INTELLIGENT MANUFACTURING SYSTEMS

Dr. Surender Kumar

Professor & Head
Department of Production Eng.
B.I.T. Mesra, Ranchi

1. Introduction

The goal of intelligent manufacturing system is akin to any normal manufacturing

system satisfying customer needs at the most efficient level for lowest possible cost. The
involvement of computers as in the computer-integrated manufacturing has been for more
than 20 years; the incorporation of computer technology does not necessarily result in
intelligent manufacturing system. It is the introduction of human like decisions making
capabilities into the manufacturing system that makes it indeed intelligent. In particular
knowledge base systems have dominated the manufacturing landscape of late 1990’s
CIM was all rage in 1980s, FMS in 1970s,

Manufacturing technology will play an important role in future human

developments if it is based on new knowledge based applications. The impact of
knowledge-based system on manufacturers is already in front and all of us to see and
experience. Any organization must make full advantage of the knowledge at its disposal.
This goal translates into the effective use of knowledge ranging from design to
production and maintenance. To this end, the knowledge captured in disparate modules in
the organization must be liberated and directed synergistically to support integrated
systems for engineering and fabrication. There is a lack of awareness of tools and
techniques to be used among manufacturers. Manufacturer’s often use existing tools
rather than investigate benefits of modern intelligent manufacturing techniques.
Intelligent system vendors can help the situation by improving their shells to fit defined
requirements for successful manufacturing applications.

2. CONCEPT OF MANUFACTURING
Manufacturing is defined very broadly as the process by which material, labor, energy,
and equipment are brought together to produce a product having a greater value than the
sum of the material put in. This can be shown as a system, as indicated figure 1. Here the
input is shown as material labor, energy and capital. The capital input provides the
equipment and facilities required for combining the material, labor, and energy. Output
includes product, but there is always some undesirable output -waste and scrap—which
should not be forgotten. Also shown in Figure 1 are external influences that should not be
ignored. External influences can include government action, natural occurrences (e.g.,
storms, floods), and of course competition. Most people directly involved in
manufacturing do not normally consider steel mills or refineries or textile mills as ‘
manufacturing’. To most people the production of automobiles, airplanes, refrigerators
and electronics would be better accepted as manufacturing. This understanding of
manufacturing would include mechanical and electromechanical products, including all
components and parts. With the focus on mechanical components, metal forming and
cutting are central to many of these

operations, and those processes will be considered here to offer an appropriate view of
manufacturing.

Before delving into manufacturing technology and application issues, it is important to

have an understanding of the mission of manufacturing functions. What are the inherent
objectives of manufacturing? In a broad sense they are (1) production, (2) productivity
and (3) quality.

The first and most elementary objective of the manufacturing function is of course,
production. Implied is the simple objective to innovate and apply tools methods, and
processes using diverse engineering and scientific skills through which prototypes can be
made and then replicated as commercial products. The second objective is to attain
productivity efficiencies in order to minimize production costs. In this case economic
consideration and process value analysis dominate. The third objective of manufacturing
is to assure the highest product quality. Finally, a new objective cold be added –
flexibility, i.e., the ability to produce more than one product with high productivity and of
superior quality.

To attain these inherent objectives, a number of operational factors have to be considered.

The opportunities to coordinate and control them have to be planned. For a situation
requiring attention to technological change the operational input factors for
manufacturing can be summarized as attitudes, resources and technology.

Is it the rapid growth of computer products? or the result of new electronics technology?
or is it an effect of a new silicon semiconductor stemming from materials technology? or
perhaps it is the result of process technology, which has allowed for rapid chemical vapor
deposition of integrated circuits on a chip? Automated assembly of chips and
microprocessors into a personal computer may have had something to do with it. It all
blurs into advanced technology as far as the general public is concerned. In line with the
practical definition of manufacturing, to professional practitioners or those studying the
subject more deeply, manufacturing technology has distinct role and can given its own
definition.
Manufacturing technology is the embodiment and integration of needed element
of science and engineering into a functioning whole to convert product concepts and
specification into practical reality for the product’s in tended use. As a practical matter,
manufacturing technology, as distinct from chemical process technology, deals with
fabrication from available materials rather than synthesis of a new one. Manufacturing
generally refers to discrete parts and operations. However, as batch operation are handled
with greater automation, the overall effect begins to blend individual stages into a
continuous flow from input to output just as in a continuous fluid stream in chemical
process. Therefore , although batch process technology dominates, continuous –flow
concepts are becoming a part of the scope of manufacturing technology.
Manufacturing technology does not necessarily follow the scientific disciplines and
technology implied in the product being manufactured within the industry. An example is
that of the new microelectronics and semiconductor industry. Here electronic engineers
and mathematicians develop the product features, but the production process depends on
chemical engineers and material scientists.

