THESIS 2 000 llllfllllllllllllHUI!IUIIHIIHMIHHlllllllllllllll 302074 21 63 Michigan State LlflRARV University This is to certify that the dissertation entitled THE APPLICATION OF THE MODULAR MODELING METHOD TO FINITE ELEMENTS ANALYSIS presented by XinLi has been accepted towards fulfillment of the requirements for MS Datfléz/q/flf 2am MS U is an Affirmative Action/Equal Opportunity Institution degree in Mechanical Engineen'ng 0-12771 PLACE IN mum BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 11/00 chlfiClDateDquS—p.“ THE APPLICATION OF MODULAR MODELING METHODS IN FINITE ELEMENT ANALYSIS BY Xin Li A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Mechanical Engineering 2000.05 ABSTRACT The Application of Modular Modeling Methods in Finite Element Analysis By Xin Li Structural modeling analysis is important to engineering design and development. An important issue is the interconnectivity problem for multi degree of freedom subsystems. The traditional finite element method requires re— meshing subsystems for mesh compatibility at interconnections. As the number of subsystems increase, the computational burden of re-meshing increases rapidly, sometimes exponentially. The Modular Modeling method is applied here to develop a new finite element analysis approach. Modular Modeling through the use of input 1 output ports on the boundaries of interconnected parts is introduced for this purpose. Using input I output ports and the Modular Modeling method to connect parts in system design does not require expensive remodeling and re-meshing. This approach uses the original FEA model and combines parts with the input I output ports among their boundaries. As the number of available alternate parts in a structural system increases, the modular F EA method minimizes the time and cost for the system integration. A two-dimensional example of FEA elastic strain-stress problems is used to illustrate this technique. This example shows it is not necessary to re-mesh components to assemble a structural system. Computation error is evaluated to show that this modular finite element method produces accurate system models for engineering design. Copyright by Xin Li 2000 TABLE OF CONTENTS List of Figures ....................................................................................................... v Introduction ........................................................................................................... 1 Multi-Component Modular Modeling for Finite Element Analysis .......................... 4 2.1. FEA Modular Modeling and the related DOF analysis ............................... 4 2.2. Visual Component Boundary Port — FEA Input/Output Connector ........... 7 2.3. FEA Modular Modeling Procedures and Formulations ............................... 9 2.3.1. Sub-system Equation Combination .................................................... 10 2.3.2. Applying Boundary Displacements Constraints ................................. 12 2.3.3. Applying Boundary Forces Constraints ............................................. 21 2.3.4. Combine the Supplemental Boundary Constraint ............................... 24 2.4. Result and Discussion ............................................................................. 27 Two-Dimensional Example for Elastic Strain-Stress Problem ............................ 28 3.1. F EA static stress-strain analysis by modular modeling ...................... . ...... 2 8 3.1.1 2-D Elastic Plane A: ............................................................................ 29 3.1.2. 2-D Elastic Plane B: ........................................................................... 30 3.1.3. Using Modular Modeling To Integrate 2-D Elastic Components ......... 31 3.2. Error Estimation and Analysis .................................................................. 32 3.2.1. Standard Reference Models for Accuracy Comparison ..................... 32 3.2.2. Modular Modeling Method Accuracy Comparison .............................. 33 Conclusions ........................................................................................................ 37 4.1 Contributions ............................................................................................. 37 4.2 Future work ................................................................................................ 38 References ......................................................................................................... 39 Appendix I. .........42 Appendix II..... ......46 Appendix III.. .49 Appendix IV.... ...51 Appendix VI...... ...67 LIST OF FIGURES Figure 1.1. The modular modeling using input / output Ports (joint) ............................... 3 Figure 2.1. The modular modeling using input / output Ports (connectors) ...................... 5 Figure 2.2. The modular modeling method for the FEA application ................................ 7 Figure 2.3. The modular modeling input / output Ports along the FEA boundary ............. 8 Figure 2.4. At least 6 constraint are needed for 2-D FEA modular modeling ................. 11 Figure 2.5. Nodal displacements along the boundary elements ....................................... 13 Figure 2.6. Nodal compatible situation along the boundary elements ............................. 14 Figure 2.7. Nodal displacements in the boundary elements ............................................ 15 Figure 2.8. Linear transformation on the boundary elements (case 1: interpolation) ....... 16 Figure 2.9. Linear transformation on the boundary elements (case 2: extrapolation) ....... 18 Figure 2.10. Linear transformation on the boundary elements (case 3: equality) ............. 18 Figure 2.11. The power conservation along the Boundary ............................................. 23 Figure 2.12. The “solvable system” and related equations. ............................................ 24 Figure 3.1. Two-Dimensional Example for Elastic Strain-Stress Problem ..................... 28 Figure 3.2 Plane A meshing details and nodal distribution ............................................ 29 Figure 3.3. Plane B meshing details and nodal distribution ........................................... 30 Figure 3.4. Re-meshed reference model with higher or lower meshing resolution .......... 32 Figure 3.5(a). Displacement contour lines for reference model (7x7). ........................... 33 Figure 3.5(b). Displacement contour lines for the Joint Modular Model ........................ 33 Figure 3 .6. Displacement difference between the Joint Modular Model and reference model (7x7) ............................................................................................................. 34 Figure 3.7(a). Displacement contour lines for reference model (3x3). ........................... 35 Figure 3.7(b). Displacement contour lines for the Joint Modular Model ........................ 35 Figure 3.8. Displacement difl‘erence between the Joint Modular Model and reference model (3x3) ............................................................................................................. 35 Figure 3.9. The difference of nodal displacements along the boundary ........................... 36 Figure 4.1. Comparing the Modular Model and other models ........................................ 37 INTRODUCTION During the design of complex systems, reliable predictions of detailed stress states for coupled components are often critical to correctly predict the performance of those systems. The finite element analysis (F EA) yields accurate, detailed stress predictions for individual system components with the resolution of mesh distribution. When more precision is needed, a higher precision mesh needs to be generated. In engineering practice, system components can be so complex that the finite element method is mandatory for the design phase. It is computationally intensive and often requires extensive re-modeling effort. Coupling system components into assemblies is a time intensive task frequently requiring reformulation for finite element analysis models. For example, the Chrysler large car models used for mechanical geometry studies contain representations of over 5500 interconnected physical subsystems [Computers in Engineering: “Chrysler designs paperless cars”, 1998]. To reduce the length and cost of the design process for a complex system, efficient modeling techniques that guarantee reliable predictions of detailed stress states are needed. These techniques must be not only computationally efficient, but also dramatically reduce or eliminate the necessity of re-modeling. According to [Computers in Engineering: “Chrysler designs paperless cars”, 1998], Chrysler engineers solved design issues by developing a computer model instead of physical prototypes, reducing the cycle time from 39 to 31 months and saving the company more than $75 million. Finite element analysis of system of components typically requires compatible node geometry at each component connection. When constructing the two-dimensional FEA models, it is a mandatory requirement to keep mesh nodally compatible [Zienkiewicz and Taylor, 1989]. At present, the methods available to ensure FEA nodal compatibility include the global llocal analysis method, the coupling analysis method, and the interface element method, [M. Aminpour, 1995; V. S. Kothnur, 1999; J. B. Ransom, 1993]. These methods have drawbacks that prevent them from being widely accepted as standard design tools, [2. Zhao, 1998]. The common limitation of these methods is that they can only be applied to mesh discretizations with a one-to-one nodal correspondence across boundaries between structural subsystems. Methods that do not require nodal compatibility on the component boundaries have much more modeling flexibility, and eliminate the need for re-meshing. A new finite element analysis approach is introduced in this paper which employs the fixed input / output structure "modular modeling method" [8. Byam and C. Radcliffe, 1999]. In Modular Modeling, the mathematical model that describes each component remains the same independent of the system model in which it is used. Fixed input / output structure means the input and output ports are standardized, therefore the internal equations of mathematical subsystem models of engineering systems have the same modularity as the engineering system. Fixed input I output structure modular modeling is a power-based, physically intuitive, top-down methodology systematic modeling methodology that eliminates equation reformulation from large model development across multiple energy domain [8. Byam and C. Radcliffe, 1999] This paper demonstrates the modular modeling method in FEA. The work presented here developed an analytical framework with the Modular Modeling analysis method, a solution strategy, and a demonstration of this solution on a F EA example. Finally, the accuracy and convergence of a Modular FEA Model will be evaluated by comparison with various conventional FEA solutions to a 2-D plate problem (Fig, 1.1). Ub=S-Ua Inlemal boundary Ea'+Eb"S=U constrains Figure 1.1. The modular modeling using input / output Ports (joint) MULTI-COMPONENT MODULAR MODELING FOR FINITE ELEMENT ANALYSIS 2.1. FEA Modular Modeling and the related DOF analysis Modular modeling is a new modeling method designed to eliminate the model reformulation and enhance model performance verification. It is defined by a standard input-output structure, in which connection model physical interaction between components via power transfer. Power Transfer between components requires the product of two signed variables such as force times velocity, pressure times flow rate, and voltage times current. The two physical variables form a component input-output pair. For each type of physical interaction there is a standard input-output definition