Robotica (1985) volume 3, pp 205-214

FMS in CIM
(Flexible Manufacturing Systems in
Computer Integrated Manufacturing)
Paul G. Ranky
Industrial Technology Institute, College of Engineering, University of Michigan, P.O. Box 1485, Ann Arbor,
Michigan 48106 (USA.)
(Previously with Trent Polytechnic, Nottingham, UK)
(Received: May 7, 1985)

Computer Integrated Manufacture (CIM) is concerned * designing and analysing components for machining,
with providing computer assistance, control and high inspection, assembly and all other processes relating to
level integrated automation at all levels of the manufac- the nature of the component and/or product, (i.e.
turing industries, including the business data processing welding, cutting, laser manufacturing, presswork,
system, CAD, CAM and FMS, by linking islands of painting, etc.),
automation into a distributed processing system. The * evaluating and/or determining batch sizes, manufactur-
technology applied in CIM makes intensive use of distri- ing capacity, scheduling and control strategies relating
buted computer networks and data processing tech- to the design and fabrication processes involved in the
niques, Artificial Intelligence and Database Management particular product.
Systems. FMS in this aspect plays the role of a highly * analysis and feedback of certain selected parameters
efficient and "ready to react to random requests" man- relating to the manufacturing processes, evaluation of
ufacturing facility (e.g. machining, test, assembly, weld- status reports from the DNC (Direct Numerical Con-
ing, etc.) in this "total concept". (Figure 1) trol) system, (source data monitoring and machine
CIM technology offers a range of benefits to managers function monitoring in real-time) and
too, when taking decisions. In such cases computing and * analysing system disturbances and economic factors of
human decision-making skills cannot be ignored either, the total system.
but
* if the decision-makers in the organization have up-to- It is quite obvious from this list that the aim of CIM is
date information on the requested processes and data, to let the advanced information processing technology
* if there is the possibility of controlling and analysing penetrate into all areas of the manufacturing industry in
large amount of business and technical data by compu- order to:
ters, * make the total process more productive and efficient.
* if design and production engineering are integrated; in * increase product reliability and
other words, if the designers can obtain guidance and * decrease the cost of production and maintenance relat-
feedback from databases, knowledge bases of expert ing both to the manufacturing system as well as to the
systems, from the shop-floor and from other levels product.
(e.g. production engineering) of the organisation, on
how to solve certain design tasks to suit the manufac-
turing facilities best, LEARN FROM PAST BUSINESS
* if manufacturing in its broad sense (i.e. not only FAILURES/SUCCESSES

machining, but also inspection, test, assembly, etc.) is

done mainly by Flexible Manufacturing Systems pro-
ducing the necessary workparts on order rather than
for stock
then the company will have a better chance for survival LEARN FROM PAST PRODUCTION
PLANNING AND CONTROL PROBLEMS
in the age of CIM.
Computer Integrated Manufacturing, in this respect,
covers all activities related to the manufacturing business,
including: LEARN FROM PAST
MANUFACTURING PROBLEMS
* evaluating and developing different product strategies,
* analysing markets and generating forecasts, Fig. 1. Computer Integrated Manufacturing System (CIM) con-
cept integrating four major modules. Distributed database man-
* analysing product/market characteristics and generat- agement and communication as well as expert learning and
ing concepts of possible manufacturing systems (i.e. decision-making systems are important modules of this "total
FMS cells and FMS systems). concept".

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 206 FMS in CIM

A SYSTEM ENTRY
POINT FIXED BUDGET
ASSETS ACCOUNTING

SALES ORDER
PROJECT INDUSTRIAL MARKETING AND PROCESSING SALES NOMINAL
ENGINEERING ENGINEERING FORECASTING LEDGER LEDGER

DESIGN
I
PROCESS
I
SALES
ESTIMATING PURCHASE
MASTER
ENGINEERING PLANNING SCHEDULING LEDGER

1
NC MACHINE,
INVENTORY
ACCOUNTING

ROBOT,ETC.
PROGRAMMING CONFIGURATION M.R.P.
CONTROL
INVENTORY
MANAGEMENT
COMPUTER
AIDED QUALITY BILL OF
MATERIAL (BOM)
-TIME STUDIES
•ROUTING,ETC. z_» >
OTHER ANALYSIS..
1.,
ENGINEERING
DESIGN DATA BASE