3. DEVLOPING AREAS OF A MANUFACTURING SYSTEM

The treatment of the subject matter follows a general rule that technology that became
commercial within the past five years is probably still new to many, On the other hand,
technologies expected to be available within the next five years need attention.
Awareness of the yet- to-come may prepare for quicker technology transfer when this do
become current. The focus of the technologies appropriate for manufacturing is on
mechanical products and their components. However, a great deal of overlap is expected
with all manufacturing industries. A schematic outline of the topic, which requires
support and manufacturing development tools, is given in Figure.2.
For uniformity in presenting batch manufacturing technology features in this text, the
individual operation are divided into one primary and three secondary processes. The
principal process is supported by preprocess finishing operations, and the generally
grouped secondary processes, as shown in Figure.3. This break down of manufacturing
best fits the definition of manufacturing described earlier.
Within the context of development and adoption of new technology into the
manufacturing section, manufacturing technology must be considered in its broadest
context. In medium sized manufacturing companies, the focal point manufacturing
technology can be the production, sales engineering, quality control, development,
research, or other functions. Large industries are more likely to have an organized plan to
innovate manufacturing processes. Whatever the actual name of the technology
department – Development, Applied R&D, Advanced Engineering, Manufacturing
Technology, Manufacturing R&D justification is an issue. Stimulus can come from
technological argument or market situation.

Direct Computer Application

To Manufacturing
Factory date, parts, workflow organization

CIM & FMS

Hardware Innovations:
Controls & Material Handling
Numerical controls & Robotics

Materials Selection & Processing

Metals & Plastics, principal and
secondary processes, cutting tools

Quality Assurance:
Quality concepts, inspection techniques
, Data handling procedure

Decision and Design tools:

Simulation and AI

Figure.2. Outline of the various topics, which requires intelligent manufacturing

Development tools
 Secondary Processes
· Heat Treating
· Welding
· Chemical Treatment

Principal Processes
Preprocesses
Finishing
· Casting Metal Cutting · Assembly
· Forging and/or Forming
· Painting
· Sizing
· Packaging
· Cleaning

Figure.3 Extended View of machining

4. INTELLIGENT MANUFACTURING SYSTEM

An intelligent Manufacturing process has the ability to self regulate and/or self
control to manufacture the product within the design specifications. An integrated
concept with factories of future, where products are produced in an artificial life
environment adds value to this. Researchers working on implementation of EXPERT
SYSTEM,, have come out with this concept of adding intelligence. On a experimental
basis it has been tried out on computer aided drafting system at University of Texas.
Recent developments in feature based solid modeling techniques for design
representation and AI applications have shown a potential to substitute EXPERT
SYSTEMS for some decisions currently handled by a designers knowledge of design
rules and practices.
The first effort in this endeavor through the Knowledge Based Systems, and expert
systems were not sufficient enough to generate the required level of intelligence owing to
their dependence on symbolic representation of the knowledge and the human expertise
required to encode them efficiently. However the advent of ANN in 80’s as self
organizing dynamic systems and model free estimators, emergence of Fuzzy Logic
Systems for modeling of uncertainty and human like reasoning and the acceptance of GA
as global optimization technique for hard problems has led to the realization of Intelligent
Manufacturing system as a feasible proposition in the current decade.

Intelligent manufacturing can be achieved in three basic ways.

™ Existing manufacturing processes can become intelligent by monitoring and

controlling the state of the manufacturing machine.
 ™ Existing processes can be made intelligent by adding sensors to monitor and
control the state of product being processed.

™ New processes can be intelligently designed to produce parts of desired quality

without the need of sensing and control of the process.

Intelligent Manufacturing system

1. Uses technology which can minimize the use of human Brain
2. Regulation for product mix and priority production, self regulated.
3. Self controlled operations with automatic feed back mechanism.
4. Monitoring and control of the manufacturing machine.
5. Monitoring and controlling the state of product being processed.
6. New processes with intelligence can be made to produce parts of
desired quality with out the need of sensing and control of process.

