www.nature.

com/scientificreports

OPEN Process knowledge graph modeling

techniques and application
methods for ship heterogeneous
models
Jianwei Dong1, Xuwen Jing1, Xiang Lu2, Jinfeng Liu1*, Haipeng Li2, Xuwu Cao1,
Chenxiao Du1, Jun Li2 & Lei Li1
In the process design and reuse of marine component products, there are a lot of heterogeneous
models, causing the problem that the process knowledge and process design experience contained in
them are difficult to express and reuse. Therefore, a process knowledge representation model for ship
heterogeneous model is proposed in this paper. Firstly, the multi-element process knowledge graph is
constructed, and the heterogeneous ship model is described in a unified way. Then, the multi-strategy
ontology mapping method is applied, and the semantic expression between the process knowledge
graph and the entity model is realized. Finally, by obtaining implicit semantics based on case-based
reasoning and checking the similarity of the matching results, the case knowledge reuse is achieved,
to achieve rapid design of the process. This method provides reliable technical support for the design
of ship component assembly and welding process, greatly shortens the design cycle, and improves the
working efficiency. In addition, taking the double-deck bottom segment of a ship as an example, the
process knowledge map of the heterogeneous model is constructed to realize the rapid design of ship
process, which shows that the method can effectively acquire the process knowledge in the design
case and improve the efficiency and intelligence of knowledge reuse in the process design of the
heterogeneous model of a ship.

In order to face the competition from world-class shipyards, large shipyards must increase the application of
information ­technology1. The design process of ship assembly and welding technology is one of the key fac-
tors that determine product cost, quality and cycle. On the one hand, after years of accumulation, the relevant
knowledge in the field of ship process design has been very rich. However, as far as the extensive production
management mode of the shipyard is concerned, the engineering knowledge lacks effective summary and sorting,
and a large amount of knowledge is scattered in the local storage medium or in the minds of domain experts,
reusability and sharing are poor, so this also reflects how to effectively manage and reuse this knowledge to
achieve rapid design of assembly and welding processes has become a problem to be ­solved2. On the other hand,
in the process of ship design and construction, a large number of heterogeneous models containing assembly and
welding processes will be generated. If these heterogeneous models cannot be effectively expressed and managed,
then the heterogeneous models containing process design information will be regarded as implicit Knowledge is
stored in the database, it is difficult to be reused, and it is a huge waste for enterprises. Therefore, it is necessary
to study how to construct a unified expression model for heterogeneous models.
In order to overcome the above problems, the process knowledge graph modeling method oriented to het-
erogeneous models is proposed. First, the process knowledge graph is constructed by using heterogeneous
model information and assembly welding process knowledge. Then, the preliminary unified semantic represen-
tation is obtained, through the multi-strategy similarity ontology mapping method guided by the knowledge
graph. Finally, invisible semantics are obtained by case-based reasoning, and the similarity of matching results
is checked.
The rest of this article is arranged as follows: In “Literature review” section provides related work. In “Over-
views of the proposed method” section, the basic concepts of the design of this paper are introduced, and the
proposed methods are described. In “Knowledge model construction based on process knowledge graph” section,

1
School of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212100, People’s
Republic of China. 2China Merchants Cruise Shipbuilding Co., LTD., Nantong 226000, People’s Republic of
China. *email: [email protected]

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 1

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

a knowledge model based on the process knowledge graph and a case-based reasoning system is constructed
respectively. In “Case study and discussion” section, the effectiveness of the method is verified according to the
selected model. Finally, we discuss directions for future research and present conclusions about the proposed
method in “Conclusion and future works” section.

Literature review
Along with the continuous development of CAD system in the field of ship construction and the widespread
application, a large number of heterogeneous model and the associated process data and knowledge continu-
ously generated and stored in the knowledge base of the enterprise, and for the unity of the heterogeneous model
representation method is on the increase, our work inspired by recent progress in several different a­ reas3, such
as the knowledge map and knowledge reuse research, we will review below.