8 FMS CONTROL
-TOOL DATA
-FIXTURt DATA
-PART PROGRAMS.
ETC.

Fig. 2. An overall CIM model highlighting FMS and its interrelationship with other module of the system.
PLANT
MAINTENANCE

* reduce the number of hazardous jobs and increase the tory, since by integrating them into CIM one can:
involvement of well educated and able humans in the * increase productivity (often by a factor of 2-3.5),
manufacturing activity and design. * decrease production cost (often by 50%).
In comparison with FMS, CIM is mainly concerned * manufacture (i.e. not only machine, but test, assemble,
with the information processing tasks at all levels of the weld, paint, package, etc.) single parts and/or batches
factory and its management, whereas FMS provides the in random order, i.e. on order, rather than on stock,
essential computer controlled manufacturing tools and * decrease inventory and work in progress (WIP) to a
systems for CIM to execute the computer generated lower level than ever before,
plans and schedules that take account of a total system * provide 100% inspection, thus increasing the quality of
rather than just one cell or shop. One could also say that the product,
CIM integrates "FMS islands" with the overall computer * decrease the amount of often repetitive, or hazardous
network, the business system, with different design and physical work and increase the need for intelligent,
manufacturing databases of the company, and allows the human work, and
optimisation of the data flow and eventually all activities * provide a reprogrammable, often almost entirely un-
at a much broader level than FMS. manned manufacturing facility (again underlined not
Figure 2 illustrates the CIM concept by showing some only for machining, but for many other processes and
of the most important modules of the business system, for a wide range of products).
CAD (Computer-Aided Design), CAM (Computer-Aided The manufacturing industry and all those involved in
Manufacture) and FMS. It also indicates the most impor- its research and development have tried hard to imple-
tant data paths and the macro level relationship between ment cost-performance effective manufacturing facilities
the indicated modules and subsystems. in the past.
In the 1960s NC machines reached a stage when they
FLEXIBLE MANUFACTUING SYSTEMS (FMS) became reasonably reliable and productive. Computers
Flexible Manufacturing Systems offer the most fascinat- and machine controllers were developed further, to-
ing production method for the computer controlled fac- gether with production concepts such as "group technol-

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 FMS in CIM 207

ogy". By the mid 1960s the first DNC (Direct Numerical

Control) system had emerged in Japan and the first in
1973 in Europe (in Hungary). Their control systems have
become more powerful along with the development of
tool, material handling and transportation systems (e.g.
automated tool changing systems, industrial robots, Au-
tomated Guided Vehicles (AGV), tool magazines, pallet
pools and pallet changing devices).
The first major step towards FMS was made in 1975
when the first NC machining centre successfully operated
unmanned, utilising an automatic tool changing (ATC)
system as well as a 10 station pallet pool and automatic * • • •

pallet changing (APC) facility.

Since then, but particularly in the 1980s, several
hundred different FMS systems have been installed
world-wide. Despite the fact that FMS has emerged from
machining and the traditional machine tool industry, this
flexible production method and know-how is invading
every other field of manufacturing, including assembly,
test and inspection, welding, packaging and many other
production and other fields. Fig. 3. Implementation phase of an FMS designed by Dedel.
Considering that the number of people employed by (Note the pallet transfer system design).
the manufacturing industry is decreasing because of auto-
mation and because most young people tend to prefer
the service industries where the speed of profit-making 2. Link these cells into a system by providing preferably
seems to be faster, and that by the year 2000 the direct access, or random material handling systems
manufacturing industry is going to replace its current (e.g. AGV), rather than serial access (e.g. conveyor
production methods and equipment, FMS technology is line) between them.
of extreme importance. 3. Create a part, tool and pallet (fixture and clamping
device) storage facility (i.e. warehouse), and
FMS SYSTEM ARCHITECTURE 4. Provide high level computer control "inside" and
To be able to achieve "truly" random production and the "outside" the system, based on a distributed proces-
above listed economic benefits, the FMS system architec- sing system, on databases and on the necessary links
ture must conform to the following basic rules: to other subsystems, such as CAD, CAM and the
1. Utilise highly automated and programmable cells (i.e. business system.
CNC machines, robots, etc.) capable of "taking care 5. Ensure that if any cells break down the production
of themselves" (i.e. incorporating powerful controllers planning and control system can reroute and re-
and self diagnostic systems), able of changing their schedule the production. In other words, design in the
tools and parts, preferably unmanned, and keeping in system not only programmable (i.e. flexible) produc-
touch with a central computer or with a node control- tion facilities, but also ensure that the routing of parts
ler from where the production plan, the part programs can be dynamically altered.
and further necessary data arrive and are fed back. One should realise that FMS is not only applicable to