for modular modeling components. [3 Byam and C. Radcliffe, 1999] Modular modeling can be applied to F EA model components. In the discussion below, a basic 2-D elastic stress-strain FEA problem will demonstrate the basic approach for connecting different FEA component models together, even when their nodes are geometrically incompatible. The following figure shows the basic concept for modular modeling in FEA. In this approach, a single, multi-node input/output port is defined to connect two F EA models. The power, force and displacement transfers through this connection port have standard definitions [8. Byam and C. Radcliffe, 1999]. leattomtch,U Figure 2.1. The modular modeling using input / output Ports (connectors) The analysis of system degrees of freedom (DOF) is the foundation of all other analysis. We can not begin any static or dynamic analysis before we decide the system DOF. In FEA modular modeling, the connection of two independent FEA components produces a modular model with a unique solution if it is the zero DOF. In another words, modular connectors provide enough constraints for whole system to generate a non-singular system equation. For example, if we combine two independent FEA components, each with 6, unconstrained DOF, we need to find at least 2x6=12 constraints for displacement to generate a system model with stable, unique solutions for different force inputs. These constraints should provide both external boundary conditions and internal boundary conditions. In Figure 1.1, a constraint set would fix one component to the ground (external constraint) and the other component is fixed on the first one (internal constraint). This combination of constraints will remove all rigid body degrees of freedom and generate system equations with a unique solution. Supplemental boundary constraint equations are critical to control the FEA modular model DOF. In the first step of the proposed analysis, we combine system equations and their degrees of freedom. For example, in 2-D stress-strain analysis, the three degree of freedom unconstrained components have a singular stiffness matrix, whose rank is N—3 where N is the dimension of the stiffness matrix. When we combine two three degree of freedom components together, the combined stiffness matrix has a rank of N+M—6, Where N and M are the dimensions of the two component stiffness matrices. External boundary conditions combined with connection constraints must increase the combined system rank to N+M to generate a model with unique solutions to external force inputs. This idea can be easily understood if we consider this with a physical model. In two-dimensional space, a 3—DOF subject can be held fixed if 3 boundary conditions are provided. When we connect two 3-DOF subjects together, an additional 3 boundary conditions prevent relative motion between the two components. When the 3 boundary conditions on one component are combined with 3 internal constraints, the resulting stiffness matrix is non-singular and yields unique solutions to input forcing. 2.2. Visual Component Boundary Port — FEA Input/Output Connector The modular modeling method generates non-singular equations by using the original component system equations and the modular input/output equation. Actually, this is one of the advantages of the modular modeling method. With this method, a great deal of effort can be saved because no remodeling and re- verification is needed [8. Byam and C. Radcliffe, 1999]. When applying this idea into F EA, the benefit is clear: an FEA model can be used directly in system re- mesh and re-verification of component models are not necessary. 2w: 2f*u=0 I ————— i ~~~~~~ / I "I K [: 3:. a: a ‘2 :> mitt] J/f‘w.‘ ——————————— ’TIN/ U//f u\\f Constrain m n ... ..-... Figure 2.2. The modular modeling method for the FEA application Definition of an l/O port on each component connection will provide proper constraint equations to make the combined system equation non-singular. The component system equations will be applied directly to generate the combined system equations. The Modular Modeling Method (Fig. 2.2) allows the development of standard integration and computer automation methods. Development of IIO port constraint equations begins with the basic properties of FEA model. Force and displacement transferred across a component connection port must satisfy basic energy and work conservation requirements [Yang, T.Y. 1986,]. The total work transferred through the IIO port must sum to zero. 6WA+6WB =0 2, 6W =‘5W3 .......................................... (21) A in which : 6W4 = ZfaN ' uaN = fzai-t ° ”2am +f2al 'uzai +fzaiz-i '"2a2-1 +fzaz 'uzaz + +fdll 'uaN 5W3 = Zfau 'uaM = bel-l 'uzbi—i +f2bl '“2bi +f2bZ-l '“2b2-1 +f2b2 'uzbz +"'+me '“Bm In the upper equations, the indices a1, a2,... aN, and b1,b2,... bN refer to the nodal number along the connection forming the IIO port. FEA m FEA m B A System Equation ‘ for Modular A + comm“ Constroin Equation --— System ( for IIO Port Equation \ System Equation for Modular B It \ :10 port room F. u / Comoction pocflon Figure 2.3. The modular modeling input I output Ports along the FEA boundary Theoretical solution is provided here for a two-dimensional, straight-line boundary and three degree of freedom elastic stress-strain analyses. Future work will be needed for three-dimensional, curve-surface boundary, different type of elements and six degree of freedom finite element analysis. 2.3. F EA Modular Modeling Procedures and Formulations New formulations for the modular modeling need to be developed for the two- dimensional F EA models. The procedures and formulations for one-dimensional FEA modular modeling have been demonstrated [8. Byam and C. Radcliffe, 1999]. The two dimensional problems are different and need updated procedures and formulations. Two-dimensional F EA modular modeling does not require mesh nodal compatibility along the connection boundary. To make F EA modular modeling, we must develop new procedures and formulations for making the F EA modular model work under the situation of mesh nodal non-compatibility along the connection boundary. Methods to connect modular F EA models without compatible nodal point geometry are given below. In this paper, four steps develop a two-dimensional FEA modular modeling formulation, which is demonstrated in section §2.3.1, §2.3.2, §2.3.3 and §2.3.4. The combined system equation of FEA models is provided by step one. Step two applies the boundary displacement constraint equations to transform the system equations. Step three puts the system equations into standard linear algebraic equation form by moving the unknown force into the unknown variable vector. In step four, the boundary force constraint equation will supplement the system equation, and make the system stiffness matrix square. The system model equations are now complete, non-singular and have a unique solution given a set of independent inputs. 2.3.1. Sub-system Equation Combination The first step is to get the combined system equation. Consider 2 FEA components: A and B. A has N nodes and B has M nodes and their stress-strain equation is following. To avoid confusion, The displacement and force on subject A and B are represented by uA, us and fit, f3. They have same definitions and units. K A -UA = FA 0r _ku kn km - pun — rf1 q kn kzz km “2 = f2 (2 2) Lkm km kNN A J'N _ A LfN 3». KB -UB = FB or _ku kn km - -ul - -f1 — kn kn km “2 = f: (2 3) _km km kMM_B _“MJB LfM_B In both equations, KA and K3, are singular. For two dimentional problems the subsystem models allow “free body" motion with two orthogonal deflection and one rotation. 80 the ranks of Kit and K3 are N—3 and M—3. For every individual equation could be solved if boundary condition such as following provided 11. 0 vll 0 u;= u1, = 0; and u;= vb = O ........................... (2.4) uc A 0 VC 3 0 Any 3 U value equal to zero could eliminate the rank of K by 3, and make it non-singular. Now we combine both components together without any re-mesh work. From this equation, we can see that it contains all of the old elements and nodes. 10 K -u =f ................................................ (2.5) syn sys sys in which K A E 0 u A f A Km = . ; (um): ; (1'8”): 0 5 KB uB fB Obviously, the rank of K3,, could be available from following equation: Rank(Km)=Rank(KA)+Rank(KB)=(N-3)+(M-3)=N+M—6 ........ (2.6) From equation 2.6 we know that the combined system needs at least 6 internal and external boundary condition to become solvable. FEA Modular A FEA Modular 8 Figure 2.4. At least 6 constraint are needed for 2-D FEA modular modeling 11 2.3.2. Applying Boundary Displacements Constraints The boundary displacement constraint needs to be defined at first, like most regular FEA procedures. This procedure will make the system equation (2.5) a full-rank linear equation. Therefore, this is the crucial step to make the system solvable. According to the previous analysis, the internal (boundary between 2 components) displacement constraint equation must be more than rank of 3. Assume the following equation: (“21)an =Smxn .(u8)mxm u............................... (27) is the boundary displacements constraint equation, we need to make sure that: Rank(smxn)23 (2.3) The simulation result with the FEA model is a linear approximation for real situation [Zienkiewicz, O. C. and Taylor, R. L., 1989]. Suppose we are using linear elements, the boundary displacements constraint analysis should be based on every nodal value on both components boundaries linear approximation. The finite element method provides a displacement value within every element only according to the nodal value in itself. Therefore the boundary constraint equation could be found by the local boundary elements interpolation. Figure 2.4 demonstrates the mean of the local boundary elements nodal displacement relationship. 12 FEA Modular B 2-Dimentional Displacement Figure 2.5. Nodal displacements along the boundary elements The boundary displacement in Figure 2.4 at node “i” was defined as “u2i-1” and “u2i”. If the boundary nodes on modular A and B is a1, a2,...aN and b1, b2,... bM, then the boundary displacement constraint equation could be written as. (“bl “b2 ace “WK =.'So(llal “02 can uaM);ooooaaoaaaoaoooooo(2.9) specially, the 2 - D equation situation : r (um—i u2bl "2122—1 "21:2 ' ° ' uZW—l ”2w )4 : Smxzrv ' (”201-1 “2m u2a2-l "2a2 ' ° ' uZaN—l uzwv) In most cases the rank of equation (2.9) will be more than 3, otherwise at least one component has only one node on the boundary. So, if we can develop boundary displacement constraint relationships, equation (2.9) will have enough constraints as required. For both subjects (Fig. 2.5), the nodal displacement value along the boundary should satisfy the boundary interpolation [Langhe, K. De, 1995]. 13 2-Dimentional Displacement Case 3 Case 1 Figure 2.6. Nodal compatible situation along the boundary elements Figure 2.5 shows the position of node “b1” has only 3 possibilities relative to the corresponding elements. Case 1 is located on the edge of corresponding boundary element. Case 2 is located on the straight extended line of the edge of corresponding boundary element. Case 3 is coincided with one node of the corresponding boundary element. Independent of which component has more boundary nodes than another, both sides of the port can be used to develop boundary displacement constraint equations. In other words, equation 2.9 has 2 equivalent forms and they are equivalent constraints. Here we present results for nodes on module 8 along the boundary. CASE 1. From the definition of finite element, any internal point value can be represented by the linear transformation of nodal values [Segerlind, L. J., 1984]. For the boundary nodes in another component, we also use similar linear interpolation approach to figure out their values [Langhe, K. 09., 1995 and Kothnur, V.S. , Yu Xie, 1999]. In this paper, the discussion was focused in the basic theoretical solution for two-dimensional, straight-line boundary and three degree of freedom elastic stress-strain problem. So we can apply the similar approach from reference [Langhe, KDe., 1995]. 