5. STAGES OF DEVELOPMENT OF SUBSYSTEMS OF IMS

Subsystems Stages Tools used

1. Design Stage Development in progress CAD modeling tools

i.e. Auto CAD
2. Prototype Stage Developed according to demand CASE tools with
CBR.
3. Procurement Stage Developed with on live Any vendor
technologies to cater for JIT interactive software
with vendor certification.

4. Process Stage Scheduling tried out with Genetic algorithms

. Genetic algorithms * Case tools.
. case based reasoning * Fuzzy composite
logic tools.
Artificial intelligence
5. Machining tried out, in progressive
CENTRE stage of development.

6. Material handling Developed with incorporation

Systems of AGVs, guidance
of AGVs and sensors being
developed
7. Storage System Developed with inventory
Expert systems.
8. Marketing Developed according to Fuzzy logic +
inputs from marketing simulation algorithms
parameters i.e. demand
6. Tools used for intelligent manufacturing

Following are the tools generally used in intelligent manufacturing:

™ Fuzzy logic
™ Genetic Algorithms
™ Neural Net Works
™ Case tools
™ Simulation Algorithms.
™ AI

a) SCHEDULING WITH NEURAL NETWORK

Artificial Neural Networks aimed towards the modeling of networks of real neurons (
i.e. animate brains) motivated by their robustness, fault tolerance, flexibility and the
learning ability of the biological brains, the building blocks of the ANN variously
called as neurons, processing elements etc. do have striking similarities with their
biological counterpart.
An interactive tool for short term production scheduling with system features and
graphical interface that allows operator to interactively control the schedule generation
and see influence on key parameter.
A constraint management subsystem checks that the current schedule agrees with
the restriction imposed from the production environments and if not, the reason for
conflict is identified and presented to planner for rectification. This knowledge base
system is able to effectively react to unexpected events or delays. Neural network
techniques come into play, repairing an inconsistent schedule towards a consistent or
an optimal schedule.

b) CASE BASED REASONING

This is extensively used as a tool to interpret the process parameters. Human experts
need to be consulted to interpret analytical results of the process,. Providing an
opportunity to improve on the procedure with an expert system. Case based reasoning
represent knowledge as ‘cases’ i.e. examples of past problems and solutions.

C) INTELLIGENT DESIGN AND ANALYSIS SOFTWARE

The expert system produce4s optimal design parameters without violating any material
or machine constraints. This is actually an extension of CBR (case based reasoning)

d) FUZZY LOGIC
Fuzzy logic system provides a means of expressing the linguistic variable in suitable
form for processing using a Computer Fuzzy logic Control of processes offer
flexibility by which process states and control actions can be described directly from
the experience and advice of the human operators, thus making it possible to apply
practical operational experience in the computerized control of complex multi-variant
process. It provides a mathematical framework to capture the uncertainties associated
 with human cognitive systems such as thinking and reasoning. The control rule is
formulated as linguistic expressions involving every words like High medium, low etc.

e) GENETIC ALGORITHMS
GA is powerful probabilistic heuristic procedure for global search and optimization in
multi-parameter search spaces, based on the mechanics natural genetics. It is to exploit
historical information to locate new points in the search space with expected improved
performance.
This is also used as a tool for optimal assembly planning. Genetic algorithms referred
to as an ADAPTIVE COMPUTATION are based on the evolutionary concept of
natural selection and survival of fittest.
In simple terms, genetic algorithm generates new rules to replace the least useful rules
already in place. These software tools allow user to solve complex problems, such as
scheduling large number of conflicting tasks, finding the shortest route that connect a
number of location, or streamlining communications network.
The genetic algorithms are used to optimize the search routine used in assembly
planning with the goal of improving the assembly process of mechanical product
thereby minimizing time and cost.

f) ARTIFICIAL INTELLIGENCE
This tool is an attempt to increase the number of human characteristics, computer and
computer-controlled systems. It is basically an ability to imitate the human intelligence.
The sub tools used are

™ An algorithm is a computer program that solves selected problem within given

time frame.

™ Early vision Computer calculations that allow systems to see by providing low-
level data.

™ Higher level vision Computer calculation that allows systems to accomplish higher
level such as smart improvement within an environment object recognition and
reasoning about objects.