Knowledge graph. The concept of knowledge graph was formally proposed by Google in May ­20124. Its
original purpose was to improve the capabilities of search engines, improve the quality of search results and
enhance the user’s search experience. In essence, knowledge graph is a way of revealing the relationship between
entities. The semantic network can continue to formally describe things in the real world and their relationships.
After 2013, with the continuous development of intelligent information services and applications and the effec-
tive management of heterogeneous and dynamic data, knowledge graphs have received widespread attention in
specific fields such as academia and industry.
In academia, H ­ u5 and others studied the application of knowledge graphs in information science and intro-
duced the necessity, current situation, development trend, and knowledge graph receipt tools and method pro-
cesses of information science knowledge graph research. Hu ­Fanghuai6 studied the construction method of the
Chinese knowledge graph based on multiple data sources; Giovanni Adorni and Frosina K ­ oceva7 constructed
knowledge representation tools for educational theory based on subject structure and representation method.
Jacopo Urbani and Ceriel J­ acobs8 studied the method of enriching query results by top-down rule-based reason-
ing on large RDF knowledge bases.
In other specific fields, research on knowledge graphs has also appeared. Jintao ­Liu9 proposed a new method
based on knowledge graphs to explore railway operation accidents, aiming to reveal the potential rules of acci-
dents by describing accidents and hazards in heterogeneous networks. M ­ aya10 et al. extract medical concepts
from a large number of patient records, and use the maximum likelihood estimation of three probability mod-
els to automatically construct a knowledge graph. Said ­Fathalla11 described how investigations in the research
field are represented semantically, thus forming a knowledge graph that describes various research questions,
methods, and evaluations in a structured and comparable way. Some companies have used knowledge graphs to
improve existing products and have developed prominent examples of large-scale knowledge graphs, including
­DBpedia12, Google knowledge ­graph13, Microsoft’s ­Satori14, Freebase, YAGO, and ­Wikidata15. Jayaram and Khan
et al.16 established a "Graph Query by Example” system, automatically finds the weighted hidden maximum
query graph based on the input query tuple to obtain the user’s query intent. Then, the top approximate match-
ing answer graph and answer tuple can be effectively found and arranged.
However, the above knowledge graphs are mainly concentrated in the public domain, and the research on
the field of ship construction is relatively limited. There is still no unified expression form for heterogeneous
models based on knowledge graphs, but it is foreseeable that this technology will be used in the field of ship
­construction17. The knowledge representation and application aspects of the company have great potential. We
will use the existing research results to realize the construction of process knowledge graph models for hetero-
geneous models.

Domain ontology and case‑based reasoning. Ontology18 is derived from the ontology of philosophy
and is used to describe the objective existence in the world. It is a systematic explanation or explanation of
objective existence. Ontology can formalize the semantic expression of concepts, which is a common method
of knowledge base construction. ­Blythe19 uses the idea of ontology to research the knowledge association of
the knowledge owned by the enterprise. This research can help the enterprise realize the acquisition of domain
knowledge more quickly. Maedche et al.20 proposed the use of ontology technology to develop an integrated
enterprise knowledge management framework. ­Wang21 proposed a knowledge management (KM) framework
based on the Web ontology language DAML + OIL, which makes knowledge sharing more flexible and efficient.
Li ­Zhimin22 proposed a simpler ontology development and expansion and a two-stage mapping method to inte-
grate virtual enterprise ontology. These methods can improve the comprehensiveness, scalability, reusability, and
sharing requirements of the ontology knowledge organization of virtual enterprises.
Case-based reasoning (CBR) first appeared in Roger’s description of dynamic memory in 1982. The core
idea of CBR is to help us solve the current problems with the help of cases that have solved similar problems in
the past. Its core work is to retrieve historical cases that are most similar to the current case in the case library,
and then retrieve the specific information based on the current problem the solutions reached are adjusted and
applied to the current situation. Chang ­Liyun23 proposed a method combining discernible matrix and math-
ematical logic operation to get the best attribute reduction results. Sun Yanqing et al.24 redefines the importance
of attributes so that the algorithm can simplify attributes while maintaining expert experience. Lin and ­Chen25
proposed the use of artificial immune algorithm for attribute feature selection. After the feature is selected by
artificial immune algorithm, the accuracy and effectiveness of case retrieval are improved. ­Shen26 combined
grey relational analysis with a genetic algorithm to optimize the feature selection process and used the result
of grey relational analysis as the initial population of genetic algorithm heuristic search. Chen Hong and Mao
­Hongbao27 used knowledge entropy to determine the feature weights of cases and verified the credibility of the

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 2

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

method through examples. N ­ orri28 conducted a comparative study on three feature weighting algorithms of
genetic algorithm, rough set theory, and fuzzy inference system, and the research results proved that the fuzzy
set theory is the best feature weighting method. ­Cunningham29 shows that in case-based reasoning, evaluating
the similarity between cases is a key aspect of the retrieval stage.
In summary, the combination of ontology and CBR can effectively improve the performance of knowledge
representation and improve the performance of the CBR s­ ystem30. The calculation of semantic similarity is one
of the hotspots of ontology research. Introducing it into CBR research can effectively improve the drawbacks of
traditional CBR systems that only use attribute similarity as the main matching reference basis, thereby improving
the accuracy of matching results. However, the existing research on semantic similarity has not considered the
influence of concept relevance on similar results, which leads to limited knowledge matching effect. Therefore,
based on existing research in the industry, this paper introduces concept relevance to improve semantic similarity
to obtain comprehensive semantic similarity, to improve the accuracy and objectivity of knowledge similarity
calculation; and in the ontology representation of case knowledge After that, it is clustered according to the com-
prehensive semantic similarity to compress the knowledge matching space and improve the matching efficiency.

Overviews of the proposed method

In this section, we first define some basic concepts, and then briefly outline our approach.

