Session IT 302-041

Using Knowledge-based Solid Modeling Techniques and Airfoil Design Data: A Case Study
in Developing an Airfoil Seed Part Generator

William D. O’Brien, Michael Chandonais, R. Scott Henderson, Mark

Prezbindowski, Nathan W. Hartman, Joseph A. Rasche

Purdue University ([email protected]) / Purdue University / Purdue

University / Purdue University / Purdue University ([email protected]) /
Rolls-Royce Corporation ([email protected])

Abstract

This paper details a research project which encompassed the development of corporate best
practices surrounding the use of company specific compressor and turbine design data in the
generation of solid model geometry. Current solid modeling processes at Rolls-Royce
Corporation for compressor and turbine engine airfoil geometry require a great deal of time and
are prone to human error in the replacement of CAD geometry when design changes are made. A
team of Purdue University students and faculty collaborated with Rolls-Royce Corporation to
improve this process using knowledge-based tools within the Unigraphics (UG) NX 2 modeling
tool set. A graphical user interface to import and manipulate aero data was developed and tested
to ensure that it met design standards and maintained airfoil geometric integrity within the Rolls-
Royce environment. Times for airfoil model replacement using the new application were
recorded and compared to the previous method. Results demonstrated that the knowledge-based
modeling applications had an average of 88.3% time savings as well as allowed for automation
of the solid model creation process of the airfoils. Suggestions for integration of such techniques
into engineering and technology education are also presented, along with implications for
education and industry.

Introduction

This project represents a project specification undertaken by four students in a senior project
course (CGT 411) in the Computer Graphics Technology (CGT) department at Purdue
University. This course provides students the opportunity to work in a team-based environment
to complete a project for an outside client. The client can represent any combination of the four
areas of the CGT department: virtual product integration, construction graphics management,
interactive media and imaging, and animation and gaming. Three of the four students on this
project were specializing in the virtual product integration area, while the fourth student
specialized in interactive media and imaging. They brought a unique blend of technical skills and
competencies that included 3D geometric modeling (surface and solid), scripting and
programming, technical writing, and prior industrial experience. One of the goals of the CGT
411 course is to have student teams use all of their knowledge acquired in their previous
scholastic experience and develop a solution to a client-based problem. In doing so, the students

Proceedings of The 2006 IJME - INTERTECH Conference

establish a fictional company for doing the work and administer the work to be done, including
financial reports, business plans, staffing, and legal issues. A substantial portion of the course
grade is based on how well they accomplish these things, in addition for developing a research
strategy to determine the effectiveness of their problem solution. The remainder of this paper
details the experiences of four CGT students in the Spring 2006 semester working on a project
with Rolls-Royce Corporation – Indianapolis.

The design of turbine engines at Rolls-Royce Corporation involves the modeling of airfoil
blades. This is a very iterative process centered on the analysis and revision of airfoil CAD
models. After revisions are made, geometrical data is re-engineered and recreated within
Unigraphics NX2. This process ranges from hours to days because the current methods of
creating the airfoil models in the CAD system are not parametric, i.e. the geometry is not
associated with the engineering definition of the airfoil after the model is created. An alternative
method of creating the geometry has been developed using Knowledge Fusion (KF), a scripting
language within NX2. This new method kept the airfoil geometry associated to the engineering
airfoil definition which allowed Rolls-Royce modelers to replace the airfoil within the existing
engine rather than deleting geometry and rebuilding it. A graphical user interface (GUI) was
developed using the User Interface Styler within NX2 to easily interact with the Knowledge
Fusion plug-in. By incorporating existing knowledge of design rationale and intent into an easy-
to-use plug-in, geometric models of airfoils can be created with the click of a button. The
resulting models are feature-based and parametric, allowing modifications to be made
automatically through the application or later manually within NX2, resulting in 88.3% time
savings. A glossary is included at the end of this paper to aid the reader in understanding the
technical terminology involved with this project.

Overview of Airfoil Design

Airfoil blades are a crucial component within a turbine engine, and their design covers many
engineering disciplines such as thermodynamics and statics. For both analysis and manufacturing
these airfoils are modeled in a CAD system. However, the complex shapes of airfoils make this
difficult. They are typically modeled using b-splines or NURBS and the development of methods
to do this has been ongoing for decades [1, 2]. At Rolls-Royce Corporation the definition of an
airfoil is determined by custom computer programs that take engineering parameters and
compute points at various sections of a potential airfoil. The points are saved in a text file
(known as an airfoil definition) for modeling. After the airfoil itself is modeled, the geometry
that is used to attach the airfoil to the engine (the stalk and platform) is added to the airfoil.
Figure 1 shows an example of an airfoil.

