Product Lifecycle Management (PLM) System for Shipbuilding Industry Automation of the process of conversion of Engineering BOM to Manufacturing BOM S.Abhilash Sharma Final year Dual Degree student, Dept of Ocean Engineering and Naval Architecture, IIT Kharagpur, Kharagpur, India O.P.Sha Professor, Dept of Ocean Engineering and Naval Architecture, IIT Kharagpur, Kharagpur, India Rajiv Sharma Assistant Professor, Dept of Ocean Engineering, IIT Madras, Chennai, India Abstract— PLM (product life-cycle management) has become something like „a magic wand‟ for various industries because of its capability to integrate different product modules via online network through the product‟s complete life-cycle, and hence providing one window access; thereby making the whole processes of product conception, design and manufacturing, delivery, maintenance and disposal integrated with a reduction in product development time and cost. However, heavy industries (i.e. shipbuilding) are different from consumer product industries because of high customisation in design process, and engineering software, widely varying scales of operations and less compatibility between different design and production processes, e.g. ship production is planned in activity driven network scheduling system in general, and is assumed more as a construction process or assembling process rather than a production process. One of the key elements of PLM is the Product Data Model (PDM), which is also termed as the Bill of Material in many shipyards. Bill of Material is a list of all the materials used to make the product. Different requirements will require different BOMs to be made by different departments. In shipbuilding industry a major problem is developing the manufacturing BOM (M-BOM) from the engineering BOM (E-BOM). This is because the E-BOM is structured in a “systems based” manner to suit the designer, whereas the M-BOM has to follow a “block/zone” based hierarchy as the shipbuilding process is an assembly of intermediate products. At present the conversion of E-BOM to M-BOM is done manually by utilizing the experience of shipyard personnel. Automation of this process will lead to considerable decrease in the design process time and hence in overall delivery time too. In this paper we present the development and the basic building concepts for a PLM system for shipbuilding industry and a case study in “Automation of Conversion of E-BOM to M-BOM” Keywords; Engineering BOM (E-BOM); Manufacturing BOM (M- BOM); BOM conversion; PLM Module; E-BOM to M-BOM I. INTRODUCTION In the present era where the demand for new ships is on the rise, shipyards need to be competitive. To maintain their competitive edge over the others, shipyards are continuously trying to produce more economical ships within less period of time. In this endeavour, most of the shipyards are now relying on product life-cycle management (PLM) systems. A PLM system has been in existence for a long time, in most of the manufacturing sectors and primarily acts as the common platform between the various software resources being used by the industry and allows for easy exchange of data across these systems. Designing is primarily a decision-making process in which the designer decides the various aspects and attributes of the product. Integrating the various processes and managing the resources of an industry across a common platform in real time helps the designer make more informed decisions and thus leads to an effective design in lower time. But in light of the shipbuilding industry, PLM has more requirements than a general manufacturing industry. This difference arises primarily due to the nature of the shipbuilding industry, which is quite different from a general manufacturing industry. Unlike an assembly line production prevalent in most of the manufacturing segments, shipbuilding industry, owing to its high level of customization is forced to follow a unit assembly production. Thus the PLM system for the shipbuilding industry should be able to integrate the processes and resources of different product types. Similarly, the design work in the shipbuilding industry is iterative where most of the work needs to be repeated to get a detailed design. Because of this nature of iterative work, the design process of the ship takes a lot of time. At present, many of the shipyards are trying to use the data of the previous vessels built to reduce the iterative work of design. This is also
12
Embed
A PLM System for Shipbuilding Industry · shipyards. Bill of Material is a list of all the materials used to make the product. Different requirements will require different BOMs to
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Product Lifecycle Management (PLM) System for
Shipbuilding Industry Automation of the process of conversion of Engineering BOM to Manufacturing BOM
S.Abhilash Sharma
Final year Dual Degree student,
Dept of Ocean Engineering and
Naval Architecture,
IIT Kharagpur,
Kharagpur, India
O.P.Sha
Professor,
Dept of Ocean Engineering and
Naval Architecture,
IIT Kharagpur,
Kharagpur, India
Rajiv Sharma
Assistant Professor,
Dept of Ocean Engineering,
IIT Madras,
Chennai, India
Abstract— PLM (product life-cycle management) has become
something like „a magic wand‟ for various industries because of
its capability to integrate different product modules via online
network through the product‟s complete life-cycle, and hence
providing one window access; thereby making the whole
processes of product conception, design and manufacturing,
delivery, maintenance and disposal integrated with a reduction in
product development time and cost. However, heavy industries
(i.e. shipbuilding) are different from consumer product industries
because of high customisation in design process, and engineering
software, widely varying scales of operations and less
compatibility between different design and production processes,
e.g. ship production is planned in activity driven network
scheduling system in general, and is assumed more as a
construction process or assembling process rather than a
production process.
