Anda di halaman 1dari 19

TEP

TEPE

Twinning Engineering Programmes and Thammasat English Programme of Engineering

Thesis Format Guideline

Revision 0.0 2008

Preface
Congratulation students. If you are reading this instruction then you are about to graduate from TEPE. You will leave behind your thesis which may be used as a reference by students or the publics in the future. Hence it is very important that your thesis be correct, complete and well-organized

Your thesis is the document that tells the whole story of your Senior Project. It must be planned carefully so that the people who did not participate in your work can understand, continue and even repeat your work. The story telling plan can be laid out by first design the step to tell, which is to write the contents. Then work out the detail from there. Make sure that the story is continuous.

The main body of the thesis shall be divided into equal-sized chapters. Chapter 1 Introduction is where you write the background, the purposes and the scope of your project. You may also discuss the literature survey in this chapter. Then introduce the readers to the rest of the thesis. Other successive chapters continuously inform the readers of you projects: theory, design, procedure, result, etc. The last chapter is then the conclusion.

The complete thesis composed of the title page, half title page, approval page, abstract, list of figures, list of tables, list of symbols, table of contents, the main body, appendix and bibliography. This Thesis Format Guideline instructs the format in the form of examples and comments on each component. Please review and follow the instruction carefully.

Finally, from all the lecturers and staffs of TEP/TEPE, we wish you a happy and successful future.

TEP/TEPE 2008

Title Page 50.8 mm

Optimum Geometric Design of Pivot Arm

Font: Times New Roman Effect: Bold Alignment: Center Size: 16

By

Font: Times New Roman Effect: Bold Alignment: Center Size: 14

Mr. Visi Rattanajit Mr. Kolawa Suwi

4710750086 4710750622

Font: Times New Roman Effect: Regular Alignment: Center Size: 14

Advisor

Submitted in partial fulfillment of the requirements for the degree of Bachelor of Engineering in the Thammasat English Programme of Engineering Faculty of Engineering, Thammasat University 2008
Font: Times New Roman Effect: Bold Alignment: Center Size: 14

Half Title Page

Optimum Geometric Design of Pivot Arm


By Mr. Visi Rattanajit Mr. Kolawa Suwi 4710750086 4710750622
Same as Title Page

Advisor Asst. Prof. Dr. Jarum Pong

Co-Advisor Dr. Nopad Uchachat


Font: Times New Roman Effect: Bold Alignment: Center Size: 14

Committees Asst. Prof. Dr. Samjit Homros Assoc. Prof. Dr. Pongchan Luangpai

Font: Times New Roman Effect: Regular Alignment: Center Size: 14

Submitted in partial fulfillment of the requirements for the degree of Bachelor of Engineering in the Thammasat English Programme of Engineering Faculty of Engineering, Thammasat University 2008
2

25.4 mm

Approval Page

Project title: By:

Optimum Geometric Design of Pivot Arm Mr. Visi Rattanajit Mr. Kolawa Suwi 4710750086 4710750622

Font: Times New Roman Effect: Regular Alignment: Size: 12

Thammasat English Programme of Engineering, Faculty of Engineering, Thammasat


38.1 mm

University approves this thesis toward partial fulfillments of the requirements for the degree of Bachelor of Engineering in Industrial Engineering

25.4 mm

_______________________________ Director, TEP/TEPE (Dr. Thira Jiasiripongkul)

_______________________________ Project Advisor (Asst. Prof. Dr. Jarum Pong)

______________________________ Project Co-Advisor (Dr. Nopad Uchachat)

_______________________________ Committee (Asst. Prof. Dr. Samjit Homros)

_______________________________ Committee (Assoc. Prof. Dr. Pongchan Luangpai)

25.4 mm

Abstract Hard Disk Drive Assembly Process Improvement


Mr. Visi Rattanajit Mr. Kolawa Suwi
38.1 mm

The main purpose of the abstract is to quickly give an idea of what this thesis is all about to the reader. The abstract shall briefly discuss the followings: background, purposes, scope of work, procedure, result and conclusion. Overall it should be about half page or not more than 500 words. Use Times New Roman size 12 font. Line spacing shall be 1.5 lines.
Font: Times New Roman Effect: Regular Alignment: Size: 12 Line spacing: 1.5 lines

Key words: Hard Disk, Assembly, Designed Experiment

25.4 mm Font: Times New Roman Effect: Bold Alignment: Center Size: 14 Font: Times New Roman Effect: Regular Alignment: Center Size: 12