™ Knowledge Engineering. A process by which knowledge is collected from experts

7. IMAPCT OF ARTIFICIAL INTELLIGENT IN

MANUFACTURING
The inventions of automation and artificial intelligence have brought about a
technological revolution in the operator-machine system relationship. These technologies
have made it possible to reduce safety risks and to implement cognitive capabilities, such
as speech understanding, logic deduction, picture understanding, reasoning about expert
knowledge, problem solving, and decision making in machines.
With automation most of the physical human work is transferred to machines. Optimal
controllers are developed to take over skilled-based activities. The human role shifts from
the doer to that of a monitor and problem-solver, awkward postures of the human body,
excessive manual force, and high rates of manual repetition are less impacting.
In recent years, numerous articles have appeared in the trade press that discusses the
manufacturing application of AI. The quantity of information alone has been
overwhelming, and the fact that it involves concepts with constantly changing parameters
has created even more confusion. The popularity of AI underscores the need for a better
definition of the discipline. The problem of defining AI is its multidisciplinary origins,
with no single fundamental concept having clearly set the initial limits. One definition
that often appears states that AI encompasses any technology that involves making a
machine behave in way that would be considered intelligent human behavior. Of course,
this requires a further definition of what is meant by “intelligence” Another definition
state that AI is any technology that enables a computer to respond to situations that were
not anticipated by its programmer
Artificial Intelligence (AI) is the science of making machines do things that would
require intelligence if done by humans. Techniques such as expert systems, fuzzy logic,
and artificial neural networks are quite common. AI technology offers the potential to
transfer rule-based and knowledge based functions from humans to machines and
computers. Examples of these functions are monitoring the state of the machine and
environment; if operational conditions of the machine change, self-adapting mechanisms
enter to modify goals; when goals change a set of suitable plans is developed to react and
alternative plans are evaluated; after simulating the effect of alternate plans the optimal
solution is implemented.
A description of an intelligent operator-machine system is outlined in Figure.4 Several
points are key to a successful realization of such systems.

1. Reliable symbolic representation schemes must be developed.

2. Appropriate levels of operator alertness must exist.
3. When the system works in the automated mode, the operator must stay in the
loop to minimize reaction time for take-over in case of system failure.
4. The operator must be able to intervene at all three levels of interaction (skill-rule
and knowledge-base) to serve as a backup in case of failure of the automation
system.
5. The display and interface system must be laid out to ease communication between
worker and machine on the three levels of interaction.
 Goals Goals

Symbol’s Knowledge-base activities

Intell
-igent Artificial
Signs Intelligence
Ruled-based activities
Inter
-face
Signals Optimal
Skill – based activities Controller

Operator

Interface
Sensors Actuators

Tool / workstation / machine

Figure.4. Description of an intelligent operator-machine system

8. USE AND APPLICATION OF AI IN SPECIFIC MANUFACTURING AREAS

The use of AI is increasing very rapidly in manufacturing. Manufacturing will change
more in next 15 years than it has in the last 75, largely due to the application of
computer-aided technology that has been developed during the last 10 yr. The tools of
this technology include:
™ Computer-aided design( CAD)
™ Generative process planning
™ Robotics / material handling
™ Computer-aided manufacturing ( CAM )
™ Automated Standards
™ Manufacturing resources planning
™ Group technology
AI is indicated as the technology that will tie these tools together. The generic application
of AI, including the following items directly related to manufacturing:
™ Fault diagnosis and repair (machines and systems)
™ Operation of machines and complex systems
™ Management (Planning, scheduling, and monitoring)
™ Design (systems, equipment, intelligent design aids, and inventing)
™ Visual perception and guidance (inspection, identification, verification, guidance,
screening, and monitoring)
™ Engineering (chemical and biological synthesis planning, and intelligent design aids)
™ Industrial (factory management, production planning and scheduling, intelligent
robots, process planning, intelligent machines, and computer aided inspection)

AI Technology Application

Expert system Design

Maintenance
Process control
Monitoring
alarm analysis
Equipment diagnosis
Process planning
Scheduling

Machine Vision Inspection

Identification
Measurement

Robotics Welding
Material handling
Part positioning
Assembly
Spray painting

Natural language Database information retrieval

understanding

Voice recognition Data entry

Inventory control
Quality Inspection
NC programming
Robotics

Speech synthesis Control room alarms

These references provide a general view of how AI can be applied.

