Computer Integrated Manufacturing Systems 1994 7(4) 269-278
Design for automating the inspection
of manufacturing parts
Michael D Reimann and Joseph Sarkis
College of Business Administration, The University of Texas at Arlington, Information Systems and
Management Sciences, Box 19437, Arlington, TX 76019~9437, USA
The integration of quality into management of advanced manufacturing systems has not
typically included automated approaches. With the advent of automated inspection and quality
tools such as coordinate measuring machines and other dimensional measuring equipment, the
linkage with integrated manufacturing systems seems natural. This paper shows the functional
relationship of computer aided process planning to automated inspection process planning. The
series of functional activities within each of these frameworks are mapped in this paper. The
operational environment to include computer aided quality control in an integrated manufacturing system is also defined.
Keywords: automated quality control, CAPP, quality information systems, inspection process planning
Introduction
Computer aided process planning (CAPP) is used in
computer integrated manufacturing (CIM) environments to automate the linkage between the design and
manufacturing phases of product development. A
general approach for CAPP developed by the
Consortium for Advanced Manufacturing - International (CAM-I) is described in this paper. The CAMI Advanced Numerical Control (ANC)' processor
specifies the structure and defines the functions necessary for automating the process planning of machined
parismatic parts. The ANC processor also provides the
foundation for developing an architecture for automating of inspection process planning. First, the overall
structure and functional components of the ANC
processor are described. The next section delineates a
comparable framework for automating inspection
process planning. The framework for automated
inspection process planning is compared to the ANC
processor. External components required for the
inspection process planning framework are presented.
The final section summarizes the main issues identified
in this paper and draws conclusions relative to practical
implementations.
The ANC processor
The objective of the Advanced Numerical Control
(ANC) processor is to automatically generate the
necessary numerical control (NC) instruction sequence
and supporting information for producing a finished
part. In theory, all that should be required from the
ANC user would be the identification of a part which
exists in an external part model database. From the
part definition, a complete process plan to machine a
finished part is derived for the removal of volume from
an initial raw stock. Thus, the ANC processor is
intended to:
• provide a structure which facilitates automation of
NC processing functions,
• bring order to the NC planning and programming
function,
• provide user-friendly tools for ANC functions which
are not fully automated.
The structure and functional components of the ANC
processor are illustrated by the IDEF0 model shown in
Figure 1. Each of the functions in this framework will
be described in the remainder of this section.
Obtain operation from the process plan
This function obtains basic information about a
machined part from the ANC user. Data that is
requested from the user by this function includes:
• identification of the part model before NC machine
operation,
• identification of the part model after NC machine
operation,
• specification of a NC machine tool to be used
(optional).
This information is used by subsequent functions in the
ANC processor. The part identification numbers are
used by the Material Decomposition function to
0951-5240/94/04/0269-10 (~) 1994 Butterworth-Heinemann Ltd 269
, a~qtanN Z ~utnl°A su2ats,~S ~u~..tntooJnut~lq paw.ta~tUl .talnazao,9
OL~
~a
t
G
'
li~--
~
,=,1
Z
i:-J0
i
P~
|.
i
i
a
iIT-~
Modifications from N C P ] m R e v i e ~
~~
q~lavS f p u v uumul.a~t G l q :slaDa p d a m o v j n u m u j o uol.loaasu~, attl a u l . m t u o m v
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
retrieve various geometric models of the part and
determine delta volumes necessary to produce intermediate and final part shapes. Identification of the NC
machine tool is conveyed to the Machine Tool and NC
Tactics function, where it is used to determine machine
tool characteristics and select an appropriate machine
tool.
Decompose material to be removed
Basic operational process plan information is generated
by this function to establish models and delta volumes
for intermediate part geometries. The technique of
volume decomposition is employed to remove individual (delta) volumes of material from a model of
stock material to produce an intermediate or final part
model.
The Material Decomposition function uses 'before'
and 'after' part model geometries which are made
available to the ANC processor from an external part
design database. The raw stock, from which the part is
produced, is not modelled in the geometric part
database. Specification for raw stock must be obtained
from a Stock Definition library, which is external to the
ANC. Various volume decomposition algorithms may
be used by this function. These algorithms are also
external to the ANC processor. Geometric models for
the delta volume geometries are developed by an
external geometric modeller.
