Sie sind auf Seite 1von 15

Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

Intelligent planning of Grafcet charts


L. Castillo*, J. Fdez-Olivares, A. GonzaH lez
Departamento de Ciencias de la Computacio& n e Inteligencia Artixcial, E.T.S. Ingeniern& a Informa& tica, Universidad de Granada, 18071 Granada, Spain

Abstract

Manufacturing systems are entering a new age in which the demands of a market in continuous change impose their conditions on
production processes. In this new generation, both the response capability to these demands and a reliable production process are
aspects that are receiving most attention and study. One of the stages that can most a!ect these terms of reliability and time response
is the development of the control program which drives the operation of the manufacturing system, although this subject has received
little attention in the literature. This work provides a new e!ort in this direction and presents MACHINE, an arti"cial intelligence
planner able to automatically obtain Grafcet charts for manufacturing systems.  2000 Elsevier Science Ltd. All rights reserved.

Keywords: Intelligent manufacturing systems; Arti"cial intelligence planning; Control programs; Grafcet charts

1. Introduction This can be a very di$cult, and sometimes, painful


problem during the life-cycle of a manufacturing system
Manufacturing systems are entering a new age in which can become a bottleneck in the achievement of the
which the demands of a market in continuous change goals of reliability and low time response:
impose their conditions on production processes, from
the design to the operation and maintenance of manufac- E It is usually developed by hand by specialised control
turing systems. In this new generation, both the response engineers. This means that it is a slow and expensive
capability to these demands and a reliable production process, which is always subject to human error.
process are two of the aspects that are receiving most E It is hardly adaptable to the life-cycle of the manufac-
attention and study. International forums such as re- turing system. When the manufacturing system layout
search and development programs [1] or special interest or control speci"cations change, new control pro-
groups [2,3] are being deployed to study and provide grams must be rewritten in order to take these changes
solutions in this direction, looking for fast and sound into account. This rewriting process must be very
manufacturing processes. careful because the quality of the control program may
These issues are receiving special attention from the decrease or, even, new errors may appear due to
arti"cial intelligence research community arising in interactions with previously written pieces of the
a new, rich and rapidly expanding research area covering program.
most of the aspects of manufacturing systems, such as Only very recently, this problem has been approached
intelligent product and process design, production plan- by means of arti"cial intelligence planning techniques
ning and scheduling or systems engineering and architec- [5}7]. The strong similarity between the process of
tures [4]. However, little attention has been devoted to planning and the process of programming, and also
the subject of the implementation of the control program between the concept of plan and the one of algorithm,
which drives the operation of a manufacturing system. make these techniques actually suited for the automation
of the design of control programs simply by using the
"nal plans of action as control program skeletons. The
aforementioned drawbacks of a design process by hand
* Corresponding author. Tel.: #34-958-24-4258; fax: #34-958-24-
3317. are overcome by the use of these knowledge-based tech-
E-mail addresses: l.castillo@decsai.ugr.es (L. Castillo), faro@de- niques, providing an additional e!ort for achieving lower
csai.ugr.es (J. Fdez-Olivares), a.gonzalez@decsai.ugr.es (A. GonzaH lez). time response and higher reliability:

0736-5845/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 7 3 6 - 5 8 4 5 ( 9 9 ) 0 0 0 5 9 - 9
226 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