Proceedings of The 2006 IJME - INTERTECH Conference

 Figure 1: Airfoil with stalk (Rolls-Royce Corporation)

If the airfoil is a turbine blade, as opposed to a compressor blade, additional geometry is created
to hollow out the airfoil and add air passages that are used to cool the airfoil while the engine is
running. Due to the geometric complexity of these features, compressor blades were used as a
starting point for the evaluation of this modeling technique. After an airfoil model is completed it
is then analyzed by computer programs to determine if it meets its design requirements. Various
analyses, such as FEA and CFD, are run to determine the properties of the airfoil such as
strength, vibration, and thermal characteristics. Based on the results of these analyses, engineers
determine a new airfoil design and the process repeats - often dozens of times for each airfoil.
For compressor blades, each airfoil iteration requires 1 – 2 hours using the existing methods
described previously. For turbine blades, each iteration requires 8 – 16 hours using existing
methods. These values illustrate the need to explore alternative methods for iterating the design
of airfoils and the use of CAD geometry in the process.

A turbine engine can contain upwards of twenty different airfoils, so any improvement in the
process of one design iteration will have a beneficial effect on the total design process. Rolls-
Royce Corporation uses the Unigraphics NX2 CAD package as its primary modeler and has
developed several methods of creating airfoil models from the airfoil definitions. All but one of
the current methods for airfoil modeling in NX2 do not create parametric airfoils [3]. The
exception is a program used in the German division of Rolls-Royce; however, that program does

Proceedings of The 2006 IJME - INTERTECH Conference

not conform to U.S. design standards. This means that after the initial geometry is made, no links
to an airfoil definition exist. For every design iteration the complete airfoil model must be torn
down, the airfoil recreated, and the model rebuilt. In some instances, such as the much more
complex turbine airfoil, the entire model may have to be recreated.

Knowledge-Based Engineering (KBE) “has gradually gained prominence as a major tool to

speed up product development…” [4]. Essentially, KBE is the capturing of knowledge from
designers and engineers and automating it within a program [4, 5, 6, 7]. This knowledge is then
used to assist designers while they create products within the CAD system [8]. KBE systems are
capable of automatically creating objects [9, 10], assisting designers while they create objects
[11], and comparing the cost versus efficiency of created objects [12, 13]. There are many
examples of companies that successfully used Knowledge-Based engineering systems for the
development of their products. For example, Park et al. [13] developed a method that produced
multiple variations of blade designs for a VTOL aircraft, performed finite element analyses, and
then chose the optimal blade based on the results. The ability of KBE systems to reduce design
and analysis time is an optimal reason why these systems are being implemented in companies
around the world [14]. In a case closely related to the Rolls-Royce scenario, an aerospace
company implemented a system which allowed them to reduce design iteration from 80 hours to
four hours [15].

Unigraphics (UGS) NX2 (and other versions) contains several methods of modeling automation
which are already integrated with NX and included with the software. These include macros,
GRIP, Knowledge Fusion (KF) and NX/Open. Macros and GRIP have the lowest level of
functionality and no associativity with the geometry they create [16]. NX/Open has the greatest
capability but also requires the most specialized knowledge of any of the tools and has less
associativity than KF [16]. Knowledge Fusion and NX/Open can be combined and also interact
with the Unigraphics’ TeamCenter PDM software [16, 7].

KF is an object-oriented, hierarchical, declarative and rule driven scripting language for KBE
[7]. Geometric features are available as classes which allow for easy creation of dynamic
geometry. A KF developer is able to create custom “applications” for NX that provide the ability
to modify characteristics of the product without having to perform a complete redesign.
Widespread usage of KF appears limited. As noted by Carleton [11]:

Although the knowledge rules were relatively easy to apply to geometry during a
design, a high degree of training was involved in developing such rules as they
require following a strict procedure. (p. 38)

