Satchitanand Kashyap

Lucent Technologies/Bell Laboratories

Engineering & Environmental Technologies
PR 2-3067, Princeton NJ 08542
Phone: (609) 639-2279
e-mail: [email protected]

Contributions to CAD/CAM/CAE, DFMA and

Design Optimization Research and its
Applications in Telecommunications

In broader terms my research and work experience can effect the three most impor-
tant metrics that are a measure of a manufacturing industry today, 1. Time-to-market
in product/process realization, 2. Quality of a manufactured product, and 3. Cost of a
product/process. My research contributions (Academic & Industrial) and its implications
on Lucent Technologies product/process realization process is categorized into following
disciplines.

1. Solid, Geometric modeling, CAD/CAM/CAE integration, and optimization theory and

practice.

2. Design for Assembly, Manufacture, Cost, Environment, etc.

3. Network Systems Engineering/Re-engineering.

Solid, Geometric modeling, CAD/CAM/CAE

Integration, and Optimization Theory and Practice
The field of mechanical design has been revolutionized by the advent of modern day com-
puters. Designs which were worked and drafted on paper have been replaced by computer
software. My contributions to this field of evolving CAD/CAM/CAE over a decade have
been significant both in theory and practice. My contributions to CAD/CAM/CAE can be
described further as follows:

1
Contributions to CAD/CAM/CAE, DFMA, and Design Optimization Research and its Applications in
Telecommunications

Solid Modeling: In the past most designs were worked out and drafted on paper, My
work [13][14] in solid modeling will help translate these designs into valid com-
puter models. The algorithm that I developed will help 1. Realize these scanned
design drafting information, 2. Represent this information in a valid mathematical
form on the computer, and finally 3. Render this information on a computer. The
actual process being in scanning these 2D legacy draftings, building a valid 3D rep-
resentation from this disparate 2D information with the developed algorithm as a
Boundary Representation, and finally rendering this 3D model on a computer. The
software that I developed on a PC using C language can also be useful in realizing,
representing, and rendering data from Coordinate Measuring Machines, Computer
Tomography data, or as a visualization tool for other scientific purposes.

Geometric Modeling: Modeling curves and surfaces is an important technical require-

ment in ship design, aircraft design, automobile design, and aesthetic design of
products. Blending is a process involved with Curve & Surface modeling that in-
volves developing continuity between two separate curves or surfaces. Curvature
continuous blending is of technical importance as it affects the hydrodynamic and
aerodynamic needs in ship body, aircraft, and automobile design, besides creating
a better look and feel of a product for aesthetics. Blending is a very difficult and
numerically intense operation. With my novel approach in variable sweep blending
[11][12] we can develop a blending patch between the two surfaces using a swept
patch between the two surfaces to be blended ensuring curvature continuity at the
linkage surface. The mathematical criteria that I have developed ensures curva-
ture continuous blend between the surface patches to be blended. My approach can
handle a versatile array of implicitly, explicitly defined parametrized surfaces. My
approach handles singularity conditions that arise during blending surfaces, and is
a significant improvement over existing numerically intensive algorithims based on
the sub-division of the surfaces to be blended. My tool incorporated in a geometric
modeler gives a user a flexible choice choosing trimming curves ensuring curvature
continuous blends at a low processing cost.

CAD/CAM/CAE Integration: As a part of my doctoral research [9][3][10][7][1][2] I

set out to develop a commercial level CAD/CAM/CAE package to facilitate fixture
design to realize a product's manufacture at the design stage of the product. This
involved developing, integrating, and implementing disparate methodologies. The
methodologies implemented on Pro/ENGINEER CAD/CAM package and Patran 1.4 CAE
package address 1. Manufacturability evaluation of a product design [3] (ie., Cost
of manufacture, Process capability to ensure quality), heuristics for fixture location
synthesis [9][10], finite element analysis of fixtures[9][7], and finally optimization
[1][2] of the final fixture design for minimal part deformation during machining.
Fixturing is usually performed by expert tool room personnel after the product de-
sign before the products manufacture. With the use of my developed tool, designers
can easily realize fixture design satisfying cost, quality, and time considerations in
the product development effort.

2
Contributions to CAD/CAM/CAE, DFMA, and Design Optimization Research and its Applications in
Telecommunications

Design Optimization: Fixture optimization [9][1][2] is a non-linear process, because of

the changing boundary conditions, and the formulation of the objective function
both the finite element analysis and design optimization become non-linear. The
complexity on this problem is further increased with complexity in the geometry.
My work in design optimization of fixtures addresses these complexities using a suite
of non-linear optimization routines. The optimization is performed using different
non-linear approaches that converge to a valid solution faster. Given the complex-
ity of the problem, a combination of non-linear optimization routines as gradient
based approaches, penalty function approaches, sequential quadratic programming,
and simulated annealing are combinatorially used to find the optimal fixture lo-
cation for minimal part deflection. The suite of optimization routines developed
and implemented adapt to changing boundary conditions as imposed by the finite
element model. This poses the problem as a multi-level multi-generation problem
which is solved by the best optimization routine at each function evaluation.
My current research in design optimization involves non-hierarchical optimization
techniques which are still in their infancy and their application in multi-disciplinary
optimization(MDO) which is different from the hierarchical optimization techniques
mentioned in the preceding paragraph. MDO is a methodology for optimizing large
systems which have many contributing disciplines with conflicting objectives. These
disciplines often interact with one another during the design process, requiring
and/or providing analysis results to and from each other. For example in electronic
packaging the trend is to wards increasing the functionality, lowering the compo-
nent size and cost. The disciplines involved can be electrical, loss, and thermal.
There are complex coupling between electrical, loss, and thermal performance of
the electronic component being optimized. My focus is towards formulating such
MDO problems that are applicable to telecommunications and solving them using
appropriate non-hierarchical optimization routines.