Some of the specific areas are explained below:

a) AI in process planning

The process-planning system consists of two components

1) a knowledge base and
2) an inference engine.
Declarative knowledge (facts) is represented in the form of frames, and procedure rules
are represented in the form of production. The fact in the knowledge base may be
broadly classified as follows:
(1) Facts about workpieces and machined surfaces,
(2) Facts about machine tools,
(3) Facts about machining operations, and
(4) Facts about tooling.

The rules in the knowledge base may be categorized as follows:

(1) Rules for identifying machined surfaces on a workpiece,
(2) Rules for consistency checking of tolerancing,
(3) Rules for selecting operations,
(4) Rules for sequencing operations,
(5) Rules for selecting machines,
(6) Rules for selecting tools,
(7) Rules for selecting machining parameters, and
(8) Rules for adding new rules in the knowledge base.

Basically Turbo-CAPP employs a backward-chaining inference mechanism for

(1) collecting candidate rules (the conflict set),
(2) Scheduling the rule-firing sequence,
(3) Selecting an appropriate rule,
(4) Processing command(s), and
(5) Modifying the knowledge base accordingly.
Both sequential and parallel processing of candidate rules are employed in Turbo-CAPP
(Wang, 1986).
In the process of creating process plans and NC code, the system must acquire knowledge
from the user from time to time. Several modules performing knowledge acquisition are
(1) the tolerances input module (TIM). Which is used to acquire geometric dimensioning
and tolerancing data for the planned part; (2) the machine-description module (MDM),
which is used to acquire machine-dependent facts; and (3) the process-manipulation
module (PMM), which allows the user to add new process (es) into and/or delete existing
process (es) from the knowledge base.
Turbo-CAPP is currently implemented in PROLOG to run on an IBM PC Based on
various input conditions, the geometric features, the machine capabilities, available tools,
workpiece, and tool material data, etc., Turbo-CAPP generates reasonably good process
plans for the machining of symmetric rotational workpieces. The system allows the user
to manipulate the knowledge base. The change is then reflected in the planning results.
Several rotational parts have already been process planned (including NC tape
generation). The process typically takes 2 to 5 minutes per part of planning time (Wang
and Sysk, 1987).
Their are several other AI-based process planning systems.

™ CMPP (Sack, 1982), which was developed at United Technologies Research Center,
is process-planning system for planning cylindrical parts (although some
noncylindrical features are allowed). In addition to generating process plans, CMPP
presents a unique feature. It performs dimension, tolerance, and stock-removal
analysis based on a sophisticated algorithm with the objective of optimizing
tolerance capabilities of shop equipment. Moreover, it also selects appropriate
dimension reference for each machining operation. Perhaps the only limitation of
CMPP is the geometric coverage-cylindrical parts.
™ AMPS (Chang et al., 1988; Kanumury et al., 1988) is an automatic process-planning
system for prismatic parts. It was developed as part of an integrated design-
manufacturing-inspection system called QTC (quick turnaround cell). AMPS utilizes
a feature-based-design solid model as input. A feature-refinement module analyzes
the design model and reasons the final machining features and their precedence
relationships. Setup planning and fixturing method planning are part of the process
planner. The end result of the system is the preparation of a detailed process plan and
part programs for the designed part. No human decision-making is necessary for the
entire planning process. In AMPS, the feature-refinement module converts the solid-
model information into a symbolic representation in a frame structure. Process
knowledge is represented in rules. As ES (expert system) shell, KEE, a computer
program, is used for the implementation.

b) AI in manufacturing planning and control

Manufacturing planning and control involves managerial decision making on three main
levels of activities. Decisions at
(1) the strategic level of production and inventory planning and control are really crucial
and should be made with utmost care.
(2) At the tactical level of manufacturing operations planning and control, the batching
and loading problems are the central focus.
(3) The operational level of process and material flow planning and control is concerned
with the detailed decision making required for real-time operation of an automated
manufacturing facility. The dynamic nature of the manufacturing environment stems
from the highly variable nature of factors such as customer orders influencing the
strategic level, material and capacity constraints in the tactical level, and machine or
workstation availability in the operational level. In order to aid manufacturing
managers in achieving productivity goals, various concepts of information systems
and computer control have between developed.
Traditionally, the manufacturing environment has been controlled by process controllers,
shop-floor computer controllers, and other computer aids for various planning tasks.
Although computer technology can rapidly process large quantities of data using
sophisticated logic, many of the necessary decisions must wait until human operators can
sift through the data, become familiar with the system status, and select proper actions.
AI techniques can provide better planning and control and a higher productivity.
Many AI-based systems have been developed to deal with various planning and control
problems in manufacturing systems. In this section, several systems are categorized and
described in the following: AI in part design, AI in process design, and AI in process
execution.

