIGURE 7 12 SIMULATION EXAMPLE VESSEL CRANE CABLE LOAD AHC 56 FIGURE 7 13 LOAD PEDESTAL HEAVE VESSEL CRANE CABLE LOAD OO 57 FIGURE 7 14 LOAD amp PEDESTAL HEAVE VESSEL CRANE CABLE LOAD AHC 1001 58 LIST OF TABLES TABLE 2 1 EFFORT AND FLOW VARIABLES IN DIFFERENT PHYSICAL DOMAINS 14 TABLES VESSELAND WAVE DATA nG mn Ag isama kaa AN dea e dd th 24 TABLE 5 li COMPARISONOFFEM CABEEMODELS urna eae o a Aasa LU do ai 33 TABLE 6 1 6X36 WS HIV RG PARAMETERS 2 2 2 20657 ld ee 42 TABLE 6 2 PARAMETERS OF CABLE AND LOAD 42 TABLE 6 3 VERTICAL OFFSET WITH LIFTED LOAD M a 43 TERMINOLOGY added mass damping tensor restoring force tensor added mass coefficient drag force coefficient coriolis and centripetal forces of added mass coriolis and centripetal forces of rigid body gravitational acceleration moment of inertia tensor rotation matrix for 6 degree of freedom lagrangian function inertia tensor mass density effort in bond graph flow in bond graph rotation matrix for linear velocity centre of gravity vector kinetic energy rotation matrix for angular velocity time
53. n equation are defined outside the 3D Mechanics block 20 The I element in the model represents the Inertial force and Coriolis force Mp3 M OCI Crp Ca w At The R element represents the damping force TU Bo J nt The Effort Source Se and a MTF acting as coordinate transformation Jacobian matrix The Jacobian matrix is set as a global variable J Only wave force is included in this model but other hydrodynamic forces can be added by the same modelling approach sah J p Texc The C element represents the restoring force with Jacobian matrix J in the code JC a nG The vessel in the 3D Mechanics block has a negligible inertia thusly can be called as shadow vessel By connecting the bond graph to an actuator located on the origin of body fixed coordinate system the outside bond graph is acting equivalent to a Flow Source to determine the vessel motion inside the 3D Mechanics block y A Figure 3 3 Vessel Model in 3D Mechanics 21 Actuator Properties Actuator Tupe Force Actuator Bl Actuator Name WesselPower Connection Points Connection Connect 13 ConnectionPointl Icebreaker Break Connection Edit Name Position IO O 7 m Orientation Bruant U 0 0 Degrees Port Properties As Power Port EB Sos Body Coordinates AAA Ch World Coordinates CJ Internal Coordinates no transformation Help Lancel Figure 3 4 Actuator in 3D Mechanics Toolbox
54. n extreme cases bending moment in local area Because of the compound material inside the cable some may have non linear elastic modulus and unevenly distributed stress level in local area Offshore crane often use multi strand steel wires as lifting cable in order to enhance its strength and minimize axial rotation for operational purpose When loaded steel wire will generate torque if both ends are fixed and turn if one end is unrestrained The torque or turn generated will increase as the load applied increases The degree to which a cable generates torque or turn will be influenced by the construction of the rope All cables will rotate to some degree when loaded 5 2 Categories of Cable Model Choo and Masarella classified current analytical and numerical methods of solving the equations of cable motion in four categories Johansson 1976 e Method of characteristics e Finite element methods e Linearization methods e Other Methods Different categories suit different modelling scenarios and in this thesis finite element methods are used in 3D Mechanics 5 2 1 Method of Characteristics The method of characteristics is a direct mathematical interpretation of the physical behaviour of the cable E J Routh proposed an early formulation of a dynamic cable model in 1905 Johansson 1976 A small segment ds of the cable is considered with its parametric trajectory x s y s z s in terms of Cartesian coordinates where s is the trajectory