4710750086 4710750622

25.4 mm

25.4 mm

Lower case Roman number

Acknowledgements

Optional. Use this page to acknowledge the organizations that provide funding for the project or to thanks the people who assist you. Keep it formal. The format of the text shall be Time New Roman 12pt font.
38.1 mm Font: Times New Roman Effect: Regular Alignment: Size: 12 Line spacing: 1.5 lines

25.4 mm 25.4 mm

Font: Times New Roman Effect: Bold Alignment: Center Size: 14

25.4 mm

ii

Table of Contents
Abstract Acknowledgement List of Figures List of Tables List of Symbols and Abbreviations Chapter 1: Introduction 1.1 Principles of engineering design 1.2 Overview of design optimization problems Chapter 2: Representations of Freeform Geometry 2.1 Interpolation splines 2.2 Bezier curves
38.1 mm

Font: Times New Roman Effect: Regular Alignment: Size: 12

25.4 mm

Font: Times New Roman Effect: Bold Alignment: Center Size: 14

i ii iv v vi 1 2 3 3 4 5
25.4 mm

Chapter 7: Concluding Remarks Appendices Bibliography

97 98 99

25.4 mm

iii

List of Figures
Figure 1.1 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Product development process. Natural cubic interpolation spline. Convex hull property of a Bezier curve. Degree elevation of Bezier curve. Linear interpolation of convex tolerance zones. 2 4 9 11 12

Figure 6.20

Thickness of robust-optimum designs compared to the thickness of the feasible designs and the trial designs. Natural frequency of robust-optimum designs generated using the proposed design rules compared with the natural frequency of the trial designs and flat plates. Convergence study of the signal-to-noise ratio of natural frequency. Neighborhood radius = 10%.

81

Figure 6.21

82

Figure 6.22

87

iv

List of Tables
Table 1.1 Table 4.1 Table 4.2 Classification of optimization problems. Comparison of the rate of convergence. Valued of design variables during the trial. 3 52 53

Table 6.3 Table 6.4

Design rules for robust-optimum corrugated panel structures. Coefficients of the response functions of the feasible designs.

77 85

List of Symbols and Abbreviations


C F I k P q, Q U v V capacitance force current springs constant or thermal conductivity pressure charge energy velocity voltage coefficient of static friction coefficient of kinetic friction density of liquid
Font: Times New Roman Effect: Regular Alignment: Size: 12

Font: Times New Roman Effect: Bold Alignment: Center Size: 14

s k

vi

50.8 mm (First of chapter page only)

Chapter 1 Introduction

Font: Times New Roman Effect: Bold Alignment: Left Size: 14 25.4mm

The thesis shall be printed on A4 sheets with top margin 25.4 mm., bottom margin 25.4 mm, left margin 38.1 mm. and tight margin 25.4 mm. Heading of each chapter shall be Times New Roman (Bold) Size 14 font. The section headers shall be numbered consecutively in each chapter and shall be of the font Times New Roman (Bold) Size 12 font.
38.1mm

The first page of each chapter must have a 50.8 mm. margin at the top of the page. The word chapter and the chapter number must appear on the first page of each chapter preferably on a separate line from the chapter title. The text of the chapter must start a line space or two below the title, not on a separate page. The content of the thesis shall be typed with Times New Roman Size 12 font with line spacing of 1.5 lines. References shall be mentioned with the authors name and year of publication and must be listed in the Bibliography section. See the Bibliography section for examples of reference format. All figures and tables must be numbered and named. A figure or a table shall be placed after the paragraph mentioning it. All equations must be typed with proper equation editor software. If referenced in the text, the equation must be numbered consecutively. All writing in the thesis must be formal. Each chapter shall present the continuous flow of the story. It is recommended that students write the table of content first and then adding detail describing and connecting chapters and sections.
Font: Times New Roman Effect: Regular Alignment: Size: 12

1 Regular page number start at this page


25.4 mm

1.1 Principles of engineering design

Engineering design is a part of the product development process. According to Ulrich and Eppinger (1995), a generic product development process can be drawn roughly as shown Fig. 1.1. Our work fits within the engineering function of the detail design stage.

25.4 mm Font: Times New Roman Effect: Regular Alignment: Size: 12

Concept development

System-level design

Detail design

Protyping, testing and refinement

Production Ramp-up
Font: Times New Roman Effect: Bold Alignment: Center Size: 12

Figure 1.1 Product development process.

