Simulation of Block Assembly Process in Shipbuilding by Petri-nets

O. P. Sha (Professor, Department of Ocean Engineering and Naval Architecture, IIT Kharagpur, India)
S. C. Misra (Professor, Department of Ocean Engineering and Naval Architecture, IIT Kharagpur, India)
Ashish Gupta (Graduate student. Department of Ocean Engineering and Naval Architecture, IIT Kharagpur, India)

Shipbuilding Industry has undergone major changes in the past two decades. Maritime sector has grown rapidly and
so has the competition. In order to survive in this competitive environment shipyards have to organize and utilize
their resources at optimum capacity and efficiency. Planning and management of resources is a very crucial aspect
of shipbuilding industry. There are many constraints on production planning in a factory. The most efficient
production schedule should be planned taking these constraints into account. Professional skill is necessary to plan
the best schedule. A scheduler must deal with a large amount of information on product planning, so as to evaluate
the best schedule. In order to design efficient factory model, which is capable of utilizing its resources to the full
extent, it is necessary that we understand various types of flow taking place in a manufacturing factory. Production in
shipbuilding can be classified from the viewpoints of product flow and job flow.

To improve efficiency of manufacturing factory it is necessary that the flow should be smooth. To simulate the flow
of product a simulation package called Petri-nets [4] is used. This software has the ability to simulate various
dominated system in shipbuilding process. It is very tedious job to simulate the complete ship fabrication in a single
model so simulation is done by dividing the ship into smaller blocks which can then be simulated individually.
Simulation of the block assembly process improves understanding of the manufacturing process and also helps in
determining hindrances like bottlenecks in the assembly line flow.

In the present work, a midship section (Nomenclature Block 04 here) of a container vessel weighing 212 tonnes is
used as a test to simulate the block assembly process. The simulation exercise also estimates the down-hand and
overhead welding requirements at different stages of the block assembly process Block 04. Following shipbuilding
practices, Block 04 divided into smaller sub-assemblies and each one the sub-assemblies are simulated individually.
Further extending the concept of simulation to actual shipbuilding practices, simulation for two and than three
blocks is carried out with appropriate phase lag. The approximate total time required by the Block-04 for fabrication
by this simulation exercise is presented along with occurrence and location of bottlenecks.
The paper also discusses how the simulation of the two and three block assembly process exercise is helpful in
locating the bottle-necks over longer time span of the production process and at what rate the blocks can be accepted
at the erection stage of shipbuilding. In order to achieve faster production it is necessary to reduce the occurrence as
well as time span of bottlenecks. Simulation of multiple block assembly process also gives an idea whether the
present assembly area is adequate to meet the production schedule.

KEY WORDS department proceeds with the design. At this stage, the designer
Shipbuilding; Assembly; Block Assembly; Petri-net; bases his calculations on the functional system of the ship. Here
Simulation. the design and drawings are produced to satisfy the owners
requirements, classification society and other statutory bodies.
INTRODUCTION This is followed by a transition stage where information
Ship production is a complex process and cannot be classified as generated for functional groups is mapped in terms of
a typical manufacturing industry in the strict sense because of constructional groups based on the constraints imposed by
high customization. It is more often a one of kind production resources and production practices of the shipyard. Resource
and therefore has significant differences from a mass production requirements in terms of manpower, facilities and due date of
based industry. Also in the recent past, shipbuilding has completion of each construction group is calculated which in
gradually evolved from a systems based infrastructure to a turn can be related to the total material requirements and the
product and process based infrastructure. cost of the product.