Fig. 4. Trumpfflexiblesheet metal processing FMS.

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 208 FMS in CIM

machining, but is a universal flexible production method

which is relatively easy to control, schedule, operate and
maintain if the above outlined system design rules are
followed.
It should be mentioned that there is some level of
confusion in the way different manufacturers and authors
use and/or define terms such as "FMS cell", or "FMS"
itself. Most manufacturers define the cell as the smallest
building block of the FMS system consisting of a compu-
ter controlled (i.e. reprogrammable) machine (note: not
necessarily a machine tool, but other machines too) capa-
ble of executing a single, or a series of operations on
parts automatically loaded and unloaded on the machine,
using tools also automatically changed. (Figures 3, 4 and
5.) Fig. 6. The work mounting station (all) of the Cincinnati-Mila-
Also note that an FMS cell is preferably unmanned but cron FMS.
does not necessarily exclude some level of human super-
vision if the economics of the automation of the involved
process, or processes demand that the designer to employ Note that in data processing terms an FMS "cell" is a
human operators, rather than 100% computer controlled node in the distributed control system, or in a LAN
equipment. (Figure 6) (Local Area Network). Sometimes the terms "cell" and
FMS may be defined as a system dealing with high- "unmanned workstation" are mixed too. We prefer the
level distributed data processing and automated material word "cell" simply because an FMS work station is often
flow using cells (i.e. computer controlled machines, as- but not necessarily unmanned. For example, consider a
sembly cells, industrial robots, inspection machines and manual work mounting station, or a co-ordinate measur-
so on), together with computer integrated materials and ing machine with some level of human assistance and
storage systems. In this sense FMS is a production know- supervision, etc., which are "cells" in our general ter-
how offering "on-order" rather than "on-stock" man- minology, but are not "unmanned workstations".
ufacturing facility for a large variety of components.
MATERIAL HANDLING AND STORAGE
SYSTEMS IN FMS
Generally, requirements set against the FMS materials
handling system include part transportation, raw material
and final product transportation and storage of work-
pieces, empty pallets, auxiliary materials, wastes, fixtures
and tools.
Materials handling systems in FMS provide palletised
component transfer between cells and include
* Automated Guided Vehicles,
* conveyors,
* industrial robots,
* special purpose pallet changing and pallet transporta-
tion systems (e.g. AGV, automated pallet carrier on
rails, etc.),
* forklift trucks and other solutions.
Since computer controlled material handling systems
offer a wide range of solutions in FMS and because the
inadequately designed materials handling system can
limit an otherwise highly flexible manufacturing environ-
ment, it is important to analyse this problem very care-
fully when designing the FMS.
Probably due to the strong influence of conventional
transfer line developments, which are less flexible than
FMS, between 1974 and 1980 the serial access work-
piece transport systems (i.e. conveyors) dominated the
FMS scene. The direct access, or random materials hand-
ling systems are not only more flexible than the closed
Fig. 5. Flexible assembly cell incorporating an Automated
Robot Hand Changing system (ARHC) designed by the author loop, or sequential workpieee transport systems, but can
capable of automatically reconnecting electric and pneumatic also simplify dynamic scheduling and flexible operation
power. (Tool changing time less than 4 seconds). control of the system.