Another turbine engine manufacturer worked with UGS consultants to develop an intelligent
master model of an entire aircraft engine heavily based on KF. They developed this in order to
reduce iteration time and increase collaboration between component specialists [17]. By using
GUIs (Graphical User Interfaces), designers were then able to change components of the engine
so that new versions of that engine were created. However because of the advantages of KF as
discussed and a proof-of-concept study conducted by Rolls-Royce the decision was made to use
KF to improve the airfoil design process. By using this KBE technology, the time required to
remodel an airfoil for each design iteration was reduced. Previous design iterations required the

Proceedings of The 2006 IJME - INTERTECH Conference

tear-down and rebuilding of the airfoil model for each airfoil definition while the airfoil based on
KF remained associative to the airfoil definition. This meant that a new airfoil definition could
be specified and the model would update accordingly. Reducing the time of a single design
iteration created a significant total time savings for the entire design process.

Application Overview

In order to meet the needs of Rolls-Royce, an NX2 plug-in was developed using KF to read and
interpret ASCII files containing airfoil definitions and generate the geometry described by the
3D information contained in the files. KF technology within NX is an object-oriented scripting
utility. In KF, each object class was written as an individual text file (.dfa file extension), using
the included NX object classes as a starting point. The NX plug-in created during this project
was designated the Airfoil Seed Part Generator and was divided into compressor and turbine
components. A simplified hierarchy of these classes used in the compressor component of the
Airfoil Seed Part Generator is as follows:

RR_Compressor
RR_Parser_XDC
RR_Parser_RRAF
RR_DC_Flowpath / RR_Flowpath
RR_SectionSplines
RR_CapLines
RR_Caps
RR_RadialSplines
RR_FaceSurfaces

RR_Compressor served as a parent, or container, class for the compressor component. It

contained a number of child classes responsible for the creation of individual elements of the
airfoil geometry, as well as the code that enabled interaction with the GUI. The child classes
RR_Parser_XDC or RR_Parser_RRAF were used to open and parse an airfoil definition file.
Airfoil definitions were provided to modelers in an ASCII text file using the XDC or RRAF
format developed by Rolls-Royce Corporation. RR_Flowpath drew a partial flow path if one
was contained in the XDC file. The RR_DC_Flowpath class optionally parsed an external flow
path definition file and generated the resultant paths (see Figures 2 and 3).

Proceedings of The 2006 IJME - INTERTECH Conference

 Tip Flowpath

Airfoil

Hub Flowpath

Figure 2: Airfoil model with flow paths

RR_SectionSplines drew each of the stacked sections of the airfoil while RR_RadialSplines
drew the four radial splines at the tangency points. The geometry of the two end caps (leading
and trailing edges) was created by RR_Caps and RR_CapLines. Finally, RR_FaceSurfaces
created four surfaces: the leading and trailing edges and the pressure and suction sides of the
airfoil. These features are illustrated in Figure 3.

Cap

Leading
Edge

Radial Suction Side

Spline

Trailing Section Splines

Edge

Partial
Flowpath

Pressure
Side

Figure 3: Airfoil model details

Proceedings of The 2006 IJME - INTERTECH Conference

Although many other classes comprised the full compressor component, those discussed here
represent the core functionality of the compressor utility. The turbine component used a similar
set of classes to perform the same operation for constructing turbine geometry.

To simplify the use of the plug-in, a GUI was created using the NX User Interface Styler (UIS).
Users opened the Airfoil Seed Part Generator dialog from a menu option or toolbar icon (see Fig.
4).

Figure 4: Application menu options and items

Once opened, the dialog gave users the options to select the airfoil definition file, select flow
paths, and change modeling options and tolerances. The dialog also provided options to
configure the continuity settings and layer color options (see Fig. 5). The turbine component GUI
operated in a similar fashion, though it contained a number of features specific to turbine
geometry.

After an airfoil file was generated, the user could then proceed to attach additional (stalk)
geometry to it. The KF input parameters in the GUI remained associated to the geometry created
by the plug-in. In this way, the airfoil geometry was now associative to the airfoil definition files.
The ability of KF to create a dynamic model defined by rules made a plug-in based on this
technology a viable solution for Rolls-Royce. When newer design revisions of the airfoil became
available, the existing model could be regenerated to conform to the specifications of the new or
updated definition file without requiring reconstruction of the attached geometry.