One of the key elements of PLM is the Product Data Model
(PDM), which is also termed as the Bill of Material in many
shipyards. Bill of Material is a list of all the materials used to
make the product. Different requirements will require different
BOMs to be made by different departments. In shipbuilding
industry a major problem is developing the manufacturing BOM
(M-BOM) from the engineering BOM (E-BOM). This is because
the E-BOM is structured in a “systems based” manner to suit the
designer, whereas the M-BOM has to follow a “block/zone” based
hierarchy as the shipbuilding process is an assembly of
intermediate products. At present the conversion of E-BOM to
M-BOM is done manually by utilizing the experience of shipyard
personnel. Automation of this process will lead to considerable
decrease in the design process time and hence in overall delivery
time too.
In this paper we present the development and the basic building
concepts for a PLM system for shipbuilding industry and a case
study in “Automation of Conversion of E-BOM to M-BOM”
Drawings of the Blocks E81, B51, B12 (adjacent blocks –
Figure 11), Other documents (as per requirement) etc.
Thus when a designer logs into his system the PLM system
provides him with the requisite reference data. The next
activity is the initiation of the CAD program where the
designer would prepare E11 block diagram, E11 nesting plan,
E11 hole plan, E11 outfitting etc. Once the designer has
submitted the drawings, the PLM system would automatically
direct the drawings for inspection which would verify the
drawings. Once the inspection is over the final drawings need
to be updated in the product data model. This updating the
product data model is an automatic process and needs no
manual intervention. Thus the next process calling upon the
data for Block E21 would also include the drawings of Block
E21.
Thus the Process Management Module not only defines a
task but also assigns the processes of a task, controls the
product data and defines the logic of the processes involved.
The BOM management system and Process Engine
together constitute the PLM system which effectively helps in
reducing the design time as well as design cost. The next
section proposes a module for PLM system which would
streamline the process of conversion of E-BOM to M-BOM
which is presently done manually by experienced personnel of
a shipyard.
III. AUTOMATION OF CONVERSION OF ENGINEERING BOM
TO MANUFACTURING BOM
A. Problem Statement
Shipbuilding is primarily an assembly activity which
involves the assembly of various components to form interim
products and the assembly of these interim products yields the
final ship. The production is therefore product oriented or in
other words focuses specifically on a particular interim
product at all times. Hence the manufacturing personnel need
a BOM which is structured in a product oriented hierarchy.
On the other hand the design involves breaking up the ship
into its constituent systems and designing each of these
systems and their components individually. A designer at all
times would be working on a specific system of a ship viz.
hull, outfitting, piping, ballast water systems, fresh water
systems, electrical cables, HVAC etc. Therefore, a designer
requires the BOM to be oriented in a system based hierarchy.
In terms of the data, both the engineering BOM and the
manufacturing BOM share the same components at the lowest
levels of hierarchy. Essentially a plate or a stiffener is a part of
the hull structure from a designer‟s point of view. At the same
time it is a part of a particular block too. Thus, the components
in both the BOMs are same with the difference only in the
grouping of these components and their hierarchy. Thus
conversion of engineering BOM to manufacturing BOM is
essentially re-grouping the components in the engineering
BOM into interim products keeping in mind the assembly
process. Thus the problem statement of automation of the
conversion of engineering BOM to manufacturing BOM is
primarily a re-grouping problem.
Automation of this conversion process will ensure lesser
design time as compared to the manual process which takes a
long time. It will also lead to lesser cost as this process saves
on the manpower being utilized for the designing process.
B. Solution Approach
As already mentioned before, the solution lies in providing
an algorithm which can efficiently re-group the components of
the engineering BOM into the groups of the manufacturing
BOM. The algorithm must however decide on how to group
the various constituents and how many groups are to be
formed. A solution to a similar problem in a manufacturing
sector was proposed by Wang and Li [4] [5]
. They developed an
algorithm to decide the assembly groups of a product on the
basis of assembly sequence.
Consider a product which has eleven components, each
numbered from 1 to 11. The physical connectivity between
each of these 11 components is shown in Figure 13. This
figure is known as the connectivity graph of the product.
Each of the nodes represents a component and the link
between two components denotes the “physical” connectivity
between the two components. Thus, from the figure we can
see that component 5 and 6 are connected to each other
physically but component 1 and 10 are not connected to each
other, even though all the 11 components are a part of the
same product.