E When the knowledge used by these techniques is cor- systems providing a more constrained domain of
rect, the use of arti"cial intelligence planning tech- study.
niques guarantees that "nal plans will also be correct Historically, one of the most appealing applications of
and, therefore, free of human errors. arti"cial intelligence planning techniques in the literature
E Due to the growing speed of modern computers, the has been its application to machining procedures of dis-
time taken to obtain plans is very low, allowing for crete parts, what is known as computer-aided process
a considerable decrease in cost and duration of the planning (CAPP) [13]. There has been a great deal of
design phase of control programs. work around this issue [14 }19] including many com-
E They allow for an easy adaptation to the life-cycle of plementary techniques such as machine learning tech-
manufacturing systems because if something changes in niques, hierarchical representations and procedures or
a manufacturing system, then now plans are completely nonlinear planning techniques, but it must be recognised
rewritten from scratch in order to take these changes that, in practice, these systems generally focus on compo-
into account, avoiding a revision process by hand. nents or pieces of the overall manufacturing process [20]
without taking into account some of the operations usu-
In this sense, this work presents MACHINE, a three-
ally needed to build a whole control program.
level system, based on arti"cial intelligence planning
For these reasons, a greater e!ort is being demanded
techniques which, starting from a knowledge representa-
for developing alternative approaches able to consider
tion of a manufacturing system in terms of :
a whole manufacturing system, not only a piece of it [13].
E a high-level description of the manufacturing system In this direction, the work reported in [21] presented
layout and start up state of devices, a system able to obtain Grafcet charts for manufacturing
E a description of raw products, and systems by means of search techniques and a knowledge
E a manufacturing problem, speci"ed as an ordered base which contains pieces or skeletons of previously
sequence of transformations on raw products, i.e., developed control programs. However, the entire process
a control recipe at phase level [8], must be supervised by a human operator, and, therefore,
the system is not completely autonomous.
is able to automatically design the "nal control program,
Other integral approaches have also been developed as
expressed as a Grafcet chart [9], tailored for this prob-
[22], applied to assembly operations, or as in [23] under
lem, this layout and these raw products.
the general name of operation procedure synthesis in
Section 2 reviews previous works in the literature
chemical domains (an extensive review of this issue can
around the subject of automatic programming in manu-
be found in [24]), but the level of detail of "nal plans of
facturing systems. Section 3 presents MACHINE, a
actions is insu$cient since they may lack some actions
three-level knowledge-based system in which the top
which can be essential for the execution of the control
level de"nes a knowledge representation of a manufac-
programs.
turing system, the intermediate level (the `intelligenta
These problems, i.e., an integral approach with a su$-
level) builds the control program skeleton and the lower
cient level of detail in the results, were initially addressed
level translates this skeleton into a Grafcet chart. Section 4
in [25,26] where authors present a new planning system,
presents a case study involving several manufacturing
called MACHINE, able to automatically obtain more
problems which are solved by MACHINE, specially two
complete plans, containing all of the actions that could
real size manufacturing systems. Finally in Section
be essential during the execution of the control program,
5 some conclusions and future directions are drawn.
and it is continued in this work, in which authors present
an extension of this system able to translate these control
sequences into a reliable and lower level representation:
2. Related work a Grafcet chart that can be used to implement the "nal
control program. This is explained in the next section.
In the late 1960s, the subject of automatic program-
ming was studied under the general topic of automated
program synthesis [10}12] and the techniques used were, 3. A three-level knowledge-based system
mainly, theorem proving techniques, but however, very
little advance seems to have been made since then. Per- MACHINE is an arti"cial intelligence planning system
haps, the problem the of automatic building of programs whose model of action has been speci"cally designed to
is too hard to be faced from a general viewpoint and describe the transformations that take place in a manufac-
research should have been focused on more speci"c do- turing system and whose architecture is shown in Fig. 1.
mains. It works at three di!erent abstraction levels: the "rst
In this sense, the domain of manufacturing systems one is a knowledge-based representation of a manufac-
is becoming an area of growing interest for researchers turing problem as will be seen later, the second one
from the "eld of arti"cial intelligence planning represents a possible solution to this manufacturing
L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 227

Fig. 1. Three-level architecture of MACHINE.

Fig. 2. A simple manufacturing system.


228 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

problem expressed as a control program skeleton, that is, of action of MACHINE. This model of action [27] is
an ordered sequence of actions that must be executed and a sort of speci"cation language, based on the "rst-order
which is named as a control sequence, and the third level predicate calculus, which has been designed as an exten-
represents this program skeleton as a Grafcet chart. Let sion of STRIPS [28]. It is a structured representation of
us look at these three levels in detail. the actions that take place in the manufacturing domain,
their requirements and their consequences in the envi-
3.1. The top level ronment. By means of this model of action a knowledge
base is built which contains the description of every type
The most important piece of the top level, which is of device that may be found in a manufacturing system: it
used to describe a manufacturing problem, is the model is a library of devices.

Fig. 3. A control sequence which solves the manufacturing problem stated in Example 2.

Fig. 4. Algorithm for translating a control sequence into a Grafcet chart.