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 FMS in CM 209

pendent lift and fork movements. Positioning systems

usually employ two different steps, i.e. the primary and
the fine positioning steps. Primary positioning is usually
carried out by dual pulse generators and a work reader
system. Fine positioning needs precise photocell align-
ment with a reflecting label at each pallet stack.
For material storage purposes automated warehouses
are integrated into flexible production systems, covering
functions such as:
* communication with the FMS material handling
system,
* materials handling system control (e.g. stacker crane
control, AGV control, conveyor control, etc.)
* real-time control of the warehouse input/output sta-
Fig. 7. Flexible machining cell incorporating a direct access pallet tions,
loading cart and an expandable buffer store. * recording individual pallet loads and locations,
* rearrangement of stored items in the warehouse,
In systems designed to manufacture prismatic parts,
* optimization of crane and truck movements for max-
the workpieces are located on pallets by means of fix-
imum throughput,
tures and clamping devices. The workpiece must there-
* stock control, order picking,
fore be loaded, unloaded and transported between FMS
* packaging and shipping of the product.
cells and workstations with these reasonably heavy addi-
From the CIM point of view the most important sys-
tional elements. (Figure 7).
tem objectives of automated warehouses in FMS can be
In the case of rotational part manufacturing systems,
summarised as follows:
workpieces are usually carried in small batches or stored in
* shorter manufacturing throughput times, because the
workpiece magazines, in a similar way to tools. (Figure 8).
parts are available within the system, as well as being
AGVs (Automated Guided Vehicles) are trucks which
available for access by the FMS to FMS material
usually pick up loads from a work mounting station
handling systems.
(where components are fixtured and clamped on pallets),
* reduced inventory because of the better organization
from the machine table, a buffer store location or a
and the minimum (i.e. almost nil) work in progress
warehouse input/output location. Components of AGV
(WIP) level required in FMS,
systems include the carts, the battery-powered driverless
* elimination of costly repetitive handling and data
vehicles capable of selecting their own path and the cart
management of material found in conventional sys-
and traffic control system.
tems,
An automated warehouse is a system providing ad-
* increased productivity.
dressable storage locations for one or more types of
To summarise, automated material handling and stor-
pallets on which material may be moved and handled
age systems are essential in FMS as well as in the overall
under computer supervision. Automated warehouses are
CIM model because they affect all major processes in-
required when the raw materials, semi-finished and
cluding receiving raw material, inspection of raw mater-
finished products, and the necessary tools, fixtures, spares
ial, part manufacturing, test and inspection, assembly,
etc., have to be stored and retrieved using computer
packaging, shipping and linking together different FMS
controlled stacker cranes, drives and lifts.
systems, warehouses and different shops in a factory.
Computer controlled stacker cranes integrate compu- (Figure 9)
ter controlled drives and lifts, fine positioning and inde-

*-•• * | L

Fig. 8. Gantry-type part loading robot and a past magazine for

turning cells. Fig. 9. Computer controlled warehouse.

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 210 FMS in CIM

AUXILIARY DEVICES IN FMS

Auxiliary devices in FMS include buffer stores, swarf
removal or part washing stations, work mounting stations
and other stations such as AGV recharge, tool mainte-
nance room, tool setup stations, etc.
Pallet mounting and fixturing are complex, but not
impossible tasks to be resolved automatically. Auto-
mated fixturing (utilising hydraulically operated devices,
or mechanically operated elements moved by robots) and
general purpose work-holding devices can solve this
problem, but because of the weight, required precision
and the complexity of movements in most cases it is done
manually, especially within a family of prismatic parts.
(Figure 6).
Even when aiming at a minimum level of WIP buffer
stores cannot be avoided, partly because of production
control constraints, and partly because of technological
reasons (for example the part must cool down before the KFMS OFF-LINE LEVEL
next operation). In this case the palletised part must wait f»MS REAL-TTHB 1.KVF.1
CAPACITY CHECK

in a buffer store, which can of course be used at the same (ALSO CHECK CELL
AVAILABILITY IN REAL-TIME)
time as an in-process pallet rack, or small warehouse in RESCHEDULE CELLS DYNAMICALLY
ACCORDING TO REQUIREMENT
order to save investment costs. (Figure 7).
Special puspose workholding devices can be fully auto-
mated, but they place great restrictions on flexibility if
one hopes to manufacture a large variety of parts in a
random order. The designers of the component can also FACTORS WHICH I
INFLUENCE FMS FACTORS WHICH CAS
contribute a great deal to solving this problem, by con- OPERATION AND INFLUENCE FMS
OPERATION AND
SCHEDULING
structing appropriate surfaces on parts by which they can SCHEDULING

be clamped and located on the pallet.