Basic concepts. In order to visualize the dimension of each process model in machining process, the basic
concepts of process dimension, dimension completeness and standardization are condensed.

Definition 1 Process Element (PE). Process Element is used as a set of constraint conditions for assembly and
welding process. It can be expressed as:
PE = {< F > | < PEi >}, i = 1, 2, ......k (1)
Among, F represents the feature in the assembly and welding process, PEi represents the ith process element
related to process decision-making, and k is the process element category, such as assembly feature type, part
material, welding material, etc.

Definition 2 Process Knowledge (PK). Standards, rules and experience knowledge used in welding process
design, such as welding equipment selection rules, process analysis rules, etc. Remember to:
PK = {P1 ,P2 , ...Pi , ...PR }, i = 1, 2, ... (2)
Among them, Pi represents rule knowledge; R represents the category of rule knowledge.

Definition 3 Process Knowledge Node Relationships (PKNR). Given the process element node Pi and the process
knowledge node Pj, suppose the mutual entity relationship between them is:
ij = PE(Pi ) ∩ PK(Pj ) (3)
If the λij ≠ ∅, given the relationship between two nodes. In the field of assembly and welding technology, the
nodes of process knowledge can be extracted by the 3D model in the process of process design. This paper con-
structs three types of knowledge node relations, which are hierarchy relation, reference relation, and example
relation respectively. Among them, the hierarchical relationship reflects the hierarchical structure and composi-
tion relationship between the three-dimensional model feature concepts; the reference relationship reflects the
concept and the concept can be matched together through the reference rule; the exemplified relationship is the
instantiation of the class concept.

Definition 4 Process Knowledge Graph (PKG). Process knowledge graph is the entity expansion and enrich-
ment of various process knowledge under the concept nodes and relationships, and it is a structured semantic
knowledge base. Process knowledge graph K defines a set of (O, I), where O stands for ontology and I stands for
instance. Both O and I triple with the form (s, p, o), s and o belong to concepts or entities, and p is an attribute.

Overview of the proposed method. Figure 1 shows the overall flow of the proposed method, which is
guided by the process knowledge graph, the unified structured representation of heterogeneous models is real-
ized from the perspective of process semantics. The framework consists of two parts: the process knowledge
graph for building heterogeneous model of ships and the knowledge reuse system based on case-based reasoning
(CBR).
Based on the analysis of the heterogeneous three-dimensional model and the process information model
instance database, the process knowledge graph is constructed, and the concept mapping, semantic mapping,
and other multi-strategy mapping methods are used to realize the unified structured semantic representation
of the heterogeneous model.
The purpose of a case-based reasoning system is to reuse the knowledge stored in the case base with the help
of a heterogeneous model process case base so that new process planning can be generated quickly. Usually,
after a new case is entered, the system is a cyclic process that includes four steps: retrieval, reuse, modification,
and retention.

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 3

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

3D model analysis
Build process
knowledge graph
Process information
model instance Semantic Concept
library analysis mapping mapping
Ship heterogeneous
Construct a unified
model
representation of
heterogeneous model
Heterogeneous model
semantics
process case library

Keep the newly

Enter new case Case modification
generated cases

Design a new assembly

Similarity analysis and welding process
design plan
N

Search for similar process Y

Similarity ≥ε Case reuse
design cases

Figure 1.  A general flow chart of the method presented in this paper.

Ethical approval. The authors state that this paper is an original work, it has not been published in any
journals, and this research does not involve any ethical issues of humans or animals.

Consent to participate. The authors declare that this research involves no human participants and/or
animals.

Consent for publish. This work described has not been published before and has not been under consid-
eration for publication anywhere else. Authors are responsible for the correctness of the statements provided in
the manuscript and consent to be published.

Informed consent. Consent was obtained from all individual participants included in the work.

Knowledge model construction based on process knowledge graph

Construct process knowledge graph of heterogeneous model. Process knowledge graph is the
entity expansion and enrichment of various process knowledge under conceptual nodes and relationships and
is a structured semantic knowledge network. In the field of ship process design, in order to solve the existence
of heterogeneous models with conceptual mismatches and semantic inconsistencies, it is necessary to build a
general process knowledge graph model to achieve a unified structured representation of ship heterogeneous
models.
The process knowledge graph is divided into two levels: data layer and model layer. The model layer is usually
built on the data layer and stores abstract knowledge, rules, and patterns. The actual data is stored by the data
layer. The overall architecture of building process knowledge graph for heterogeneous ship model is shown in
Fig. 2. The part in the dashed box in the figure is the process of constructing the process knowledge graph and
also the process of updating the process knowledge graph. The construction process of knowledge graph is based
on the acquired heterogeneous model of ships, and a series of automatic or semi-automatic technical means
are adopted. The knowledge elements are extracted from the heterogeneous model and then the knowledge
is processed. After knowledge fusion and knowledge processing, it is stored in the process knowledge graph.
Finally, the model layer and the data layer are linked through services to form a large process knowledge graph.
The construction of the process knowledge graph is an iterative construction process. As can be seen from the
figure, each round of iteration includes the following three stages.