55. namic behaviour of the crane and the load as negligible Lloyd 1989 while people who study the AHC control treat vessel motion as unaffected 10 by the crane hence an independent variable outside the equation Moreover people who study the hydrodynamic behaviour of a submerged load usually only apply simple motion to the load in CFD calculation Fackrell 2011 or experiment Also cables are sometimes a one dimension spring characterized with only heave motion However ordinary computers have long reached the capability of processing complex mathematical model of a marine operation if not in real time at least Bond graph modelling technique enables a generic approach to build a physical system whereas control system can be modelled by traditional signal units Also the calculated results can be demonstrated by 3D visualization Therefore a system integrated with major subsystems can now be established and simulated expectedly to have a better approximation of the reality This thesis introduces a generic modelling approach for heavy lifting operation by using bond graph technique signal blocks and 3D mechanics toolbox in simulation software 20 sim The input data resources are from ShipX strip theory DNV rules and CFD calculation managed in MATLAB panel and communicated with 20 sim A fully interactive system of vessel crane cable and control system is realized and able for easy parameter selection If real time simulation can be reach
56. nt Tupe C3 Rotation Screw C weld 0 DOF Translation Free Rotation Free B DUF Direction using order of Connection Points Translation Asis 1 e Help Swap Connection Points BA H UY Cancel Forwards Backwards KE Figure 4 6 Boom Cylinder Setup in 3D Mechanics 4 3 Control System The control system of an offshore crane consists of two kinds manual control and auto control The crane can be controlled by solving its kinematic model The resulting signal for each hydraulic system can be input into hydraulic model hence the crane model 4 3 1 Manual Control The manual control part can be realised by joystick or other input device e g haptic control Sanfilippo et al 2013 The joystick sends signals of the crane tip velocities or cylinder velocities into kinematic model where corresponding joint velocities and cylinder velocities are calculated by jacobians and fed to the hydraulic system using PID controller Joint angles and vessel motions can be obtained by sensors in 3D Mechanics model 30 Joystick Inputs Crane Tip Jacobian Phan 3D Mechanics Velocity y Figure 4 7 Crane Tip Velocity Control Joystick Inputs Hydraulic Cylinder 3D Mechanics Cylinder Velocity Figure 4 8 Direct Cylinder Velocity Control jb E 1 AibCyfinder boom E 1 AbsomCylinder base mm 1 A base Motor PID base EaseHydraulics Figure 4 9 Cylinder Velocity Input and PID Control 4 3 2 Auto C
57. o the steadily translating coordinate 40 Buoyancy MSe MSe curentspeed cable 00 3 F A TF STEF MSe Addedhiass Damping Figure 6 5 Cable Load Model 6 4 Model Assessment To assess the accuracy and authenticity of the cable and load model an evaluation between 3D Mechanics model and standard DNV rules calculation is conducted The 3D Mechanics model will have a series number of discretized FEM parts n and all parameters are set according to the distribution of physical properties without manual tuning 6 4 1 DNV Rules Calculation According to DNV RP H103 5 2 DNV 2010 the stretched length L of a cable L is W 5 wL Le L 1 TEA m where L stretched length of cable m L original length of cable m W Mg pgV fully submerged weight of lifted object N w mg pgA fully submerged weight per unit length of cable M mass of lifted object kg m mass per unit length of cable kg m g acceleration of gravity 9 81 m s p density of water kg m E modulus of elasticity of cable N m A nominal cross sectional area of cable m V displaced volume of lifted object ms For an axially stiff cable with negligible bending stiffness the offset of a vertical cable with a heavy weight at the end of the cable in an arbitrary current with unidirectional x direction velocity profile is W w z L Sr Ie where 1 Fpo 5 PCpxAxUc L 1 N is the hydrodynamic damping f
58. of Vessel Hydrodynamics 0000an0000annnnnnnnnnnnnnannnnnnnnnrrnnnnnnrnnnnnnrrnnnnnnrnnnnnnne 18 3 2 2 Coordinate Transtiormaiion 19 3 3 BOND GRAPE MODEL EEN 20 34 IPARAMETER Scola aa 22 3 4 1 Uer E deene EE 22 S42 EIERE deeg eet 23 343 20 SiM Parameler ee 23 3 5 MODEL ASSESSMENT rene nana 24 5 4 A AA PAA AN PAA EC 24 3 5 2 DPIECHESTO A mana E II nA AA II E 26 A GRANE MODEL AA AA 27 4 1 DYNAMIC SN EE EE 27 All SOOWOKS MOGEN WEE 27 4 12 Converting Steps and TIPS 1 ccccscccccsccesccncccrcossencccrconseneccsconsentessoaseenecssoansenscsscansescennoases 28 di VRIY DRAWING SE aa aan AA ka aaa aaa bn Pa kaaa naala 28 4 2 1 MANDES Ci a adan laat bata 29 4 22 20 Sim e e Ee EE 29 A3 KONO OS TEM ice id AA o o ld