L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 229

3.1.1. The library of devices Requirements. They are facts that must be satis"ed, i.e.,
This knowledge base must be created and maintained they must be true, for a correct execution of the action.
by a control engineer (whose main task is to provide the They are divided into three di!erent kinds of require-
necessary knowledge to describe the operation of every ments:
device) and a knowledge engineer (whose main task is to
structure that knowledge in terms of the model of action Table 1
of MACHINE). In this sense, every device is represented Logic formulae of transition conditions for the Grafcet in Fig. 5
as an agent connected to other agents and able to carry CT1 (STATE C1 OFF)  (LOCATION BLOCK P1)  (FREE
out a set of actions. The description of an agent has BLOCK P1)  (STATE C2 OFF)  (STATE DRILL
several components and the most important ones are STOP)  (STATE DRILL UP)  (STATE BELT OFF)
these: CT2 (STATE BELT ON)  (LOCATION BLOCK STORAGE)
Name. The identi"cation of the agent.  (FREE BLOCK STORAGE)
CT3 (STATE DRILL DOWN)  (HAS-HOLE BLOCK)
Variables. They represent a set of objects related to the CT4 (STATE DRILL UP)
operation of the agents: raw materials, connection points, CT5 (STATE BELT OFF)
etc. CT6 (STATE DRILL DRILLING)
Actions. This consists of a description of every action that CT7 (STATE C2 ON)  (BLOCKED BLOCK P2)
the agent is able to carry out. Every action of every agent CT8 (STATE C1 ON)  (LOCATION BLOCK P2)  (FREE
BLOCK P2)
is composed of the following components. CT9 (STATE C1 OFF)
Name. An identi"cation of the action. CT10 (STATE C2 OFF)  (FREE BLOCK P2)
CT11 (STATE DRILL STOP)
CT12 TRUE
CT13 TRUE

Fig. 5. The "nal Grafcet chart for the manufacturing problem of


Example 2. Fig. 6. Case 1. A batch system.
230 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

Fig. 7. Grafcet chart for case 1.

Previous. They must be true before the action starts.


They act as a kind of preconditioning of the action.
Simultaneous. They must be true during the execution
of the action. They are used to avoid any type of
harmful interaction during the execution of the action, in
such a way that every action which could interfere
with it executes either before or later, but never simulta-
neously.
Later. They must be true after the action. They are
used to leave every device in a safe state every time they
are used to carry out an operation.
Hence, these requirements are a kind of invariant that
must be true before, during and after the action.
E4ects. They are the consequences in environment of the
execution of the action. They are divided into two di!er- Fig. 8. Grafcet chart for case 2.
ent e!ects.
Addition. They are new facts that become true.
Deletion. They are old facts that were true but become
false. Hence, the facts included in the deletion e!ects are
also included in the previous requirements.
L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 231

Fig. 9. Case 3. A discrete manufacturing system.

Example 1. Let us consider the small system shown in Simultaneous. The drill must be on and the object must
Fig. 2. It is composed of two pistons, a mobile drill and be grasped at the base of the drill.
a belt. Its operation is clear: piston C1 pushes the block Later. The drill must be up again.
into the drill base, then C2 grasps it, the drill makes
a hole, the block is released and the belt carries the block E4ects
into the STORAGE. Addition. The drill will be down and the object will
have a hole.
In this example, every device is represented as an Suppresion. The drill is no longer up.
agent, but in particular, the drill can be represented
as an agent able to carry out the following four actions: As can be seen, the key aspect of this knowledge base is
turn on, turn o!, go down, go up. Let us consider the knowledge representation and structure of the opera-
the action of going down. Its representation would be as tion of every device. The role of the control engineer is
follows: very important, but the role of the knowledge engineer is
Name: DRILL-DOWN essential since, as noted in [20], the techniques used to
Requirements avoid the existing gap between theory and practice in
Previous. The drill must be up. manufacturing domains tend to be knowledge intensive in
order to de"ne a sound start point in the search of
solutions.
Another important feature of this knowledge base is its
 This is a literal description. Actually, it is a "rst-order logic repres-
entation which uses a structure based on the LISP programming reusability. It is parameterised for many known types of
language and where every stated fact is represented as a simple literal. devices (valves, pumps, belts, mixers, etc.) and, once
232 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

E The initial state. That is, a description, by means of the


"rst-order predicate calculus, of the start up state of
both devices and raw products.
E The goal of the manufacturing problem. That is, the
recipe [8] of the manufactured product. It is an
ordered sequence of transformations, de"ned on the
formerly represented system, that must be made
on the raw products mentioned above.

Example 2. In Example 1, the system layout would be


represented by means of an agent for every piston, the
drill and the belt, respectively, extracted from the library
of devices. The raw products are very simple: there is
a block next to piston C1. And "nally, the goal is also
very simple: the block must have a hole and it must be
into STORAGE.

In real life, these manufacturing problems are solved


by control engineers, who have a similar knowledge-
based representation, by designing a control program
which would coordinate the operation of the devices of
the manufacturing system to meet the goal speci"ed in
the recipe and, so, to obtain the desired manufactured
product. In this work, the role of the control engineer is
played by the intermediate level of MACHINE, that is,
a planning system which contains the `intelligencea
needed to solve these problems and automatically
design a skeleton of the control program starting from
the knowledge-based description of a manufacturing
problem.