Assembly and welding process knowledge extraction. The extraction of assembly and welding process informa-
tion is the first step in the construction of process knowledge graph. The key issue is to extract entities, attributes
and relationships from heterogeneous models of ships with different structures, and convert them into struc-
tured data. The process knowledge extraction in this paper includes attribute extraction, relation extraction, and
entity extraction. Entity extraction refers to automatically identifying named entities from heterogeneous model

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 4

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

Ship
heterogeneous
model

Knowledge Knowledge
Structured data
fusion reasoning

Attribute
extraction
Semi-structured
data Process
Relation Entity Ontology
Knowledge
extraction disambiguation construction
Unstructured Graph
data
Entity
extraction Quality
assessment

Process knowledge
Knowledge fusion Knowledge processing
extraction

Figure 2.  Process knowledge graph architecture.

Lap joint
Angle
joint T joint

Butt
joint
Material
Welded specifica-
Welding joint tions
groove Welding
parameters current

Welding Welding
Y-shaped method Welding voltage
groove process
parameters Welding
Welding
X-shaped groove speed
groove
Heat Number of
input welding passes
I-shaped Wire Filling Wire
groove diameter height of type
V-shaped solder
groove

Figure 3.  Entity extraction of welding process parameters in heterogeneous model.

data. The quality of entity extraction has a great influence on the efficiency and quality of subsequent knowl-
edge acquisition, so it is the most basic and key part of the knowledge extraction of the assembly and welding
process. As shown in Fig. 3, take the welding process parameters as an example. In order to link the extracted
knowledge to the process knowledge graph, it is also necessary to extract the association relationships between
entities and connect the entities through the relationships. Relation extraction is to solve the basic problem of
how to extract the relations between entities. The purpose of attribute extraction is to collect attribute informa-
tion of specific entities from different heterogeneous model information. The method of data mining is used to
directly mine the relationship pattern between entity attributes and attribute values from heterogeneous model
information to realize the positioning in the model. The object-oriented method is used to establish the process
knowledge data model of the heterogeneous models in the CAPP system, and the knowledge base, knowledge
logic and the process database driven by the object model are combined to realize the analysis and acquisition
of process knowledge. The foundation of its application lies in the establishment, application and maintenance
of the process knowledge base. This data mining method has low requirements for objects and is suitable for
heterogeneous models in ship construction. As shown in Fig. 4, it is a data mining method based on process
knowledge base and heterogeneous models.

Knowledge fusion and knowledge processing. Through the knowledge extraction of the assembly and welding
process, the goal of obtaining entity, relationship, and entity attribute information from unstructured and semi-
structured data is realized. However, this knowledge can contain a lot of redundancy and error information,
and the relationships between data are flat and lack logic and hierarchy. Therefore, it is necessary to clean and
integrate it to form structured data in the knowledge base.

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 5

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Definition of process
Process data mining Process data mining
data mining based on
results cleaning knowledge output
object pattern

CAPP process knowledge data mining engine

Model
Heterogeneous model Process Knowledge
Base Process Knowledge Base

Figure 4.  Process data mining method based on object mode.

Knowledge fusion includes two parts, namely entity link, and knowledge fusion. Through knowledge fusion,
ambiguity and misconceptions can be eliminated, thereby ensuring the quality of knowledge. Entity linking refers
to the operation of linking the entity object extracted from the text to the corresponding correct entity object in
the knowledge base. The process of entity linking is: extracting entity referents from model information through
entities; Whether the entity with the same name in the knowledge base and it represent different meanings is
judged, and whether there are other named entities in the knowledge base and it represents the same meanings;
After confirming the correct entity object in the knowledge base, the entity reference necklace is connected to
the corresponding entity in the knowledge base.
Through the knowledge extraction of the assembly and welding process, knowledge factors such as entities,
relationships and attributes can be extracted from heterogeneous models. After knowledge fusion, the ambiguity
between entity referents and entity objects can be eliminated, and basic fact expressions can be obtained. How-
ever, the fact itself is not equivalent to knowledge. It needs to go through the process of knowledge processing
to obtain the final structured and networked knowledge system.
Taking the heterogeneous model in Fig. 5 as an example, the process knowledge contained in the heteroge-
neous model is expressed through processes such as knowledge extraction, knowledge fusion, and knowledge
processing.