aT 30 4 3 1 IE EI AAA A ee 30 Ee E eier A EEN 31 5 CABLE MODEL O KGG kn BABU gean faisin cat ee 32 5 1 CABLE CHARACTER STE o BAO pa a eel el LG Ga 32 52 CATEGORIES OF GABLE MODEL o a Lo ies ba 32 5 2 1 Method of Characherisiice 32 52 2 Finite Element MethOdS ccccecececececccececececececececesecececacececaecacacacaenenenenenenenenenesnenas 33 52 3 ee ele MEMO aoire EA LE Jaca sos ceed e RA AE A aes A N Et ee ees SE 34 5 3 CABLE MODEL IN 3D MECHANICAL 34 6 LOAD MODEL lt A 37 6 1 KEUNG OPERATION EE 37 625 Eet e EE 37 6 2 1 MOTTON FUNCION OTE LOAD ana a e 37 6 2 2 Velocity amp Acceleration Dependent Coeitcient 38 6 3 20SM MOD EIN e Ga do 38 64 MODEL ASSESSMENT tania iii
59. on of the crane pedestal and the load without AHC system from transient state at the beginning to the steady state at the second half Figure 7 14 is the heave motion of the crane pedestal and the load with AHC system from transient state at the beginning to the steady state at the second half 56 2 4 E Heave Motion without AHE Load Pedestal 100 200 EUR 400 500 600 ZU time 5 Figure 7 13 Load 8 Pedestal Heave Vessel Crane Cable Load 100 t DL 57 Heave Motion with AHC m oad m Pedestal 0 100 200 300 400 500 600 700 800 time s Figure 7 14 Load 8 Pedestal Heave Vessel Crane Cable Load AHC 100 t 7 3 3 Discussion As the plot shows 1 Without AHC system the heave motions of the load and crane pedestal are identical because the crane does not have relative motion between them and the small deviation caused by rotational motion can be ignored 2 With the effect of AHC system the load does not have good compensated motion at the beginning of transient state but the amplitude gradually decreases while entering to the steady state This is because the AHC system used in the model so far is based on crane kinematics which has limitations and response time lt has better performance towards regular and small motions rather than irregular and drastic motions Winch based AHC system may have different performance in this case which can be investigated in the future 58 8
60. ontrol The auto control system may include heave compensation system anti sway system or the combination for all three translational degrees of freedom The heave motion of the vessel can be compensated by crane kinematics or winch The sway motion can only be compensated by crane kinematics There are many algorithms to realise the function of compensation One simple solution is to add compensating motion velocity on opposite direction into crane kinematics or winch motion velocity Hydraulic Figure 4 10 Compensating Motion velocity on Opposite Direction One thing should be noticed is that all manual auto control algorithms are signal flow model or simple codes in Jacobian Control Box The physical entities hydraulics and dynamics are modelled as power flow model hydraulic library 3D Mechanics bond graph which remains intact from the variations of control system 31 5 CABLE MODEL 5 1 Cable Characteristics Offshore cable is usually treated as flexible body with infinite degrees of freedom It has physical features of elasticity and plasticity in both axial direction and radial direction In practice the external forces offshore cable may experience includes concentrated force on both ends gravity force along the entire length buoyancy force on the submerged part and damping forces caused by relative motion between cable and the fluid around Those external forces can cause axial tension axial torque radial shear and i
61. or at the end of cable Figure 5 4 Cable Model n 3 L 10 35 Figure 5 5 External Force Input 36 6 LOAD MODEL 6 1 Lifting Operation Offshore crane is designed to be multifunctional The loading objects include subsea construction materials ROV platform supplies personnel carrier and all other different objects Each loading object has its unique physical property and installation requirement The lifting operation can normally be divided in five phases e Lift Off The object is fixed to the cable and snatched from the deck The major external forces are friction from the deck supporting force and lifting force e In Air The object has been lifted and being moved to the designated position before being lowered down into the water The major external forces are air resistance inertia force and lifting force e Splashing The object has been lowered close to the free water surface and experiencing splashing impact until it is totally submerged and unaffected by the water surface The major external forces are splashing force buoyancy force and lifting force eo Deeply Submerged The object is deeply submerged in the water yet still has a certain distance from the seabed The major external forces are buoyancy force hydrodynamic force and lifting force Landing The objected has been lowered close to the seabed and ready to land The major external forces are buoyancy hydrodynamic force lifting force and possibly landing impa