The swarf clearance and retrieval is done at the cells
FINISHED
themselves as well as at the washing station of the FMS. COMPONENT
Perfect workpiece cleaning is important, especially be-
fore inspection. Unremoved swarf can cause several un- Fig. 10. FMS operation control tasks and links to the CIM
modules.
expected problems during the automated inspection cy-
cles. Let us now briefly extend the already discussed princi-
In some systems the pallet is loaded by an AGV, or ples with some new rules regarding FMS operation con-
conveyor line with the fixtured part on the washing trol and summarise them in order to understand more
station, where it is docked. Then it is tilted by a mechan- details of the ways FMS systems can be operated and
ism, while being rinsed under high pressured coolant or controlled.
air. In other designs the washing station follows a shape A summary of the most important operation control
similar to that of a bell. This is then lowered onto the activities is shown in Figure 10, identifying three levels at
part from above, utilising vacuum whilst it is waiting, on which simulation and optimization is required prior to, or
a wired cart for example. during FMS part manufacturing, the three levels being:
In order to improve washing quality recently robots * the factory level, or business level handled by the
have been programmed to perform this job utilising business system of CIM,
vacuum tools and/or washing liquid jet guns. This solu- * the FMS off-line level, representing simulation and
tion is providing a better result than those previously optimization activities prior to loading a batch or a
discussed since the robot can execute a washing cycle single component on the FMS, handled sometimes by
which was programmed for the particular part, or batch. the CAM system, sometimes by the FMS part prog-
(To raise another important aspect of FMS programming ramming computer,
one could generate part washing programs automatically * and the real-time controlled level, handled by the
using the CAD/CAM system and an expert system to FMS operation control system, representing a situation
ensure high quality of cleaning). Once the part is clean it where the parts are already physically as well as logi-
can be taken away by a robot or Automated Guided cally in the DNC controlled environment.
Vehicle together with its pallet. One can also see from Figure 10 that to satisfy the
demand generated by the business system modules of
FMS OPERATION CONTROL CIM (i.e. the MRP and the MPS programs) the first
So far we have discussed the major components of CIM major task is the selection of the appropriate part mix for
including those used by the FMS subsystem of the com- a defined period of time, which can be as short as a few
puter integrated factory. hours, a shift (normally eight hours), a day, a few days, or