As discussed earlier, the primary considerations in engineering design are functionality, manufacturability and reliability. However as technological advancements and market competitiveness urge designers to set higher standards; better performance, time-to-market, quality and cost also become significant. To deal with these issues,
38.1mm
25.4mm

various engineering design principles have emerged. Some of them are: Optimal design: Design to maximize performance under limited resources (constraints). Probabilistic design (Haugen, 1968): Using means and standard deviations in the computation instead of nominal values for design variables and/or operating conditions, providing designers with probabilistic information.
A reference a publication

Robust design (Phadke, 1989): Design to reduce sensitivity to variations. Axiomatic design (Suh, 1990): This design principle is at conceptual or system level. In an ideal design, functional requirements (or design objectives) and design variables should have a one-to-one relationship. Key characteristics or KCs (Lee and Thornton, 1995): Key characteristics are features whose variations have the most effect on the overall product. Designers should identify and pay special attention to those features. Single objective optimum design without consideration of other issues may leads to a design that is very sensitive to variations (not robust) or is not manufacturable. This research is built around the use of design optimization. The next section discusses some of the well-known categories of optimization problems and solution techniques.

2
25.4 mm

1.2 Overview of design optimization problems

A table is mentioned in the text

To optimize the design is to find values of the design variables that minimize or maximize the objective function while satisfying the constraints (if there are any). For example one might want to maximize the surface quality of the machined parts by treating turning speed, tool tip radius and feed rate as design variables. Design optimization is a very broad area. Table 1 provides some of the possible classifications of the optimization problem by different criterions.
A table is numbered Font: Times New Roman Effect: Bold Alignment: Center Size: 12

Table 1 Classification of optimization problems. Criterions Constraint Continuity of design space Number of Objective function Type of objective function Classes Constrained*/unconstrained* Continuous*/discrete/mixed-type Single-objective*/multiple-objective*

Linear/non-linear*; differentiable/non-differentiable*; Deterministic*/stochastic

Objective Application specific

Local*/global Geometric*/topology*/network/combinatorial/queuing/ space planning/scheduling and etc.

Different priorities be set for the objectives. Usually a single solution is expected. If, however, the priorities are not clear, exhaustive methods such as parameter space investigation are used to generate a feasible solution set and a set of extreme designs called Pareto optimum solutions are selected from the feasible solution set. More details of the indirect optimization techniques that are related to this work can be found in Chapter 3.

50.8 mm (First of chapter page only)

Chapter 2 Representations of Freeform Geometry

Font: Times New Roman Effect: Bold Alignment: Left Size: 14

The representation of part geometry should allow the design space to be explored by as few parameters as possible in order to minimize computational effort. It would also be desirable for this representation to be compatible with CAD tools. Interpolation splines and NURBS such as B-spline and Bezier curves provide economic ways to describe freeform geometry. With a few control points these representations allow the shape of complex freeform geometry be controlled. This enables the development of an approximate mathematical model of the relationship between an objective function and the control point locations. Let

Pi = {xi , yi }, i = 0,1,2,K, n

be a set of n + 1 control points and

P(t ) = {x(t ), y (t )} be a point on a curve defined by the control points at a parametric


value t. Different representations can be used to relate control points to points on a curve. One of the frequently used representations is shown in Fig. 2.1. The mathematical formulations for those curves are discussed in Sections 2.1 to 2.3. Section 2.4 is an additional section that discusses the relationship between the movements of the control points and the geometry of a NURBS curve.
A figure is mentioned in the text

P3

P4 P1 P2 P0
A figure is numbered

P5

Figure 2.1 Natural cubic interpolation spline.

Font: Times New Roman Effect: Bold Alignment: Center Size: 12

2.1

Interpolation splines An interpolation spline is a set of piecewise curves. The most common form is a

cubic spline, which has continuous first and second derivatives. Between two consecutive points a piece of curve defined parametrically by a cubic polynomial is used to connect the two points. Following Ferziger (1998) a cubic polynomial in parametric form joining Pi and Pi+1 can be written as
Equation is numbered

P (t t )3 (t ti )3 Pi i P (t ) = Pi i +1 i + Pi 1 + Pi (ti +1 ti ) + i +1 i Pi 1 (t ti ) + + 6 i 6 i 6 i 6 i

(2.1)

where i = t i +1 t i . Note that the index i starts from zero. The parameter t i can be assigned to Pi as proportion to the total distance of straight lines running from P0 to Pi, i.e. t 0 = 0 , t i = t i 1 + ( xi xi 1 ) 2 + ( yi yi 1 ) 2 or as uniform value proportional to the index value i.e. t i = i .
Numbered equation is mentioned in the text

To solve for the second derivatives, Pi s, differentiate the polynomials in Eq. 2.1 and match the values of the first derivatives at the control points; this provides the set of linear tridiagonal equations:
( + i ) i 1 P + Pi Pi + Pi 1 + Pi1 + i 1 Pi+ i Pi+1 = i +1 6 3 6 i 1 i (2.2)

Different end conditions can be specified in Eq. 2.2; for examples with n + 1 control points P0 and Pn = 0 provides natural spline, P0 = Pn1 and Pn = P1 is a periodic spline.