As subsequent operations are performed in the
ANC processor, adjustments in the results developed
by this function may be needed. These requests for
modifications are fed back as inputs to the Material
Decomposition function as they occur.
Determine machine tool and NC methods
In this function an appropriate machine tool and NC
method to remove each delta volume is determined. A
simplifying assumption made for this function is that
only one machine tool may be used for machining of
the part. The machine tool can be optionally specified by
the user when the model was initially identified. If the
machine tool is predetermined, suitable NC methods
for each delta volume are established by this function.
When the machine tool is not pre-specified by the user,
this function will first determine an appropriate NC
method for each delta volume. Once the NC methods
for all delta volumes are established, the most suitable
machine tool is then selected.
The design for this function allows for either a
manual or automatic mode for establishing machine
tool and NC methods for the part machining process.
This could potentially require the user to examine and
assign NC methods for every delta volume. An intelligent decision tool would be needed to automate the
decision process for determining an appropriate
machine tool and NC methods. Such a decision aid
would be external to the ANC processor.
Subsequent ANC processor operations may require
adjustments to the results developed by this function.
These requests for modifications are fed back as inputs
to this Machine Tool and NC Methods function as they
occur.
Determine setup
The necessary setup characteristics for the part machining process are established by this function. Some of
the specific issues that are addressed by this function
include:
• Orientation of the Part Information from the
previous functions is used to develop appropriate
orientations for the part. Models of the part
geometries are obtained by the Material Decomposition function. These models are used to
determine the physical characteristics of the part that
influence fixturing and clamping of the part.
Constraints imposed by physical characteristics of the
part, the NC methods employed and the selected
machine tool are used to determine suitable part
orientations. A geometric modeller is used to
construct orientations for the part, from which
desired features can be machined. These orientations
are compared to machine tool capabilities to ensure
compatibility with the machine tool.
• Fixture design The geometry of the part, delta
volume removal requirements and part orientation
are used to choose appropriate fixtures for holding
the part. Features of the machined part are compared
with the fixture feature capabilities of available
fixtures. Based on this comparison, a suitable fixture
is selected.
• Clamp design and selection Clamp design and selection uses part orientation and fixture information to
choose clamps for holding the part. Candidate
surfaces for clamping are compared to clamps
defined in a clamp table. Clamp selection can be
performed manually or by an automated decision
process. Clamp selection data is fed back to the
fixture design sub-function to ensure compatibility
between fixture and clamp selection.
• Setup evaluation All support information and determinations made up to this time are presented to the
user for review and modification. Information that
can be reviewed includes:
intermediate geometric part models
- NC methods
machine tool identification
part orientation
fixturing
- clamping.
-
-
-
-
Based on this review, the user can initiate modifications to previously obtained results.
Determine cutting tool and homer
This function establishes the appropriate sub-delta
material volumes, suitable cutting tool and cutting tool
holder for the removal of each sub-delta volume. Delta
volumes are used to define the required volume to be
Computer Integrated Manufacturing Systems Volume 7 Number 4 271
Automating the inspection o f manufactured parts: M D Reimann and J Sarkis
removed from a 'before' part model to produce a
feature on an 'after' part model. The actual types of
machining steps (work elements) are defined by a
specific NC method. Based on the total delta volume
and the corresponding NC method, sub-delta volumes
for each work element are established. Information
obtained from previous ANC processor functions
which are used by this function includes:
•
•
•
•
individual delta volumes
NC methods for producing the delta volumes
manufacturing parameters and standards
machine tool setup data.
Cutting tool selection requires sub-delta volume and
NC method information to choose appropriate cutting
tools for removing each desired sub-delta volume.
Available cutting tools are specified in a cutting tool
library. Individual cutting tools are compared against
each required sub-delta volume to make a choice.
Tool holder selection uses setup and cutting tool
information to choose the holder. The cutting tool and
machine tool are compared with available holders.
Compatibility must exist between the machine tool,
cutting tool and cutting tool holder to result in a holder
selection. Special tool holders may be supported by this
function.
Feedback from the NC Plan Review and Tool Path
Simulation function may necessitate adjustments in this
function.