3.2. The intermediate level

The most important piece of the intermediate level is


the planning algorithm which is described in detail in
[26,29]. It is a partial order planner [30] which follows
a basic means end analysis in order to "nd an ordered
sequence of actions of the agents that meet the goal of the
Fig. 10. Grafcet chart for case 3. manufacturing problem starting from the initial state
described in the problem. This is called a control se-
de"ned, it can be used for the representation of a quence and may be represented as a pair S"1A,(2
manufacturing problem and reused again as new manu- where A is a set of actions and ( an order relation
facturing problems are faced. de"ned on it.
The representation of agents and actions in the manu-
facturing problem allows for an easy reasoning process. If
3.1.2. Manufacturing problems
the goal is a transformation which must be solved and
A manufacturing problem consists in carrying out an
this transformation is an e!ect of an action, then include
ordered set of transformation operations on raw prod-
that action in the control sequence to achieve this trans-
ucts in order to obtain manufactured products. These
formation, but, in order to do that, all of the requirements
problems are represented in MACHINE by three com-
of the action, must also be solved by other actions.
ponents:
Every time an action solves a requirement of another
E The domain. It is the manufacturing system layout action, this requirement is protected and the planner
represented by means of the aforementioned model of
action. Every device in the manufacturing system is
translated into an agent by instantiating its corre-
sponding type in the library of devices.  A strict order relation and not necessarily a total order relation.
L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 233

Fig. 11. Case 4. A reactor-based batch system.

guarantees that none of the actions in the sequence, A control sequence is not actually a control program,
neither now nor in the future, will produce a mutual but it may be considered as a skeleton of the control
contradiction with respect to this achieved transforma- program since it contains all the knowledge necessary to
tion. This protection mechanism is usually known as translate it into a lower level representation. In fact,
a causal link [26,30,31]. a control sequence has several of the most important
The causal link protection mechanism is the most features that control programs are expected to have:
important means to guarantee that, not only all of the
requirements of the actions in a control sequence have E Discrete state space, that is, the states of the system are
been satis"ed, but also that none of the actions contra- represented in a logic or symbolic form instead of
dicts the requirements of other actions. If a contradiction a numeric form.
between an action and a causal link is detected (this is E Event driven, that is, its evolution can be described by
called a threat) then it is solved by reordering the action means of a sequence of discrete events. Every event
and the causal link, giving priority to the causal link or to may be considered as the execution of an action in the
the action, respectively, in such a way that the contradic- control sequence, which produces a perfectly known
tion disappears [26]. set of qualitative changes in the environment repre-
This process is repeated following a best "rst search sented as the e!ects of the action.
process [6] until there are no pending transformations or E Asynchronous operation. Events do not appear re-
requirements to be solved. The "nal result of the plann- lated to an exact notion of time, instead, they take
ing algorithm is a control sequence S"1A,(2, that is, place following a relative order between them: the
an ordered sequence of actions of the agents of the order speci"ed in the control sequence.
manufacturing problem which achieves the goal. E Parallelism, that is, several actions may execute simul-
taneously. MACHINE builds parallel sequences
Example 3. Let us consider again the manufacturing whenever it is possible and the actions in parallel
problem presented in Example 2. The control sequence branches do not interfere, in other cases, actions will be
designed by MACHINE for solving this problem is in sequence. For example, once piston C1 has been
shown in Fig. 3. This control sequence has been auto- extended, and therefore, the block is pushed into the
matically designed by MACHINE in 5.85 s. base of DRILL, then it is retracted again and C2 is
extended, both simultaneously. On the other hand,
since C2 must be grasping the block during the drilling
process, it cannot be retracted until DRILL is up
because, in that case it will produce a violation of
 Running on a Pentium II under Linux operating system. a simultaneous requirement.
234 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

Fig. 12. Grafcet chart for case 4.