Unified semantic expression of heterogeneous models. The unified semantics of the ship heteroge-
neous model is an expression form in which the concepts in multiple ontologies are established correspondingly,
and the multiple ontologies are linked together through mapping. At the same time, the similarity of concepts
is evaluated, the mapping rules are formulated according to the similarity results, and the mapping rules are
corrected to ensure that the unified semantic representation of heterogeneous models is realized in the process
knowledge graph.
As there are related concepts and semantic connections between the various process knowledge ontology
contained in the ship heterogeneous model, the mapping relationship between the ontology is established. Taking
the welding tool resource ontology and the feature welding set resource ontology as an example, the mapping f
between the welding tool resource ontology W and the feature welding set resource ontology T is represented by
the set M = {m}, where m represents the basis of a five-tuple, the mapping unit can be written in the form e = (g, u,
v, r, s). g is the unique identifier of the mapping unit; u and v are elements in W and T respectively; r represents
the relationship between u and v; s is used to identify the similarity of the ontology mapping. Thus, the mapping
relationship between the concepts and parameters in the welding tool resource ontology and the feature welding
set resource ontology is established. The mapping relationship among connection tools, feature welding sets,
and instance ontology is shown in Fig. 6. The semantic similarity of the concepts in the heterogeneous model is
calculated through the corresponding mapping rules, to ensure that the unified semantic expression of the ship
heterogeneous model is realized in the process knowledge graph.

Knowledge reuse technology based on case‑based reasoning. After the construction of the het-
erogeneous ship model case base is completed, the assembly and welding process decisions extracted previously
are recorded and stored. A case-based reasoning (CBR) method is needed to retrieve and reuse assembly and
welding process knowledge stored in heterogeneous ship models. In the CBR system, the existing welding pro-
cess information is retrieved by calculating the similarity degree. Similarity can be divided into local similarity
and global similarity. Local similarity refers to the similarity between two attributes, while global similarity refers
to the similarity between two cases.
Figure 7 shows the calculation method of local similarity and global similarity between the new case A and
the existing case B, where j represents the serial number of influencing factors and n represents the total num-
ber of influencing factors. Once the influencing factors of ship heterogeneous model cases (­ D1, ­D2,…Dn), the
similarity of these attributes can be calculated by the nearest neighbor algorithm. Through the analytic hierarchy
process, according to the importance of each factor to the case, each factor is assigned a weight, and then all the
local similarities and their respective weights are considered, and the overall situation of the new case A and
case B in the knowledge base can be calculated Similarity. In the process of using the analytic hierarchy process,
a comparison matrix is created based on the input of the decision-maker, which gives the relative importance
between the two influencing factors. Experts need to make judgments on factor X relative to factor Y. The more

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 6

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 5.  Based on the process knowledge graph to express the process knowledge contained in the
heterogeneous model.

important the factor, the higher the score. The score is generally set to 1–9. If the importance of X relative to Y
is set to 3, the importance of Y relative to X is 1/3. By calculating the geometric mean and standardizing it to
obtain the relative weight associated with each influencing factor.
The similarity between the two cases is the global similarity, which can be defined as the weighted sum of
the similarity of each influencing factor. The global similarity between the new case and the existing case can be
calculated by the following formula.
n

Sim(A, B) = θp Sim(Xj , Yj ) (4)
j=1

Among them, Sim(Xj,Yj) represents the local similarity between the j-th influencing factor Xj,Yj between the
new case A and the existing case B, and Sim(A,B) is the local similarity between the new case A and the existing
case B The global similarity. n represents the number of influencing factors. θp represents the weight of the p-th
influencing factor.
For the calculation of local similarity, there are different calculation methods for different types of influencing
factors: enumerated factors, numerical factors, and irrelevant factors.
For numerical influencing factors that can be identified by numerical values, the following formula can be
used to express:
1
Sim(Xj , Yj ) =
1 + |Xj − Yj | (5)

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 7

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Concepts Concepts
Ontology Mapping Ontology
and and
module relations module
parameters parameters
Welding Characteristic
parameters Parameters

Function
Welding Spatial Relations
identification Welding planning
characteristics

Body Engineering
Semantic feature Component size
information semantics
Welding tool Feature welding set
resource ontology resource ontology
External
Engineering Weld Welding
parameters
semantics constraints

Working surface Semantic feature

Welding Welding
information interface

ĂĂ ĂĂ

Mapping
module
Welding
Welding tool
component
information
information
Instance
ontology

Figure 6.  Mapping relationship between ontology.

Set j=1

Get the influencing factors j in the new case A

Type of influencing factor j

Calculate the local similarity of influencing factor i between case A

and existing case B

j˘n˛ Y j=j+1

Determine the weight of each influencing factor

through analytic hierarchy process

Calculate the global similarity between case A and

case B

Figure 7.  The method of calculating local similarity and global similarity.

The attribute domain of enumerated influencing factor similarity is generally enumerated, that is, any two
values of influencing factors correspond to a local similarity. Attributes such as assembly requirements, welding
methods, and construction types have this type of local similarity.
|Xj − Yj |
Sim(Xj , Yj ) = 1 − (6)
M
The irrelevant factor means that there is no connection between the different values of the influencing factors.
If the two influencing factors are the same, the local similarity is 1; otherwise, the local similarity is 0.