c) AI in part design
Engineering design may be defined as a process in which scientific principles technical
information, and creativity are all combined in order to produce an optimum and product
that will serve its intended purpose. Design engineering involves a number of distinctive
phases, beginning with the definition of particular problem and ending with the selection
of an optimum solution. Various approaches to engineering design have produced
different decompositions of this process. Commonly, engineering design is broken into
three main phases, as illustrated in Figure.5.

Problem Synthesis
Definition
Analysis

Evaluation &
optimization End
Product

Figure.5. Decomposition of engineering design.

Synthesis (preliminary design) is an essential feature of all design work; this task deals
with the formation of design alternatives. Synthesis involves searching and checking of
subsystems. The end result of this phase is the selection of one or, at most a few
preliminary design alternatives that satisfy the key constraints of the particular problem.
During the analysis phase a selected design alternative is studied using mathematical and
other scientific procedures in an effort to determine the suitability of the design to the
intended environment. Important aspects of this phase are the selection of the proper
analysis procedures, the correct use of these procedures, and the appropriate
interpretation of the results.
The evaluation and optimization phase of the engineering design process involves the
evaluation of the analyzed design alternative. At this point, backtracking and repetition of
previous design phases are often required to produce a feasible, acceptable, or optimal
design solution to the specified problem.
With the introduction of the computer as a powerful tool for the engineering design
process, attention has shifted from pursuing study on design methodology to the
development of software to aid engineers in the design process. Traditionally, these aids
have been limited to the well-structured aspects of design such as ANALYSIS and
graphics. Advancements in AI research and the subsequent emergence of expert systems
provide a new powerful tool for the development of computer programs that can be used
as aids for the solution of ill-structured phases of the engineering design process-
preliminary design.
The first step in part design is a conceptual one dealing with synthesis, which is ill
structured. Here the basic ideas of the product from the design standpoint are generated
and tested. The second step in part design is a detailed one, involving analysis. Each part-
design concept is placed under the engineering microscope for detailed consideration.

Their are several intelligent simulation software’s for part design some of them are as
follows:

™ The I-DEAS drafting and solid modeling program (from S D R C U.S.A)

The I-DEAS software is an integrated drafting, solid modeling, design and
manufacturing solution, which provides you with greatest design flexibility.

™ The ANSYS Finite Element Analysis ( FEA) program ( From ANSYS Inc. U.S.A.)
The ANSYS program from ANSYS provides the full range of engineering design
analysis and optimization capabilities- solid modeling, preprocessing, solution, post
processing, graphics and design optimization features. The ANSYS program enable
users to generate innovative design options by helping determine all aspects of design
: displacements , stresses , forces , temperature distribution, magnetic field strength ,
fluid flow, pressure distribution , material saving and other important design
considerations.

™ The FLOTRAN Computational Fluid Dynamics program.

The FLOTRAN Computational Fluid Dynamics ( C F D) program is a powerful
engineering tool that enables users to solve fluid flow and heat transfer problems.

™ The ADAMS mechanical simulation and prototyping program (From MDI U.S.A)
ADAMS enables the user to predict the behavior of mechanical system undergoing
large displacement motion. ADAMS empowers you to perform four important types
of analysis: Kinematics, Static, Quasistatics and Non-linear and linear Dynamics.

™ The LS-DYNA program (From Livermore Software Technology Corp. U.S.A)

LS_DYNA is a general purpose, explicit FEA program used to analyze the nonlinear
dynamic response of three-dimensional inelastic structures. The various application
areas of LS-DYNA are – Crashworthiness, Metal forming, Metal cutting, Penetration,
Glass Forming, Plastics forming etc.
™ The DATAMINE mining program (from DATAMINE U.K)
The DATAMINE is typically used in data capture and analysis, exploration, geology,
geochemistry, surveying, ore body modeling, underground and open pit mine design
and production planning and also in related areas such as environment studies.