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 FMS in CM 211

a few weeks, but usually not longer. (Note that this The input to the loading sequencing program can be
period of time very much depends on how random is the the MRP, or directly the MPS output, or alternatively if
part arrival at the FMS, what the batch sizes are, whether higher level optimization is possible, the result of the part
there are batches at all, etc.) mix evaluation procedure using lotsize analysis and oper-
If there are a sufficient number of different batches of ation balancing with capacity checks. Loading sequencing
components and if the time span allows, (i.e. due dates should also take care of some real-time disturbances,
are not too close) it is worth while doing a batchsize, or such as which cell is not operational, etc. but is not a
lotsize analysis to group the components into economic real-time program, although it is valid for a very short
batches and to establish economic cycle times. Batchsize time span only.
analysis is less important in FMS compared to conven- Once the loading sequence is known parts can be sent
tional methods, because the aim is a single component, in the calculated order to the FMS for processing. At this
rather than batch production, because the setup and point there is a need for a DNC program which deals
down times are low anyway, and because the entire with data communications tasks, part program editing
production is better organised, there is practically no and downloading and others.
WIP and the inventory holding times are much shorter FMS must be able to react to real-time changes; to
too. However, it can prove to be useful and can save mention a few, what should happen to its operation con-
production cost and time, depending on the components, trol
on the setup costs, on inventory holding costs and on * if the processing priorities of a part or a batch must
other less important factors. be changed?
Balancing can also be useful in particular if there are * if a part, or a batch must be deleted from the schedule
many operations to be done on the part, and if the and removed from the system?
resources allow different operations to be performed in a * if a cell breaks down?
random order at no additional cost (e.g. if machines have This is an area when fast "secondary optimization" is
an adequate selection of tools in their magazines, or if required using a dynamic FMS scheduling program.
robots can automatically change their hands, if machines (Note that further aspects of this and the software re-
and assembly robots can change parts automatically, quirements are described in detail in Ref. 1).
etc.). Balancing has a potential in the machining oriented
FMS too, when one cell is more "popular" than the FMS VERSUS JOB SHOP OPERATION CONTROL
others, and thus becomes overloaded. The operation control of the job shop is guided by a
Balancing in robotised assembly systems (or FAS- schedule which is more often a desired, or theoretical
Flexible Assembly Systems) is generally very beneficial plan only, rather than a real one because of the continu-
because of the large amount of different operations as- ous disruptions and unplanned discontinuities in produc-
sembly systems usually perform and because it helps to tion. Eventually when the "panic" is intense enough the
decide how the operations should be distributed between job shop is partly controlled by (usually wrong) decisions
the assembly robots. Finally, in the case of conventional taken under heavy pressure, or as an equally bad alterna-
mechanised assembly line development balancing is an tive, a new schedule is prepared in order to re-establish
essential design tool to evaluate the number and arrange- the theoretically correct plans.
ment of different assembly heads and the overall cycle The scheduling algorithm used in the job shop envi-
time of the line. ronment is off-line, because it is applied at the beginning
To avoid unrealistic planning, capacity planning is im- of the scheduling period and the results are prepared for
portant, but capacity checking is even more important the entire shift, or for an even longer period of time, for
shortly after the above discussed processes in order to example a week or a month. If an unexpected event
ensure that the established part mix can be processed on happens, such as a tool break, machine tool or robot
a given system. This process is indicated by feedback break down, a part not setup on planned time, etc., then
loops at the three major modules of the off-line process. because of the deterministic scheduling methods used,
Loading sequencing, or FMS off-line scheduling is a the production is disrupted.
simulation, or optimization method to establish the best A properly designed and implemented FMS need to
order in which parts should be processed by the system in and can perform operations under the control of a
a short period of time, e.g. few hours, or up to the length dynamic scheduling system.
of a shift. This means that decisions concerning which workpiece
If parts arrive at the FMS in a random order, in very is manufactured next on which cell are made close to the
small batches, or as single components and if they must end of the operation currently being performed by the
be executed as soon as they arrive, there is not much particular cell. In other words, in FMS part scheduling
point in selecting the appropriate part mix, nor in estab- need not necessarily be carried out in full detail and has
lishing an optimum loading sequence, because there is not necessarily to be made in advance, because it is capable
simply no time for that. However, in the case of most of responding to real-time decisions.
FMS parts arrive in a fairly well planned order, there is Cells in FMS can be accessed by part and tool carrier
time to load sequence and then it is essential not only for robots, or AGVs in random order, or in other words in
utilising the system at a level close to 100%, but also for any order (as long as the actual order is programmed).
processing the parts without delay. To demonstrate this very important principle refer to

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 212 FMS in CIM

LINK TO OTHER FMS

SYSTEMS

AUTOMATED GUIDED VEHICLES CARRYING PARTS

Fig. 11. Flexible Manufacturing System design with AGV part transport between the cells and the warehouse.
Figure 11 which shows an FMS layout where cells have pallet
pools (Figure 12) and to which the workpieces are taken
from the warehouse and loaded onto pallets and then
carried into the DNC system.
Figure 13, represents a different approach. In this case
the parts are mounted at the warehouse and a combina-
tion of fork-lift trucks and AGV(s) are used to take the
components directly from the warehouse to the cells and
back.
Cells in FMS are also capable of sending and receiving
data and generally acting as an intelligent node of the
distributed data processing system. When following a
fixed schedule prepared off-line regardless of real-time
changes on the shop-floor, it is virtually impossible to
fully load all processing stations. This kind of imbalance
Fig. 12. FMS cell incorporating a chain-type pallet pool. (At a r e s u h s m hJ h e r ti c o s t s d u e t o u n d e r u t i l i z a t ion
distinguished location the AGV can dock for loading/unloading or
of t h e s stem
parts to the cell.) y -