The good property of this type of spline is that the control points lie on the curve but the main disadvantage is that the curve exhibits oscillating or wiggle behavior that can be hard to control. One possible solution is to use tensioned spline but then the curve become computationally more expensive. Moreover, different amount of tension must be determined for different geometry.

50.8 mm (First of chapter page only)

Chapter 7 Concluding Remarks

This research focuses on the development of an optimum and robust design methodology by integrating existing CAD and CAE tools. We have demonstrated such integration in Chapter 4 for design optimization. In the CAD part of the work, use of NURBS curves (Bezier and B-spline form) allows the design space to be explored using only a few design variables. Moreover, the degree elevation property of Bezier curves enables us to explore the design space with minimal number of control points at first and progressively increase the number as needed. The CAE tools (governing equation solvers such as the flow solvers) are used as black boxes. A generic optimization algorithm was developed based on the black-box approach with the use of design of experiments (DOE) and steepest descent techniques and a computer program (optimizer) is written based on this algorithm. We were able to integrate the optimizer with several black-box solvers through the use of previously described CAD tools. In all the examples there are 3 to 5 design variables and they require direct solutions numering O(100) to converge to an optimum solution. These numbers can be decreased by improving the optimizer. One of the possible ways to increase efficiency of the optimizer is to develop some means (e.g. some form of a trust region analysis) to track the behavior of the local response function during the search in order to automatically select a suitable DOE technique (finite difference, factorial design or small composite design) for each design iteration. For further improvement, one can also use the information obtained from the DOE such as the effects of each control point to identify the most sensitive region along the curve and specifically add more control points to that region instead of simply using a degree elevation algorithm. One can see from the work presented in this thesis that complex design problems can be solved with integration of relatively simpler engineering tools. As CAD, CAE and CAM tools become more mature it is important for engineers to realized the capability of these existing tools and make most use out of them by using them together.

97

Appendix
A. Multi-dimensional response function This section provides a methodology to find coefficients of a response function by a least square approach. Consider a scalar value y that is a function of design variables x = (x1, x2, x3,, xn). The exact relationship is unknown and is represented with a function
F; y = F(x). We try to approximate the response of the function F with a response

function F ; y = F(x) F (x) .


Let F (x) = ci Gi (x) where ci are unknown constant coefficients and Gi are
i =1 k

known functions and F is called the response function. The basic example of a linear
response function for n = 3 is k=n+1=4 G1 = 1, G2 = x1, G3 = x2 and G4 = x3 To obtain the coefficients of the response function, first there must be a sufficient number of samples, m k. A set of m samples can be written in the form of: F(xi) = di, I = 1,2,3,, m. Then match the response function to the data set and write it in the matrix form as: GC = D
G1 (x1 ) G2 (x1 ) G3 (x1 ) ... Gk (x1 ) c1 d1 2 2 2 2 c 2 G1 (x ) G2 (x ) G3 (x ) ... Gk (x ) 2 d 3 3 3 3 Where G = G (x ) G (x ) G (x ) ... G ( X ) C = c and D = 3 d k 1 2 3 3 M M M M m G (x m ) d c k ... Gk ( x m ) 1

The coefficient C can be obtained from C = (GTG)-1GTD

This approach is called least square fitting because the response function obtained this way will minimize the square of error.

98

Bibliography

Ferziger, J. H., 1998, Numerical methods for engineering applications, John Wiley & Sons. Hougen, E., 1968, Probabilistic approaches to design, John Wiley and Sons. Lee, D. J., A. C. Thornton, and T. Cunningham, 1995, Key characteristics for agile product development and manufacturing, Agility Forum 4th Annual Conference Proceedings, March 7-9, 1995, Bethlehem, PA, pp. 258-268. Phadke, M. S., 1989, Quality engineering using robust design, Prentice Hall. Suh, N. P., 1990, The principles of design, Oxford University Press, New York. Ulrich, K. T. and S. D. Eppinger, 1995, Product design and development process, McGraw-Hill, Inc

99