14 U082) l U(2a2-1) '0 .- .0 Boundary 2-Dimentional Displacement Figure 2.7. Nodal displacements in the boundary elements As it shows in figure 2.6, we can get the shape function of an arbitrary point in the element which has the coordinates of (x, y) [Yang, T.Y. 1986]: “21“: y) = u2al +clx+c2y u2i—1(x’y)=uzal-l+c3x+c4y .............................. (2.10) p The c1, c2, c3, c4 is determined by the shape of element and can be derived from the following equation: (“ZaI-l I (C, I fb, -b, o b, o -b,. o ‘ um c2 = 1 a, -aj 0 -ak 0 a}. 0 . “2a2-1 ...... (2.“) c3 ajbk -akbj 0 b; -bk 0 bit 0 -b1 “zaz Kg) _ 0 a, —a1 0 -a,, 0 a, _ uh}, ( “243 ) From equation (2.10) and (2.11), we get a general interpolation form for any point located in the boundary element. Because the values of c1, c2, 03, c4 are determined linearly only by properties of the corresponding boundary element, we could conclude that the u2b1-1 and u2b1 in (2.10) are linear functions of u2a1, u2a1-1, u2a2, and u2a2-1, u2a3, u2a3—1 correspondingly. 15 The transformation matrix 81, 82 in the upper equation is decided by the corresponding boundary element and the x, y coordinate value of node “b1”. After we figure out all of the equations (2.12) for all boundary nodes on component B, we combine them and get the equation (2.9). r u2b1-l =81 '(uzai-i u2a2-l “us—1) u2bl=sz.(um “202 “203)7 (2.12) In this paper, we are discussing the straight-line boundary problem. So all boundary node “b1”, “b2”, ...... “bM” is locate on the edge of boundary elements. In this situation, the equation (2.12) could be simplified and decided only by nodal value on boundary edge [Segerlind, L. J., 1984]. M2“) UQa1-1)'Na1(x y) “um Uaflmam'” : +UQa2-1)‘Na2(x,y) +UQa2)‘Na2(x,y) i I : I IUWJ) :Ule-l) ' . :U b1 Euaaz) : E :urzaz-n 1UP“) :9 ’ : .1? J1 a2I : at! :m ia2 «0—0- 8 *——_"' 3 .1 t 82 : 81 f 82 s: Figure 2.8. Linear transformation on the boundary elements (case 1: interpolation) Figure 2.7 shows the linear interpolation of node “b1”. Suppose the distance from node “b1” and node “at” is $1, and from node “b1” and node “a1” is $2, which could be derived from the their coordinate values, we can get the linear interpolation equation from the shape function [Langhe, K.De., 1995]: _ _ 32 S1 u2b1-l — Na1(s)'u2a1—1 +Na2(5)'“2a2-1 " ' ' _' ' 'uzai-i 1' '“2a2-1 sl +s2 13-1-32 u2b1 =Na1(s)'u2a1+Na2(3)'u2a2 = ”u2a1+ 'uzaz (2-13) 31 +s2 sl +s2 16 The distance from node “b1” and node “at”, which is $1, and from node “b1” and node “a1”, which is 52, could be derived from following coordinate computation. St = Jam —xal)2 +0511" yal)2 32 =J(be-xa2)2+(ybl —ya2)2 .............................. (2.14) It is necessary to mention that the s1 and s2 values are always positive in case1. CASE 2. In this case, node “b1” is located on the straight extended line of the edge of corresponded boundary element. This happens because node “b1” is the terminal point on the boundary, and it does not contact with the corresponded component. In the following figure (2.8), we can see that equation (2.13) is still valid. The only thing that needs to be modified is that we should to re-define st and s2 according to the situation of “b1 ” location. If the “b1” location comes with the sequence of “a1, a2, b1”, the s1 and 52 should be: SI = ‘J(xbl -xal)2 + (ybl " yal)2 32 = “/09,l "xa2)2 + 0’»: - ya)? .............................. (2_15) If the “b1” location comes with the sequence of “b1, a1, a2”, the st and 32 should be: 31 = -J(xbl —xal)2 +(ybl "' yal)2 52 zfixbl ‘1‘”): +0,“ _ yd)? .............................. (2.16) 17 “UQi-I) U(2a1-1)'Na1(x,y) +U(2a2-1)'Na2(x ,y) uaaz) I I :Ule-I) :UQbI) :UQOI) 82: lbl _ b1! :81 A2 L 81 a 8 : SI s2 p—-—----— I, d on I Figure 2.9. Linear transformation on the boundary elements (case 2: extrapolation) CASE 3. In this case, node “b1” is located on the exact same nodal point of corresponded element. Clearly, this is a “nodal compatibility” situation. It is easy to conclude that if “b1” is located on same nodal point of “a1” _ _ _ 52 SI u2b1-l -uZal-l ‘1'“2ai-i +o'“2«2-i " '“2a1-1 '1' ' 'uzaz-i 31 +32 31 52 “2b: =“2a1 =1’“2ai +0'“2a2 = '“2a1 1' '"2a2 S, +32 S1 +32 If “D1” is located on same nodal point of “a2” - _ _ 52 31 u2b1-l —u2a2-l ‘0'“2a1-I +1'“2a2-1 "' 'uzai +—'_'u2a2 1 52 31 +32 3 s — — . . -—2 e 1 ' u2b1 -u202 ‘0 “2111'” “2a " uza 1' “2a2 .t',+s2 ' sl+s2 U0“) ”90 0 111-1 u 1.1 I U(2b1-1)=UQa2-1) ll ung’Jafi'; ) r UleFUQaZ) : : I ' I : ' i :Uaazi I I ' ' I : : : a1; ‘2: ”1 min ; a2 ”32:0 :7 -H--sl=0 82 t 31 ._ g E i _. E II Figure 2.10. Linear transformation on the boundary elements (case 3: equality) 18 Therefore, equation (2.13) is valid too for this case. When the condition of “s1=0 or 51:0” was represented in (2.13), the output gets the simplified result as equation (2.17) or (2.18). Overall, we could conclude that in the 2-dimentional elastic problem, the displacement constraint equation along the straight-line boundary is equation (2.13). Based on different nodal location, the interpolation parameter s1 and 82 should be computed according to equations (2.14), (2.15), (2.16), (2.17) or (2.18). From these equations, we could get all interpolation functions for “b1, b2,...bM” in figure 2.4. In equation (2.19), sb11 and sb12 represent the s1 and s2 value for “b1”, and Lb1 represent s1+s2 value for “b1”, bl bl bl bl .. -2— . ...L. . — ...2_. . _I_ . u2b1-l " L "201-! ‘1' uZaZ-l . u2b1 " “2a: 1' L “zaz In La La b1 b2 b2 b2 b2 u -—2—-u +‘—-u u -—2--u +—‘—-u 2112-1 "' L 2a2—i 203-1’ 2b2 ‘ 2a2 L 2a3 b2 Ln Lbz b2 saw by saw an .— ._2_ . I_ . — -2— . _I_ . ...... quM—l - L “um-1H '1' uZaN—l ’ “21w " L uZaUV-I) 1' “2m (219) an an m m If we define displacements by node as following: u2bl-l “2122-1 “zen-1 u = u = ’ ...... u = bl ( um } b2 [ “252 ] bM [ “25M ) Equation group (2.19) could be gotten if all boundary nodal coordinate values are available. Also, the equations of (2.19) could be rewritten as the boundary displacement constraint equation (2.20). 19 “bl {Cbl’al “b2 = 0 um I 0 Cbl’aZ 0 Cb2‘a2 Cb2‘a3 0 0 CW‘aUV-I) CW'aN J (2.20) In upper constraint equation, the constraint elements was defined as: Cbl'al = i=1 or 2; (31: 0 I L“ s” I Cbl'a2= 0 _2_. I L») decided by Ii Lu 0 I aY position \ 0 bl EL. Lu) Or the equation (2.20) could be write as: I “251-: \ If}: 0 bl b] 0 it. bl 0 0 0 0 0 0 0 0 \ — 0 o Lbl bl o 3‘— o 02 L“ b2 3.2_ 0 §I_ L112 b2 2 o ‘2 o L,2 o o o o o 0 20 ”ME." of r” If": C O (if: ___ Lax 0 Si— I N 0 o o o o o o o i 0 Lb” 0 ii I“) \ “201-1 “201 “zaz-i u2a2 “203-1 “243 “um-D4 “zerN-i) uZaN-l I“2a~) 2.3.3. Applying Boundary Forces Constraints Boundary force constraints analysis is required to solve the system equation. From the description of §2.3.1, we know that after boundary displacement constraint equations convert the system equation to full rank, the boundary force constraint equations and initial conditions are also required to solve the system. In fact, the boundary force constraints equation is very important to enable the system equation to get the final solution. The corresponding mathematical Km'“1xm =fm; n) m. After applying the description is, if we need to solve system boundary displacement constraints equation, the equality of the rank of stiffness matrix K' and f ' can only guarantee the solution exist, but it can not guarantee that the solution is unique. It is absolutely necessary to have the force constraint equations for the system equation. The physical sense of this statement is: both the boundary displacement and force constraint equations are required to ensure that the total power (for dynamic model) or work (for static model) flow through the boundary is conserved. Modular modeling ports provide the power constraint. Power is conserved across modular ports because modular ports are power (or work) transfer mechanisms. The power at the connected modular modeling ports always sum to zero. For the elastic stress-strain problem, we need not make dynamic analysis work. Therefore, the work at the connected modular modeling ports always sum to zero. The work along the boundary defined as: (WV = 23W, = 2:6, -u,.) =1"r ~u Because work across the connection of component A and component B will be mA+mB=o z, MA=_MB ....................................... (221) -. (rT-u),+(rT-u),, =rI-u,+r;-u,,=o 21 from Equation (2.20), we get (ff -u,,)+(f; -C.uA)=0=> (f;w +f; -C)-uA =0 Because the power conservation is valid for any displacement under elastic range, we can say that the equation (2.22) is valid for any of those displacement. Therefore, we could conclude r,’ +r; -C=0 and r,+c-r,, =0 Especially, for 2-dimentional case: r T I f 2a1 \ (“201—1 \ f bel I (u2b1-1 \ f 2a1-1 “201 f 2b1-1 “2121 f 2a2-1 “2a2-1 f 21,2-1 “21:2-1 .0 6wx=2fm'uuv= fzaz ' “zaz ma=2fsu'ueu= f2b2 ' “21:2 fZaN-l uZaN-l beM -1 “nu-1 (fzalv) (“2w ) (fun) (“21:14) The equation (2.1) should be the beginning step of the constraint equation for nodal forces. From equation (2.1), we get ( f2al NT (“201—1 \ f f2“ NT r“zen-1 \ f 241-1 “m f zen-1 “2b: f 2a 2-1 “2a 2-1 f 2b2-1 “21:2-1 mx+ma=2fm'um+£fm'um = f2a2 ' “2.22 '1‘ fun ' “2b: :0 fZaN-l uZaN—l fun-1 quM-l (fun; (“zany (fsz 1 (“2w ) When we place equation (2.20) into the above equation, 22 I f 241 )7 r“zen-1 \ I f 251 IT (“w-1 \ f2a1-1 “2411 f 2b1-l “2a: f2a2-l u2a2-l f 2b2-1 “2.:2-1 5W4 + We = fzaz ' “2n '1' f2b2 ' 8' “zaz = 0 """"" (2°22) f2aN—l uZaN-l f 2m -1 “zen-1 If“) Iuw) Ifwl Iuw) Is“ s“ \ A 0 —‘—- 0 O 0 0 0 0 0 La La s“ s“ 0 J— 0 l 0 0 O O 0 0 Lb1 1%] Std Sta 0 0 J— 0 —I— 0 0 O 0 0 Ln [12 in which 8: s” s” 0 0 O J— 0 ' 0 0 0 0 All work is done by elastic force and displacement 2-Dimentional Displacement FEA Modular 8 Figure 2.11. The power conservation along the Boundary ‘ Figure 2.10 shows the sum of work transferred along the boundary equal ZERO under the situation of the elastic range. 23 2.3.4. Combine the Supplemental Boundary Constraint From the boundary constraint equations (2.20) and (2.23), we get the total 2M constraint equations (if MN). From figure 2.12, we can see that when the system DOF is 0, there must be M>3 and N>3, and the system equation is mathematically non-singular. UFS'UP lntemal boundary External U2i-1=0; 3431310 constrains boundary U2i=0; conditions: U2k-1=0; \T Figure 2.12. The “solvable system” and related equations. » For any solvable system like figure 2.12, the system equation, boundary condition and boundary constraint equation should be combined through the following procedures. Step 1. Use the FEA models to get the combined system equation (2.5), the boundary displacement constraint equation (2.20), and the boundary force constraint equation (2.23). All of these constraint equations could come out from nodal coordinate values and HG port shape. If KA is pxp, KB is qxq matrix, then K A 5 0 uA “,4 f4 =KAB° —_- .............................. (2.23) 0 KB “3 “3 f3 KAB is a (p+q) x(p+q) matrix. 