DFMA, and System Driven Design

My research, development, deployment and consultancy in the last two years has been
towards applying concurrent engineering principles in realizing manufacturing related
problems early in the design stage. This can be designing a product for assembly, cost, dis-
assembly, or other down-stream applications. I initially complied the Design for Automated
Guidelines as part of Lucent-wide best practices. I have developed and extended the
Design for Simplicity design critiquing tool for design for automated assembly [8] which
has now been implemented as a web-enabled tool. As an out come of this research product
designers within Lucent have successfully evaluated their design for assembly, and cost.
My work in snap-fastener design [6] facilitates faster physical design of consumer products
as a CAD-enabled design tool. My work [4] in modular design of cabinets has facilitated
the Network Wireless business at Lucent to design, manufacture, a modular cabinet for
evolving product platform family.

3
Contributions to CAD/CAM/CAE, DFMA, and Design Optimization Research and its Applications in
Telecommunications

Presently I have been involved with systems level design and optimization of Lucent
communication products focusing on reducing costs. This concurrent engineering ap-
proach involves the use of competitive information in designing a product to meet cus-
tomers requirement in Lucent's products for maximum functionality at minimal cost of
manufacture. The is achieved by deploying processes like Target Costing, Conjoint Analy-
sis, & Quality Function Deployment within Lucent's business units.

Network Systems Engineering/Re-engineering

With the de-regulation in the telecommunications market, and the revolution in the
telecommunication industry towards video, data, and voice convergence, the opportu-
nities for novel means to cater to communications are boundless. The advances in trans-
porting data as packets as in Voice over IP (VoIP) or as cells as in Aysnchronous Transfer
Mode (ATM), or as dense light pulses as in optical networking creates many possible ways
of configuring, operating, and managing a network. My experience [5] has been lever-
aged in examining the technical feasibility and business possibilities in leveraging these
advances in communications for new business ventures. I am presently examining the
possibilities of configuring a transport network (Optical, ATM, & IP) to cater to increased
need of band-width in the future with potential for business growth. Technical criteria
that are of importance in configuring networks are in feasibility, interoperability, scal-
ability, reliability to transport video/data/voice seamlessly across different networks is
the major focus. Underlying reasoning being low cost deployment, and operation of the
network. My consultancy work in this arena is helping Lucent's businesses as in Data
Networking, Optical Networking and Global Service Providers realize their objectives.

References
[1] S. Kashyap, & W. R. DeVries, ``Achieving Design Specifications Through Finite Ele-
ment Analysis and Structural Optimization During Product Realization'', International
Journal of Process and Product Design Optimization.

[2] S. Kashyap, & W. R. DeVries, ``FEA and Structural Optimization in Fixturing Design'',
Journal of Structural Optimization.

[3] S. Kashyap, & W. R. DeVries, ``An Agent-based Methodology for Manufacturability in

the Product Realization Process'', Transactions of XXVII North American Manufacturing
Research Institution of SME.

[4] S. Kashyap., `` Modular Cost Effective design of Wideband CDMA Macro Cell for Do-
CoMo'', Technical Report(Restricted), Lucent Technologies/Bell Laboratories 1998.

[5] S. Kashyap., `` Confluence of Voice, Video, and Data on Packet Switched Networks,
QoS, Reliability, Scalability: Business Propositions and Technical Feasibility of Small-

4
Contributions to CAD/CAM/CAE, DFMA, and Design Optimization Research and its Applications in
Telecommunications

end Terminals'', Internal Technical Document(Restricted), Lucent Technologies/Bell

Laboratories 1998.

[6] S. Kashyap., `` A Feature-based Attachment Level Snap-Fastener Design for Shorter

Design Cycle-time'', Internal Technical Document ITD-98-34629P, Lucent Technolo-
gies/Bell Laboratories 1998.

[7] S. Kashyap, & W. R. DeVries, ``An Integrated Concurrent Engineering Framework for
Manufacturability Evaluation, Fixture Design, Analysis, and Optimization in Virtual
Prototyping'',31st International CIRP Seminar on Manufacturing Systems, Berkeley
1998, page 517-522.

[8] S., Kashyap., `` Product Design for Automated Assembly'', Internal technical Document
ITD-97-324945, Lucent Technologies/Bell Laboratories 1997.

[9] S., Kashyap., `` An Integrated Methodology for Fixture Design, Analysis, and Opti-
mization'', Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy NY, 1997.

[10] S., Kashyap, & W. R. DeVries, ``Kinematic Constraint of Workpiece During Machin-
ing, A Feature Based Approach", Transactions of XXV North American Manufacturing
Research Institution of SME.

[11] S., Kashyap., ``A Survey of Multiplicity of Orthogonal Projections on to Free-form

Surfaces'', Technical Report # 94040, Design & Manufacturing Institute, Rensselaer
Polytechnic Institute, Troy NY, 1994.

[12] S., Kashyap., `` Curvature Continuous Blending Using Variable Sweeps'', Masters The-
sis, Rensselaer Polytechnic Institute, Troy NY 1994.

[13] S., Kashyap., ``Realizing, Representing, and Rendering 2D Design Data as B-REP Solid
Models'', Bachelors Thesis, Bangalore University, 1987.

[14] S., Kashyap., ``Automatic Generation of 3D Solids from 2D Orthographic Views'', Tech-
nical Report #: 94039, Design & Manufacturing Institute, Rensselaer Polytechnic In-
stitute Troy NY.