Proceedings of The 2006 IJME - INTERTECH Conference

 Figure 5: Application GUI

Test Methodology

In order to determine the value of the airfoil seed part generator, the product was tested based on
the amount of time it could save over the current methods employed by Rolls-Royce. These
methods created geometry that was not associative to airfoil engineering data, resulting in a static
model that could not be easily updated to reflect changes in that data. With the associative
geometry produced by the Airfoil Seed Part Generator, it was hypothesized that the use of the
plug-in would yield significant time savings. Times to create various airfoils were obtained
through the usage of a survey.

Subjects
Two expert participants were used in this study. Both were Rolls-Royce designers that were
involved in the airfoil modeling process. These participants were expert users of NX and were
identified as potential users of the Airfoil Seed Part Generator. Participant 1 has twenty-nine
years of industry experience as a Designer in industry, with twenty-one years of experience at
Rolls-Royce. He graduated from a technical institute with an Associate’s degree in Tool Design
Technology and has been using 3D CAD tools since 1987. Participant 2 has similar experience.
He is a Designer at Rolls-Royce, with twenty-two years of experience with the company. Similar
to Participant 1, he has been using 3D CAD tools since 1987.

Procedure

Proceedings of The 2006 IJME - INTERTECH Conference

Data on the previous methods to creating airfoils was obtained through experts in airfoil design
at Rolls-Royce. These values measured the airfoil creation time from start to finish for a single
iteration. The participants were shown an airfoil part and asked to estimate the time to model the
geometry using traditional Rolls-Royce design methods. Once that time was recorded, they were
then asked to use the Airfoil Seed Part Generator to create the given airfoil geometry.
Participants were given a set of instructions to perform using the Airfoil Seed Part Generator
while a researcher measured the length of time to complete the test. Participants opened the seed
file and then launched the Airfoil Seed Part Generator. A text data file containing mathematical
airfoil design data from the engineering department was loaded and all options to create the
airfoil model to Rolls-Royce specification were selected. After the airfoil was generated, the
researcher recorded the time to successfully generate an airfoil and any errors or shortcomings
encountered in doing so. If the airfoil did not successfully generate on the first try, the time taken
to edit the airfoil and create a usable model was recorded. Data was collected in this manner for
thirty (n=30) airfoils. The time was recorded for each iteration, and a comparison was made
between the iterations as shown in Figures 6 and 7. In order to provide the most conservative
estimate of significance, values reported at 0 minutes were assumed to be 59 seconds and values
reported at 1 minute were assumed to be 119 seconds (1 minute 59 seconds).

Results

Data obtained from experts at Rolls-Royce that estimated the time to create an airfoil iteration
using current methods were found have a mean of x = 700 seconds (11 minutes, 40 seconds)
with a standard deviation of s = 458 seconds (7 minutes, 38 seconds). Time estimates ranged
from 300 seconds (5 minutes) to 1200 seconds (20 minutes). Time estimates were used because
Rolls-Royce did not keep a detailed record of airfoil creation time. Furthermore, exact times
could not be recorded during the project due to time and monetary considerations.

Data measured to create the airfoil times using the Airfoil Seed Part Generator method ranged
from 21 seconds to 442 seconds (7 minutes, 22 seconds). Lower times tended to appear when the
model successfully generated on the first attempt and higher times appeared as outliers when the
model failed to generate and a manual fix was applied. The mean for airfoil iteration creation
time was found to be x = 81.93 seconds with a standard deviation of s = 96.56 seconds.

It was hypothesized that the airfoil seed part generator would have a significant savings in time
over the current methods in use at Rolls-Royce. To test the hypothesis (Ha: µ > µ0), a one-sided t-
test was used, resulting in values of t = 2.337. This result shows levels of significance at α = .05
with a p-value between .02 and .01. These results are displayed in Table 1.

Table 1: Test Statistics

Airfoil Seed Part

Current Method
Generator
Mean ( x ) 700 sec. 81.93 sec.
Std. Dev. (s) 458 sec. 96.56 sec.

t-value = 2.337 Alpha (α) = .05 .01 ≤ p-value ≤ .02

Proceedings of The 2006 IJME - INTERTECH Conference

After removing outliers, the t-value increased to t = 2.432, however the p value remained
between .01 and .02 for a highly significant result. This suggests that there is a significant
savings in time by using the airfoil seed part generator as opposed to the current methods
employed at Rolls-Royce. The average time savings of airfoil iteration remodeling, based on
mean data from the control and test samples, was found to be 88.3%.

