A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No.

A/2008, University of Debrecen, Debrecen, 2008.

An Essay to Introduction of Differential Geometry and Applications Abstract This paper is the result and a brief summery of the studying the available papers-which contain a series of Differential Geometry, Computational Keywords such as Computer Geometry,

Graphics, CAD, Curve and Surface Modelling- with respect to the main research proposal which looks forward the new, accurate and speedy ways of Surface modelling. Presented paper tries to answer following question; where and what is the current role and application of differential geometry in Surface Modelling in its modern style. Differential geometry as the fundamental tool of defining curve and surfaces seams to be the best starting subject of the three years research around

Automotive surface Modelling and Design.

For approach this

reason many definitions in Mathematics, CAD and also Computer Available paper also shows the knowledge of the researcher around principle definitions of defined subject and also proofs his efforts of studying detailed course sources. Introduction One of the advantages of either surface modelling toolbox of CAD packages or professional surface modelling packages is that using its high level functionality doesnt require in-depth understanding of the mathematics. Although higher level of understanding helps to have more from system. This is exactly why some initial definition of mathematics are reviewed.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

Differential geometry Differential geometry is a branch of mathematics that uses the methods of differential and integral calculation to study problems in geometry. The theory of space curves and surfaces in three dimensional space. Differential geometry has grown into many fields and branches concerned with complex geometric structures which have wide applications in different area of science and mathematics. Below are some examples of how differential geometry is applied to other fields. In physics with Einstein's general theory of relativity, in study of electromagnetism and mechanics, in economics with monitoring diagrams, in engineering solves problems in image processing, in structural geology for analyze the geologic structures and finally in geometric modelling and computer graphics including computer aided geometric design which powered by ideas from differential geometry. This paper up to projects need briefly talks about differential geometry of smooth curves and surfaces which are classified in class A surfaces. Differential Geometry of surfaces and curves is a huge subject with so much equations which covers all types and shapes of curves and surfaces. Automotive Surfaces are kinds of class A surfaces which are a set of Free Form and high quality surfaces. Class A surfaces are those aesthetic Freeform surfaces that are visible for human eyes which have an optimal aesthetic smoothness shape. Below comes some foundation definitions which are necessary in this regard.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

Differential calculations Differential Calculation, a field in mathematics which studies the function changes when their inputs change. The first object in study of differential calculation as later in application of NonUniform Rational B-spline (NURBS) for creation surfaces explains is derivative. Solving the Derivative function of NURBS algorithms or other surface definition algorithms is totally vital for creating surfaces. Derivatives are usually used to find the maximum and minimum of a function. Equations for solving derivatives are named differential equations .The derivative of a function at a chosen input shows the best linear approximation and behaviour of the function near that input. The reverse process of differentiation is integration which has also much application in surface and curve modelling of CAD and other computer graphic packages. Applications of derivatives Taking derivative and solving for critical points is a way to find minimum or maximum values of function which are useful in optimization applications. Finding the shortest curve between two points which lies on a surface and also finding smallest aria surface filling in a closed curve in 3D SPACE are other applications of derivatives. Graph sketching is other application of derivatives which nowadays widely are used in data information monitoring. Many physical processes are also described by derivative

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

equations. For example Velocity is the derivative of displacement and Acceleration is the derivative of an object velocity. Studying the classic differential geometry

Differential Geometry Curves-Surfaces-Manifolds, Wolfgang Khnel, 2002 is the name of the main reference book which has
reviewed for this course. The intent is to present basic notations which gives ability to define curve and surfaces and goes on to divide a surface in to different simple defined parts. According this book curves and the local theory of surfaces are based on Euclidean space as the ambient space. Topological structure in the form of limits, open sets, differentiation and integration is used. Curve In practical word curves are created in many different ways for example as a profile curve of a technical object. With this definition there are too many regular curves available in space, many of them have not any special names and application instead others have special applications with defined names. Frenet Curve Frenet Curve is a regular, length parameterized and continuously differentiable curve which at every point of its vectors is linearly independent. There are two different types of Frenet curves; Plane Curves which are identify in 2D space and Space Curves which are identify in 3D space. The Frenet curve determined by unique conditions and with special vectors and values 2D and 3D curves