The main phases of the product life cycle of ships are concept Today in the world of advanced computing, there are
/tender design; preliminary/contract design; detailed design; comparably fast computers which have facilitated the
production; trials and commissioning; operations (service); and development of several software packages on planning and
disposal (scrapping). On obtaining the contract, the design management. Many manufacturing systems have been

WMTC 2009 Last name of first author Page number

developed and are used to deal with a large amount of available during the design development. The initial
manufacturing information efficiently. Various design systems, build strategy addresses the whole ship and then an
such as computer-aided design systems (CADs) have been interim product.
developed and are used to generate production information. In Scheduling decides the time order in which the jobs
addition, computer-aided engineering systems (CAEs) have are to perform so as to complete the production process
been used aggressively to evaluate information on the designed
products. Recently, information technology has given designers Resource Allocation decides the material and
a work environment so they can plan the production more manpower requirements at each stage in the production
easily. In fact, to evaluate the possibility and effectiveness of the process
generated schedules, the professional skills, experience, and
3. Materials Planning and Procurement: Information from
intuition of a schedule planner are necessary. A scheduler must
production engineering and production planning provide the
deal with a large amount of information on product planning, so
basis for materials planning. Procurement of major and long
as to evaluate the best schedule. In order to design efficient
lead time machinery and equipment get decided at this
factory model which is capable of utilizing its resources to the
stage. Drawing and work instruction are generated for
full extent it is necessary that we understand various types of
subcontracted work.
flow taking place in a manufacturing factory.
4. Production Control: At this stage, the progress of the actual
SHIPBUILDING production process is monitored and necessary feedback is
Shipbuilding is an assembly process involving hundred of provided to production planning when the manufacturing
thousands of individually prefabricated parts and items of deviates from the plan so as to take corrective measures in
machinery, equipment and outfit. The productivity of the ship the production planning.
production process depends on how efficiently one can assemble
them together. Figure 1 shows a flow diagram of a typical In modern shipbuilding, product-oriented work breakdown
modern shipbuilding process. Modern shipbuilding practice structure classifies the ship on the basis of its interim
requires that larger proportions of the man-hours be spent products/blocks. Group technology concepts have been
planning for production. This is to ensure that not only less time successfully used in integrating the interim blocks with zone
is needed to build a ship, but also more ships can be built from a outfitting and zone painting [2]. Hull Block Construction
single position (dry dock). Economically, this can be justified as Method forms the basis by which the ship is divided and
return on investment on modern facilities (dry dock and subdivided into blocks (Figure 2). These blocks form the basis
block/grand block assembly areas) can be recovered by for control of the production process and also have an impact on
spreading the fixed cost over a large number of ships. the Zone Outfitting Method and the Zone Painting Method. The
size, shape and the number of blocks will determine the
As seen in Figure 1, the ships are now built by assembly of subsequent nature and work package of the interim products and
blocks or grand blocks having reasonable outfit content. thus will have a large influence on the productivity of the yard.
Therefore, the role of a modern design department would now The division of the ships into manageable blocks should be such
include the following four tasks: that:
1. Design and Engineering: The traditional task of the design The blocks are simple and have logical boundaries.
department was functional design which includes the The number of blocks should be minimised, if
detailed specification of all structure, material and necessary by joining the blocks, so as to exploit the
equipment, meeting all the relevant regulatory largest capacity crane at the erection site. There may be
requirements. The additional task now is production sometimes an additional manufacturing level in which
engineering which aims to map the functional groups into the blocks are joined together into grand blocks.
constructional groups. This involves the development of all The shape of the blocks should be such that they have a
the detailed drawings, sketches, instructions and other stable configuration and can stand on their own with
documentation needed by the shipyard to build the ship. minimum or no external support.
2. Production Planning: Based on the information obtained An attempt must be made to minimise scaffolding,
from production engineering there are three steps in lifting and turning of the blocks during its fabrication
production planning: process.
Build Strategy is a document that attempts to capture The blocks should be configured such that the various
design, planning, production methods and practices for sub-assemblies that make up to form the block can be
a specific task into an integrated system. The assigned to one of a minimum number of work package
document spells out objectives, constraints, potential groups considering similarities in problem area and
difficulties, and other information that needs to be need to minimise variations in working times.
developed up front before the main task can
Identify assembly and erection process of the blocks
commence. The production engineer develops and
consistent with safety and need for block accuracy and
refines the build strategy as more information is
rigidity.