OTHER FMS CELLS CAN

FMS CELL BE EASILY ADDED TO
WITH TWIN PALLET THE SYSTEM
STATION

AGV RECHARGE
STATION AGV R O U T E v / 1 \
\
SINGLE PALLET

'M\
AGV DOCKING
STATION STATION

OCH HIGH-LIFTINC AUTO-TRUCK

( I . E . A FORK-LIFT TRUCK
BUILT ON THE TOP OF AN AGV)
FMS CELL
WITH TWIN PALLET
STATION

LINK TO OTHER FMS

SYSTEMS
AUTOMATED PALLET RACK/WAREHOUSE

UULH MOUNTING/UNLOADING
STATIONS

Fig. 13. FMS design utilising a high-lifting auto-truck capable of accessing not only the cells in the system but also the pallet rack,
where parts are mounted and stored.
Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 FMS in CM 213

When applying the more advanced dynamic scheduling BEGIN FMS_Part_prograa/CodaiABC08

system and the "variable route" FMS part programming BEGIN Pallat_prograa/CodaiNo.l,Pallat_codaiPll
Exacuta.oparation/CodaiNo.1, Tool.filaiPlNo.1TO1I
method (as described in Ref. 2) offering alternatives for Exacuta_oparation/CodaiNo.2, Tool.filaiPlNo.2T02l

both loading scheduling as well as for secondary optimi- IF <Condition trua> THEN
BEGIN
zation as mentioned above, a much higher level of flexi- Exacuta o p a r a t i o n / C o d a i N o . 3 , Tool f i l « i P l N o . 3 T 0 3 l
E x . c u t . oparation/CodaiNo.4, T o o l . f i 1 « I P 1 N O . 4 T 0 4 I
bility and equipment utilization can be achieved even in END <• PALLET PROGRAM END •>
ELSE
cases when part processing priorities change, or cells BEGIN
Exacuta oparation/CodaiNo.A3, T o o l _ f i l a i P l N o . A 3 T 1 3 I
break down, etc. Exacuta_oparation/CodaiNo.A4, T o o l . f i l a i P l N o . A 4 T 1 4 I

Figure 14 shows a sample structure of an FMS part IF <Condition t r u a > THEN

Exacuta_oparation/Coda:No.A5, Tool f i I . I P I N O . A 6 T 2 6
program providing "variable routing". Figure 15 de- ELSE
BEGIN
scribes this part program using a language, in structure Exacuta_oparation/CodaiNo.A&, T o o l . f i 1 B I P 1 N O . A 6 T 1 6 I
Exacuta_oparatlon/CodatNo.A7, T o o l ' f i l a i P l N o . A 7 T 1 7 )
similar to PASCAL, developed by the author. This part ENDI ( • 2nd ELSE • )
END! ( • lmt ELSE • )
program description is very important because it contains ENDI ( • PALLET PROGRAM No.1 • )