Once the connectivity graph has been obtained, a
connectivity matrix for the same is defined. The connectivity
matrix will always be a symmetric matrix, with all the
diagonal elements as zeros. This is because no component is
assumed to be connected to itself. Connectivity between two
nodes is represented by 1 while the absence of connectivity is
represented by 0. Each node necessarily must have a
connection to at least one other node in the graph i.e. the
matrix cannot have any row or column with only zeros.
Let the connectivity matrix be represented by M. If there
are n nodes in the connectivity graph, the size of M would be
n x n. For representing the connectivity between node 5 and
node 6 the value of M (5, 6) and M (6, 5) are both set to 1.
(5 , 6 ) (6 , 5 ) 1M M (1)
Since there is no connectivity between node 1 and node 10,
the corresponding matrix values are set to 0.
(1,1 0 ) (1 0 ,1) 0M M (2)
The connectivity matrix for the connectivity graph shown
in Figure 13 would be:
(3)
The next step is the decomposition of the assembly into
sub-assembly groups. This process starts by finding the
articulation points in the connectivity graph.
An articulation point of a graph is defined as the node
which when removed will fragment the graph into two
separate sub-graphs. For example, consider the node 1 to be
removed from the graph and all the links with node 1 to be
broken. This will lead to two separate sub-graphs {5, 6} and
{2, 3, 4, 7, 8, 9, 10, 11} which do not have any common
nodes. Thus node 1 is an articulation point.
On the hand if node 2 is removed from the graph. It can be
seen that the connectivity throughout the graph still exists and
there are no sub-graphs. Thus node 2 is not an articulation
point.
In the given example there are four articulation points {1 4
7 8}. Based on the articulation points, the 11 components are
grouped to form sub-graphs of the original connectivity graph.
Each of the groups formed represents a sub graph of the
original connectivity graph and none of these groups can be
further decomposed further i.e. there are no articulation points
within the sub-graph. The sub-groups formed are:
Figure 13: Connectivity Graph
Group 1 = {1 5 6}
Group 2 = {1 2 3 4}
Group 3 = {4 7 8}
Group 4 = {7 10 11}
Group 5 = {8 9}
The hierarchy is shown in Figure 14. This division results
in the formation of a 3 level product hierarchy which is
product oriented.
In the actual manufacturing process, the product is divided
into smaller interim products on the basis of ease of division.
An interim product must have minimum connections to other
interim products to make it suitable for independent
concurrent assembly. A block which has a large number of
connections to other blocks would be difficult to assemble
independently as it shares many connections and managing
each of them is difficult. On the other hand if a block is
connected to only one other block, it is possible to assemble it
independently without much consideration for its connectivity
to other blocks. Since the articulation points in a graph are
single elements connecting two sub-graphs, the groups formed
by the algorithm share the same concept of least connections
to other sub-graphs. Thus, the groups formed by the algorithm
are similar to those formed on the basis of ease of division.
There is a minor difference which needs to be accounted for
while adopting this model for the Hull Block Construction
Method (HBCM). The connectivity in case of HBCM does not
indicate physical connectivity. Instead it indicates a structural
connectivity. The method described above finds the
articulation points in the connectivity graph where the
structure can be divided into separate parts. In case of the
structural components of a ship, the division into blocks is
done at a structural joint so that the assembly of these blocks
does not cause a structural weakness at the joint. For this
reason it is necessary that for the HBCM only the structural
connectivity be considered. The algorithm when applied to
structural members provides accurate results only for a
structural connectivity matrix. An example to differentiate the
Product
Group 3 Group 4 Group 5 Group 1 Group 2
1
5
6
1
2
3
4
7
8
7
10
11
8
9
4
Figure 14: Initial Product Oriented BOM
Figure 15: Panels of a typical bulk carrier
structural and physical connectivity has been provided in the
next section.
The structural connectivity is relevant only in the case of
HBCM. For other components such as the piping, outfitting,
HVAC ducts etc a physical connectivity is sufficient to
provide the product oriented hierarchy. This is because the
aim of the breaking up the piping and outfitting structures into
smaller zones does not have any structural issues related to it.
The next section provides an illustration where the present
model has been applied to relevant data for shipbuilding.
IV. ILLUSTRATIVE EXAMPLE
Consider the cargo hold of a typical bulk carrier shown in Figure 15. The problem is to find the different groups into which the components must be grouped as per the above approach and compare its closeness to the actual grouping done in the shipyards.
The typical grouping adopted by a shipyard is shown below in Table 2.