24 ‘ Step 2. Apply the external boundary conditions and boundary displacement constraint equation (2.20) to transform system equation. This step will replace [u2H u2i u2k_,]=[0 0 O]; and [UZbI-l Um UZbZ-l Um U2bM-l Ule 37 =S' [Um-t U2al UZaZ-l Um U2aN-l UZeNK the displacement of some boundary nodes. For example in figure 2.11 the external boundary conditions and boundary displacement constraint equation is Using upper equations, we can get rid of u2i-1; u2i; u2k-1; and u2b1-1; u2b1; u2b2-1; u2b2; U2bM-1 and U2bM-1 from UA and U8, and get following equation (assume N>M): uA u’A KA E 0 u, uA fA u g! cI-l ll ’3 a II (2.24) In which TU is a (p+q-3-M) x(p+q) transformation matrix, and K'AB is a (p+q- 3-M) x(p+q) matrix that is transformed from KAB. Step 3. Move the unknown forces (external and internal boundary forces) to the left side, and leave the right side of the equation as a totally known force vector. For example, if 1], fk, fl, fm in [FA FB]T is unknown, we can simply replace it by “0". This will make the system equation become a standard linear algebra equation, in which TF is a (2M+3) x(p+q) matrix and K’AB is a (p+q+M) x(p+q) matrix. {u’A \ f u" 1 “’4 {A f’A K’u' = (K'AB 5 Tr)’ “’3 :10”. “’8 = (2.25) “’3 f8 f’B Kf’A+B ) \f’A‘l-B ) 25 Step 4. Apply the boundary force constraint equation. Because all boundary forces in equation(2.23) is unknown forces that have been moved to left side, we can directly apply the boundary force constraint equation into the system equation. (u’A \ (u’AI (PAN fital f2M K’ T .. f f AB ' F 2.02 + 2:112 ’SF=Q =9 . . “’8 =K"’AB. “,8 = f’A °-- (2.26) 0 : S’ r r ’ 2aN 2»! (rant) (rho) KO) Because the boundary force could provide M constraint, the system equation matrix becomes a square matrix with full rank, which SP is a (2M+3) x(M) matrix, and K’”AB is a (p+q+M) x(p+q+M) matrix. Therefore, the final integrated system is (u’A I (PA) K’AB T; “’3 = f’B ..................... (2.27) o s 8', If”) I“) By solving equation (2.27), the system internal forces f’A+B and displacement distribution u’A and u’B are available. In following chapter, I am going to demonstrate this approach with a example. Through this example, the system modeling and integrating details could be shown clearly. 26 2.4. Result and Discussion Modular modeling system FEA equations can be developed from the subsystem equations, external boundary conditions, and internal constraint equations. If enough constraint equations are available and the DOF is zero, the modular modeling FEA equations can be solvable. The general form of the combined system equation (2.27) has been developed here for the two- dimensional elastic problem. Modular modeling FEA equations could be realized automatically with a proper algorithm. The reason is that all of parameters used in combined systems come from individual FEA model and connection description, such as nodal coordinates and boundary conditions. This point is valuable for the modular modeling method because this means that the system could be integrated by the user's selection of parts and connection type. Solution of the modular modeling formulation can be shown to be equivalent to solving the associated Lagrangian formulation [Appendix I, Radcliffe and Diaz, 2000]. Although the modular modeling formulation is derived based on only work constraint at the boundary, its solution is identical to the solution of the Lagrangian formulation which analysis minimum potential energy for the system. It is gratifying that the solution for the modular modeling method coincides with the traditional energy-based Lagrangian approach. 27 TWO-DIMENSIONAL EXAMPLE FOR ELASTIC STRAIN-STRESS PROBLEM 3.1. FEA static stress-strain analysis by modular modeling . To demonstrate the last chapter work about integrating the modular modeling FEA systems, a simple example for a 2-D elastic plane force-displacement problem will be solved by the approach and procedures in the chapter 2. After the combined system has been solved, the result was compared with same object but derived from FEM using two different mesh densities. In figure 3.1, there is a combined plane willbe a subject of load. The two triangular components have FEA models: plane A has a FEA stiffness matrix KA, plane B has a FEA stiffness matrix K3, 3] /’ $1 Plane A Load I /,’ y Plane B Load Figure 3.1. Two-Dimensional Example for Elastic Strain-Stress Problem From figure 3.1, the subject has 3 external boundary conditions so it is O DOF, using the approach in last chapter and original equation (3.1) and (3.2), applying the external and internal boundary constraint (port) equations. The system equations are singular. 28 3.1.1 2-D Elastic Plane A: We need to get the individual plane meshing and system equation. From the following procedures, it could be solved like any traditional examples [L. Segerlind, 1984]. The related system properties are: Young’s modulus: E = 15x106 (N/cm2); Poisson’s ratio: u = 1/4; Plane thickness: t = 0.1cm. Finite element summary: eleme i j k (6 (8 (1) 1 3 2 (2) 3 4 5 (2 (4 (3) 2 3 5 (3 (4) 2 5 6 2 Y (5) 4 7 8 (1 f ' (6) 5 4 a ,x (7) 5 8 9 ¢ 1 (8) 6 5 9 (9) 6 9 10 Nodes’ coordinator values: Nod 1 2 3 4 5 6 7 8 9 10 e X 0 2.0 0 0 2.0 4.0 0 2.0 4.0 6.0 Y 0 2.0 2.0 4.0 4.0 4.0 6.0 6.0 6.0 6.0 Figure 3.2 Plane A meshing details and nodal distribution ‘ By the traditional method [L. Segerlind, 1984], we get all element stiffness matrixes in plane A: [K( 1)], [K(2)], [K(3)], [K(4)], [K(5)], [K(6)] and combine them together as following. See the Appendix II for the detail of computer software. 29 50 1 X 00000000003000004603 00000000000200006480 00000000506M00685fl£6 00000000056500622586 000000506fi0044522400 00000005650064253600 0000006600005H4£0000 000000420000fl5£60000 00300000665fl00006fl20 OOOZOOOOMfiflSOOOOfiéOS 006%506640IM66006M5000 oossoswsumwsoossosoo OOOOGB5£6£004£500000 00006425M600£8050000 £8665fl££500000000000 64%6a566050000000000 305266006M2000000000 02%54Aw600660300000000 08202600000000000000 30038600000000000000 = m 3.1.2. 2-D Elastic Plane B ~ The meshing and system equation for plane B could be done by similar way. It has same static properties as plane A: Young’s modulus: E = 15x106 (N/cm2); Poisson’s ratio: u = 1/4; Plane thickness: t = 0.1cm. Finite element summary: eleme nt 0) m m IA\ Loa Nodes’ coordinator values: 6.0 6.0 6.0 3.0 3.0 3.0 6.0 3.0 1 Node Figure 3.3. Plane B meshing details and nodal distribution 30 By the traditional method [L. Segerlind, 1984], we could get all element stiffness matrixes in plain A: [K(1)], [K(2)], [K(3)], [K(4)], and combine them together as following. See the Appendix III for the detail of computer software. can -8 -2 0 -3 -3 0 (DO rods and: N ”'ro {301 II coco m¢$oo I or I d om x105 I N I CO 01 .5. .5 o o I 0: I m (00600000 I O) I 0| 0 00 N N 01 I d or I 01 ONOOOOOO I (D I I ..I. a) I U! N N U'l I 03 CO I I [0 doc: I 01 cl: (1,01 MS “I I care OOOONOOO 00000000 0 CO GOD I N I a: O 000 3.1.3. Using Modular Modeling To Integrate 2-D Elastic Components We are using this example to demonstrate the details of the modular modeling FEA approach. We need to follow the steps one by one according to the description in previous chapter. The final solvable system equation is too large to be shown here. For detailed computing procedure and result, please review Appendix V. 31 3.2. Error Estimation and Analysis To validate the result from §3.1, we need to compare it with a traditional model. Both higher and lower meshing resolution models are needed for the comparison to verify the modular modeling result. For a linear problem, higher resolution meshes yield a more accurate result. The modular model mesh has more resolution than then low resolution reference model but less resolution than the higher resolution model. We expect the modular model to be more accurate than the lower resolution mesh model and less accurate than the higher resolution mesh model. Comparing these models will provide useful information about the detail in the model and component connections at the boundary. 3.2.1. Standard Reference Models for Accuracy Comparison Two standard reference FEA models (Fig. 3.4) will be used to evaluate the accuracy of the 13 element modular modeling example (Fig. 3.1). The higher resolution standard reference model has 72 elements while the lower resolution standard reference model has only eight elements. We use the same coordinate systems, load and FEA equations to ensure the accuracy of the formulations are directly comparable. /l 43 44 45 46 47 4s 49 /l (61) (63) (65) (67) (69 (71) 7 a 9 36 (62) (64) (66) (68) (70) (72) 42 (49) (51) (53) (55) (57) (59) (5) (6) (7) (a) 29 (50) (52) (54) (56) (58) (60) 35 (37) (39) (41) (43) (45) (47) 22 (33) (40) (42) (44) (46) (48) 28 (25) (27) (29) (31) (33) (35) 15 (26) (28) (30) (32) (34) (36) 21 (13) (15) (17) (19) (21) (23) (1) (2) (3) (4) s (14) (16) (18) (20) (22) (24) 14 (1) (3) (5) (7) (9) (ll (2) (4) (6) (8) (10) (12) 1 2 3 4 5 6 7 1 2 Load Load Figure 3.4. Re-meshed reference model with higher or lower meshing resolution 32 3.2.2. Modular Modeling Method Accuracy Comparison Displacement contour lines were used for comparing the overall system accuracy. Displacement means the total displacement at every node. If a node has “a” displacement on X direction and “b” displacement on “Y” direction, its total displacement is (Ja’ +b2 ). Because we are focusing on the linear case and the load is same, the linear static elastic problem should have the same and repeatable result. Therefore this comparing is reliable. Predicted displacement for the modular model were compared to the displacements predicted by the higher resolution standard reference model. EXCEL was used to compute and plot total model displacements (Fig. 3.5). The 3—‘D displacement difference surface (Fig. 3.6) shows the largest model differences are less than 10%. In particular, nOte that the modular model results are almost universally smaller displacement than the higher resolution reference model. This indicates that the modular model is stiffer than the higher resolution reference model. 3031...... . 9);;ng U, ’ 47' ...-W l- W SE \ 0!. ’\\m “‘ I.., 'r "u I I“:’,. ,1. . 9., . .~. “.‘_‘ h. ' . "I. "'2’”, ' to :33: .. % Load Figure 3.5(a). Displacement contour Figure 3.5(b). Displacement contour lines for reference model (7x7). lines for the Joint Modular Model 33 Nodal Dleplecem ant Difference j; _ ' .1500); 40.00% 45.00% ‘ ., .0056 45.0056 .0056 «10.00% Figure 3.6. Displacement difference between the Joint Modular Model and reference model (7x7) Predicted displacement for the modular model were next compared to the displacements predicted by the lower resolution standard reference model. EXCEL was again used to compute and plot total model displacements (Fig. 3.7). The 3-D displacement difference surface (Fig. 3.8) shows that in this case, the largest model differences approach 25%. In particular, note that the modular model results are almost universally larger displacement than the lower resolution reference model. In contrast to the comparison with the higher resolution reference model, this results indicates that the modular model is less stiff than the lower resolution reference model. Figure 3.7(a). Displacement contour Figure 3.7(b). Displacement contour lines for reference model (3x3). lines for the Joint Modular Model Nodal Displacement Difference n20.00%-25.00% I15.00%-20.00% l10.00%-15.00% I5.00%-10.00% D0.00%-5.00% s7 n-5.oo%-o.oo% I-10.00%--5.00% [moose-40.00% Figure 3.8. Displacement difference between the Joint Modular Model and reference model (3x3) Comparison of the three models (fig. 3.9) indicates that nodal displacements along the boundary of plane A and B uniformly converge to the higher resolution standard model result as the number of elements in the model increase. Because the number of elements in the modular model lies between the lower and higher resolution reference models, its displacements are bracketed by the 35 two reference model displacements. As expected, the model stiffness uniformly decreases as the number of elements is increased, indicating the modular model accuracy is comparable to traditional FEM models with the same mesh resolution. Bet-day III-placement Load 1 0.0014 ‘3‘ 0.0012 z' \9\ 0.001 \4 \\5 ; 0.0000 \ \ 0.0000 \5 1 %] \ 0 0004 \ 7 Figure 3.9. The difference of nodal displacements along the boundary 36 CONCLUSIONS 4.1 Contributions Application of the modular modeling method in FEA has been discussed; its advantages and efficiency have been shown in examples and