1, Xj = Yj
Sim(Xj , Yj ) =
0, Xj �= Yj (7)

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 8

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

Heterogeneous
CATIA System CAD System
model analysis
Concept
Model Analysis
Extraction
Terminology Process
Extraction Analysis
Relationship Feature
Extraction association

ĂĂ ĂĂ

System Ontology Process knowledge

Case knowledge reuse
Model graph schema layer
System Ontology
Model

Process Knowledge Graph Representation

Model layer Data Layer Visualization

Figure 8.  Process knowledge graph modeling system for ship heterogeneous model.

Case study and discussion

In the construction of modern ships, heterogeneous models occupy an important proportion. The double bot-
tom segment is a typical heterogeneous model. Through the study of double bottom segmentation, the process
knowledge graph of the double bottom segment is constructed to verify the proposed feasibility of the method.
Figure 8 shows the process knowledge graph modeling prototype system for ship heterogeneous models
constructed in this paper, it mainly includes two main stages: construction of heterogeneous model process
knowledge graph and unified semantic generation, and case knowledge reuse. It also includes CATIA platform
information extraction, heterogeneous model, and process information model analysis, which provides support
for heterogeneous model information extraction and process knowledge graph construction. It can be seen from
Fig. 7 that, firstly, the system ontology model is constructed by extracting the inherent concepts, terminology,
and relationships of the system through the CATIA platform. Secondly, by analyzing the heterogeneous model
case library and mapping related concepts, the model layer of the process knowledge graph is established. Then,
the multi-strategy ontology mapping of the Protégé platform is used to generate a process knowledge graph with
a unified semantic expression of heterogeneous models. At the same time, through the mutual transformation
of process knowledge graph, can realize knowledge reuse and data exchange.

Build heterogeneous model process knowledge graph. The construction of the process knowledge
graph is divided into two stages. First, in order to ensure the accuracy and uniformity of concepts in the model

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 9

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Figure 9.  Double bottom section.

layer of process knowledge graph, a top-down approach is adopted to map the features and process knowledge
concepts to the process knowledge graph model layer to form the basic model layer of the process knowledge
graph model. Since the assembly process decision is closely related to the coupling relationship and the hetero-
geneous model contains the macro process. Therefore, to ensure the integrity of the process knowledge graph,
the relationship attributes resolved by the heterogeneous model are added to the basic model layer of the process
knowledge graph to form an extended model layer of the process knowledge graph. Secondly, with the increase
of process information model instance data, the process knowledge graph can be updated and expanded by
extracting process entity concepts and process entity relationships.
Taking the heterogeneous model of the ship’s double bottom segment as shown in Fig. 8 as an example, the
process knowledge graph is constructed through the following steps:

Step 1. Identify N assembly welding process characteristics Pi from the heterogeneous model.
Step 2. In the process file of the heterogeneous model, M assembling and welding process information are
traversed.
Step 3. Through semantic analysis and lexical analysis, traverse M assembly and welding process informa-
tion, and extract welding method information, welding equipment information, and assembly and welding
strategy information. In this step, the relationship between the geometric features and the welding equipment
involved in the current assembly and welding process will be obtained.

Taking the heterogeneous model shown in Fig. 9 as an example, the process knowledge graph constructed by
the above method is shown in Fig. 10. In the process knowledge map model layer, the process elements (PE) of
the double bottom segment are included, such as welding material, welding speed, welding position, etc. In the
data layer, there is process knowledge (PK) such as the selection of welding equipment and the type of welding
groove. The Process Knowledge Graph (PKG) is composed of process knowledge nodes relationships (PKNR)
composed of various process elements and process knowledge.

Case knowledge reuse. Figure 11 shows the case retrieval interface. The designer extracts the semantic
words according to the retrieval intention and enters them into the semantic retrieval column, and enters the
attribute value of the case knowledge in the attribute similarity check layer. The semantic concept words in this
example are "plate thickness", "groove type", "welding method" and "weld seam length".
In hierarchical retrieval, the semantic similarity optimization algorithm is first used to calculate the semantic
relatedness, and the semantic relatedness between domain ontologies is used to retrieve a subset of cases mapped
by concept nodes, and then the numerical similarity is calculated by the attribute similarity algorithm retrieval to
improve the retrieval accuracy of case knowledge. In the process acquisition interface, the system gives a sorted
list of the most relevant cases, and the case structure and corresponding process design principles can be viewed
through the case process information interface to realize the reuse of case knowledge.