1400

1200
Time (seconds)

1000

800

600

400

200

0
Current Method Seed Part Generator

Figure 6: Box Plot of Test Data

Proceedings of The 2006 IJME - INTERTECH Conference

 Mean Time To Create Airfoil

800
Current Method
700

600
Time (seconds)

500

400

300

200
Seed Part
Generator
100

0
Method

Figure 7: Comparison of Estimated Time Using Old Methods and

Recorded Time with Seed Part Generator

Discussion

The results of this test suggest a significant time savings for Rolls-Royce when using the airfoil
seed part generator instead of the current methods. When the seed part generator failed to create
geometry correctly the first time, it was usually due to a fillet feature not correctly forming.
Making the model usable and generating usable geometry was easily done through a quick fix of
removing the fillet, generating the next airfoil iteration and replacing the fillet manually. This
same procedure is commonly done on models in the current Rolls-Royce airfoil modeling
method. The occurrence of a non-generating iteration was relatively low (3 out of n = 30
samples). The variation in airfoil modeling time appears to have been reduced as well (Figures 6
and 7). However further research is needed to determine whether this is true. It should also be
noted that because half (n =15) of the values from the test data that were measured in minutes,
these results were rounded up to the nearest minute to provide a conservative estimate. Even with
these conservative estimates, very high levels of significance were still found. In actuality, the
level of significance may be higher than the numbers reported.

In addition to the technical details of this project, curriculum implications for this project also
exist. Current constraint-based CAD tools all create geometry in a very similar fashion: a profile
is sketched on a 2D plane, dimensional and geometric constraints are added to the profile, and
the 3D form is applied to finalize the geometry [18]. Given this process, the students were able to
apply the basic modeling fundamentals and strategies they learned while using other CAD tools
and quickly adapt to the structure and syntax of UG NX 2. By having been exposed to these
modeling fundamentals in their course work, the students were able to quickly come up to speed

Proceedings of The 2006 IJME - INTERTECH Conference

on the NX 2 software used by Rolls-Royce. This is an example of the application of domain-
specific knowledge in solving the problems encountered in the typical design process – the
ability to apply strategic knowledge and usage of the CAD tool when engaged in a specific
context [19, 20, 21]. By distilling the CAD tool into its common elements, it is not necessary for
an academic institution to have the latest release of software, nor does it require them to have
expensive maintenance. Three of the four students that worked on this project had never used
NX 2 before this experience, but they were able to use existing knowledge of how CAD tools
work in order to be knowledgeable of the software in a matter three weeks and complete the bulk
of the project in a matter of about five weeks.

Anecdotally, this project was also beneficial to the students for gaining industrial experience on a
project with a tight timeline and significant deliverables. As an added incentive, it was suggested
at the beginning of the project that if the students met all of the major milestones of the project,
they would be taken to Rolls-Royce’s corporate office in the UK to present their results. While
this was a good reward for the students for a job well done, it also exhibits the nature of a
company that has invested itself in partnering with higher education to solve particular problems.
It also demonstrated the benefits of academic / industry partnerships in an area that does not
otherwise have many of these opportunities: CAD data management and engineering design
graphics.

Future Areas of Research

Several areas of the project have the potential to be further developed. The addition of several
basic features to modify the created geometry such as the ability to shift the airfoil sections and
the ability to tilt the airfoil would increase the value of the plug-in as a time-savings tool. These
operations normally require new airfoil definitions and remodeling. Further development of the
turbine core geometry would increase the associativity of this plug-in to the engineering data and
increase the speed at which thermal engineering changes can be made.

Knowledge Fusion also has the ability to interact with TeamCenter, the UGS PLM/PDM
software used at Rolls-Royce. Capability could be added to track design changes and read in
previous designs from the database. PDM integration would allow for airfoil designs to be shared
in real-time globally between all divisions of Rolls-Royce. Another area this plug-in could effect
is the engineering optimization process which uses various computerized analyses to refine the
airfoil design. This plug-in could allow batch analysis processing of these analyses.

References

[1] Korakianitis, T., & Pantazopoulos G. I. (1993). Improved turbine-blade design techniques
using 4th-order parametric-spline segments. Computer-Aided Design, Vol. 25(5), 289-299.

[2] Corral, Roque, & Pastor, Guillermo. (2004). Parametric design of turbomachinery airfoils
using highly differentiable splines. Journal of Propulsion and Power, Vol. 20(2), 335-343.