WMTC 2009 Last name of first author Page number

 The blocks should provide maximum accessibility, hierarchical framework so as to optimise the hull structural
both in terms of area and volume, for on-block production process. The number of hierarchical divisions may
outfitting and on-block painting so as to: vary depending on the shipyards facilities and the type of ship
install machinery and other components in the it is building. The interim products manufactured at each level
engine room in the hull block construction method (except for the Grand
Block and Erection Level) are examined for similarities in their
arrange deck machinery, mooring gear, piping, production aspects. Accordingly, the interim products at
fittings, etc., and different manufacturing levels are grouped so as to:
perform as much painting as possible before further modularise the production process
erection of the block justify the presence of expensive but highly efficient
facilities achieve manpower savings
The hull block construction method can have a 7-tier

Figure 1: Stages in Shipbuilding Process [1]

SHIPBUILDING PROCESS AND PETRI-NETS product. Ideally this activity should be avoided or
In shipbuilding there are three important production activities, minimized, but sometimes it is necessary to incorporate this
namely : so as to have a smooth product flow.
1. Operational Activity: In this activity the form of the product Ship production process can be viewed from two aspects:
changes during each production stage, e.g. cutting, welding product flow and job flow [3]. Product flow of an interim
and fabrication of sub-assembly, assembly, block assembly, product can be defined as logical and sequential movement of a
erection, etc. product through different production stages, whereas the job
2. Transportation Activity: In this activity only the position of flow is the sequence of jobs at each assembly stage. For interim
product changes. Products such as piece-parts, block production, the activities under product flow are sub-
subassemblies and assemblies are transported between assembly, assembly, block assembly and erection, whereas the
various production stages. job flow activities will mainly include fitting, tack welding,
welding and inspection.
3. Storage Activity: In this activity neither form nor position of
product changes. This activity controls the progress of the The production of an interim product (block) at different stages

WMTC 2009 Last name of first author Page number

can be viewed as a series of cells each having the above 2. Transition: It shows the starting and ending of activity or
mentioned production activities. The product flow and job flow we can say duration of activity. Transitions are shown by
at all stages must be decided so as to ensure high production rectangles.
efficiency. 3. Token: It shows the movement of a product. Token
(modeled as a product) is shown by big black dot.
Simulation of ship building process provides deep insight of the
fabrication and assembly process. The hull block construction 4. Arcs: The arcs show the movement of tokens (i.e. product
method divides the ship into manageable blocks using the top- flow). There are two types of arcs.
down approach. The simulation of the complete ship can be seen Input arc: These arcs start at place and end at transition
as sum of simulation of the individual blocks. Simulation of Output arc: These arcs start at transition and end at place.
each block is further divided into sub-assemblies like flat panel,
Petri-nets have undergone various changes in the past few
curved panel etc. Each sub-assembly is simulated individually
decades; it has been modified at various levels to suit modeling.
and finally all these sub-assemblies are combined into a single
By adding more attributes or parameters of modeling it is
model with suitable time lag given to each sub-assembly
become easier to model the systems more efficiently. In
resulting in the simulation of complete block production
shipbuilding, time is an important parameter and in this work
process.
Timed Petri-nets has been used wherein time is the additional
A Petri-net (also known as a place/transition net or P/T net) is parameter in Petri-net modeling [5, 6].
one of several mathematical representations of discrete As we know, a shipyard has to deal with high level of
distributed systems. As a modeling language, it graphically uncertainty and complexity during production. Activities in
depicts the structure of a distributed system as a directed shipyard are dependent on each other so if there discrepancy in
bipartite graph with annotations. Petri-net was invented in 1962 one activity it will effect the whole production. Shipyard
by Carl Adam Petri [4]. The four important components of petri- activities can be divided into sequential, parallel, dependent and
net shown in Figure 4 are: limited resource activities. Petri-net is suitable tool for modeling
1. Places: It shows the state of activity and they are these type of activities efficiently.
represented by a circle.