Discussion. Table 1 shows the comparison between the method in this paper and other related methods.
It can be seen that the method in this paper and the method i­n31 both use structural elements of different
granularities (for example, parts, features,); Both of them are oriented towards the manufacturing domain and
have multi-level detail characteristics. The approach ­in31 is to perform a geometric decomposition of complex
CAD models, aggregating technical data and topological relationships into geometric features, by proposing a
machining process model for expressing the connection between machining features and machining strategies.
The approach in the ­literature31 structured the representation of multi-detail level process knowledge in terms
of manufacturing feature carriers, but did not consider heterogeneous models. The ­literature32 proposes the
use of a metamodel as a basis for modeling manufacturing process information, which defines the information
and relationships between the process of manufacturing process planning, the resources used and the products
manufactured. This method is oriented to the field of machining and does not systematically consider macro and
micro process knowledge. The ­literature33 proposes an ontology-based semantic retrieval method for heteroge-
neous 3D models, using hierarchical feature ontologies and ontology mapping techniques to generate a unified
description of heterogeneous models in the form of semantic descriptors. Although the unified structured rep-
resentation of heterogeneous model features is solved by the methods ­in33, the sharing and reuse of heteroge-
neous models cannot be realized. After comparison, this paper derives semantic expressions between process
knowledge maps and heterogeneous models by proposing a multi-strategy ontology mapping method to obtain
implicit semantics and check the similarity of matching results through case-based reasoning, thus realizing the
reuse of case knowledge and fast design of processes.

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 10

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 10.  Constructing a heterogeneous model process knowledge graph.

Figure 11.  Case search and reuse module interface.

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 11

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Comparison factor Our approach 31 32 33

Degree of structure Parts/Features Part Manufacturing features Design features

Model representation PKG UML UML Ontology
Level of detail ▲ ▲
Micro process knowledge ▲ ▲
Macro process knowledge ▲
Application field Manufacture Manufacture Manufacture Design
Data set Heterogeneous model Process data Isomorphic model Heterogeneous model

Table 1.  Comparison of the method in this paper with other methods.

Conclusion and future works

Aiming at the problem that the knowledge contained in the existing heterogeneous models is difficult to express
the influence on process knowledge reuse and sharing, a structured modeling method that uses process knowl-
edge graph to uniformly represent heterogeneous models is proposed in this paper. Ontology is used to organize
and formalize various process-related information that exists in the heterogeneous model. It provides a structured
way to express and manage the process knowledge in the model. At the same time, case-based reasoning (CBR)
can process the process knowledge stored in the semantic model, so it is used to retrieve and reuse the existing
process knowledge to obtain the candidate welding process information.
The paper takes the construction of the process knowledge graph model of the heterogeneous model as the
starting point. First, the heterogeneous model analysis is extracted through feature recognition, manufacturing
semantic. And then the extracted information is fused and processed to obtain the process knowledge ontology.
The method based on multi-strategy ontology mapping is used to generate a unified semantic expression of het-
erogeneous models. This technology can not only express the process knowledge contained in the heterogeneous
model, but also effectively improve the efficiency of knowledge reuse, which is of great significance for shorten-
ing the design cycle of the ship component welding process. Overall, the main contribution of the method is
to provide a knowledge model containing implicit and explicit process knowledge with multiple levels of detail
based on process knowledge mapping for describing the process design intent of heterogeneous models. The
method has the following features and advantages.

1. With the help of ontology modeling, ontology mapping and semantic reasoning, a unified semantic repre-
sentation is established for heterogeneous models to support knowledge sharing and reuse in process co-
design. This helps to compensate for the insufficient consideration of implicit process knowledge in existing
feature-based models.
2. Supports multi-mode retrieval by utilizing an improved similarity metric scheme. Can promote the effective-
ness of process knowledge reuse.
3. The knowledge embedded in the heterogeneous model is expressed using process knowledge mapping, thus
enabling rapid reuse of assembly and welding process knowledge. Process knowledge is not only limited to
the semantic level, but the process attributes are closely linked to the manufacturing semantics.

In the future, we still need to further explore some issues to improve the practicability of our method: More
semantic rules will be used to improve the accuracy of semantic description; Process design information and
knowledge are highly complex, and uncertainties caused by fuzzy information in the process design process
need to be considered.

Data availability
In this submission, all the data are transparent.

Received: 5 October 2021; Accepted: 9 February 2022

References
1. Wang, L., Shen, W., Xie, H., Neelamkavil, J. & Pardasani, A. Collaborative conceptual design—state of the art and future trends.
Comput. Aided Des. 34(13), 981–996. https://​doi.​org/​10.​1016/​S0010-​4485(01)​00157-9 (2002).
2. Ulonska, S. & Welo, T. Product portfolio map: A visual tool for supporting product variant discovery and structuring. Adv. Manuf.
2(002), 179–191. https://​doi.​org/​10.​1007/​s40436-​014-​0077-​yO (2014).
3. Mohamed, O. A., Masood, S. H. & Bhowmik, J. L. Optimization of fused deposition modeling process parameters: A review of
current research and future prospects. Adv. Manuf. https://​doi.​org/​10.​1007/​s40436-​014-​0097-7 (2015).
4. Nickel, M., Murphy, K., Tresp, V. & Gabrilovich, E. A review of relational machine learning for knowledge graphs. Proc. IEEE
104(1), 11–33. https://​doi.​org/​10.​1109/​JPROC.​2015.​24835​92 (2015).
5. Hu, Z. Research review on application of knowledge mapping in china. Lib. Inf. Serv. 57(03), 131–137.
CNKI:SUN:TSQB.0.2013-03-025 (2013)
6. Li, M. X. & Wang, S. Research context and theme analysis of mapping knowledge domains in china in rencet ten years. Doc. Inf.
Knowl. 04, 93–101. https://​doi.​org/​10.​13366/j.​dik.​2016.​04.​093 (2016).
7. Paulius, D. & Sun, Y. A survey of knowledge representation in service robotics. Robot. Auton. Syst. 118, 13–30. https://​doi.​org/​10.​
1016/j.​robot.​2019.​03.​005 (2019).