numerical results. Most importantly, the 2-dimensional boundary constraint equation that is used to connect FEA modules has been developed and verified in examples. Modular modeling method is shown to use the original FEA data, removing the need for re-meshing and verification, so it saves cost and time. From a general description and example, we believe that the equation (2.5), (2.20) and (2.23) is the proper way to get the combined system equation and boundary constraint equations. For actual engineering problem, just like the example, we can sequentially use the equations (2.5), (2.24), (2.25), (2.26) and (2.27) to get the solution. For 2-Dimentional elastic stress-strain problem, if the system have 0 DOF, we can conclude that the developed system equation of FEA modular model is non-singular and could be solved with the boundary constraint equations, external boundary conditions and loads. The modular model has good computational accuracy and quality that is close to the re-meshed higher resolution model and is much better than the relatively low-resolution model. The modular modeling method was shown to provide accuracy comparable to traditional FEA models with comparable mesh resolution. The ability to achieve this accuracy without remeshing is a significant improvement to FEA modeling technology. Result Preclslon and Quality improving \ Low Meshing Resolution Modular Model High Meshing Resolution Reference Model Assembly Reference Model Figure 4.1. Comparing the Modular Model and other models 37 4.2 Future work The equation developed in this paper can be only applied to a simple 2-D problem, the accuracy of this approach will be decided by the equality of boundary equation to make this approach practically useful in industrial applications, we must develop the following topics: 9 Convergence of the stiffness matrix 0 All kinds of meshes o The 3 dimensional flat boundary 0 The 3 dimensional arbitrary boundaries ‘ o The Boundary Constraint Equation development automation, '. Apply modular modeling method into the nonlinear FEA problem. There are 2 ways to develop the boundary constraint equation. From equations (2.20) and (2.23), we can see that component B was treated as the subject of interpolation from component A. It is easy to understand, the contrary way could also provide similar constraint equations. Those equations should also be enough to make the system solvable. But the problem is, “which one is ‘ better?” I believe that this is a more complex and practical problem and needs research through great number of actual case, cause the meshing and constraint have so many kind of types that it is very hard to simply identify them. Theoretically, the boundary interface input / output equation between components will be only decided by their shape function. Therefore, the automatically formulation should be practically realized by proper programming. This feature should be one of the critical block in the future modular modeling FEA software. 38 REFERENCES Byam, B. P. and Radcliffe, C. J., 1999, “Modular Modeling of Engineering Systems Using Fixed Input-Output Structure”, 1999 IMECE, Proceedings of The Symposium of Systematic Modeling, ASME, New York, November “Computers in Engineering: Chrysler designs paperless cars”, 1998, Automotive Engineering lntemational, Volume 106, Number 6, (June): Page 48 Segerlind, L. J., 1984, Applied Finite Element Analysis, 2nd Edition, John W&S, New York Zenkiewicz, O. C. and Taylor, R. L., 1989, The Finite Element Method, 4th Edition, McGraw-Hill, New York Yang, T.Y. 1986, Finite Element Structural Analysis, Prentice-Hall, New Jersey Meirovitch, L., 1967, Analytical Methods in Viberations, MacMillan Publishing, New York Rosenberg, R. C. and Kamopp, D. C., 1983, introduction to Physical System Dynamics, McGraw-Hill, New York Langhe, K.De., Finite Elements with boundary interpolation applied to the solid mechanics field problem, Computers 81 Structures, Vol.54, No.5, pp.851-857, 1995. Aminpour, M.A., Ransom, J.B., and McCleary, S.L., “A Coupled Analysis Method for Structures with Independently Modelled Finite Element Subdomains,’ lntemational Journal for Numerical Methods in Engineering, Vol38, pp. 3695- 3718, 1995. Kothnur, V.S. , Yu Xie , “T wo-dimensional linear elasticity by the boundary node method,” lntemational Journal of Solids and Structures, Vol.36, No.8, pp.1129- 1147, Mar. 1999 39 Ransom, J.B., McCleary, S.L., and Aminpour, MA, “A New Interface Element for Connecting Independently Modeled Substructures,” AIAA Paper Number 93- 1503, 1993. Belytschko, T., Lu, Y.Y., Tabbara, M., “Element-Free Galerkin Methods for Static and Dynamic Fracture," lntemational Journal of Solid Structure, Vol.32, No.17/18. PP 2547-2570, 1995 Hesthaven, J.S., Gottlieb, 0., “Stable Special Methods for conservation laws on triangles with unstructured grids.” Computer Methods in Applied Mechanics and Engineering, Vol.175, 361 -381 , 1999 Zhao, Zhiye, “A Simple Error Indicator for Adaptive Boundary Element Method.” Computers and Structures, Vol.68, 433-443, 1998 Besseling, J.F., Gong, 0.6., “Numerical Simulation of Spatial Mechanisms and Manipulators with Flexible Links,” Finite Elements in Analysis and Design, v.8 n1- 3. 139121-128, 1994 Elkaranshawy, H.A., Dokainish, M.A. “Corotational finite element analysis of planar flexible multibody systems,” Computers and Structures, Vol.54, n5, pp881- 890, 1995 40 IV. VI. APPENDIX Lagrangian FEM Modular Modeling notes The Mathematica® program for plant A. system stiffness matrix computation. Filename: Plane_a.nb The Mathematica® program for plant B. system stiffness matrix computation. Filename: Plane_b.nb The Mathematica® program for re-meshed plant. Work as the reference for error analysis. Filename: Plane_all.nb The Mathematica® program for combined plant. System integration. Filename: Plane_a+b_connect_6.nb The MS Excel data table for the result of FEA computation and error analysis. 41 TO: Mr. Xin Li From: Clark Radcliffe Subject: Lagrangian FEM Modular Modeling Date: April 7, 1999 Per discussion with A. Diaz 4/6/00 at your MS defense, I have outlined below the relationship between the Modular Modeling Method and the Lagrangian approach to the joining of two bodies with constraints. The Modular FEM Modeling Problem (Xin Li MS thesis,4/6/00) is formulated as K, 0 . u, = 1’, (1a) 0 K, n, f, ‘,=S-fi,and (1b) W,+W,=0 (10) where u, =[2‘]. u, =[2’], f, =[g‘], f, =[;"] A B A B and the work done on the plates A and B by the connection are W, = f, -1’i , and W, =f',-1‘1, respectively. Substitution of the connection forces and displacements into (1c) yields the relationship between internal joint forces A fl=$flg (my subject to the constraints All variables denoted ”bar" variables (7) refer to 'extemal” quantities not involved in a joint while all “hat“ variables (it) refer to 'intemal" joint variables. Partitioning the equations and rewriting them K, E, o 0 'fi,‘ 1,, K5 K4 _" 3 .‘I‘ = 5‘ (2a) 0 O K B K a “a fa I. O O E; KBJ 1.6 B .I ..f3 .. subject to constraints Ii , = 90, (2b) and r, = -s’r, (20) Question: Is the solution of this problem equivalent to the solution of the Lagrangian formulation? The Lagrangian Problem (A. Diaz, 4/6/00) Given the potential energy function 42 H =-;—u:K,u, Jig, +-;—u;K,u, 41$, (3a) subject to displacement constraints 11, = S «11, (3b) The Lagrangian is L=II+(fi, -S-0,)T J. (4) =%u:K,u, 4111?, +%u;K,u, -fi;f, +(1’i, -3-1’i,)T VI The solution of (3a) and (3b) is a stationary point of (4) that minimizes the potential energy (3a). Taking the derivative of L with respect to each of the independent variables u,,u,,}». 3f =0=K,fi,+f(_,fi,—f, (4a) Bu, 3L Ar-- “ a “r .. =0=K,u,+K,u,—S/1 (4b) 3“) 3_L =0=K,fi,+fi,fi,—f, (4c) an, a]. =o=f<';i,+1‘i,0,—2 (4d) Bun 8L . ., -a—A—=0=u,-Su, (46) Putting (4a)-(4e) into matrix form yields, F— :— ‘ -_- --- If, K, o o o u, f, K’,’ K, o o —s’ fi, 0 o o Ki, 0"fin=fa (5) 0 0 K; K, I fin 0 _o —S o I 0_-/1-_0. Note that (5) is a symmetric formulation. 11 can be shown that solutions to (4a-e) rewritten as (5) are also solutions to (2a- c). First note that the Lagrange formulation equation (4e) is equivalent to Modular Modeling constraint (2b). 0:0, —Sfi, 4:: 0, =50, (6) 43 Let the Lagrange multiplier, 2. = —f,, the intemal boundary force on plate B. Then, the Lagrange formulation equation (4d) is equivalent to the fourth equation in the Modular Modeling matrix equation (2a). 0: K ,u,+K,fi,+/1¢=IK,u,+K,fi,=f, for/1: (7) The Lagrange formulation equation (4b) is equivalent to the second equation in the Modular Modeling matrix equation (2a). Use the force constraint (2c) from the Modular Modeling formulation to redefine the internal force f, appearing in the second equation in the matrix equation (2a) to write o= K (u,+K,u,—s’2eKTu,+K,u, =f, forf, =-s’r, =S’A (8) The Lagrange formulation equations (4a,c) are identical to the first and third equations in the Modular Modeling matrix equation (2a). 0 = K5, +‘K’,a, —f, :1 Km, +K‘,0, =1”, (9) o = K5, +K,a, —f, :1 Km, +K,a, = f, (10) The conclusion is that the two formulations have identical solutions with the recognition that the Lagrange multiplier it = —f,, the joint internal force. General Form of the Displacement Constraint A more general form of the displacement constraint is the linear relationship SAuA =8,“, (11) with this relationship, the Lagrangian formulation becomes 1 r1 = 5.15191, -..51, +—;-u;1<,u, —u,T,f, (12a) subject to displacement constraints S,u,=S,u, (12b) Note that in this case, I have not partitioned the displacements u, and u,. The body forces 1‘, and I, represent all external forces applied to the two bodies and are simply a superset of the previous external forces f, and i, that also allow external tractions to be applied at the position of joint connections. The Lagrangian is L=H+(S,u, —S, -u,)T J1 13 =%U:K,u, -u:fA +%u;KBuB "uirafa +(Saua _SA ~u,)T '2' ( ) whose stationary value occurs at the solution to 31L =0=K,u, —f, -s§2 (14a) A 3‘1" =0=K,u, -f, -s;2 (14b) B %=0=S,u, —S,u, . (14c) Putting (14a)-(14c) into matrix form yields, ' K, 0 —S§ u, f, 0 K, Sf, 11, = f, (15) -S, S, 0 ,1 0 Note that (15) is a symmetric formulation with u, and 11, defined as in (1c). With this definition, the joint constraint matrices in (15) can be used to define the original Modular Modeling formulation by defining, 0 0 S,=[0 5,] (16a) 0 0 and S,=[0 1] (16b) 45 APPENDIX II 1 II 8 15000000; u I 73 I: a 0.1.3 'ayeten propertlea: Young'a modulus (II/4:119); Poieeon'a ratio; Plane thickneaa («)3 '3 II- {10 3: 2: 2: ‘1 5r 5: 6: ‘}}W8 {3: ‘0 3: 5: 7p ‘1 B, 5: 9}, “I {2' 5: 5: 6r .1 a! 9: 9! 10}, final: for plane A: nodee diatrihution, pick up the 13:] 3k: of In element: by: “In-filial]: :lnafldifnllr haullnll'r '3 X:- {0, 2, 0, 0, 2, 4, 0, 2, A, 6}3 Y:- {6, 4, 4, 2, 2, 2, 0, 0, 0, 0}; up: value tor plane A nodee, pick up the In a 1nd of In node by: '3 'anx[[n]]3 m-![[n]] '3 ”fulfil: ”IMIYIBL Bltmeyllk, 9];Array[c:l., 9];Array[c:j, 9]3Array[¢k, 913 9013143] 'I(Y[[(W[[n]])ll)-(Y[[(NKIIDI])]]): {nu 9}]: DOIDSIDII(YIUNKIIIIIJHH'(Yllflullnnnl): {no 9}]: I”[33:01]3(YIIWIIIIIIDII)-(Y[[(N~T[[n]])l])n (n: 91]: DOICi-lnl-(IIHflKIIDIIHI)-(3[[(U[[n]l)]])a (n. 91]: Dolcilln]I(XIIINIIIBIIHJ):(IIHNRIIDIIHIL (n. 9}]: no[c1:[n]a(X[[(m[[n]])]])-(1[[(u1[[n]])]]), (n: 9}]: 'Detlne 31,83,319 c1,cj,c1: for all eleuente, cowute every value by: " 3 '31 [n] “3 [n] Jilin] 3 3:1 in] “Hill ~11nt I It in] .n In] -!:I in] ' 3 'c1[n]-n[n] -xj [n] 3 c: [n] .11 [n] -n [n] 3 Ch [n] :13 [n] -n [n] " 3 't - “bl-[{IIHNKIIBII)llorliflmlfnllll]I. {n.9}]3nlatrlot:[t:,vlot:aolned -> True, Plotnabel->'e ']'; 31 ('0 1: .).-I C1(", 1: 'I"! (C1[1]), Do[Pr:l.nt['Bi(', 1, -)--, (31(1)), - (EH11): ' “('I 1: 'I"! (”(11): ' ' Cj(', 1: ')"1 (C3 [1])! 