E Nondeterminism, that is, from a starting state, there As can be seen, a control sequence has two dummy
can be several di!erent, but valid, evolutions. In this actions called START and END, respectively. These ac-
case, actions in parallel branches execute independent- tions are used to delimit the control sequence, but, in
ly of each other and they can follow di!erent evolu- addition, action START is used to generate the start up
tions. state of both raw products and devices in the following
L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 235

Fig. 13. Case 5. A real-world manufacturing system.

way. This initial state consists of a set of facts that stage(s). Every literal could be represented as a logical
describes what is true at the beginning of the problem, variable which can be implemented either as an input
then this set of facts is `codeda as the e!ects of the action of the control system (when the information provided
START. by the variable comes from a sensor in the manufactur-
As said before, a control sequence is not a control ing system) or as an internal or timed variable of such
program. It is rather a high-level representation of such a control system.
programs, but it contains all the knowledge needed to E Additional stages, containing a dummy waiting
build a real control program, i.e., the representation of action, are included in the chart at the end of parallel
every action, its requirements and e!ects. The next sec- branches in order to synchronise the end of these
tion shows how to translate these control sequences into branches.
a lower-level representation exploiting these sources of
knowledge. The algorithm for translating a control sequence
S"1A,(2 into a Grafcet chart is shown in Fig. 4.
3.3. The lower level
Example 4. Finally, the Grafcet chart for the example
presented in the beginning is the one shown in Fig. 5.
The algorithm for translating a control sequence into
A detailed description of transition conditions is shown
a Grafcet chart is very important since it allows to
in Table 1. The "rst action can be considered as an
represent a control sequence, and all of the knowledge
initialisation action which de"nes the start up know-
embedded in it, as a formalism which could be directly
ledge, as can be seen in the description of condition CT1.
employed by a control engineer. It builds up a Grafcet
Next, two parallel branches have been designed by
chart in the following terms:
MACHINE in order to push and grasp the block and,
E Every stage contains only one action of the sequence, at once, to turn on the drill. Once the block is grasped
and the order relation between the stages is induced by (CT7), the drill goes down and up again, making the hole,
the order relation in the control sequence. The action and then, only when piston C2 has retracted (CT10), the
in every stage could be represented as an output signal belt carries the block into the storage. When the block is
to the corresponding device. into storage (CT2) the belt is stopped and the program
E Transition conditions after a stage are built by means ends.
of the addition e!ects of every action meaning that the
action has achieved all of its e!ects and that the execu- As can be seen, this detailed Grafcet chart could be
tion of the chart may continue with the following directly implemented in most control systems.
236 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

Fig. 14. Grafcet chart for case 5.