Detail~optimize N C p l a n
The preparation of the detailed NC plan is carried out
by this function. Many of the previously determined
NC production characteristics are used by this function
to carry out its objective. This information may
include:
• individual delta volumes for material removal from
the before part model to produce the after part
model;
• corresponding NC methods required to remove the
delta volumes to produce the desired features;
• machine tool identification;
• part orientations, fixture and clamp specifications;
• sub-delta removal volumes;
• cutting tool and holder specifications.
This function is comprised of the following subfunctions:
o f strategy Identifies the overall
strategy the NC process will employ to produce the
part. Possible strategies may include the minimization of tool path travel lengths, minimization of the
number of fixture changes, minimization of the
number of part re-orientations, or minimization of
the number of cutting tool and holder changes. A
number of decision steps are performed to determine
the most appropriate strategy to be used by the NC
Plan Generation sub-function.
1. Identification
272 Computer Integrated Manufacturing Systems Volume 7 Number 4
2. N C plan generation The chosen strategy is used to
sequence the work elements in the NC plan.
Information provided about each work element to
this sub-function includes setup data, NC method,
cutting tool and cutting tool holder specifications.
The result produced by this sub-function is a NC
plan within which work elements have been ordered
based on a designated hierarchy of strategies.
3. N C tactics The ANC user may review each work
element in the NC plan through graphical techniques. NC tactics for the removal of sub-delta
volumes can be altered by the ANC user through
direct interaction with the system. Seven stages are
associated with each NC tactic:
• A p p r o a c h - movement from a clearance plane to
an operative geometric position.
• Enstage - movement from an operative geometric
position to initial contact with part material.
• E n t r y - movement from an initial part material
contact to full part material contact.
• M a c h i n i n g - movement through the continuous
engagement with part material.
- movement away from full part
material contact to exit from material.
• Destage - movement away from final contact with
part material to an operative geometric position.
• Retract - movement from an operative geometric
position to a clearance plane.
• Departure
The ANC user accesses the geometric modeller to
determine the cutting tool movements. A tactics
table, which includes coordinate points and other
pertinent machine tool process information, is
developed by this sub-function.
4. R e v i e w o f N C p l a n Once the NC plan and associated
tactics have been determined, the user can review
the part generations with the graphical interface.
The geometric modeller can be used to display the
position of the cutting tool and holder as they move
in relation to the stationary part. Detection of
proper coverage and collision avoidance can be
observed with the NC Plan Review sub-function. In
addition, interference with fixtures and clamps may
be examined. Because this sub-function can detect
conflict situations, it will produce feedback to
previously executed functions in order to rectify any
problems.
Generate~simulate tool path
The objective of this function is to generate tool path
commands for each NC work element. Tactics records
from the NC plan are used as an input to this function.
Tactics specify the NC steps and characteristics which
are necessary to produce the sub-delta volumes.
Information in the tactics record is formatted for
internal use by the ANC processor. A tool path
generator sub-function converts a two-dimensional
representation into a three-dimensional parametric
equivalent. Graphical simulation of this approach
enables the user view the tool path motion. The
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
simulation uses the geometric modeller to produce the
visual images.
Produce machine control information
This function generates the appropriate Cutter Language (CL) instructions which will be used by a NC
machine tool to machine the part. The tool part
information that was generated in the previous function
is converted into compatible CL data. This translation
may use various techniques that exist in the research
literature, many of which utilize various forms of expert
systems. Once these basic instructions are determined,
graphical simulation will be used to represent the NC
machining process. The generic CL instructions are
then translated to machine dependent NC code.
Produce support information
The last function simply provides a user-friendly
interface, for the machine tool operator, to obtain
information from the ANC processor. Components of
this function include:
• generation of operator instructions
• visual aids
• material orders.
The finalized NC plan and associated tactics form the
basis for information that is presented to the operator.
This information is translated from an internal representation to an operator understandable form.
A N C research issues
A number of research opportunities can be identified in
the ANC processor. Many of the functions which
comprise this framework rely on some form of selection
and sequencing decision process. Appropriate tools
presently exist that would address many of these
decision processes from an automation perspective.