HULL ERECTION
M
A
N GRAND BLOCK
U JOINING
F
P A BLOCK ASSEMBLY
L A
A C
N T
N U SEMI BLOCK
I R ASSEMBLY
N I
G N
G SUB BLOCK
L ASSEMBLY
E L
V E
E V PART
L E ASSEMBLY
L

PART FABRICATION

Figure 2 Hull Block Construction Method [2]

WMTC 2009 Last name of first author Page number

 Figure 3 Division of Ship into Blocks [3]

Place
Input arc Output arc
Token

Transition

Figure 4: Model of an Activity by Petri-net

REPRESENTATION OF SHIPBUILDING as the state of an activity, the scheduler is made aware of the
PROCESS BY PETRI-NETS state of an activity by a tokens position. In addition, a history of
In order to model an activity it is necessary to define activitys activities is acquired by collecting histories of tokens
parameters in terms of Petri-nets parameters. To simulate movements.
product flow it is necessary to define each and every activity in
As mentioned earlier, there are three major activities in
job shops and movement of product through these job shops.
shipbuilding, i.e. Operational, transportation and Storage. The
Representations of event and activity: There are two types of principle behind modeling of these activities using the Petri-net
events in production activities. One is an event at the activity components (transition, place and token) is briefly explained
start and the other is the event at the activity end. Activity is below:
denoted by Transition which shows the duration of an activity.
Operation activity. Operation activity is produced when
The state of the activity that exists between activity is the products are carried in the workplace. Product flow
modeled by place, as is the condition of an event. Figure 4 between workplaces is represented by transaction firings
illustrates the connection of transition with place. To simulate of transporting in and transporting out of the
production, it is necessary to consider the passage of time during workplace.
the activity. Hence, a time Petri-net is introduced. A place Transportation activity: Transportation activity is used to
modeled as the state of activity holds reign until a specified time move a product in and out of workplaces. The activity
passes. begins by placing a product on transporting equipment.
Storage activity. Storage activity is used for keeping a
Representation of production. A product is modeled as a token
product in a storage space intentionally to produce
and therefore product flow is represented by a tokens
efficient production activity. This activity is modeled by
movement. Information on the product model of parts, sub
the Petri-net, similar to operation and transportation
assembled modules, and so on is linked to a token. A token
activities, with consideration of the limited storage space.
moves from one place to another. Because the place is modeled