L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 237

Table 2
Logic formulae of transition conditions for the Grafcet in Fig. 14

CT1 (STATE MP OFF)  (STATE MV OFF)  (STATE MV2 OFF)  (STATE EP OFF)  (STATEEV OFF)  (STATE MIX1 OFF)
 (STATE HEAT1 OFF)  (STATE SV OFF)  (STATE ST1 OFF)  (STATE ST2 OFF)  (STATE V1 OFF)  (STATE B1 OFF)
 (STATE V2 OFF)  (STATE HEAT2 OFF)  (STATE MIX2 OFF)  (STATE AV OFF)  (STATE AP OFF)  (STATE ASI
OFF)  (STATE AS2 OFF)  (STATE BELT1 OFF)  (STATE VB2 OFF)  (STATE BP OFF)  (STATE BOTTLER OFF)
 (STATE BELT2 OFF)  (CONTAINS MILK MILK-TANK)  (TEMP MILK LOW)  (CONTAINS ENZYME ENZYME-
TANK)  (CONTAINS FLOUR ADDITIVE1)  (CONTAINS COCOA ADDITIVE1)
CT2 (STATE BOTTLER ON)  (BOTTLED MILK)
CT3 (STATE AP ON)  (ADDED MILK COCOA)  (FLOW COCOA RTANK TANK2)  (FLOWIING COCOA RTANK TANK2 AP)
CT4 (STATE AP OFF)
CT5 (STATE BOTTLER OFF)
CT6 (STATE AV ON)  (OPEN-FLOW COCOA TRANK TANK2)  (OPENING-FLOW COCOA RTANK TANK2 AV)
CT7 (STATE MIX2 ON)  (MIX MILK TANK2)  (MIXING MILK MIX2)
CT8 (STATE B1 ON)  (CONTAINS MILK TANK2)  (FLOW MILK TANK1 TANK2)  (FLOWING MILK TANK1 TANK2 B1)
CT9 (STATE HEAT1 ON)  (TEMP MILK HIGH)
CT10 (STATE HEAT1 OFF)
CT11 (STATE B1 OFF)
CT12 (STATE MIX2 OFF)
CT13 (STATE AV OFF)
CT14 (STATE ST1 ON)  (OPEN-FLOW STEAM BOILER HEAT1)  (OPENING-FLOW STEAM BOILER HEAT1 ST1)
CT15 (STATE MIX1 ON)  (MIX MILK TANK1)  (MIXING MILK MIX1)
CT16 (STATE MP ON)  (CONTAINS MILK TANK1)  (FLOW MILK MILK-TANK TANK1)  (FLOWING MILK MILK-TANK
TANK1 MP)
CT17 (STATE MP OFF)
CT18 (STATE MIX1 OFF)
CT19 (STATE ST1 OFF)
CT20 (STATE MV ON)  (OPEN-FLOW MILK MILK-TANK TANK1)  (OPENING-FLOW MILK MILK-TANK TANK1 MV2)
CT21 (STATE MV2 OFF)
CT22 (STATE V2 ON)  (OPEN-FLOW MILK TANK1 TANK2)  (OPENING-FLOW MILK TANK1 TANK2 V2)
CT23 (STATE V2 OFF)
CT24 (STATE MV ON)  (OPEN-FLOW MILK MILK-TANK P1)  (OPENING-FLOW MILK MILK-TANK P1 MV)
CT25 (STATE MV OFF)
CT26 (STATE SV ON)  (OPEN-FLOW STEAM BOILER P2)  (OPENING-FLOW STEAM BOILER P2 5V)
CT27 (STATE SV OFF)
CT28 (STATE BELT2 ON)  (CONTAINS BOTTLES BOTTLER)  (FLOW BOTTLES BOTTLES-STORAGE BOTTLER)  (FLOW-
ING BOTTLES BOTTLES-STORAGE BOTTLER BELT2)
CT29 (STATE BELT1 ON)  (CONTAINS COCOA RTANK)  (FLOW COCOA ADDITIVE1 RTANK)  (FLOWING COCOA
ADDITIVE1 RTANK BELT1)
CT30 (STATE V1 ON)  (OPEN-FLOW MILK TANK B1)  (OPENING-FLOW MILK TANK1 B1 V1)
CT31 (STATE V1 OFF)
CT32 (STATE BELT1 OFF)
CT33 (STATE BP ON)  (CONTAINS MILK BOTTLER)  (FLOW MILK TANK2 BOTTLER)  (FLOWING MILK TANK2
BOTTLER BP)
CT34 (STATE BP OFF)
CT35 (STATE AS1 ON)  (OPEN-FLOW COCOA ADDITIVE1 BELT1)  (OPENING-FLOW COCOA ADDITIVE1 BELT1 AS1)
CT36 (STATE AS1 OFF)
CT37 (STATE BELT2 OFF)
CT38 (STATE VB2 ON)  (OPEN-FLOW MILK TANK2 BOTTLER)  (OPENING-FLOW MILK TANK2 BOTTLER VB2)
CT39 (STATE VB2 OFF)
CT40 TRUE
CT41 TRUE
CT42 TRUE
CT43 TRUE
CT44 TRUE

4. A case study have been instantiated from the library of devices of


MACHINE. The "nal Grafcet chart is shown in all of the
This section shows the performance of MACHINE in cases but, since a detailed description of every transition
"ve selected examples including batch and discrete condition in a chart would be di$cult to show due to its
manufacturing and two real-world examples (case 4 excessive length, only a detailed table of transitions con-
and 5). All of the devices which appear in these problems ditions is shown for case 5. Although some charts seem
238 L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239