™ The MENTOR GRAPHICS electronic design program

The MENTOR GRAPHICS Corp. (MGC) is the world’s leading provident of
Electronic Design and Automation tools. The MG software are available for all
application in electronic design PCB, IC, ASICFPGA, DSP, Mixed Analog Digital
Design, Air flow, Heat transfer and so on.

™ The MOSS Infrastructure development program (from MOSS SYSTEMS Ltd. U.K)
The MOSS software can be used extensive for feasibility study, planning and design
of basic infrastructures like highways, tunnels, bridge and railways.

™ The MARc K7
The MARC K7 allows the user to perform a wide variety of structural, fluid and
coupled analyses using the finite element method. These procedures provide solutions
for simple to complex linear and nonlinear engineering problems.

™ DEFORM systems include Automatic Mesh Generation (AMG)

DEFORM systems include Automatic Mesh Generation (AMG) which allows users
to specify remeshing. Criteria and optimization criteria for an intelligent mesh, based
on solution behavior. The AMG system is optimized to meet the severe requirements
of large deformation modeling. Forming equipment models are coupled into the
simulation engine. The energy loss of a hammer, screw press or power limits of
hydraulic press can be included in the analysis.
Corner unfill, load requirements and die pressure are available output. Graphical
output includes deformed mesh (to demonstrate material flow), nodal velocities (to
quantify local material flow at any time during the process) and field variables, which
are easily available. Graphics can be exported to word processors or other programs
for reports or presentations.
DEFORM include sophisticated animation creation and reply capability, which
allows users to graphically demonstrate material flow and the change in field variable
in a “real time display format.

™ MSC / Superforge
MSC / Superforge accepts geometry data from all major CAD system providing a
seamless transition from CAD to analysis.
The Graphical User Interface speaks the language of the process engineer, bridging
the gap between the shop floor and the engineering analyst process characteristics
such as metal flow, die filling, flash and tooling loads can be predicated by computer
simulation, then passed along to the shop floor for recommended process changes.
With its finite volume technology, MSC / Super Forge has been successfully used in
industry to simulate many 3-D forging processes which are beyond the scope of finite
 element based forging simulation technology. These processes include the multi-stage
forming of complex 3-d parts such as crankshafts, connecting rods, valve craps and
axles.

d) AI in Process Design

The process-design stage exhibits a high degree of similarity to the part-design stage
except that the focus of attention is on designing processes by which to manufacture parts
rather than designing the parts themselves.

In the context of AI-based process design, the following systems have been developed

™ The Mdonnell Douglas Research Laboratory is investigating AI-based parts to

assist process designers. An example can be taken from the study of the metal-
joining problem. When a part designer has decided a weld is required, a process
designer must specify the details of a welding-process plan. This plan is based on
decisions about the welding process to be used., heat treatment requirements, joint
preparation before welding, cleanup after welding, and parts jigging, to mention but
a few.

™ Embodies a constraint-directed reasoning approach to job-shop scheduling in the

process-design stage of manufacturing. The system has been developed and
implemented by Fox and co-workers at Carnegic-Mellon University (CMU) (Fox,
1981), ISIS was developed for factory scheduling. The system uses a variety of
representation and search techniques for reasoning with constraints. ISIS provides a
general solution to the shop-scheduling problem and has been developed within the
environment of the Westinghouse turbine component plant.

™ In another project at CMU, Fox and co-workers have developed a knowledge-based

system (KBS) capable of interactively modeling and simulating factory organizations
(Fox, 1984). The system includes a number of features aimed at producing
simulations that are easier to build and that produce more understandable results than
currently available simulators.

™ Wysk et al. (1986) presented an AI-based control and scheduling system for FMS.
The system is able to (1) generated potential scheduling alternatives based on real-
time shop information and scheduling, (2) update performance rules based
“simulation experience,” and (3) affect the control on a variety of FMSs.

e) AI in Process Execution

The first step in process execution is one in which raw materials and tooling are brought
together with manufacturing equipment such as NC machines to produce parts. This step
also includes finishing the parts, inspecting and/or testing, and placing the parts in storage
until required. The second step in process execution includes bringing together parts and
tooling along with other manufacturing facilities such as robots to assemble products.
Again, finishing, inspecting, testing, and storing may be required.