1 llHNIIInIIHI 1 IIIOWHDJDII 1 no[a- -xDet[( ’ 1 xrumrnnm 1 Arraypl, 9]; Do[nn[n] a fi[ -pr1ne[ae[n])-, (n, 9} ]; Dn- Arraymel-ent, 9] t not“ - (Trauma-[(881111)] -DD) 2 31 [n] 0 0 (:1 [n] 0 (:1. [n] 31 [:1] c3 [n] l-ua ' “('1 1, 'I": (“[1])10 {1, 9} 13 YIHNIIIDJIHI Y[[(W[[nll)ll )13 'Prlnt[h]', {n a 9}]; YIHNKIIDIIHI a: [n] 0 urn] 0 cjfn] 0 Chill] 3 Ella] Chin] 3Mn] 133;.) m 2 Hana] 3 x a 1: 015m; Keleuenfln] :- nationallaeflt] 3 I'l’rint: [xeluent [n] ] ', (n, 9}] 46 3 (000000000000000000001 00000000000000000000 00.000000000000000000 00000000000000000000 00000000000000000000 00000000000000000000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O 0 0 O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O 0 0 0 0 0 0 0 0 O 0 0 0 0 O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (0 O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O 0) 8.11 I '31-1-11kfll.21-13k[21.25-1-1dkf3l.21-1-13k[4].zk-1-13k[s],2x-1jk[6],-, Do[Print['elenent: ', e]3 ijk[1] - (2xfl1[[e]])-l3 Arrayfljk, 6] 3 1:139] . (”WING“): 13kI3I-(3XNJIIOII)-13 13k“) ' (3XWIIOII): 1211:1518 (3XM[[O]])-1r 11k”) 8 (”WING”): Mann-80101311140]: DO [mini-llliiklilo ilklillll 8M11II13RI1]: ilklllll «tutti. 5]]: {1. 5}]. (j! 6}]! {.0 9}], 'rill the element K into the integral 8(xall), baeed on every corrdinated nodee'; 447 000) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ~3 -3 ~2 ~8 3 0 2 22 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 -5 ~6 5 ~16 -5 22 2 0 0 0 0 -5 -6 0 0 —5 ~16 2 ~3 —3 —2 —8 22 5 -3 —2 ~16 5 0 —5 -6 0 5 5 -5 -6 0 0 0 5 22 ~3 ~2 ~5 ~3 —8 -3 ~8 0 0 -3 ~3 ~2 ~8 0 -5 5 ~16 S 22 22 0 0 -5 ~5 ~16 0 0 -6 0 0 0 0 0 0 0 ~5 ~5 ~16 ~5 44 0 0 -6 -6 0 0 10 ~16 10 44 ~5 ~5 ~6 5 ~16 0 0 0 0 5 0 0 0 0 -5 5 —6 -5 22 0 0 -3 -3 0 0 0 0 0 0 5 22 -5 -6 0 0 ~16 0 0 —3 ~2 ~3 -8 —5 ~16 2 0 5 ~5 5 ~8 5 11 -2 ~8 -2 22 11 0 0 0 0 0 ~6 0 0 0 0 0 0 0 -5 5 ~8 —3 0 0 -3 22 -2 0 0 —6 0 0 0 -3 ~3 S —5 -8 ~5 ~16 0 0 ~3 0 0 0 0 0 0 0 0 ~5 ~16 5 0 5 -8 ~3 22 -2 ~3 -2 22 -3 0 0 ~2 -3 5 -8 -3 5 0 0 0 0 3) 3 0 I 48 APPENDIX Ill 1 n 8 150000003 u 8 T; t n 0.1; 'syeten propertiea: Young'a noduluem/cm’); Poisson's ratio; Plane thickneea(an)3 '3 III-{1, 2, 2, 4);”: {4, 4, 5, 6km: {2, 5, 3, 5}; 'neah for plane A: nodea diatribution, pic): up the 1333):; of in element by: '3 ~1n=x1[[n]]: :ln-Milni]: manila]?! 13(01 3' ‘3 3' 6' 6}; ’3 {0' o, o. -3, '3’ -6}; -r,! value for plane A nodea, pic): up the n a m3 of tn node by: '3 annual]: tn-Yflnii '3 marl”. 4]: multi. 4]; Arreyfnk. 4]: Muriel. 4]; Arr-tried. 4]: Arreylck. 4]: 0018113] . (YIIMJIIDIIHI) - (Y[[(NK[[n]])l]). tn. 4} ]3 0013113] - (Yllflmllnnnll - (Yllffiiiniinl). (n. 4} I} 001315an 1* (Y[[(N1[[n]])]])-'-(YIIWJIInIIHII. {no 4} l: bowl-in] - (xmmrtnnm) - (HUN-7111111)”). (n. 4} i: DOlczllnl-(1[[(NI[[n]])i])-(311013111111th (n. A} 13 Miami] - (111(m11n11111)-11). (n. 4} ]: 'Detine bi,bj,3k,¢i,¢j,¢k tor all elemente, coupute every value by: '3 “31131-15 In] 41513] 3 3:1 [11] .“Mill -Y=|- [n] 3 33! [n] .n [n] -¥:l In] " 3 “C1 In] .13: In] 41 in] 3 Cd [DI-111111415111] 3 Chin] =le [111-11111] ' r '1: ' TIblOHXlHl-fllnllnlflll("011311)”). (31.4)]3LiatPlot[t,PlotJoined -> True, Plothabel->'e ']'3 Do[Pr1nt["B:l(', i, ')I", (Bi[:l.]), " 330', :l, ')s", (Bj[:1.]), ' “('1 10 '1"! (“111)! ' Ci(', 1' '1"! (Ci-[1]): " CM": 1: ')='a (C3111): " CH“: 1. ')-'- (035111)]: {1. 4}] 1 1 xmmrnnm YIHNIIMD]! Do[A=-2—xnet[[1 111014111411)“ YIHWIInllllll]:'Printlhi'. {n41}: 1 xmmrnum trumrrnnm 1 Bi [n] 0 33 [n] 0 h):[n] 0 Arr-”[33, 9]; Do[nn[n] - — [ 0 (:1 [n] 0 c: [n] 0 Ck[n] 3 3“ c1 [n] 31 [n] c: [n] 3:1 [n] C):[n] Bk[n] l u 0 'Print[nn[n]]', {n, 4”,». n [u 1 0 ] 1‘“: 0 0 ‘—’“— 2 Array[xelenent, 4] 3 Dom - (Tram-Ouufnill JD) 3 m - KR.BB[n] 3 R s t *Atm; Kelenent [n] s Aationalireflt] 3 'Print [xelenent [n] ] ', (n, 4}] 49 r0 0 11 o o 11 o o 1) o 0 01 0 0 1) 0 0 11 o o 1) o o o o 0 1) o 11 0 11 0 0 11 o 0 o 0 1) 0 0 <1 0 0 <1 0 o o 0 o 11 0 0 1) 0 11 o o 11 o 0 o 11 o 0 11 o 0 11 o o 0 M11'000000000000’ o o 11 o 0 11 0 o 11 0 0 o o 0 11 o 0 r) o 0 1) 0 0 0 o 0 11 0 o 1) 0 0 11 o 0 o o o 11 o 11 o 11 0 0 11 o 0 (0 0 1) o 0 1) 0 0 11 0 0 1n mulijkr ‘13 '2i-l-ijk[l],21-ijk[2],Zj-l-ijk[3],2j-l-ijk[4],2k-1-ijk[5],2k-ijkl61,'3 no[Print['ele1aent: ', e]3 1:1):[11 e (2xfl1[[e]]) -13 111312] - (3x11111011): 111513] . (3x11311011) -1:1:lk[4] - (2x11311011): 1:1):[5] :- (2xux[[e]]) -13 ijk[6] - (2 xuxflellh Rteap-xelenent[e]3 DO [DOIMIIIIUIKli]. 1111(3)” -h11[[1:lk[1]. 1111:1311] +Kt-lvlli. 3]]. {1. 61]. {:l. 6}]. {0. All: 'Pill the element K into the integral x(xall), baaed on every corrdinated nodee'; m1 / 100000 ( 8 0 —8 -2 o o o 2 o o o 01 o 3 -3 -3 o 0 3 o o o o o ~8 —3 22 5 —8 -2 ~6 —5 0 5 0 o -2 —3 5 22 -3 -3 —5 -16 5 o o o o 0 -8 -3 11 5 o 0 -3 -2 o o 0 0 —2 —3 11 0 0 —3 -s o o o 3 -6 -5 o o 22 5 -16 -5 o 2 2 0 —5 ~16 o 0 5 22 -5 ~6 3 o o o o 5 -3 —3 —16 -5 22 5 -3 —2 o 0 5 o -2 —8 -5 -6 5 22 -3 -8 0 o o 0 o 0 o 3 -3 —3 3 0 ( o 0 o 0 o o 2 o —2 -8 o s) 50 APPENDIX IV 1 II I 150000003 u s -‘-3 t a 0.13 Isyaten propertiee: Young'a modulus (II/Inf); Poieeon'e ratio3 Plane thickneee(cn)3 '3 I! a {l, l, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 6, 8, 9, 9, 10, 10, ll, ll, l2, l2, l3, l3, 15, 15, 16, l6, 17, 17, 16, 18, l9, 19, 20, 20, 22, 22, 23, 23, 24, 24, 25, 25, 26, 26, 27, 27, 29, 29, 30, 30, 31, 31, 32, 32, 33, 33, 34, 34, 36, 36, 37, 37, 38, 36, 39, 39, 40, 40, 41, 4l}3 NJ- {0, 9, 9, 10, 10, ll, ll, l2, l2, l3, l3, 14, 15, l6, l6, l7, 17, 10, 10, l9, 19, 20, 20, 21, 22, 23, 23, 24, 24, 25, 25, 26, 26, 27, 27, 26, 29, 30, 30, 31, 31, 32, 32, 33, 33, 34, 34, 35, 36, 37, 37, 36, 36, 39, 39, 40, 40, 41, 41, 42, 43, 44, 44, 45, 45, 46, 46, 47, 47, 40, 40, 49}3 IR- {9, 2, 10, 3, ll, 4, l2, 5, l3, 6, l4, 7, 16, 9, 17, 10, 10, ll, l9, 12, 20, 13, 21, 14, 23, 16, 24, 17, 25, 16, 26, 19, 27, 20, 20, 21, 30, 33, 31' 24. 33, 25, 33, 26, 34, 27, 35, 26, 37, 30, 30, 31, 39, 32, 40, 33, 41, 34, 42, 35, 44, 37, 45, 30, 46, 39, 47, 40, 46, 41, 49, 42}3 'neeh for plane A: nodea diatribution, pic): up the i3j3k3 o! in element by: '3 '13-'1118111 :InIl-‘Illnllr anIIRIInll'i r. {0, l, 2, 3, 4, 5, 6, 0, l, 2, 3, 4, 5, 6, 0, l, 2, 3, 4, 5, 6, 0, l, 2. 3, 4. 5. 5. 0. 1. 2' 3. 4. 5. 5. 0. 10 3. 3: 4. 5: 6. 0. 1. 2. 3. 4, 5. 6}: 12(6, 6, 6, 6, 6, 6, 6, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 2, 2. 2, 2, 2, 2, 2, l, l, l, l, l, l, l, 0, 0, 0, 0, 0, 0, 0}3 '1,! value for plane A nodea, pick up the In a m: of In node by: '3 “In-1118]]: “I11!!!” '3 Arr-win. virus-earl”. “ism-313k. 72]: mulch 72lrmnr1C3. wishful“. 72]: 00131111] I (![[(N-T[fn]])]]) - (YIIOIRHDIIHI). in. '72} I: 00131111] I (YllfmffnllHl) -(![[(N1[[nll)]]). (n. 72} II not-11:11-(11101110111111)42110111121111”). in. 72} 13 hotel!!!) I (311(lelnlllill - (Xllflwllnlllll). (n. 72} l: micflnl I (31101111101111) - (811(Mllnlllll). (n. 73} l: 001C311!) I (11101111101111)«11101111141111». (n. 72} l: 'Detine Pi,bj,bk,ci,c5,c1: for all elenente, oeauute every value by: '3 '31 [n] I” [n] 415111]: 3:1 in] Infra] 41 In]: Skin] 1'11an -Y:l [nl'r “(31181-13nt 41 [n] 3 6:1 in] .n in] 41: In] 1 Chin] II! [n] 4113] " I Do[Print['bi(', 1, “)I', (BlIil), " Bj(", 1, I')-", (3311]): " “('11 1. “I": (3311]): " 01": 1o '1": (C1111): " C32”": 1. ')"o (321111): " “I“: 1: '1": (93111)]- {11 731] 1 1 xmmrtnlim rtumrnnm no[1-;xn-e[ 1 xtumrrnnm 11101710111111 ]:-mn6111-.1n.721] 1 xmmrnnm rmmtnnm 51 1 mm o Bj[n] o bkln] o m-ytu.721;no[u1n1-——l o c11n1 0 cm] 0 Cklnl . 121.721]: 3‘ c1111] 8:1[n] c1 [n] :3 [n] Ch[n] lkln] 1 u 0 ”9,, " 1: 1. o ‘“'“' o 11 3:1 16(130000 4000000 0 4000000 16000000 0 0 0 W0 maylxele-ant, 72] 3 Don! I (Tran-poeel (”[n] ) ] .w) 3 m a Kmbblnl 3 x a tom-m3 xeluant[n] a Rationallaew] 3 'Print [blunt [n] ] ', (n, 72]] 0 : 0 all I [ee 8 ..] 3 0 ‘ ° 90x90 mayfljh, 6]) '21-1-1jh[1] ,21-1jkt2] ,21-1-13313] ,23-1-131114] ,2h-1-13kl5] ,2k-1jkl6] , . 3 Do[Pr1nt['eleI1ent: ', e]3 13k[1] a (2xl1[[e]]) ~13 11k121-(3xn11-11): 13k”) ' (3XWIICN) -13 133“] I “XVIIONM ijk[5] a (2le[[e]]) -13 11k[6] - (2xllx[[O]])i Rtup-KeluaentIeh DO [mlmllliflklih 113””) I'muliilklil: ijklilll +K¢W[[1: 3]]: (1: 5}]: {’0 ‘1]: {.0 731]} 'l'ill the element 1: into the integral 8(xall), based on every oorrdinated nodea'3 'Do[Pr1nt[xall[[1]]]:{1,9011'3 'uae to llat all all e1enant'3 V 01 1 £1 1 02 £2 03 0 8 8 8 8 012 0 013 -100 014 -100 034 0 035 £35 036 0 8 8 8 8 1093 1 L 0 J 52 801WI{KI11.0 II P, 01 II 0, 02 II 0, 085 II 0}, (£1, 12, £85, 01, 02, 03, 04, 05, 06, 07, 08, 09, 010, 011, 012, 013, 014, 015, 016, 017, 018, 019, 020, 021, 022, 023, 024, 025, 026, 027, 028, 029, 030, 031, 032, 033, 034, 035, 036, 037, 038, 039, 040, 041, 042, 043, 044, 045, 046, 047, 048, 049, 050, 051, 052, 053, 054, 055, 056, 057, 058, 059, 060, 061, 062, 063, 064, 065, 066, 067, 068, 069, 070, 071, 072, 073, 074, 075, 076, 077, 078, 079, 080, 081, 082, 083, 084, 085, 086, 087, 088, 089, 090, 091, 092, 093, 094, 095, 096, 097, 098)] {{n 40.12» 100.185» 100.111 -10,U2-:0. 240759204691901504941237302080346279417244981474855212567071088‘724681749649365l U3 -1 - 347542805637642019646677141081930240935493453365697259040357719134473019431083040000 ’ U4 12401 18079660558731 1072034171491445249031 15498671995043630743888585903681914279 -1 - 65164276057057878683751963952861920175405022506068236070067072337713691 14332807000 ' U 5 205954334863 3698 80095975548959899080552964795407 3087896933 5753759 1 74282363 82627 _. - 5792380093 96070032744461 90 1 8032 1 7068225822422276 1 62098400596 1 9855745 5032385 13840000 ' U6 1481837928189712029859843475336704801854606190320752558880294466925488853591 1839 -: - 52131420845646302947001571 162289536140324018004854588856053657870170952914662456m0 ' U7 22137170005762172957942048025131657966255020675267314683432452737745739697826513 4 - 347542805637642019646677141081930240935493453365697259040357719134473019431083040000 ' U8 1022794547568881 16990243 8630364855 13441 1 101948 1 10483020096642720950587875331467 4 - 2896190046980350163722309509016085341 1291 121 1 1380810492002980992787275161925692000 ' U9 8159494195900848084706599775574783566907639926889935255106568031324591080886601 ... — 8688570140941050491 1669285270482560233873363341424314760089429783618254857770760000 ' U 10 36269361 3342089 1 9 l 625946770325936973939284296435 8328380235 19940293 85555823 38309 4 — 8688570140941050491 16692852704825602338733633414243 1476008942978361825485777076000 ' U 1 1 31 1261400596633478201 160767300198386486351955936267201508021301 1272729412897601 4 - 23169520375842801309778476072128682729032896891046483936023847942298201295405536000 ' U 1 2 2123383606361286489971817918389014384523235983528035869708374677817852767676479 4 - 434428507047052524558346426352412801 1693668167071215738004471489180912742888538000 ' U 1 3 33692102364413670352972673709569061317902354000881057659814443992150994713623803 4 - 173771402818821009823338570540965120467746726682848629520178859567236509715541520000 ' U 14 1777472296077346438132223937951948762833 -: - 29826512098930623285334063096942656507 12000 ’ U 1 5 102793473455059203877821760107041 1 31 806336805663641 1694094014054134940067620301 1 4 _ 104262841691292605894003142324579072280648036009709177712107315740341905829324912000 ' U 1 6 337439483740490893 1406024732322075 8465272862625047 1 1 6106898080268663218604372589 —: - 34754280563764201964667714108193024093549345336569725904035771913447301943 1083040000 ' U 17 1 179476595 267419637458454060288820562 147578883 85323 1 876674 1 54088755882653046 10003 —: - 104262841691292605894003142324579072280648036009709 1777121073157403419058293249120000 ' U 18 _. _ 213339190069810914120041910443423915327120387149070243633644408859972757170772107 1042628416912926058940031423245790722806480360097091777121073157403419058293249120000 ' U 1 9 15053847884544485964477221380063016033688175936937182069317454636531416384727727 _, - 1303285521 141 15757367503927905723840350810045012136472140134144675427382286656140000 ' U20 302539156432708362842088963187000869264877392172223109936881021805737136971417647 -: - 1042628416912926058940031423245790722806480360097091777121073157403419058293249120000 ' U2 1 130177900967636184385373402405152572901289037448588859903504928231878516958476929 4 - 104262841691292605894003 1423245790722806480360097091777121073157403419058293249120000 ' U22 413023 65069422922908959207 825707 82049059601 6876953 136039477074853308481029922423 -’ - 1158476018792140065488923803606434136451644844552324196801 1923971 1491006477027680000 ' U23 72852482455170235647385151398271334582444842343627923547161990691485780222679101 4 - 52131420845646302947001571 16228953614032401800485458885605365787017095291466245600“) ' U24 -: __ 1 1 164325855744279166436188539559201070554814753817030072565585032757399046996393 267340619721263092035905493139946339181 14881028130558387719824548805616879314080000 ’ 53 653490666676378545532963508454898687531 1314279710360477248414730575955298715143 41705136676517042357601256929831628912259214403883671084842926296136762331729964800 ’ 55481789961684446051 141 10790645733223831855954297049707]613975586577363495845883. 