quite complex, the facts that control sequences are free of The manufacturing problem is de"ned by the following
harmful interactions between their actions (because of sequence of transformations:
the causal link protection mechanism) and that Grafcet
1. STEP 1. Add ingredient B to ingredient A in reactor
charts have been built on the basis of these control
R-501. During this operation, the mixture must be in
sequences, guarantee that Grafcet charts are also free of
agitation.
harmful interactions between their stages.
2. STEP 2. Heat the mixture.
3. STEP 3. Add ingredient C to the mixture maintaining
4.1. Case 1: A transportation problem
the agitation. During this mixing operation a residual
gas is generated which must be evacuated through the
The layout of the manufacturing system is shown
scrubber S-501. Part of this gas is condensed and it
in Fig. 6. The problem to be solved is the transporta-
precipitates at the bottom of S-501. Once the mixing
tion of WATER and ACID to TANK3 and TANK4,
operation ends, this residual liquid must be carried
respectively.
into the external tank T-503.
In this case, this is an easy transportation problem
4. After the addition of ingredient C, the mixture must be
which can be carried out in parallel, as can be seen in the
cooled and carried into tank T-502.
Grafcet chart shown in Fig. 7. MACHINE obtains this
chart in 26 s. As can be seen, this problem de"nition is a control
recipe speci"ed at phase level [8,32]. In this case,
4.2. Case 2: Another transportation problem MACHINE solves the problem in 5477 s by designing
the Grafcet chart shown in Fig. 12.
This problem is de"ned on the same manufacturing
system as that shown in Fig. 6. However, now the prob- 4.5. Case 5: A dairy products problem
lem is more di$cult since WATER must be carried into
TANK4 and ACID into TANK3. As can be seen, this The system is shown in Fig. 13. This manufacturing
transportation problem cannot be solved in parallel, system is inspired by a real problem, and it is composed
since it would produce an undesired mixing of products. of 24 di!erent interconnected agents (pumps, valves,
MACHINE solves the problem in 255 s and it designs mixers, heaters, belts, and a bottler). The design of a con-
the Grafcet chart shown in Fig. 8 in which the transpor- trol program for such a system is di$cult, even for
tation of both products is carried out sequentially. a control engineer, due to the number of agents and the
total number of available actions.
4.3. Case 3: A discrete manufacturing problem The goal in this case consists in adding an ingredient
(initially contained in ADDITIVE-1) to the milk initially
The layout of the system is shown in Fig. 9. The contained in MILK-TANK and then proceed to bottle
problem to be solved consists of the following steps: the mixture. In this case, MACHINE obtains the Grafcet
chart of Fig. 14 in 1836 s. A detailed description of
1. Drill PIECE-1 (initially in IN-BUF1). transition conditions is shown in Table 2.
2. Mill PIECE-2 (initially in IN-BUF2). These "ve problems show that MACHINE, a
3. Carry both pieces into the assembling machine. computer program based on arti"cial intelligence plann-
4. Assembly. ing techniques, is able to automatically design Grafcet
5. Carry the "nal piece to OUT-BUF. charts for real-size manufacturing problems and that
these Grafcet charts are free of human errors (mainly due
This problem is solved by MACHINE in 187 s by to the causal-link-based protection mechanism).
means of the Grafcet chart shown in Fig. 10. The machin-
ing of PIECE-1 and PIECE-2 in parallel, the assembly
and, "nally, the operation of the robot arm, are all shown 5. Some conclusions
in this "gure.
MACHINE is a three-level knowledge-based system
4.4. Case 4: A problem involving a reactor which is able to recognise manufacturing problems de-
scribed by means of its library of devices, which is able to
This problem is extracted from [32] and the layout of automatically design a Grafcet chart which solves that
the system is shown in Fig. 11. In this batch problem manufacturing problem and which can be directly imple-
there are three types of raw products: an ingredient A, mented in most control systems. This directly implies
stored in tank T-501, an ingredient B, stored in great advantages, specially in what new generation
tank T-505 placed somewhere out of the system, manufacturing systems are concerned about.
and an ingredient C stored in tank T-504 also out of the Starting from a correct description of devices, the use
system. of arti"cial intelligence planning techniques allows for an
L. Castillo et al. / Robotics and Computer Integrated Manufacturing 16 (2000) 225}239 239