all necessary information in one single structure. It de- BEGIN Pal 1at_prograa/CodaiNo.2,Pal 1at_codaiP12I
PALLET PPROGRAM No.2 DATA STRUCTURE DESCRIPTION *>
scribes alternatives, their order, when can they be exe- ENDI ( PALLET PROGRAM No.2 •)
cuted, what fixtures and finally what tools are required. allat_prograa/Coda:No.3,Pal 1at.codaiP33I
PALLET PROGRAM No.3 DATA STRUCTURE DESCRIPTION •)
In other words, it represents the "production rules" END! ( PALLET PROGRAM No.3 •)
relating to a single component, or maybe a batch and (• FURTHER PALLET PROGRAMS CAN BE ADDED HERE... •>
must be therefore accessed by the dynamic scheduling END. <• FMS PART PROGRAM ABC08 • >
program before selecting current (alternative) production Fig. 15. Production rule base of a part to be manufactured on
routes. (Obviously an FMS part program can contain FMS providing alternative routing and dynamic scheduling.
more than one pallet program with a single operation
description, as well as can be more complex than the one
shown in Figure 14). CONCLUSION
Let us emphasize that CIM is very new, there are many
"unsolved problems", and that there are no "truly integ-
rated" CIM installations in its real sense known to the
author at the time this text was written in March 1985.
PICTURING AND CLAMPING There are many working systems incorporating some
DATA I S STORED AT THE
PALLET PROGRAM LEVEL important modules of CIM, such as factory or corporate
wide data communication and business systems, integ-
rated CAD/CAM systems and FMS, but there is still an
enormous amount of work and money to be put into
computing, communications, interfacing, Local Area
Networks, knowledge engineering,34 and manufacturing
development and research to see working installations
and the real benefits of this "total system" concept.1S
Integration and dynamic operation control problems in
CIM and in FMS will be solved easier as soon as power-
ful microcomputers will be integrated parts of all machin-
ery utilised on the shop-floor and the organization as a
whole, and when people who not only understand the
data processing aspects, but the involved processes too,
will write the software for these machines .2<s~12

References
1. Paul G. Ranky, Computer Integrated Manufacturing, an
Introduction with Case Studies, (Prentice-Hall International,
New York, Autumn 1985).
2. Paul G. Ranky, The Design and Operation of Flexible
Manufacturing Systems (IFS Publications, and North Hol-
land Publishing Co., 1983).
3. N. Shahla Yaghmai, "Expert Systems: A tutorial" American
Soc. Information Sciences 297-305 (September 5, 1984).
4. Nick Cercone, "Artificial Intelligence: underlyning assump-
tions and basic objectives" American Soc. Information Sci-
ences 143-151 (September 5, 1984).
5. "Technology, a comprehensive survey of the different com-
ponents of CIM" Tooling and Production 24-204 (March,
Fig. 14. The "variable route" FMS part programming concept 1984).
and program structure, allowing alternative operations to be 6. P. Bunce, "Planning for CIM" The Production Engineer 64,
taken (and production routes to be assigned) if necessary. (Note No. 2, 21 (February, 1985).
that an operation can mean machining, washing, inspection, 7. J. Harrington, Computer Integrated Manufacturing
assembly, welding, grinding and others.) (Krieger, 1979).
Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290
 214 FMS in CIM

8. M. Annborn "The Factory of the Future" 3rd International 12. Paul G. Ranky, "Integrated software for designing and
Conference on Flexible Manufacturing Systems, 1984, analysing FMS," 3rd International Conference on Flexible
Boeblingen, W-Germany (IFS Publications, Bedford, Manufacturing Systems, 1984, Boeblingen, W-Germany
1984). pp. 59-73. (IFS Publications, Bedford, 1984) pp. 347-361.
9. M. Mollo "A distributed control architecture for a Flexible
Manufacturing System", 3rd International Conference on
Flexible Manufacturing Systems, 1984, Boeblingen, W- Further reading
Germany (IFS Publications, Bedford, 1984), pp. 227-241. Daniel S. Appleton "The State of CIM" Datamation 66-72
10. C. Accomazzo "The integration of an inspection cell in (December 15, 1984).
FMS" 3rd International Conference on Flexible Manufactur- "Plan for Computer Integrated Manufacturing" Seminar lec-
ing Systems, 1984, Boeblingen, W-Germany (IFS Publica- tures organised by the Institution of Production Engineers,
tions Ltd., Bedford, 1984) pp. 285-297. London (November, 1984).
11. D. Van Zeeland "System software" 3rd International Con- Naoaki Usui " M m ' s "super" manufacturing system", Ameri-
ference on Flexible Manufacturing Systems, 1984, Boeb- can Machinist 86-114 (August, 1984).
lingen, W-Germany (IFS Publications, Bedford, 1984). pp. Duncan Holland, "Strategic benefits gained from CIM", The
307-311. Production Engineer 14-16 (June, 1984).

Downloaded from https://www.cambridge.org/core. University of Leeds, on 11 Jun 2018 at 09:18:57, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S0263574700002290