Special algorithms and heuristics can be tested to
determine which would be most efficient. Additional
research opportunities exist in regard to optimization of
these decisions. This area is fertile for further development of each of the ANC steps, as well as applying
these generic stages to various automated planning for
machining, assembly, inspection and material handling
processes. How the ANC process functions can be
utilized to develop the instructions and plans for an
automated inspection process planning system are
described in the next section.
Automated inspection process planning framework
using advanced numerical control concepts
The ANC processor provides a general framework for
the automated development of process plans for
numerically controlled machining equipment. This
framework can be used as a foundation to automate
other processes within a CIM environment. Competition in manufacturing environments is increasingly
becoming time-based. To improve competitive posture,
many organizations are adopting a total quality orientation. Automation of quality functions in a CIM
environment is an integral component in philosophies
such as total quality management (TQM) and quality
function deployment (QFD). Research in integrating
automated inspection equipment, such as dimensional
measuring equipment (DME) or coordinate measuring
machines (CMM), with other components of CIM has
received limited attention until recently2m.
This section serves two purposes. First, it demonstrates that the ANC framework is generalizable and
can be used to develop other components in a CIM
environment. It also describes a framework which is
oriented towards quality functions and part inspection.
This framework is referred to as the Expert Programming System-One (EPS-1) processor 7. As we shall
see in a later section, the EPS-1 is part of a larger
operational environment that integrates inspection
process planning with the remaining components of a
CIM system.
EPS-I architecture
The EPS-1 is a functional subset of the ANC processor.
As such, the EPS-1 embodies many of the same
characteristics as the ANC. Certain characteristics that
are part of the ANC are not required in the EPS-1
framework. Only the differences between ANC and
EPS-1 processors will be described in this paper. The
main differences are primarily due to the underlying
purpose of each processing framework. Whereas the
ANC processor is primarily for machining purposes, the
EPS-1 architecture is not a material transformation
process, but an inspection process. Thus, many of the
data and tool characteristics in the EPS-I environment
do not require material removal that are characteristic
of the ANC processor.
The EPS-1 requires a part identification number and
an overall process plan as inputs. The identification
number is used to retrieve a geometric model and a
dimensioning and tolerance model of a part. The DME
NC plan generator uses the overall plan to produce
dimensional measuring interface specification (DMIS)
control data and support information for a specific
DME. Figure 2 shows the relationships among nine
functional modules that form the EPS-1 architecture.
The remainder of this section describes each of these
nine functional modules.
Obtain operation plan
The EPS-1 user provides required data to the EPS-1 for
initiating the automated inspection process. This function determines the scope of the inspection process to
be performed on the indicated part. Data needed by
the EPS-1 includes:
• a valid part identification number
• an inspection type
Computer Integrated Manufacturing Systems Volume 7 Number 4 273
,.,.1
~.¢ Idcmdficafion
Inmp¢~io~Type
I DME C l ~ V l l ~ h i m ~
3"
Mo~r
I
~"~
Datglwe Inter.MachiKLibruyDME
D'a",,,l~ee
I
T"'¢"
I e~
II Deeompo~itionT
1A'211 Featut~
1
c~
I
NC Medaod~
Module
t
11
Operato~
In~etio~s
111
..
,q
.~~,
4~
11 i _ ,
,~o,,..o. I .o.
T1
Probe Holde~
Module Module
iI
Ope~tionPI~.6
DME NC Plans wilh Tactic~
1
MachineTool
IX.pmdmtCon~oKs)
ProbeP~hA.? I ProbePath
~ I D ' Module
r
Gel~elzic
t
11
l~w*merric Probe ] Produce Con~l ]
PareC,~m'ator Ii
I
li~os'mliou
A.8
DMlS Co~rol D~*
I
DMIS D m
lvll~c~i~
I Produce Support ]
Formatter
Co~a,ol
]
Module
hafoemstiou
Vi~d Aim
A.9I o,l-o
VisualAid Gee.mot
Figure 2 High level IDEF0 diagram for EPS-I framwork
IC
•
IT
,p~
m
.
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
• a D M E class or machine (operational)
• an inspection strategy (optional).