62. orce on the lifted object The parameters are defined as 41 z horizontal offset at vertical position z m CL horizontal offset at the end of cable z L m L un stretched length of cable m Con damping coefficient for normal flow past cable Cox damping coefficient for horizontal flow past lifted object D cable diameter m A x projected area of lifted object m7 U z current velocity at depth z m s z1 z2 integration variables m The formula can be transformed into Z K 1 q L q L K 1 ln gt 4 z in K 1 Ta where W Fng 1 K 7 lt pC D U2 wL wL q Eis Dn c c 6 4 2 FEM in 3D Mechanics In 3D Mechanics the length and mass of each rigid bar element for a cable length L is L AL Am m AL 2n The moment of inertia of each rigid bar element is m AL y sg AL The axial moment of inertia Al is negligible The spring and damping coefficient of each rigid bar element are Ak EA 2AL Ac 2 Ak m 6 4 3 Cable Parameters The tested cable is Lankhorst Ropes 6x36 WS IWRC standard wire rope with higher breaking strength Table 6 1 6x36 WS IWRC Parameters Diameter mm Weight kg m Elastic Modulus GPa Structural Damping Ratio per meter C 10 2 including fluid friction Table 6 2 Parameters of Cable and Load 20000 10 0 001178 effective area ratio 0 6 42 6 4 4 Results
63. ormat Janssen 2013 There are special requirement for this kind of conversion however 4 1 1 Solidworks Modelling Solidworks is a well developed 3D modelling tool which supports many general formats among which STL files can be imported into 20 sim 3D Mechanics toolbox and animation window In current model a Rolls Royce PSV 100 Crane model was simplified and modelled in Solidworks Boom Cylinder Cylinder Rod Cylinder Cylinder Piston Ly Base Fundation Figure 4 1 Solidworks model of simplified Rolls Royce PSV 100 This knuckle boom crane Rolls Royce PSV 100 has eight assemblies parts named as in the picture Each assembly contains several sub level assemblies parts e g one cylinder set has two parts a piston and a rod Solidworks mates combine those assemblies and parts to form a complete structure Mates can only be created in where physical mechanical connection exists otherwise unexpected bodies and joints will be generated because each mate represents one joint in 3D Mechanics model There are three types of mates provided in Solidworks standard mates advanced mates and mechanical mates Only standard mates are allowed if the model needs to be converted into 20 sim through COLLADA After modelling Solidworks can export its assembly into COLLADA format DAE through a plug in which preserves physical and visual information of the model 27 4 1 2 Converting Steps and Tips Solidworks Export to
64. set condition the FEM model is still applicable for simulation purpose but requires extra evidence to support Since under small horizontal offset condition the FEM model shows good compliance however the parts number is the choice of the part number only depends on the distribution of current force and nonlinearity of the cable Higher parts number will have better approximation of the shape finer distribution of the current force and more accurate dynamic simulation result But it will also increase the computation amount and time The decision shall be made under different practical conditions 48 7 SIMULATION EXAMPLE All sub models introduced in Chapter 3 6 can be assembled into one big model selectively The interactive physical entities are connected with power port whereas control loop are connected with signal port Some examples are demonstrated as follows Because of the computation limits on PC hydraulic sub model was replaced with direct PID control 7 1 Vessel Crane 7 1 1 Model GlobalParameters Environmental force Joystick Joystick Figure 7 1 Simulation Example Vessel Crane pagan KA et i b id ama Na s a ae e E Figure 7 2 Animation Vessel Crane 49 7 1 2 Results The default ship s175 from ShipX database is used The video in the appendix showed the anmimation of the target vessel cyan comparing with another vessel grey with SHIPX output movem