Table 2: Typical grouping from shipbuilding point of view
Sl. No. Group No. Component Elements
1 1 1, 2, 3, 4
2 2 4, 5, 7
3 3 6, 7, 8
4 4 8, 9, 10, 11, 12, 13
5 5 13, 14, 15
6 6 14, 16, 19
7 7 17, 18, 19, 20
For the sake of simplicity and limiting the BOM to only two levels, the assembly groups are assumed to be individual components. The individual assembly groups are numbered as shown in the Figure 15. There are 20 panels defined for this cargo hold. A uniform longitudinal extent is assumed for each of the panels shown. Similarly the floors connecting the various panels have not been included in the calculations as they would not be articulation points and would only be extra nodes connecting different panels.
The various panels under consideration have been listed in Table 3. To apply the approach suggested above, the first step
18 17
20
19 16 14
15
13
12 11
9 10
8
6
7 5
2 3
4
1
Articulation Point
Not an Articulation Point
Figure 16: Connectivity Graph for the cargo hold of the bulk carrier
is to define a connectivity graph between all the panels. Each of the panels would be a node in the connectivity graph. As already mentioned in the previous section, connectivity between these nodes is defined by a structural connectivity instead of physical connectivity. Thus there would not be any connectivity between panel 7 and panel 9 even though there is weld line connectivity between the two. On the other hand, panel 8 and 9 share a structural continuity and are considered to be connected in the connectivity graph. The final connectivity graph is shown in Figure 16.
On the basis of this connectivity graph the articulation points are evaluated. The violet nodes are the articulation points in Figure 16 and the green nodes are not the articulation points. On the basis of these articulation points the groups are formed. The various groups obtained by the application of the algorithm are listed in Table 4.
Table 3: List of panels of cargo hold under consideration
Sl. No. Panel No. Panel Description/ Location
1 1,20 Deck
2 2, 18 Upper Hooper Tank Side Shell
3 3, 17 Inner Upper Hooper Tank
4 4, 19 Tween
5 5, 16 Side Shell
6 6, 15 Lower Hooper Tank Side Shell
7 7, 14 Inner Lower Hooper Tank
8 8, 13 Longitudinal Girder Double Bottom
9 9, 11 Tank Top
10 10, 12 Bottom Shell
It can be seen that the groups generated by the algorithm (listed in Table 4) and the groups decided by a typical shipyard (listed in Table 2) are much in agreement. This shows that the algorithm is quite successful in generating most of the groups.
Table 4: The groups generated from the algorithm.
Sl. No. Group No. Group Elements
1 1 1, 2, 3, 4
2 2 4, 5
3 3 5, 7
4 4 6, 7, 8
5 5 8, 9, 10, 11, 12, 13
6 6 13, 14, 15
7 7 14, 16
8 8 16, 19
9 9 17, 18, 19, 20
V. CONCLUSION AND FUTURE WORKS
This method does not consider the assembly lead time into the grouping methodology. Thus even though the initial Manufacturing BOM has a product oriented hierarchy, it is not clear that this BOM will result in the minimum assembly lead time.
To overcome this problem one can try to perform a critical path analysis with the initial groups formed by the algorithm. Based on this analysis, decision can be taken to merge one or more groups to improve the assembly lead time of the product.
Similarly this method can be applied only to the lower most level of the BOM, because the assembly groups produced after the application of the Wang and Li algorithm cannot be further divided into smaller sub-assembly groups
An approach to overcome this problem might be in trying to apply the same method repeatedly over the groups formed. Thus once the initial groups have been formed out of the components, a similar process is repeated with the group as a whole being considered as a component for the next level. This way a hierarchy can be continued in a bottom to top approach till the final hierarchy has the entire ship in a single group.
Application of these developments would result into a manufacturing BOM which would provide the entire hierarchy and be efficient from the product assembly lead time perspective.
VI. REFERENCES
[1] M. Grieves (2005), „Product Lifecycle Management: Driving the Next Generation of Lean Thinking’ McGraw-Hill, 1st Edition, Oct 2005, pp 141-143
[2] HHI (2009), „Hyundai Heavy Industries Shipbuilding Division Selects Siemens PLM Software Technology to Implement Innovative Digital Shipyard‟, Siemens PLM Software News and Press Release
[3] Tae-wan Kim (2009), „Development of a PLM Model for Shipbuilding Industry‟Presentation in ISCSI 2009
[4] Chang, Sheung-Hung, Lee, Wen-Liang and Li, Rong-Kwei (1997), ‘Manufacturing bill og material planning’, Production Planning and Control, 8: 5,437-450
[5] WANG,H. P., and LI, J. K., 1992, Computer-Aided Process Planning, (Elsever Science, Amsterdam)