The inspection type establishes the scope of the
inspection operation (i.e. courtesy inspection, inprocess inspection, final inspection, etc.). The part
identification number and the scope of the operation
are used to determine the correct model in the
geometric model database and the corresponding
model in the D & T database. The output from this
function includes work element identifiers that are
passed on to the Task Decomposition function, and the
type of D M E to be utilized, which is used by the
Determine NC Methods and D M E function.
The EPS-1 must obtain the scope of the D M E
operation from the EPS-1 user, whereas the A N C
obtained a machining process plan. Because the EPS-1
does not remove material from a part, it only requires
the identification of a single part model. On the other
hand, the A N C process requires the existence of before
and after part models.
Task decomposition
The Task Decomposition function determines the
evaluated dimension and tolerance ( E D T ) features that
are to be measured by the inspection process. The EPS-1
automatically determines these from information
obtained by the Obtain Operational Plan function. The
part identification and the inspection type uniquely
identify the geometric and D & T models of the part. An
entire D & T model is obtained through subroutine calls
to the D & T modeler.
The D & T model is used by the Task Decomposition
function to construct a skeletal list of the work elements
necessary to perform the inspection process. For each
node in the D & T database, an individual work element
item is constructed. Subsequent functions in the EPS-1
use this information to determine other inspection
process requirements and characteristics.
The EPS-1 only requires one part model identification, and there is no product transformation feedback
required. The output data representations also vary,
whereas the EPS-1 data is a list of E D T features for the
part to be inspected, the A N C outputs geometric
models of the current state of the part and derived delta
volumes. Another major difference is the lack of the
need for a raw stock model in the EPS-I.
process initialization, selection of a D M E will be
bypassed. In this case, the NC measurement methods
required for the inspection process are compared with
those available from the specified D M E , to ensure that
the D M E is capable of taking the required measurements. If a D M E was not specified, then the EPS-1
compares the capabilities of available D M E against the
NC measurement methods required for the inspection
process in order to choose a DME.
The requirements for the ANC and EPS-1 processors
are similar for this function. Different methods of
obtaining and transforming data are required, since the
characteristics of the equipment in each of the
processes will utilize different internal representations
of the data.
Determine setup
Part setup generally applies to the part orientation and
fixture and clamp selection in a NC environment. To
simplify the complexity of this process, fixtures and
clamps were not incorporated into the initial specification of the EPS-I. Part orientation is accomplished by
alignment of part surface normals with the D M E
measurement axes. Three potentially conflicting objectives are considered:
• maximize the number of measurements in any given
orientation
• minimize the number of orientations to achieve the
required inspections
• eliminate possible collisions between the D M E and
the part due to improper orientation.
The above strategy requires the geometric modeiler to
perform boolean algebra intersections on the part and
probe models so as to detect potential collisions.
In the initial version of EPS-1, fixtures and clamps
are not required. Fixtures and clamps must be used in
the A N C environment.
Determine probe~holder
Combinations of measurement probes and holders for
the inspection process are determined by this function.
Selection of a probe can become a complex task when
the following factors are taken into consideration:
Determine methods and D M E
•
•
•
•
D M E technologies
probe relative orientation
multiple probe setups
probe shape.
This function serves two purposes: first, it determines
the appropriate NC measurement methods for the
inspection process; then it identifies the most appropriate D M E for taking the actual measurements. For
the purposes of the EPS-1, the combination of tolerance class and sub-feature class is used to establish
individual NC measurement methods. Once the
methods have been determined, a D M E machine can
be chosen. If a D M E machine was specified during
Due to this complexity, candidate probes must be
identified through a process of elimination. The
capabilities of the available probes are compared to
required measurement characteristics of the inspection
process. A probe is determined for each sub-feature on
the part. Several probes may be required to measure
the entire part. The strategy at this point may be to
select probes based on minimizing the number of
probes for the entire inspection process. A n o t h e r probe
Computer Integrated Manufacturing Systems Volume 7 Number 4 275
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
selection strategy might be to maximize the number of
measurements obtained from the individual probes.
There are four basic differences between ANC and
EPS-1 processors relative to this function:
• The ANC sub-delta volume generator is replaced by
the EDT sub-feature generator in the EPS-1.
• The EPS-1 design does not require the user to supply
manufacturing parameters.