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

are created. Frenet curves have equations in space which have relation with local theory of surfaces. The global theory of curves studies the property of closed curves, in particular their curvature properties. A closed curve Is called convex. Surfaces The local theory of surfaces study the parameterized surfaces in 3D space. Surfaces are developed theory of 2D curves which are given by a differentiable map. Local theory admits a geometric linearization in the sense that at every point there is a linear surface. In the other words each surface in 3D is a combination of infinite linear surfaces. Curvature of surface Curvature of surface just like the curves is related to the changes into the tangent planes. Each plane is determined by just one direction which is normal vector, therefore the studying normal vectors explains the curvature of surface. Revolution and Rule Revolution and rule have vide application in surface modelling which occur often and allow simple computations of geometric quantities. Surface of revolution formed from circles centred at one of the axes. Surface of ruled formed from lines along some fixed curves iv variable direction. Few useful mathematical definitions Numerical analysis

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

Numerical analysis is the study of algorithms for continues mathematics. mathematics engineering Modern doesnt have numerical exact analysis answers, for in so continues obtaining and have

approximate solution is ideal. These analysis in all fields of wide applications optimization differential equations. Numerical analysis find the techniques which gives approximate but accurate solutions for problems in engineering such as Finite Element Analysis (FEA) and Computational Floyd Dynamics (CFD) which together have effectively application in automotive body design. Spline Spline in the mathematical field of numerical analysis is a defined function by polynomials. Also in CAD and other computer graphics packages term of Spline normally refers to a parametric polynomial curve. Spline because of its accuracy, simple construction, ease of use and ability to approximate complex surfaces through curve fitting have wide acceptance in surface modelling. In the other words Spline is a wide class of data interpolation/smoothing functions in 2D or 3D. B-spline In numerical analysis a B-spline is a Spline function that has minimal support with respect to a given degree, smoothness, and domain partition. Therefore each Spline function of a given degree , smoothness and domain partition is a linear combination of B-spline with the same degree, smoothness and same partition.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

In CAD and other computer graphic packages the term of Bspline refers to a spline curve parameterized by Spline functions that are expressed as a linear combinations of B-spline. B-spline is a generalization of a Bezier curve. Bezier curve In the field of numerical analysis, a Bezier curve is a parametric curve which first time has bee used for automotive body modelling. In computer graphics Bezier curves are an important tool used to model smooth curves. Bezier curve has many control points which could be graphically used to manipulate the curve. In order to create more complex curves, simple Bezier curves are patched together. For having maximum smoothness the control point at which two curves meet and one control point on either side must be collinear. Bezier surfaces Bezier surfaces are kind of mathematical spline used in finite modelling, computer graphics and CAD. Such as Bezier curve, Bezier surfaces also are defined by a set of control points. Bezier surfaces for the first time for automotive body modelling was used and still have wide application. Bezier surfaces have different degree of complexity which each degree could be selected for special surface complexity. meanwhile there are other type of Bezier surfaces-bicubic- which generally have enough degrees of freeform for most application. NURBS Non-Uniform rational B-spline(NURBS) Is a mathematical model

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

used in CAD and computer Graphics for creating Curves and Surfaces such as aerospace exterior surfaces and automotive bodies which could be exactly or partly reproduced in situations which development is needed. NURBS are generalization of Bezier splines which could be named as uniform non-rational B-splines. In the beginning NURBS were only used in special CAD packages of car companies, later they became popular on personal computers CAD packages. NURBS have wide acceptance in CAD, FEA and CFD and are part of industry wide used standards which are IGES, STEP and ACIS. NURBS are also found in various 3D modelling and animation that allows representation of geometrical shapes in a compact form. NURBS surfaces are functions of two parameters mapping to a surface in 3D space under effects of control points. A NURBS curve is defined by its order, control points, and a knot vector. NURBS are generalization of B-spline and Bezier curves and surfaces. Nowadays normally computer graphics package applications offer NURBS technology. NURBS- control points Control points determine the shape of the curve. Basically each point of a curve or surface is computed by an algorithm of control points and governing parameter. Introduction to CAD In early CAD systems computer packages used numerical methods to model surfaces. In the 1960s CAD systems used 2D geometry to depict the height and width of the calculated objects and CAD became a tool for drafting reason.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

First CAD system was available on personal computer by 1980s. Since then numerical models functions are the that are the of basis CAD for mathematical foundation tasks.