WMTC 2009 Last name of first author Page number

Software used for simulation of the block assembly process is Simulation for all the sub-assemblies is carried out using
HPSim. HPSim was developed to support the design and HPSim. For identical sub-assemblies the simulation is done
simulation of Petri-nets, both in a graphical and intuitive only. Figures 8, 9, 10 and 11 shows the simulation for the sub-
manner. The software features Place/Transition Nets, Stochastic assemblies Block 04-01, Block 04-02, Block 04-03 and Block
Petri-nets and Petri-nets with time. The actual simulation is 04-06 respectively. Appropriate time for marking, cutting and
visualized as a Token Game Animation. This can be executed in welding including the variation of time depending on welding
single step or continuous mode. Besides that, a fast forward type and position have been incorporated in the model.
mode is also available, in which the graphical representation of
the token is not synchronized with the simulated actual position. Figure 12 shows the simulation model for the complete Block
When transition is yellow it shows that token/product is 04.
under certain activity. When transition turns green it shows the
activity is over, token is ready to fire or product is ready to A simulation study was carried out using this model to
move to the next stage. Time of activity is fed into transition a understand the ship production process assuming two identical
separate parameter. A status bar below keeps track of time as the (Block-04 and Block 05) and three identical Blocks (Block-04,
simulation proceeds. When simulation is over it shows the total Block-05 and Block-06) of the similar configuration as Block-
time elapsed which is the total time for assembly. Similarly, 04. The simulation shown in Figure 13 shows the movement of
there are counters in transition and places which keeps track of sub-assemblies through various shops during fabrication of two
number times of certain activity and number of items undergone identical Blocks (Block-04 and Block-05). In modeling the
that activity. Simulation speed is flexible which can be increase simulation of two blocks (Fig. 13) two assembly lines are
or decrease if required. It also maintains a XL Data Sheet for provided.
marking and firing request.
Simulation of assembly of three identical Blocks (Block-04,
SIMULATION OF BLOCK ASSEMBLIES Block-05 and Block-06) is show in Figure 14. In modeling the
A test simulation is done for a Block 04 weighing 212 tones of a simulation of three blocks (Fig. 14) only two assembly lines are
container vessel. This block is located amidships and shown in provided. Such studies help us identify whether the assembly
Figure 5. line is of adequate capacity or needs expansion. In these models,
stockyard, plate treatment shop and the prefabrication areas are
For simulation of block assembly process, the Block 04 has shown. The prefabrication area is further divided into curved
been further broken down into sub-assemblies of smaller size panel area, flat panel area and floor preparation area. The
and weight. The sub-assemblies of Block 04 are shown in fabrication of 2D curved sections takes place in the curved panel
Figure 6 with the nomenclature P and S to identify the port area.
starboard side blocks. The assembling order is determined after
considering the relation between sub-assemblies. The HPSim keeps a token count, which helps in determining the
assembling order and the network of activities for Block 04 are occurrence of particular activity. This allows us to determine
shown in Figure 7. how much resource is consumed for a particular activity and
how much additional resource is required further for completion
of this activity.

Figure 5: Division of Ship into blocks, Block 04 Marked with red.

WMTC 2009 Last name of first author Page number

 Block 04-01P/S

Block 04-02P/S Block 04-03P/S , Block 04-04P/S and Block 04-05P/S

Block 04-06P/S

Figure 6: Breakdown of Block 04 into sub-assemblies

WMTC 2009 Last name of first author Page number

 Block 04 Op. Block 04

Block 04-01 Op. Block 04-01 Op. Block 04-02 Op. Block 04-03 Op. Block 04-05
1

Op. Block 04-04

1 Block 04-01P
Op. Block 04-06

2 Block 04-01S Op. Block 04-01 Op. Block 04-05

2 Block 04-02 1 Op. Block 04-01P 1 Op. Block 04-05P

2 Op. Block 04-01S 2 Op. Block 04-05S

1 Block 04-02P
Op. Block 04-02

2 Block 04-02S 1 Op. Block 04-02P

3 Block 04-03 2 Op. Block 04-02S

Op. Block 04-03/04/06

1 Block 04-03P

1 Op. Block 04-03P/04P/06P

2 Block 04-03S
2 Op. Block 04-03S/04S/06S

4 Block 04-04

1 Block 04-04P

2 Block 04-04S

5 Block 04-05

1 Block 04-05P

2 Block 04-05S

6 Block 04-06

1 Block 04-06P

2 Block 04-06S

Figure 7: Assembly Order and Network of Activity for Block-04

WMTC 2009 Last name of first author Page number

 Figure 8: Simulation model for assembly of Block 04-01P

Figure 9: Simulation model for assembly of Block 04-02P

WMTC 2009 Last name of first author Page number

 Fig 10: Simulation model for assembly of Block 04-03P

Fig 11: Simulation model for assembly of Block 04-05P

WMTC 2009 Last name of first author Page number

 Figure 12: Simulation model for final assembly of Block 04

WMTC 2009 Last name of first author Page number

 Figure 13: Simulation model for assembly of two consecutive Blocks 04 and 05

Figure 14: Simulation model for assembly of three consecutive Blocks 04, 05 and 06