• The EPS-1 must select an appropriate probe instead
of a cutting tool.
• The EPS-1 does not determine a probe holder.
Detail~optimize operation plan
Inspection process information that was developed by
previous functions is used by this function to complete
the operation plan. The EPS-1 finalizes the sequence of
work elements and determines the corresponding
movement tactics for the inspection process. Three
activities are performed by this function:
• NC plan generation
• determination of NC tactics
• review and modification of the NC plan.
Sequencing of the individual work element items is
determined by this function. The NC strategy previously
specified by the user forms the basis for the ordering
process. On the basis of this strategy, an appropriate
set of logic is used to establish the specific sequencing
of the work elements. The types of logic used depends
on the sequencing strategy. Mixtures of heuristic,
optimizing and decision table logic may be employed.
Relative positioning of the probe is also determined for
every sub-feature measurement to be made.
All the processing up to this point has been automatically carried out by the EPS-1. The work elements,
NC plan and NC tactics are now complete, and the user
can intervene to make any adjustments. The user may
review a graphical representation of the inspection
process sequence. This will indicate potential coverage
and collision problems. The user can make the necessary
corrections to eliminate any problems. Simple changes
can be made by the user without any further intervention
by the EPS-1.
The primary difference between the ANC and EPS-1
processors, for this function, is in the internal representation of the data and the decision making approach
employed. These will differ due to the need for more
data items in the ANC processor.
Generate~simulate probe path
This function serves two purposes: to generate locational and parametric probe path data; and to visually
simulate that data. The primary input to this function is
the NC plan with tactics, and the primary output is an
internal representation of the probe path data.
The NC plan with tactics contains an ordered list of
the work elements required to inspect the identified
EDT features. The work elements are processed
276 ComputerIntegrated Manufacturing Systems Volume 7 Number 4
sequentially, and contain the sub-feature mean, the
sub-features class "and the tolerance class for each
entry.
The measurement process involves the generation of
initial and departing probe positions and parametric
probe path motion. Positioning of the probe for
measurement is determined by generic routines that are
geometry-specific. EDT sub-feature class/tolerance
class combinations dictate the appropriate routine. The
routines use specific D&T information to establish
actual probe position.
The parametric probe path data is mathematically
determined for each measurement. The probe path
data is composed of two curves that are parametrically
synchronized. One of the curves represents the probe
end, while the other represents the probe axis. Once
the parametric probe path is created, it is used to
graphically illustrate the movement of the probe in
relation to the stationary part. The user can review the
movements of the probe over the part to observe
coverage and detect collisions.
The type of algorithms required by the ANC and
EPS-1 processors for this function will necessarily have
to differ. For example, there will need to be a greater
diversity in the path patterns that are required for
machining when compared to measurement operations.
Generation of the probe path for taking dimensional
measurements will incorporate different characteristics
from that of a cutter tool path.
Produce control information
Numerically controlled DME inspection instructions
are generated by this function. These instructions
conform to the dimensional measuring interface
specification (DMIS) 8. The NC plan with tactics
provides the necessary tolerancing information. The
coordinates of the actual points for every movement
and measurement are retrieved from an internal representation of the probe path. All of this information is
combined to produce DMIS instructions. This function
also allows DME independent control information to
be interspersed with the DMIS instructions.
The main difference in this function between the
ANC and EPS-1 processors is the type of output data.
DMIS data will be output by the EPS-1, while APT
compatible CL data will be output by the ANC
processor.
Produce support information
Information that supports the actual NC inspection
operation is developed by this function. It is defined in
such a way that it can be tailored to support the specific
environment of an individual implementation. Information for a DME operator is generated, as well as for
other users indirectly involved in the NC inspection
operations. These individuals may include quality
assurance engineers, shop supervisors, foremen or tool
crib operators.
The main difference between the ANC and EPS-1
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
ances for the geometric features of a part 1°. Many
existing geometric modellers do not have the capability
to produce the dimensioning and tolerancing information that is needed by inspection applications. Thus, a
Dimension and Tolerancing (D&T) modeller 11 was
designed and developed to fill this void. The D&T
modeller defines various tolerance nodes, and assigns
these nodes to one or more geometric features. After
geometric features have been added to a part model,
the D&T modeller must be used to augment the
geometries with specific dimensioning and tolerancing
data.
processors for this function is that the EPS-1 will
generate less support information.