Mathematical models define the small structural parts of 3D models. In CAD packages many elements such as point, curve, line are predefined which geometrically are manipulated in CAD systems. Many mathematical models exist to handle basic and developed operations such as root finding, interpolation which allows to find objects location in a coordinate system from minimum points value. root finding techniques also can be useful for computing changes in signs and polynomial equations. Numerical Differentiation approximate the derivatives of an unknown function which have wide application in freeform surface modelling. Integration uses known and unknown values to solve equation that result in precise calculations for analytical approximations.

Objects in CAD Point represents a position in the 3D workspace so is defined by Cartesian coordinates. Point in usual way just is used as reference to snap other objects. But in my PhD research, points play basic role as digitized information which come from a reference design. In this method Cartesian Coordinates of points simply are stored as a txt format file. Wireframe objects are lines, arcs, circle and curve which have special Equations and with a single coordinated point defines.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

For creating simple wireframe objects in CAD, parametric equations in place of Implicit functions are applied. Parametric equation is a combination of trigonometric sinus and casinos functions. Meanwhile creation process of complex wireframe objects like Bezier spline, B-spline and NURBS couldnt be easy as simple ones. For the reason of defining Spline based wireframe objects which have wide applications in surface modelling other method is applied. These objects by many control points and vectors which calculate by algorithms, are defined. Surfaces Surface is the part of the outer object with zero thickness which is represented in parametric way by special algorithms. Surfaces have different types such as regular, mesh, facet, voxel, organic, smooth freeform and much more. How ever the research topic is much related to smooth freeform surfaces rather than the other type. History of surfaces One of the earliest surfacing technologies in early 60s was the Coon Patch, a four-sided, curved surface component. It was first used to design ship hull shapes which have simple curved surfaces with few changes in directions. Automotive surfaces were other surfaces to be define. Automotive surfaces are smaller than the ship hull and directions change faster with much more frequent styles. Bezier mathematics is the foundation of many systems which has had wide acceptance in the earlier 60s. Bezier curve differ from

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

splines and its curvature is controlled by passing a curve through many points. these control points are connected by a control polygon that creates a framework for the curve. With Bezier curves could easily present the automobile body. B-spline was developed in the early 70s, it is similar to Bezier curves in its use of a control polygon. But uses a succession of polynomial segments to represent a curve, while Bezier has single segment. In the late 70s rational equations were added to the Bspline. In a rational B-spline control points weighted by a fourth component and represents a segment without approximation. Then in the 80s Came Non-Uniform Rational B-splines (NURBS) which refers to the manner which curve has with segment ends. NURBS has many advantages over other surfacing methods, applying NURBS helps to have fewer surface, easier editing and accurate data transfer. Surface modelling Surface modelling creates surfaces representing the aria of the part where every position on the surface is defined by a mathematical method. Mathematical methods for defining surfaces are Bezier, B-Spline, NURBS, Coons patch. Freeform surface Modelling Freeform surface modelling is the art of creating freeform surfaces with CAD. Car bodies, consumer products, turbine blades are many example of this regard. CAD packages even developed ones for creating surfaces have two basic methods: 1. Creating Spline curves then applying Sweep of Loft along guide

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

curve. 2. Direct creating of the surfaces with managing of the surface poles and control points. Then these created surfaces by applying derived methods such as Offset or extension created new and much complex surfaces. Freeform surface modelling is used in CAD and other computer graphics packages. Most of modelling packages use nonuniform rational B-spline(NURBS) Mathematics to describe curvature smooth surfaces. Fewer surfaces for object creation Applying NURBS capabilities has much more applications than any other mathematical approach in surface modelling. Packages which are based on NURBS techniques need minimum quantities of surfaces to create a complex surface. fewer surfaces leads fewer error and fewer time in surface modelling. This advantage of NURBS directly reduced modelling efforts of maintaining the surface tangencies. For example a car bonnet which needs fifty separate Bezier surfaces, could be modelled with a single NURBS. The time saving of NURBS technique is significant which require two hours for a surface where with other techniques of deferential geometry it had taken up to two weeks to smooth all the tangencies between Bezier surfaces. NURBS packages offer most powerful editing capabilities than other regular packages such as Bezier packages. NURBS advantage is not just its ease of editing but also its power of local editing which lets designer edit a small part of a model without affecting the entire curve or surface. CAD system that supports

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

NURBS has the capability to import other geometry such as Bezier or other surfaces. This is why NURBS is identified as a standard worldwide in industry for representing curves and surfaces and supports by international Initial Graphics Exchange Standard (IGES).