WMTC 2009 Last name of first author Page number

 In this work, an attempt has been made to show how Product
RESULTS AND DISCUSSION flow can be simulated by the Petri-net. In shipbuilding,
Using this simulation model, the total time taken for the activities are dependent on each other to large extent so each
fabrication of the 212 tonnes Block-04 was 171 man-hours. The and every job shop has to work efficiently to carry out whole
focus of this exercise was to detect bottlenecks that occurred at production smoothly and efficiently.
the sub-assembly and assembly stages. In the above simulation
exercise bottlenecks were detected in the assembly of the Sub- The main advantage of this simulation model is that it allows the
block-04-01(P/S), Sub-block-04-02(P/S) and Sub-block-04- modeling of blocks of different configurations, but having the
05(P/S). In the Sub-block-04-01(P/S), the largest delay took same fabrication sequence which is the case in modern ship
place as floor fabrication took a large time, whereas for Sub- production. Simulation of two blocks and three blocks give the
block-04-02(P/S) and Sub-block-04-05(P/S) the side shell idea about phase difference between consecutive block
fabrication where bending of plates and stiffeners and their were fabrications and determines at what rate blocks can be accepted
involved caused the delay. Therefore, to improve production in the erection yard. With duplication of assembly line the rate
rate it is necessary to remove the bottleneck at the Sub-block- of fabrication increases but this may become constant after
04-01(P/S) level. setting up 2, 3 or more assembly lines for particular set of
resources. Assembly of 2 and 3 blocks shows the bottlenecks in
The total fabrication time for two blocks (Block-04 and Block- long run of fabrication process, arrival or departure material at
05) simulation was 380 man-hours with Block-04 getting various sub-assembly and assembly areas, critical paths and
completed after 217 man-hours and Block-05 at the end of 380 dependencies of job shops among each other.
man-hours. In this simulation exercise also the main bottleneck
was detected in the assembly of the Sub-block-04-01(P/S) due
to the larger time required for fabrication of floors. After elapse
of 102 man-hours bottlenecks were detected at Sub-block-04-
02(P/S) due to unavailability of resources. The fabrication time REFERENCES
for Block-05 was less than Block-04 as assembly of Sub-block- [1]. Shin, D. S., Design of Greenfield Shipyard, International
05-02(P/S) started as soon as Sub-block-04-02(P/S) got over. Seminar on Shipbuilding Opportunities in Gujarat,
The total fabrication time for three blocks (Block-04, Block-05 Ahmedabad, Gujarat, India, 29-30 September, 2007
and Block-06) simulation was 525 man-hours with Block-04 [2]. Storch, R. L., Bunch, H. M., Moore, R. C. and Hammon, C.
getting completed after 217 man-hours, Block-05 at the end of P., Introduction to Shipbuilding, SNAME, USA, 1995
380 man-hours and Block-06 at the end of 525 man-hours. In
this simulation exercise, fabrication of Block-06 took less time [3]. Aoyama, K., Nomoto, T. and Watanabe, K., Development
than that for Block-05 as because when the assembly of Block- of a Shipyard Simulator based on Petri nets, Journal of
06 started the Sub-block-06-02(P/S) was over. Marine science and Technology, Vol. 4, pp. 35-43, 1999.
Due to unavailability of data, in all the above simulation [4]. Petri, C., Kommunkation mit Automation, PhD Thesis,
exercises only the fabrication and assembly times for the steel University of Bonn, Bonn, Germany, 1962.
piece parts and sub-blocks were considered. If zone outfitting [5]. Barad, M., Timed Petri nets as a verification tool,
method and zone painting method are to be incorporated, Petri- Department of Industrial Engineering, Faculty of
net for outfitting and painting are to be modeled and Engineering ,Tel Aviv University, Ramat Aviv, Tel Aviv
incorporated at the appropriate level in the above models. 69978, Israel.

CONCLUSION

WMTC 2009 Last name of first author Page number