Proceedings of The 2006 IJME - INTERTECH Conference

[3] Rolls-Royce Corporation. (2005, August). NX2 knowledge fusion: Airfoil creation &
morphing – Initial trial. Indianapolis, IN: O’Brien, W., & Rasche J.

[4] Bermell, P. and Fan, I. (2002). A KBE System for the Design of Wind Tunnel Models Using
Reusable Knowledge Components. Paper presented at the VI International Congress on Project
Engineering. Barcelona, Spain.

[5] Prasad, Brian. (2005). What Distinguishes KBE from Automation. COE NewsNet. Retrieved
April 27, 2006 from http://www.coe.org/newsnet/Jun05/knowledge.cfm#1.

[6] Rosenfeld, Lawrence. (1995). Solid modeling and knowledge-based engineering. In D.

LaCourse (Eds.), Handbook of Solid Modeling (pp. 9.1-9.11). New York: McGraw-Hill.

[7] UGS Corporation. (2004). Knowledge fusion for users: 2004 PLM World. Retrieved on
February 15, 2005 from pdf file.

[8] Hunter, R., Rios, J., Perez, J.M., & Vizan, A. (2005). A functional approach for the
formalization of the fixture design process. International Journal of Machine Tools &
Manufacture, 46, 683-697.

[9] Clark, A. L. (2001). A solid modeling services architecture for KBE applications. ACM
Symposium on Solid Modeling and Applications: Proceedings of the sixth ACM symposium on
Solid Modeling and Applications. Retrieved February 9, 2006, from ACM Portal database.

[10] Sekiya, T., Tsumaya, A., Tomiyama, T. (1998). Classification of knowledge for generating
engineering models: A case study of model generation in finite element analysis. Research in
Artifacts, Center for Engineering, University of Tokyo, Japan.

[11] Carleton, S. (2005). Design rationale in a knowledge-based computer-aided design

environment. Unpublished master’s thesis. Purdue University, West Lafayette, IN.

[12] Susca, L.1, Mandorli, L.2, Rizzi, C1. (2000). How to represent intelligent components in a
product model. Department of Industrial Engineering, University of Parma1, Italy. Department of
Mechanical Engineering, University of Ancona2, Italy.

[13] Park, J., Park, S., Hwang, I., Moon, J., Yoon, Y., Kim, S. (2004). Optimal blade system
design of a new concept VTOL vehicle using the Departmental Computing Grid system. School
of Aerospace and Mechanical Engineering, Seoul National University, Seoul, Korea.

[14] Spitz, W., Golaszewski, R., Berardino, F., Johnson, J. (2001). Development cycle time
simulation for civil aircraft. NASA Langley Technical Report Server. Retrieved from ACM Portal
database, February 16, 2006.

[15] Phillips, R. E. (1997). Dynamic objects for engineering automation. Communications to the
ACM, 40(5). Retrieved February 9, 2006, from ACM Portal database.

Proceedings of The 2006 IJME - INTERTECH Conference

[16] Kok, Raymond. (2004). Knowledge Fusion Tips and Tricks. UGS Corporation

[17] Collaborative Product Development Associates, LLC. (2004). The intelligent master model
process for achieving functional design. Retrieved February 14, 2006, from http://www.cpd-
associates.com.

[18] Wiebe, E.N. (1999). 3-D constraint-based modeling: Finding common themes. The
Engineering Design Graphics Journal, 63(3), 15 - 31.

[19] Hartman, N.W. (2004). The development of expertise in the use of constraint-based CAD
tools: Examining practicing professionals. The Engineering Design Graphics Journal, 68(2), 14-
25.

[20] Hartman, N.W. (2005). Defining expertise in the use of constraint-based CAD tools by
examining practicing professional. The Engineering Design Graphics Journal, 69 (1), 6-15.

[21] Hartman, N.W. & Branoff, T.J. (2005). Learning theories: Applications for instruction in
constraint-based solid modeling and other engineering graphics topics. The Engineering Design
Graphics Journal, 69 (2), 6-15.

Biography

William O’Brien is an undergraduate student in the Computer Graphics Technology program at

Purdue University. He has been an intern with Roll-Royce.