65. sim the NED E xe Ye Ze coincides the hydrodynamic reference frame H X Yn Zn at time 10 The axes are x axis forward y axis starboard z axis downward the same as in Marine control toolbox It is different from ShipX axes because all ShipX data must be transformed by Marine control toolbox Fossen and Perez 2014 into MATLAB data file before being input in to 20 sim In 20 sim 3D animation the world reference only support a camera view of z axis upwards so the world reference shall be ignored and the direction of the Earth fixed reference is determined by the stationary position of body fixed B Xp Yo Zp at time 10 Figure 3 1 Body fixed reference in ShipX and 20 sim t t1 3 2 Motion Equation 3 2 1 Equation of Vessel Hydrodynamics Fossen presented the hydrodynamic behaviour of a vessel in traditional quadratic linear differential equations by utilizing the Kirchhoff equations Fossen and Fjellstad 2011 Take the second and third equations from Kirchhoff equations d T T gt MT ag ag ee EN CO any E aa a Also we have two equations from Momentum Conservation Principle OT am fm re X m m X re A OT gt gt 7 gt gt gt 7 gt Jo m Ux m xT UA IG Restate into standard form of the equation of motion Mim D h C M U N T T Where 1 v w Th F Qh and t F Of The inertia matrix M is recognized to be 18 m m X T m k m X Te I And
66. sistance Gerritsma 4 Beukelman e Generate hydrodynamic coefficient files re7 and re8 e Calculation options choose z coordinates from CO Ob The condition information for frequency domain simulations must be chosen according to eo Vessel velocities must always include the zero velocity it is optional to add more velocities that are needed for manoeuvring Wave periods it is recommended to use values in the range 2 0s to 60 0s e The wave heading must be chosen every 10 deg starting from 0 deg 3 4 2 MSS Post process After ShipX calculation put all five output result files hyd re1 re2 re7 re8 into current folder in MATLAB Install MSS toolbox and run gt gt vessel veres2vessel input MSS is using x forwards y starboard and z downwards by introducing a transformation matrix 1 1 1 1 1 1 in the function And the position of origin depends on the calculation option defined in ShipX After abovementioned process a MATLAB data file mat will be created and a data structure of vessel will be given saved to mat 3 4 3 20 sim Parameter Input 20 sim model uses data from MSS data file mat In newly released 20 sim 4 4 version a library of 20 sim MATLAB connection toolbox is published using TCP IP protocol for data exchange and control In MATALB a short script can be written for users to select the target MSS data file mat 20 sim model file emx and desired forward velocity wav
67. t BV 2013 However the library does not provide all components that are required So some components are self designed and added into the library later An example of main derrick hydraulic is briefly introduced below 28 4 2 1 Main Derrick Sketch 4 3 directional valve Compensator Pressure compensated pump o Tank Figure 4 4 Simplified Main Derrick Hydraulic Sketch 4 2 2 20 sim Modelling According to the simplified sketch of main derrick hydraulic a 20 sim hydraulic model can be built using the hydraulic components library except the cylinder The parameters can be adjusted according to the crane technical specification The 20 sim model is equipped with input signal port for control system and output power port for dynamic model in 3D Mechanics In 3D Mechanics the joint where hydraulic system is deployed acts as Power Interaction Port 1x1 and outputs position and velocity of itself Lis22M5C OPE SEO ELA Figure 4 5 Boom Cylinder Hydraulics Model 29 Joint Properties Connection Points Connection Connecte Break Connection NG ConnectionPoilnt boot vis A ConnectionPolnt2 boomt viod Edit Name Position 3 38 008 9 46 008 1 25 m Orientation Bryant 120 O 900007 Degrees Initial Translation H 40000000000001 5 im Locked Less lt lt As Constraint Joint Power Interaction Port 1x1 Spring Damping Output Position Output Velocity an Forces Joi