Operational environment of the EPS-I
The EPS-I interacts with various external components,
as shown in Figure 3. The main components in this
operational environment include a geometric modeller,
the Dimensioning and Tolerancing (D&T) modeller,
the Applications Interface Specification (AIS), the
Dimensioning and Tolerancing Applications Interface
Specification (DTAIS), the Dimensional Measuring
Interface Specification (DMIS), and support databases.
All of these components play an integral role in the
EPS-1 framework, either as inputs, controls or mechanisms for processing. Each of these components will be
described in the remainder of this section.
Applications interface specification
Due to unique capabilities that are offered by various
geometric modellers, an inspection application may
require the use of several modellers. To avoid the
necessity for developing unique interfaces for each
combination of application and modeller, a standard
interface similar to IGES has been developed. The
Geometric Modeling Program of CAM-I has defined
the application interface specification (AIS) to satisfy
this need.
The AIS is a dynamic interface between modellers
and applications software. Modeilers and applications
that have been designed and implemented to conform
to the AIS can interface directly with one another. The
AIS defines a core set of capabilities that should be
available from any geometric modeller. If a particular
modeller has the capability to perform a specific
function, then the interface acts only as a communications media. On the other hand, when the function is
not supported by a modeiler, the application interface
would be required to perform that function.
Geometric modeller
A geometric modeller is used to delineate, identify and
define the physical characteristics of a part. Various
geometric entities can be described by their boundaries.
The geometric modeller facilitates certain functions
that are specific in inspection. In particular, collision
between probe and part can be detected by performing
a boolean intersection on geometric models of the part
and probe. Even though the geometric modeller9
specified for use in the EPS-I is limited to planes,
cylinders, quadratics and cubics, the EPS-1 framework
can accomodate other three dimensional objects.
Dimensioning and tolerance modeller
Inspection of physical objects by DME and CMMs
requires specific data about the dimensions and toler-
Known DME Machine Specifleation
User Supplied
Part Identification
Expert Programming
System - One
0EPS-1)
QuaXty
Plans
User Support
Information
DME NC Plan
Generator
I
i
Dimensioning and
Tolerance Modeler
Geometric
Modeler
•
Figure 3 Operational
ment of the EPS-I
I
Dimensioning & Tolerancing
Applications Interface
Specirgatlon
Applications Interface
Specification
I
I
]
SC
environI
|
Computer Integrated Manufacturing Systems Volume 7 Number 4 277
Automating the inspection of manufactured parts: M D Reimann and J Sarkis
Dimensioning and tolerance AIS
The AIS allows a user to retrieve geometric data about
a part, and also enables applications to utilize the
functional capabilities of the geometric modeller. The
D&T model expands the geometric model of the part to
include data that is necessary for dimensioning and
tolerancing. Unfortunately, the AIS does not accommodate dimensioning and tolerancing data, nor does it
enable external applications to access the functional
capabilities of the D&T modeller. Therefore, it was
necessary to extend the AIS so that it could create,
read, modify and delete nodes in the D&T model. The
dimensioning and tolerancing applications interface
specification (DTAIS) 12 was created to satisfy this
need. The DTAIS is similar in concept to the AIS in
that it is used to construct the D&T model.
Dimensional measuring interface specification
There are three classes of interface for linking CAD
systems with DME13:
• One-to-one interfaces require users to develop their
own communication linkages that work with specific
devices in their system
• One-to-many interfaces allow one CAD system to
interface with numerous DME and CMMs
• Many-to-many interfaces allow the user to select the
most appropriate CAD systems and DME which
satisfy their particular needs.
The many-to-many interface is the most desirable
because it provides the greatest flexibility and has the
highest potential for integration of inspection processes.
the DMIS s'14"t5 was recently adopted by ANSI as a
national standard for many-to-many interfaces. DMIS
provides the linkage between the EPS-1 and DME, as
shown in Figure 3.