autonomous process that obtains a Grafcet chart which [12] Waldinger RJ, Constructing programs automatically using
is free of human design errors. theorem proving. PhD Thesis, Carnegie-Mellon University, 1969.
MACHINE and the use of modern computers allows [13] Maropoulos PG. Review of research in tooling technology, pro-
cess modelling and process planning. Part II: process planning.
for a rapid prototyping process, that considerably reduc- Comput Integrated Manuf Systems 1995;8:13}20.
es the cost and duration of the development phase of [14] Chang H, Lu W, Liu X. Intelligent case retrieval and modi"cation
control programs allowing for a faster response time. for machining process planning of axisymmetric parts. AAAI
If the layout of the manufacturing system changes or Special Interest Group in Manufacturing Workshop: Arti"cial
production needs are rede"ned, then new results can be Intelligence and Manufacturing, State of the Art and State of the
Practice. AAAI Press, 1998. p. 55}62.
completely and quickly redesigned to meet the new speci- [15] Descotte Y, Latombe JC. GARI: a problem solver that plans how
"cations. In addition, this avoids the need for a revision to machine mechanical parts. IJCAI 1981. p. 766}72.
and rewriting of old control programs which could have [16] Gil Y. A speci"cation of manufacturing processes for planning.
again new human errors. Technical Report CMU-CS-91-179, Carnegie Mellon University,
However, although these advantages are very 1991.
[17] Nau D, Gupta SK, Regli WC. AI planning versus manufactur-
interesting, MACHINE must be, in fact, considered as ing-operation planning: a case study. IJCAI-95, 1995. p. 1670}6.
a step forward in this direction and it still can be ex- [18] Park SC, Gervasio MT, Shaw MJ, DeJong GF. Explanation-
tended with some kind of arithmetic reasoning, or man- based learning for intelligent process planning. IEEE Trans Sys-
agement of uncertainty for incorporating sensing tems Man Cybernet 1993;23(6):1597}616.
information into the building of Grafcet charts with [19] Tsatsoulis C, Kashyap RL. Case-based reasoning and learning in
manufacturing with the TOLTEC planner. IEEE Trans Systems
conditional branches. Man Cybernet 1993;23:1010}23.
Since the results obtained are very good and, in fact, [20] Participants of the Workshop. AAAI Special Interest Group in
they could be directly applied in many cases achieving Manufacturing Workshop: Arti"cial Intelligence and Manufac-
sound results and very fast, these topics will receive turing, State of the Art and State of the Practice. Discussion track,
further study with the goal of achieving a full applicabil- 1998.
[21] Mizutani H, Nakayama Y, Ito S, Namioka Y, Matsudaira T.
ity of the results automatically obtained by MACHINE. Automatic programming for sequence control. In: Scott AC, Kahr
P, editors. Innovative applications of arti"cial intelligence. AAAI
Press, 1991. p. 315}31.
[22] Klein I, Jonsson P, Backstrom C. E$cient planning for a minia-
References ture assembly line. Artif Intell Engng 1998;13(1):69}81.
[23] Aylett R, Soutter J, Petley G, Chung P. AI planning in a chemical
[1] IMS. Intelligent Manufacturing Systems, an international re- plant domain. European Conference on Arti"cial Intelligence,
search and development program. http://www.ims.org/, 1995. 1998.
[2] AAAI. American Association for Arti"cial Intelligence, special [24] Soutter J. An integrated architecture for operating procedure
interest group on manufacturing. http://sigman.cs.umn.edu/, synthesis. PhD Thesis, Loughborough University, 1996.
1996. [25] Castillo L, GonzaH lez A. A nonlinear planner for solving sequential
[3] PLANET. European Network of Excellence in Arti"cial Intelli- control problems in manufacturing systems. In: Coelho H, editor.
gence Planning, technical coordination unit in intelligent manufac- Progress in arti"cial intelligence, Lecture Notes in Arti"cial Intel-
turing. http://odl.education.salford.ac.uk/dan/pims/home.html, ligence. Berlin, Springer, 1998.
1998. [26] Castillo L, Fdez-Olivares J, GonzaH lez A. Automatic generation of
[4] Luger G, editor. Arti"cial intelligence and manufacturing re- control sequences for manufacturing systems based on nonlinear
search planning workshop: state of the art and state of the planning techniques. Artif Intell Engng 1999, to appear.
practice. AAAI Press, 1998. [27] Castillo L, GonzaH lez A. MACHINE: a model of action for multi-
[5] Allen JF, Hendler J, Tate A. Readings in planning. Los Altos, CA: agent domains. European Conference on Planning ECP-97, 1997.
Morgan-Kaufmann, 1990. Electronically available at http://decsai.ugr.es/&1cv.
[6] Nilsson NJ. Principles of arti"cial intelligence. Palo Alto, CA: [28] Fikes RE, Nilsson NJ. STRIPS: a new approach to the applica-
Tioga, 1980. tion of theorem proving to problem solving. Artif Intell
[7] Russell S, Norvig P. Arti"cial intelligence: a modern approach. 1971;2:189}208.
Englewood cli!s, NJ: Prentice-Hall, 1995. [29] Castillo L, Development and application of nonlinear planning
[8] ANSI-ISA. Batch control Part 1: models and terminology (SP- techniques for control programming in manufacturing systems (in
88). Instrument Society of America (ISA), 1995. spanish). PhD Thesis, Universidad de Granada, 1998. Electroni-
[9] IEC. Preparation of function charts for control systems. Technical cally available at http://decsai.ugr.es/&1cv.
Report IEC-60848, International Electrotechnical Commission, [30] Weld D. An introduction to least commitment planning. AI Mag
1988. 1994;15(4).
[10] Green C. Application of theorem proving to problem solving. [31] McAllester D, Rosenblitt D. Systematic nonlinear planning.
IJCAI, 1969. p. 741}7. AAAI-91, 1991. p. 634}9.
[11] Manna Z, Waldinger RJ. Knowledge and reasoning in program [32] Fisher T. Batch control systems: design, application and imple-
synthesis. Artif Intell 1975;6:175}208. mentation. Instrument Society of America (ISA), 1990.

Das könnte Ihnen auch gefallen