68. the remaining terms which are known to contribute the Coriolis and centripetal forces can be state as Io a2 A3 b b b3 M n 0 0 0 0 13 Ao 0 0 0 A3 0 Q hrs 0 0 0 a a 0 eS alg d a gt 0 b b as 0 a b O b a aA 0 b bj 0 The hydrodynamic force can also be separated into two components Th Trad Texc Where Traa is the hydrodynamic radiation forces and T is environmental excitation forces from wind and waves The radiation forces can be expressed in frequency domain as follows Trad A w i B w n Cn Where A w and B w are the frequency dependent added mass and damping matrices The restoring forces are assumed to be a linear formulation Cy The resulting motion can be stated as n e and the dynamic equation in frequency domain becomes w M A w ndo jwB w nGw C nG w Terc T Naval architects usually write the equation in a mixed frequency time domain formulation Mag M4 i t Crp Ca w H t Blow H t C n t Texc T Strictly speaking this equation is not correct because the real vessel motion is not purely harmonic but it is tolerated in a dominated frequency motion The excitation forces include forces from wave current and wind can be derived from force RAO in ShipX and other forces from appurtenances such as rudder propeller foil and fin stabilizer can be either assumed or through classical calculation The motion equation is esta
69. variables To investigate the feasibility of velocity acceleration dependent coefficients a CFD experiment can be conducted to compare the difference of coefficients under same velocity and acceleration condition but with and without periodic movement Group A Group B Velocity acceleration dependent Amplitude frequency dependent m Gd yy ap Ho pee Ang Ann if n wl cos wt 8 w sin wt oU j b w O Figure 6 1 CFD Experiment to Test amp Verify Velocity acceleration Dependent Coefficient If valid the velocity acceleration dependent coefficients can then be fitted and input into 20 sim model as functions of velocity and acceleration 6 3 20 sim Modelling In 20 sim 3D Mechanics the loading object can be simply modelled as an object attached to the end of the rope with an 6x1 actuator in body fixed coordinate representing extra added mass inertia force damping force and another 1x1 actuator in world coordinate representing buoyancy force Buoyancy Force A Added Mass and Damping Force TTTT 3 Figure 6 2 Load Model in 3D Mechanics 38 Actuator Properties Actuator Type Force Actuator Panay Naa Connection Points Connection Con d ConnectionPointl Body Break Connection Edit Mame Position IO 0 0 imi Orientation Bryant 0 0 0 Degrees Port Properties As Power Port C3 Body Coordinates 8 In world Coordinates AAA 2 Internal
70. when cable experiences unsteady irregular external forces a stepwise computerized solution method is required but both over detailed and time consuming Most of the time approximation is involved so that the accuracy is compromised 5 2 2 Finite Element Methods Finite Element Methods enable the cable to be conceptually modified Johansson 1976 before mathematical formulation Comparison of different FEM cable models are given below eo 5 Mass String The cable is modelled as an inelastic string with mass lumped at several nodes e Mass Spring The cable is modelled as springs with mass lumped at several nodes e Bars The cable is modelled as a series of straight bars with distributed mass which are jointed at both ends Beams The cable is modelled as a series of beams with distributed mass which are jointed at both ends and with additional internal curvature d Mass Spring Bars Figure 5 1 Four Types of FEM Cable Model Table 5 1 Comparison of FEM Cable Models pro ee ee ma Slope Continuity Structural Damping Tension Induced Torque Y Y rm 33 5 2 3 Linearization Method In some stability and frequency response studies of cable vehicle systems the motion equations are linearized and assumed to only have small deviations from the state of equilibrium The linear partial differential equations will then be reduced to linear partial differential equations with curve length as the space variable Moreover sometimes
Download Pdf Manuals
Related Search
Related Contents
WIN100 SERIES - Lux Products dreamGEAR Power Pak 3 In 1 [Product Monograph Template - Standard] 光エステゾン WallGard GRAFFITI BARRIER データシート Towards a unifying CSP approach for hierarchical verification of Aladin Square Manual 4R unner Copyright © All rights reserved.
Failed to retrieve file