Summary and conclusions
In this paper, we have presented a review of the
advanced numerical control (ANC) processor developed
by CAM-I's ANC Program. It was shown how the
functional modules of an ANC processor could be used
to develop a processor for automated quality and
inspection control called Expert Programming SystemOne (EPS-1). To be competitive, manufacturers must
produce their products to meet customer requirements
in a timely fashion at a competitive price. The EPS-1 is
an essential component for achieving these objectives
in integrated manufacturing environments. The linkage
of the EPS-1 in an operational environment was also
described.
A description of how the EPS-1 processor would be
implemented to inspect representative parts is presented
by Reimann and Sarkis 16. Efforts addressing the issue
of automating and integrating the inspection planning
process have only recently emerged. This is not only
278
Computer Integrated Manufacturing Systems Volume 7 Number 4
due to competitive pressures, but the increased technological availability of various DMEs, specifically
coordinate measuring machines which have shown
increasing benefits 17.
Future research should focus on extending the ANC
to include other aspects of automated planning within
manufacturing environments. So far, only the framework for the ANC has been developed; operationalizing
many of these functional modules is required. Development of decision tools to help in the automation of
these steps is also needed. Numerous research possibilities in determining appropriate algorithms and
knowledge bases for these decision tools exist for both
the ANC and EPS-1.
Acknowledgements
This work was supported by a grant from the Consortium
for Advanced Manufacturing International.
References
1 CAM-lAdvancedNumericalControlConceptualDesign-Volume
I Project Notebook & Status Report, Computer Aided
Manufacturing-International, Inc., Arlington, TX (1979)
2 Brown, C W 'IPPEX: An automated planning system for
dimensional inspection', Proc. 22nd CIRP Int. Seminar on
Manuf. Syst., Enschede, Netherlands (June 11 1990)
3 EIMaraghy, H A and Gu, P H 'Expert system for inspection
planning', Ann. CIRP, Vol 36 No 1 (1987) pp 85-89
4 Menq, C H, Yau, H T and Lai, G Y 'Automated precision
measurement of surface profile in CAD-directed inspection',
IEEE Trans. Robotics Association, Vol 8 No 2 (April 1992)
pp 268-278
5 Menq, C H, Yau, H T and Wong, C L 'An intelligent planning
environment for automated dimensional inspection using coordinate measuring machines', J. Eng. Ind., Vol 114 (May 1992)
pp 222-230
6 Reimann, M D and Fowler, J W 'The expert programming
system-one (EPS-I]', Proc. Fourth Int. Conf. State-of-the-Art on
Solids Modeling, Boston, MA (May 1987)
7 The Expert Programming System-One (EPS-1), Task One Design
Review, Computer Aided Manufacturing-International, Inc.,
Arlington, TX, R-87-ANC-01 (1987)
8 ANSl/CAMllOl-1990DimensionalMeasuringlnterfaceSpecification, Computer Aided Manufacturing International, Arlington,
TX (1990)
9 CAM-I Test Bed Modeler, Computer Aided ManufacturingInternational, Inc., Arlington, TX (1985)
10 ANSI Y14.5M, Dimensioning and Tolerancing, The American
Society of Mechanical Engineers, ANSI Y14.5M (1982)
11 CAM-I D& T Modeler Vision 1.0 - Dimensioning and Tolerancing
Feasibility Demonstration Final Report, Computer Aided
Manufacturing-International, Inc., Arlington, TX (November
1986)
12 Johnson, R H Dimensioning and Tolerancing Final Report,
Computer Aided Manufacturing-International, Inc., Arlington,
TX R-84-GM02.2 (May 1985)
13 Zink, J H 'Closing the CIM loop with CMM', Automation, Vo136
No 1 (1989) pp 48-50
14 Schreiber, R R 'CMMs: traits, trends, triumphs', Manuf. Eng.,
Vol 104 No 4 (April 1990) pp 31-37
15 Stovicer, D 'CMMs key to plant with a future', Automation,
Vol 37 No 12 (December 1990) pp 24-25
16 Reimann, M D and Sarkis, J 'An architecture for inegrated
automated quality-control', J. Manuf. Syst., Vol 12 No 4 (1993)
pp 341-355
17 Smith, L 'Quality too: CMM smooths mold design', Am.
Machinist, Vol 137 No 4 (April 1993) pp 26-28