Multiresolution surface modelling and Surface patches A Multiresolution surface is a valuable framework which all higher and lower resolution details are available in surface. NURBS with Several editing operation such as translation, rotation, extension, scale, reflection, and control point manipulation are employed for modelling complicated objects.

However with only these basic operations it is difficult to model more complex editing. Several researchers have worked on improving the surface modelling. Here are many developed techniques of surface modelling:

1. Free-Form deformation for general applications 2. Non-parallelepiped lattices for more general applications 3.Variational surface modelling that allows direct manipulation by defining a set of constraints 4. Applying B-spline wavelet But the new way of presenting multiresolution complex surfaces which applies by modern design packages are surface patches. Surface patch technology provides very elegant solution for parameter space and parametric derivatives. Packages which

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

apply surface patches solve geometry creation problems with technologically advanced methods of polynomial mathematics that provide a Varity of advantages over traditional geometry definitions which gives the mathematical capabilities to handle existing curve and surface generation functions . Conclusion and future work In recent years, methods from computational geometry have been widely adopted by the computer graphics with efficient algorithms. in the other words now the 3D computer graphic packages are created with a wide range of data structures from differential and computational geometry. 3D computer graphic algorithms are strongly dependent on derivatives and integrals of surfaces and curves in special points. Classical differential geometry equations couldnt be enough to response Automotive surface modelling applications as a single tool and It is not the main mathematical tool for surface modelling. But as a useful tool, plays an important and necessary role in surface modelling and computer graphics algorithms. After all above definitions and history of surface design and modelling the future research will focus on discrete mathematics - in place of continues mathematics- which are the future of modelling and basic definition of Reverse Engineering.

References and bibliography

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

[2] David F.Roger: An Introduction to NURBS, Morgan kaufmann, 2001

[3]A.Cirelo: Numerical approach to the keel design of a sailing yacht, 2008

surfaces,2004 [6] Xiuzi Ye: Reverse Innovation design, 2007

Apendix I Brief Research Proposal - introduction Success of a product design directly is based on applied CAD software package then all optimization in CFD and (or) FEA packages. A careful CAD, CFD and EFA modelling with successful optimization leads to a stable, efficient and generalized method. Automotive body as a complex surface shape always have challenge with different aria of science and engineering such as CAD, CFD, FEA, safty and state of art. Having best results, success and innovation in all of these arias which leads a new car body shape, needs new methods of modelling and analysing at the same time which sounds complicated. Normally primary model is created from a points cloud of last product which generated by a digitizer in CAD. Then the current industrial accepted method is to apply changes in CAD after each analysing in CFD or FEA packages.

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

-objective Those kind of industrial products which CFD and FEA are necessary for design and need much developments as possible In order to having more speed, efficiency, safety and etc. -goals 1. Discover faster and much accurate method of Automotive body surface design and reverse engineering. 2. Create a high level algorithm for modelling and optimization which manages the process of design from idea to high quality products. 3.having a work space which modelling package and CFD/FEA analysers have online data exchange/communication. -Methodology First of all finding minimum quantities of points-for same results of CFD and FEA analysing- for surface modelling, secondly true communication between CAD and analyser packages are what methodology suppose to have. Primary model usually comes from a points cloud or designer base on customer needs, market requasts, state of art, economics limitations and initial CFD and FEA knowledge create primary model by CAD. primary design has a long way to be the finished model. In this regard automotive body optimization is one of the difficult ones. Points cloud which have been created by digitizer is imported in CAD. Obviously digitized points are too much and most of them

A.Mosavi, An Essay to Introduction of Differential Geometry and Applications, reports of the faculty of Informatics, Scientific Computing, No. A/2008, University of Debrecen, Debrecen, 2008.

are unnecessary and for sure few points are missed. So mathematical understanding the surface characteristics could bring the knowledge of identifying initial points for creating surfaces which later could be the base of a feedback system structure to digitizer. This method of digitizing can pick up most important points which are necessary and also enough for creating primary model. For design a new product after creating primary model firstly some changes are applied on model in order to change its shape in to the ideal CAD model. Obviously the ideal created model economically, aerodynamically and stability couldnt be the finished CAD model. Therefore some optimization of CFD and EFA seams to be necessary; but not in current common nonorganized method. For automotive surface modelling two optimization must be done in same time and as each one has effect on other makes the process complicated. Therefore creating a algorithm for this reason in todays situation of try and error method of optimization, could be a innovation.