Michael Chandonais graduated from the Computer Graphics Technology in the Spring 2006
semester. He is employed by Rand Worldwide as an Applications Specialist providing customer
support in the use of CAD technologies.

R. Scott Henderson graduated from the Computer Graphics Technology in the Spring 2006
semester. He is employed by Rand Worldwide as a Technical Training Specialist providing
customer training and technical education in the use of CAD technologies.

Mark Prezbindowski graduated from the Computer Graphics Technology in the Spring 2006
semester. He is employed in the interactive media field.

Nathan W. Hartman, Ed.D. is an assistant professor in the Computer Graphics Technology

department at Purdue University. His research areas include the development and definition of
expertise in the use of computer graphics tools, human spatial visualization abilities, and
defining computer graphics as a discipline.

Joseph A. Rasche is a Senior Engineer and Specialist for Rolls-Royce Corporation in

Indianapolis. His area of expertise is in structural analysis, and he also serves on the corporate
Engineering Methods group.

Proceedings of The 2006 IJME - INTERTECH Conference

Glossary

.dfa – (Design File ASCII) contains Knowledge Fusion classes and scripts

.dlg – (dialog) UI Styler components

Airfoil – generic term for a blade used on or in an aircraft engine

Airfoil Definition – text file which lists all the points associated with a specific airfoil design

Blade Book – airfoil definition file format used by Rolls-Royce U.S. turbine engineering
department

Computational Fluid Dynamics (CFD) – an engineering analysis tool that computes the
movement of a fluid (gas or liquid) as it interacts with a system

Compressor Blade – thinner airfoil used to compress and guide air at the front of a turbine
engine.

Engine Axis – the central axis of a turbine engine running along its length

Finite Element Analysis (FEA) – An engineering analysis tool that breaks a computer model
into many simply shaped finite elements. Mathematical equations can be run on the finite
elements for a total analysis of the entire object.

Flowpath – The path air moves through an engine along the engine axis. Bounded at the hub and
tip.

Hub – The section of the airfoil nearest to the engine axis

Inner Guide Vane (IGV) – airfoils which help to guide air into the compressor

Knowledge Fusion (KF) – “an interpreted object-oriented language that has been designed to
permit an end-user to easily add engineering knowledge to the task at hand...” (UGS NX3.0
documentation)

Leading Edge – edge that leads the airfoil through air (see Fig. 6).

Proceedings of The 2006 IJME - INTERTECH Conference

 Suction Side

Leading Edge
Pressure Side

Airflow
Trailing Edge

Figure 6. Compressor Airfoil Section Diagram

NX2/3 – versions of the Unigraphics CAD package used throughout the project.

Pressure Side – the side of the airfoil that creates pressure as it rotates (see Fig. 6)

Radial Axis/Direction – the axis perpendicular to the engine axis

RRAF (Rolls-Royce Airfoil) – a Rolls-Royce airfoil definition file format developed in the
United Kingdom

Rotor (blade) – rotating airfoil to either compress air (compressor) or be rotated by air (turbine).

Seed Part – a modeling part file with basic settings, information and possibly standard datums or
geometry already created to be used to start modeling a new part

Stator (blade) – static airfoil guides and controls airflow between rotors.

Stack – the position of airfoil sections along the radial direction to define the airfoil

Stalk – geometry that connects the airfoil to the disk.

Section – an outline of an area of an airfoil. Sections may be the result of planar or conical
intersection of an airfoil. Sections are stacked along the radial direction.

Suction Side – the area of airfoil opposite the pressure side. As this part rotates through the air it
leaves a gap creating suction. (see Fig. 6)

Tangency Points – the points on a section where the sides meet the edges (four per section)

Thermal Barrier Coat (TBC) – coating applied to the exterior of airfoils to prevent
overheating.

Tip – the section of an airfoil farthest from the engine axis

Turbine – a fan like structure that is driven by a fluid

Proceedings of The 2006 IJME - INTERTECH Conference

Turbine Blade – an airfoil on a turbine

Turbine Engine – An engine wherein compressed air is mixed with fuel and ignited to drive one
or more turbines. The turbine drives a shaft that powers the compressor and possibly a fan,
propeller or generator.

Trailing Edge – last section of an airfoil to move through the air as it is rotated (see Fig. 6)

XDC – airfoil definition file format used by Rolls-Royce U.S. compressor engineering
department.

Proceedings of The 2006 IJME - INTERTECH Conference