Several AI-based systems for the process-execution problems of manufacturing

have been reported:

™ Fox and coworkers at CMU have developed an AI system, PDS, for the online
diagnosis of malfunctions in machine processes (Fox, 1984). The operation of the
system is based on information acquired from tens or hundreds of sensors attached to
the machine(s). Although this is the only wait to obtain large amounts of timely
information, it introduces further problems since no sensor is error-free and sensor
performance degrades over time. Therefore, PDS reasons about the sources of its
information in addition to performing diagnosis based on the information.

™ Gini et al., (1983) have proposed CAMA (computer –assisted maintenance) as a

knowledge-based consultation system for automatic maintenance and repair. Further
efforts by the same researchers have been aimed at automatic error recovery in robot
programs operating

9. STEPS FOR IMPLEMENTING AN INTELLIGENT SYSTEM

SOLUTION
Steps for implementing an intelligent system solution in manufacturing are as follows:

™ Select an existing software that will first augment an expert system and then integrate
with task solution software.

™ Collect trade standards, heuristic rules, past experiences and recommendation from
books, experts and standard codes.

™ Organize a complete set of rules governing this task, and describe them in a formal
way acceptable to a list processing software.

™ Generate a list of features and other part attributes called for in the rule set.

™ Generate relevant data by extracting the features and other necessary part attributes
from the model data of an object.

™ Organize the rules governing this task into a knowledge base.

™ Implement an inference engine as a reasoning mechanism to find a solution on the

knowledge base and the external data.

™ Incorporate the results into the model and its representation to obtain a final solution.
10. CONCLUDING REMARKS

As mentioned previously, the application of AI includes (1) problem solving (2) logical
reasoning, (3) natural language processing, (4) automatic program generation, (5)
robotics, (6) computer vision, and (7) expert systems. Many, if not all of these
applications are required in manufacturing. It is believed that manufacturing will become
a dominant customer of AI technology.
Although AI has shown great potential in manufacturing applications and will definitely
continue to be a focal point for man researchers, several critical issues remain to be
resolved: (1) Flexible knowledge representation: an important issue in designing an AI
application is the selection of a suitable knowledge representation scheme. Current AI
systems, with only a few exceptions, employ either production rules or frames. Any real-
world problem, especially in manufacturing, requires a more complicated and effective
knowledge-representation stricter. This may, imply a demand of efficiently combining
the strengths of two or more existing schemes into a powerful one. (2) Learning
procedure: several AI systems are capable of learning new knowledge. However, the
learning procedure used are very limited. For example, newly learned knowledge may be
contradictory to the knowledge already in knowledge base because of the lack of a
consistency checking mechanism. (3) Handling of uncertainty: uncertainty is a norm
rather than an exception of real world. When one tries to model a real world problem, the
inclusion of uncertainty variable already exist, their weaknesses are obvious. For
example, it is well known that membership functions are the essential element in a fuzzy
set application. However, deciding a membership function so that the function can
precisely describe the problem being modeled requires much research effort. Before a
generalized scheme for constructing the membership function is available, the fuzzy set
application remain difficult.
It is not clear what application will dominate in the near future, although process control
appears a likely candidate, process control applications may be more mature at this point,
due to the possibility of making small meaningful test cases.
During the next few years many problems must still be faced in the application of AI to
manufacturing. Included in the list are:
1. Understanding how to pick problems that are within the capabilities of the available
tools and experience.
2. Real –time issues which will surface as bugs when current tools are applied in
increasingly stressful situations.
3. Integration issue, both with in-place software and with the total factory organization.
In general, more must be learned about AI by manufacturing engineers, and more must be
learned about manufacturing requirements by AI practitioners. Manufacturers should not
jump into using an AI / ES approach instead of conventional techniques without carefully
considering the trade – off. There will be cases that will be better handled with
conventional techniques.
Usually, these will be such that a solution can be expressed as an algorithm. It may be
possible, however, that the flexibility of the expert system approach will be the most
import consideration.
Bibliography.

1. Deb S.R and Chattopadhayaya,” Proc. 3rd SERC school on ‘Advanced Mfg.
Tech.’ ”, Jadavpur Calcutta, 1997.
2. Kumar Surender,” Industrial Robots and CIM”, Oxford and IBH Publication Co,
New Delhi.
3. Kumar Surender and Jha, A.K, ”Computer Aided Design and Manufacturing”,
Dhanpat Rai and Co. Ltd., New Delhi.