1 158476018792140065488923803606434136451644844552324196801 1923971 1491006477027680000 ’ 333247248023 30733327854202993 1 57205 1 36506598 1325 855053 8757304405385767539397 179 - 2005054647909473 190269291 198549597543858616077109791879078986841 1604212659485560000 ' U28 4 4514317503571215108838615008658359688408232704731517068658567 8416103878199922520901048415374801 133493646702405462737530720000 ’ 249686264884074784446725960135786590985831020991309564459748495755468063048327 1629106901426446967093799098821548004385125562651705901751676808442842278583201750 ' 29331382326278761244717801491485661579522416819081876822557832756103433442920039 1737714028188210098233385705409651204677467266828486295201788595672365097]5541520000 ' 1712420497606071306341314139004593295]5151703486905644005583443919699866762192333 1042628416912926058940031423245790722806480360097091777121073157403419058293249120000 ' 119360537160532845405338788678346836247742454898149789880532456464282008592271787 521314208456463029470015711622895361403240180048545888560536578701709529146624560000 ' 6606999509400089749620326286934535884389157889726264404579765801142636725061671 40101092958189463805385823970991950877172321542195837581579736823208425318971120000 ' 1 1721869101 108650515656352867069173442941043198800918069336184992501023624722519 - 40101092958189463805385823970991950877172321542195837581579736823208425318971 120000 ' U35 ., 17225876320008073282208547332348756548787368 1384802749943038955223056060979337759 104262841691 292605 894003 1423245790722806480360097091777121073157403419058293249120000 ' 612558 126473545622379944796547065 1 875 1284055 8030394 1 7542784485 56536721 8972801889 1737714028 18821009823338570540965 l204677467266828486295201788595672365097155415200“) ' 338729778637942638895637920852370964903421 3030675275 32032543 33 3440908477459767 l 2005054647909473190269291 198549597 54385861 6077 109791 879078986841 1604212659485560000 ' 78954039598365824976000092805774277820972948035 1379540906138290882915665241733 1 19307933646535667758148730060107235607527414075872069946686539951915167746171280m0 ' 3610481263584212696491 18301733141 15967865492388956046941073884832955188359327749 _ 20852568338258521 178800628464915814456129607201941835542421463148068381 1658649824000 ' 4207371 1821613044515068331872439857171 10173296494463209871320873923016814060071 91458633062537373591230826600507958140919329833078226063252031351 1771 1037660080000 ' 89986670214447333434445684987128919247461387734983 15 1063053774042967763 1 59563569 5213 1420845646302947001571 1622895361403240180048545888560536578701709529146624560000 ' 863037 17 14300677766509679803 66342524862941 6307 12289523 32937565728607 499364890589 - 1737714028188210098233385705409651204677467266828486295201788595672365097] 5541520000 ’ 622759379741434857490385589235 14762237208873 854653749344861 1 84980024167233 19903 3475428056376420 1 9646677 14 l 08 1 93024093 54934533 656972590403 577 19 1 344730 1 943 108304000 ’ 25924735093365357472005661275391681280608760492804995720616183795161787779498789 1 158476018792140065488923803606434136451644844552324196801 1923971 1491006477027680000 ’ 65084661062741094076799989718636365807312853834884255632851204851 1283607 34 1 5 1401 347542805637642019646677141081930240935493453365697259040357719134473019431083040000 ' 261 849897663 3069825702763 1 8948405757435 128503076143612810929237915994346033466861 104262841691292605894003 1423245790722806480360097091777121073 157403419058293249120000 ' 16356096827774845307641326560853 1936413631 15602755890710806956817582979944427523 8688570140941050491 16692852704825602338733633414243 1476008942978361 82548577707 60000 ' 307243747584624899308598237530244139919922723439417144044467342038171786071629021 104262841691292605894003 1423245790722806480360097091777121073157403419058293249120000 ' 65 1 80635 57427804737467 56855656293041 387236201 860765 54274934705 704064262923 842723 347542805637642019646677 14 1 08 1 93024093 549345 33 656972590403 577 1 9 1 344730 1 943 1083040000 ' 1202427728767036838303 13069620685351949470662277945898242934388777976820993328987 - 34754280563764201964667714108193024093549345336569725904035771913447301943 1083040000 ' 10864960485416698843397382362362667294003058836891391030235754248629037634775559 57923800939607003274446190180321706822582242227616209840059619855745503238513840000 ' U25»- U26-: - U27 -: U29-:- U30-:- U31»- U32-:— 1133-:- U344 U364- U374- U38»- U39» U404- U414- U424 U434- U444— U454— U464- U474- U484— U494- USO-1 U514— 54 10622986420959809679966565871589323923 1 1 1425272507 6323093403627757973 1 8291 1 17619 267340619721263092035905493139946339181 14881028130558387719824548805616879314080000 1041978925199901 181530481831079023294134015295146426865941 14166448548835448621 -556068489020227231434683425731088385496789525385115614464572350615156831089732864 ' 153620712632911483813983634848578240108898976430806547534238673121718086856181607 347542805637642019646677141081930240935493453365697259040357719134473019431083040000 ’ 502375029820938063123880210372885914633774698519709455014933401037934949562787 2715178169044078278489665164702580007308542604419509836252794680738070464305336250 ' 54569418839793723608839296273310983778475856721124419056235282228659359181907949 115847601879214006548892380360643413645164484455232419680119239711491006477027680000 ’ 32802769230862019067622768352095441995373570867715003541871424440731405519091 18101 1877936271885232644344313505333820569506961300655750186312049204697620355750 ' 2333442160220862713766718861440764623475665091381 131 1735704306697045571346251719 8688570140941050491 1 669285270482560233 8733 6334 14243 147600894297836 1 82548577707 60000 ’ 187166009385 152522628716269048623218506174688001 165686073862615524702587940621 990150443412085526058909233851653108078328926967798458804437946252059884419040000 ' 69736674015738563518992457174669853821091945124770471978262095166173834807040137 260657 10422823 15 1473 50078558 1 1447680701620090024272944280268289350854764573312280000 ' 3367650991638417 13801826345587444861 89368057605028803901 880398283207229145207081 - 1737714028188210098233385705409651204677467266828486295201788595672365097155415200“) ’ U62 _’ 370465798531755351641488100359750799510741127350219553695837092107646252325599 1253 15915494342074391830699909349846491 1635048193619924424366775725263291217847500 ' 17588469663004193690498]7823495618690389658017909903685245274916412983046348487 891 1353990708769734530183104664877972704960342710186129239941516268538959771360000 98421685 1677657883 8301 008230960056632444743284025625 88 1 5935483253 17620458776343 - 2896190046980350163722309509016085341 1291 121 1 13808104920029809927872751619256920000 U65 4 17365748488 109925029 1 53220874059270795 6000895007 52498 1 43 336820685 14770233 1 8 1 2201 8688570140941050491 1669285270482560233873363341424314760089429783618254857770760000 33662301624615491 1 1232741 1998360574033372080324044804413403558448832488241449209 8688570140941050491 1669285270482560233873363341424314760089429783618254857770760000 46467083 57288605 12679485 5208 1 8497484261077895 126288666504267959792606943 1628793 - 23 169520375842801 30977 847607 2 1 2868272903289689 104648393602384794229820 1 2954055 36000 U68 4 _ 37258995748189859724797941642052847058073323863856330971957776054848256671469209 8688570140941050491 1669285270482560233873363341424314760089429783618254857770760000 346231293415013597 1700892355730879589737167954561 1760630370476975129315867640763 17377140281882 10098233 38570540965 1 20467746726682848629520 1 788595672365097 1 5541 520000 ' 198133 1 14676820746196130448266398846913345 19981927871095177613899728324617962977 - 434428507047052524558346426352412801 16936681670712157380044714891809127428885380000 ' 56949190075991910940873746433683149152717845397290828762062521 1427807006623957 38615867293071335516297460120214471215054828151744139893373079903830335492342560“) ' 2 1 505592277973 83095203 1 83004428322094705 63 6527 1440607304402493 8691 86865692082941 - 69508561 12752840392933542821638604818709869067313945180807 1543826894603886216608000 ’ U73 4 193191 8282701028857010286853744044073575227891058526782793 1016430176483074792267 1158476018792140065488923803606434136451644844552324196801 192397114910064770276800“) ' 5789771835762771859276923813427477808060482516995 1266972605382449380409049959107 20852568338258521 1788006284649158144561296072019418355424214631480683811658649824000 ' 426691 171992021 15629531441 5236899703463399846293895939873149196353976450089421 - 22864658265634343397 8077066501 2698953522983245 82695565 1581300783779427759415020000 ' 6163391258779987477052804421782058246707615476195761989013167848323245639499271 1 - 20852568338258521 1788006284649158144561296072019418355424214631480683811658649824000 ' U77 4 703936851541 10708973 3485 8895948605034 1 9625323658898734752 1 882063 14 1 6619256045083 34754280563764201964667714108 l93024093549345336569725904035771913447301943108304mm ’ 23363563748425462001 1662829853 84432574609487 107 1 12025803332032179276923505785437 69508561 12752840392933542821638604818709869067313945180807 1543 826894603886216608000 ' U52-1- . U534 U544— USS-1— U564- U574- USS-:- U594- UGO-+- U614 U634- U64-+ . U66-1— . U674 . U694- U704 U71»- U72 -: U74 -: - U75 -: U764 U784- 55 12317807864873283420847671569598044351538415567453216770606309250696341927002029 -57923800939607003274446190180321706822582242227616209840059619855745503238513840000 265379672263363802569329926l827935243314666893251l164339174346119851319850152669 ‘-695085611275284039293354282]6386048187098690673139451808071543826894603886216608000 120746789515786911786578163981978144763398211020798512940026909535096184188325 U81"mu556068489020227231434683425731088385496789525385115614464572350615156831089732864 ' 9771811120624939598269543137505526241339]30585276754933023531585728313997936339 23169520375842801309778476072128682729032896891046483936023847942298201295405536000 6301005586964814724576850947136148515072766951307856649171101893054787039542873 -28961900469803501637223095090160853411291121113808104920029809927872751619256920000 ' 762904414743221335785898096792034908599766709485547776983386950482238862103661 ,1J85-4 - 1695330759208009851935010444302098736270699772515596385562720581 143770826493088000 0 U86 63259036520945 8839006295 1057915 18736886482228566]84492936607625665748085428097 17 ' 4 - 1737714028188210098233385705409651204677467266828486295201788595672365097155415200“) ' 414130314733591724756526495835433916197481647 - 36967473 1 8729 1 6955099801 7741 225752060941 85 8080000 ' 1 153249694467343 33 7405 86260427769 1 354594357 1 6796973715029756571721m109201669597 - 40101092958189463805385823970991950877172321542195837581579736823208425318971120000 ’ U89 —3 31 178220619316669531736127527703350369800299383634551681757299012001420762115561 17377140281882100982333857054096512046774672668284862952017885956723650971554152m00 ’ 155904276176392353915865267715108669026569681354812131 179811791541m2846036440981 52131420845646302947001571 1622895361403240180048545888560536578701709529 1466245610“) ' 76035913558142506533109795654271731673245897329755856739726517621868783383761973 - 347542805637642019646677141081930240935493453365697259040357719134473019431083040000 6510960936844358465641208684332022443587042597229899955906195366695392712818783 — 19307933646535667758 1487300601 07 235607 52741407 587206994668653995 19 1 5 16774617 1280030 ' 689365 87536874496286661 661 1 1 87747685484627793673963258 13782884578291 1 1739680707 - 28961900469803 50 16372230950901 6085341 1291 121 l 13808104920029809927872751619256920m0 ' U94 4 66237696282206757 1 28954283 6562002630320097959905977324 1 58627274847652763 53672537 173771402818821009823338570540965120467746726682848629520178859567236509715541520000 ' 413857284088650504348038792669732326250329859144444776998109692535591 123855871 169533075920800985 1935010444302098736270699772515596385562720581 143770826493088000 ' 73015642602569888258996879153544375079176858001 122228649545896040126032829329627 1737714028188210098233385705409651204677467266828486295201788595672365097]5541520000 471679342659904223176590588335 1650821203483553971295207642733604473733627338467 1930793364653566775814873006010723560752741407587206994668653995191516774617128011” ' 260252991266536944617225 10173328405 199981 188065579673797672572561775087656937209 }} 57923800939607003274446190180321706822582242227616209840059619855745503238513840000 U79 —0 U804 U824- U83 -1 U84—1 U87 -: U88 -3 U90»- U91 -: U92 -: U93 -: U95 -:- U964- U974- U98-9- 0115: I {01, 02, 03, 04, 05, 06, 07, 06, 09, 010, 011, 012, 013, 014, 015, 016, 017, 010, 019, 020, 021, 022, 023, 024, 025, 026, 027, 026, 029, 030, 031, 032, 033, 034, 035, 036, 037, 036, 039, 040, 041, 042, 043, 044, 045, 046, 047, 040, 049, 050, 051, 052, 053, 054, 055, 056, 057, 050, 059, 060, 061, 062, 063, 064, 065, 066, 067, 066, 069, 070, 071, 072, 073, 074, 075, 076, 077, 076, 079, 060, 061, 062, 063, 034, 035, 036, 067, 066, 069, 090, 091, 092, 093, 094, 095, 096, 097, 096} I. 9: 'Oive 0(n) value to liat of 011et. 011at[[l,n]]-0n'3 'double check 12 011st 1a correct.'3 011et[[1, 3]] m . ram-[\fmn-qu, (2x (7:: (3-1) +1))]])'+ (an-1:111, (2x (7:: (1-1) +1) -1)]])’ . 15. 71. 11. 71]: "WIN-:51] 1a the dieplacenent at poaitlon {1:51'1 56 am 0 0000397902 000055708 0.000663627 0.000776718 0000833543 000087507 0.00089985 0000890793 0000878236 0000816166 0000759257 0000740533 0.000549315 blatantourrlot [I'D] 7 0000636563 0000770334 0000872305 0000915 15 1 0.000893 197 0.00081 1204 0000709462 0000793737 0000857346 0.00091493 1 0000942044 0000931388 0000892521 0000864815 1 2 3 4 - ContourGraphics - 57 0.0009 1301 3 0000936392 0000965958 0000984314 0000985405 000097927 0000992764 000101909 000101069 000102364 000103857 000104844 000106492 000110837 0.001 12839 000107818 0.0010854 0.001 10172 0.001 1 1826 0.001 14791 000123026 APPENDIX V -3 -8 -5 -6 -3 -2 -16 -5 -16 5 22 -5 -6 0 -3 -2 22 -s 22 -3 -a -2 -16 -8 -3 -3 22 -2 -8 5 -3 -a -16 -5 -6 1o -16 -3 -3 -5 44 -16 -s -16 0 0 -5 -16 44 10 - 16 -16 0 -8 -2 -2 22 11 -3 -3 5 11 -8 -3 3 3 - . -s -15 5223 2.. -16 4 1000003 01 -5 -16 -3 22 5 -8 -2 -6 -3 22 -3 -3 -5 -3 -3 -16 -3 11 5 11 -16 -3 -3 -16 -2 -8 -3 -8 -3 -2 31 58 3"- £2 £3 £4 £50 £60 £70 £60 £90 £100 £11 £12 £13 £140 £150 £160 £170 £160 £19 £20 91 92 930 940 950 960 96 990 9100 911 1 013 1 30III Wm) 59 30all- 1V101 or m o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o 01 o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o 1 o o o o o o o o 0 o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o 0 o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o 1-2 0 o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o ...—2 o o o o o o o o o o o o o o 1.. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o . o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o o o o o 1-2 0 o o o o o 1 o o o o o o o o o o o o o o o o o o o o o o o 1-2 0 o o o o o 1 o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o 1. o o o o o o o o o o o o o o o o o o o 1. o o o o o o o o o o o 1'01I 60 3 1000000000000000000000001 00000000000000000000000 10000000000000000000000 01000000000000000000000 00100000000000000000000 00010000000000000000000 00001000000000000000000 00000100000000000000000 00000010000000000000 000 00000001000000000000000 00000000100000000000000 00000000010000000000000 00000000000000000000000 00000000001000000000000 00000000000100000000000 00000000000010000000000 00000000000001000000000 00000000000000100000000 00000000000000010000000 00000000000000001000000 00000000000000000100000 00000000000000000010000 00000000000000000001000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1] ‘1'02 I 1'0 :1 201.1023 feat Tranntomtionnatrixby : T0.0all ' 61 I I I n 8. ‘1 J 00 000000000000000000000000mm000000000000 .— I m u 1 0124567690 0 12390 1“ mmummmmmmmmmmmummnwuwwnuununnnnuunmnun 62 -3 -8 -5 -6 -3 -2 -16 -5 22 0 0 -3 -2 5 22 -16 -3 -11 -16 22 -16 -5 0 5 22 -3 -3 5 -16 -5 -16 5 10 -16 -5 44 -6 -5 44 -6 10 -5 -16 -16 -5 -16 2 -3 -3 5 22 s 11 -3 -3 -3 -3 -3 -8 -2 s -11 -3 -16 -3 -3 -3 -5 0 -16 0 -8 -3 22 22 -3 -3 -5 3 -3 -16 -3 3 0 63 MIKIII-b 0 -1 -1 100000000000000000000000 0 -1 0 0000000000 -1 -1 001-20 0001-2 0 01.30 0 A a 0 0 .2. a 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 n.6[xnr] 218493540493073842176000000000000000000000000000000000000000000000000000000000. 00000000000000000000000000000000000000000000000000000000000 801"[{m.0' II 2020], {03, 04, 05, 06, 07, 08, 09, 010, 011, 012, 014, 015, 016, 017, 018, 019, 020, V3, V4, V5, V6, V9, V10, £1, £2, £3, £4, £11, £12, £13, £19, £20, g1, g2, g7, 98, 911, 912]] {(034 _ -2041150230705 049 _ 25769698430570 05» _ 8662215172565 13317750352961700' 31723750352961700' 56517750352961700' 06» _ 79359420331675 -12067138811082 08» _ 91853071812403 110981252941347500' 66588751764808500' 110981252941347500' 09 ~105415086089231 010 269461324093314 011 " 543943758824042500' " 332943758824042500' 7 -2479244oo314217 012» 866230479658018 014a 666645560737 867943758824042500' 975943758824042500' 55490626470673750' -19912355731059 397587017928351 U157 110981252941347500' U167 563943758824042500’ U177 -26416321906499 018* 564259691405924 0199 -282201884005113 110981252941347500' 656943758824042500' 987943758824042500' 020» 500819713669427 v3» _ -678632499652849 v4» 432943758824042500' 10654200282369360000' _ 1641723224448110 v5» _ 66686394550089 V64 2697536739171 2130840056473872000' 343943758824042500’ 2663550070592340’ v9» _ -10056869673993300 v10» 83087993445792600 £14 54354200282369360000' 87867200282369360000' 84346766723995 8621943225249440 f3+ _183863918674800 £46 88785002353078’ 78646502353078 ' 44392501176539 ' _93998899125853 £11» 694453503415314 £12» _703562101587 fl” 4345401176539 ' 54654654176539 ' 142634576539’ _.469089364343894000 £194» 7609309641345 f20q_ 1721670875290200 4756754745176539 ' 456456456539 ' 88957644176539 ' 31790674735717 9733417591862140 55046800022816 91" " I 92", ' o 97" : 33463465653078 88785002353078 6534524876539 98+ 87351542020393 911% _9463422511822380 9124 _193o7074104135}} 6576579176539 ’ 567678345353078 ' 997584296102 651 NM {{034 -0.00015326539217269, 044 -0.00015326539217269. 054 -0.00015326539217269, U64 -0.00071475324324991, 074 -0.0001812188769314, 084 -0.00024856457649176903,094 -0.0001937977674698, 0104 -0.000285645697424133, 0114 —0.000285645697424133, 0124 -0.000285645697424133, 0144 -0.00036403594316897, 0154 -0.00017942089500091, 0164 -0.00029906009398439, 0174 -0.000238025082672836. 0184 -0.00038117719721275, 0194 -0.000285645697424133. 0204 —0.000285645697424133, V34 -0.00006369624013694,V44 -0.00035315173762, V54 -0.0001938874971248, V64 -0.0005959370274947. V94 -0.00018502470134319, V104 -0.00047104487235471, £14 0.95001142635067, £24 109.62907398655632, £34 -4.141778764472281, £44 -21.631811496107005,£114 12.70620981649933, £124 -4.932619555920048, £134 98.61541943476531, £194 16.670395461247498, £204 19.353827219991384. 914 —0.95001142635067, 924 -109.62907398655632, 974 8.423994255565799, 984 13.2822155220159, 9114 -16.670395461247498, 9124 -19.353827219991384}} m.o;uz-03013.0;v1.o,v2.0,v7.(034.1111)”; V! a (04 +012) [2; V11 1:019) V12 .0202 “Boundary Conditions and 131.91.00.11“: Con-train Ignition"; mm: a {01, 02, 93, 04, 05, 06, 07, no. 09, 010, 011, 012, 013, 014, ms, 016, m7, 018, 019, 020}; mm: a {V1, V2, V3, V4, V5, V6, v7, V8, V9, V10. V11, V12}; '01” diapIooon-m: volu- to 11.1: o! 011.: and V1131?) 'Cuwuto tho gourd. 41.91.43.411: 81: nodal position)” an . r-b1o[(\/(m1u[[1, ((axn)-1)]])’+(m1n[[1, (2xn)]])3), (n, 1., 101]; um {O . 000000000000000000, 0 . 000826648034066665. 0.000731311222584203, 0.000847252275786886. 0 . 000832209342095123, 0 . 000932414037579913, 0 . 000684510360745850. 0 . 000727484384202640, O . 000891287339200160, 0 . 001191523504541550} vn . rablo[(‘\/(V1:llt[[1, ((2xn)-1)]])’+(v11.:[[1, (2xn)]])'), (n, 1., 6}] 1 arm {0.000000000000000000 . 0.000773086716148328 . 0.001031 152270031470. 001087953 IOWOOOOOO, 0010963540372081521 ,0.001 191523504541550} 66 APPENDIX VI The Result of FEA computation and Displacement Difference Analysis Reference Model Result: (From re-meshed 7x7 modal. 72 Elements) 0 0.0003 0.00059 0.00079 0.00091 0.00102 0.001128 0.00042 0.00046 0.00077 0.00086 0.00094 0.00101 0.001078 0.00078 0.00083 0.00087 0.00091 0.00097 0.00102 0.001085 0.00088 0.0009 0.00092 0.00094 0.00098 0.00104 0.001102 0.00089 0.00088 0.00089 0.00093 0.00099 0.00105 0.001118 0.00082 0.00076 0.00081 0.00089 0.00098 0.00106 0.00074 0.00065 0.00071 0.00086 0.00099 0.00111 Modular Model Result (interpolation value in yellow area). Boundary Displacement: , 0 0.00026 0.00052 0.00077 0.00086 0.00095 0.001031 0.00037 0.00041 0.00067 0.00081 0.00088 0.00094 0.001009 0.00073 0.00078 0.00083 0.00084 0.00089 0.00094 0.000986 0.00079 0.00081 0.00083 0.00088 0.00091 0.00094 0.000964 0.00085 0.00084 0.00083 0.00088 0.00093 0.00099 0.00104 0.00077 0.00077 0.00078 0.00085 0.00091 0.00106 0.001116 0.00068 0.00071 0.00073 0.00_0_81 0.00089 0.00104 0.001192 570.1% 45.59% —0— Joint Modular Model -- O— -Re-meshed Reference Model (7X7) - - i - - Reference 3: I 10.00%-15.00% Model (3X3) , 3‘ I 5.00%-10.00% 0.00°/o-5.00% 67 The Result of FEA computation and Displacement Difference Analysis (2) Reference Model Result: (From 3x3 (8 elements) model). 0 0.00024 0.00049 0.00073 0.00084 0.00095 0.00106 0.00026 0.00027 0.00059 0.00076 0.00085 0.00093 0.00102 0.00051 0.0006 0.00055 0.00079 0.00086 0.00092 0.00098 0.00077 0.00078 0.0008 0.00082 0.00086 0.0009 0.00094 0.00073 0.00076 0.00079 0.00081 0.00091 0.00093 0.00099 ‘ 0.0007 0.00073 0.00077 0.0008 0.00088 0.00099 0.00103 0.00055, 0.00071 0.00075 0.00079. 0.00089 0.00098 0.00108 Modular Model Result (interpolation value in yellow area). 0 0.00026 0.00052 000077000086 0.00095 0.00103 0.00037 0.00041 0.00067 0.00081 0.00088 0.00094 0.00101 0.00073 0.00078 0.00083 0.00084 0.00089 0.00094 0.00099 0.00079 0.00081 0.00083 0.00088 0.00091 0.00094 0.00096 0.00085 0.00084 0.00083 0.00088 0.00093 0.00099 0.00104 0.00077 0.00077 0.00078 0.00085 0.00091 0.00106 0.00112 0.00068 0.00071 0.00073 0.00081 0.00089 0.00104 0.00119 Displacement difference: (result-reference)/MAX(reference) (%) 0.00% 1.16% 2.31% 13.47% 1.52% -0.43% +2.38% 9.66% 12.13% 6.73% 3.95% 2.31% 0.68%» 095%» 19.32% 15.24% 24.25% 4.42% 3.11% 1.79% 0.48% 2.11% 2.17% 2.22% 4.90% 3.90% 12:90%311.91% 10.22% 7.13% 4.03% 5.97% 2.19% 5.00% 4.66% 6.16% 3.59% 1.03% 3.62% 2.80% 6.18% 7.42% 2.09% 0.06% 4.97% 1.27% 0.27% 5.22% 10.17% Nodal Displeament Difference I -15.00%--10.00% 68 "Illlllllilillllllilli