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 12

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
www.nature.com/scientificreports/

8. Antoniou, G. et al. A survey of large-scale reasoning on the Web of data. Knowl. Eng. Rev. 33, 1–43. https://​doi.​org/​10.​1017/​S0269​
88891​80002​55 (2018).
9. Liu, J., Schmid, F., Li, K. & Zheng, W. A knowledge graph-based approach for exploring railway operational accidents. Reliab. Eng.
Syst. Saf. 207, 107352. https://​doi.​org/​10.​1016/j.​ress.​2020.​107352 (2021).
10. Rotmensch, M., Halpern, Y., Tlimat, A., Horng, S. & Sontag, D. A. Learning a health knowledge graph from electronic medical
records. Sci. Rep. 7(1), 1–11. https://​doi.​org/​10.​1038/​s41598-​017-​05778-z (2017).
11. Fathalla, S., Vahdati, S., Auer, S. & Lange, C. Towards a knowledge graph representing research findings by semantifying survey
articles. Proc. Int. Conf. Theory Pract. Digit. Libr. 10450, 315–327. https://​doi.​org/​10.​1007/​978-3-​319-​67008-9_​25 (2017).
12. Saorín, T. & Pastor-Sánchez, J.-A. Wikidata and DBpedia: A journey to the center of a web of data. Anuario ThinkEPI https://​doi.​
org/​10.​3145/​think​epi.​2018.​31 (2018).
13. Yoo, S. Y. & Jeong, O. R. Automating the expansion of a Knowledge graph. Expert Syst. Appl. 141, 112965. https://d ​ oi.o
​ rg/1​ 0.1​ 016/j.​
eswa.​2019.​112965 (2019).
14. Microsoft’s Satotr. http://​blogs.​bing.​com/​search.
15. Wang, Q., Mao, Z., Wang, B. & Guo, L. Knowledge graph embedding: A survey of approaches and applications. IEEE Trans. Knowl.
Data Eng. 29(12), 2724–2743. https://​doi.​org/​10.​1109/​TKDE.​2017.​27544​99 (2017).
16. Jayaram, N., Khan, A., Li, C., Yan, X. & Elmasri, R. Querying knowledge graphs by example entity tuples. IEEE Trans. Knowl. Data
Eng. 27(10), 2797–2811. https://​doi.​org/​10.​1109/​TKDE.​2015.​24266​96 (2015).
17. Arenas, M., Grau, B. C., Kharlamov, E., Marciuska, S. & Zheleznyakov, D. Faceted search over RDF-based knowledge graphs. Web
Semant. Sci. Serv. Agents World Wide Web 37, 55–74. https://​doi.​org/​10.​1016/j.​websem.​2015.​12.​002 (2016).
18. Rayyaan, R., Wang, Y. & Kennon, R. Ontology-based interoperability solutions for textile supply chain. Adv. Manuf. 2(2), 97–105.
https://​doi.​org/​10.​1007/​s40436-​014-​0073-2 (2014).
19. Nkiaka, E., Taylor, A. L., Dougill, A., Antwi-Agyei, P. & Warnaars, T. Identifying user needs for weather and climate services to
enhance resilience to climate shocks in sub-Saharan Africa. Environ. Res. Lett. 14(12), 123003. https://​doi.​org/​10.​1088/​1748-​9326/​
ab4dfe (2019).
20. Barao, A., De Vasconcelos, J. B., Rocha, A. & Pereira, R. A knowledge management approach to capture organizational learning
networks. Int. J. Inf. Manag. 37(6), 735–740. https://​doi.​org/​10.​1016/j.​ijinf​omgt.​2017.​07.​013 (2017).
21. Zhang, C., Romagnoli, A., Zhou, L. & Kraft, M. Knowledge management of eco-industrial park for efficient energy utilization
through ontology-based approach. Appl. Energy 204, 1412–1421. https://​doi.​org/​10.​1016/j.​apene​rgy.​2017.​03.​130 (2017).
22. Qiao, L. H. et al. An ontology-based modelling and reasoning framework for assembly sequence planning. Int. J. Adv. Manuf.
Technol. 94(9–12), 4187–4197. https://​doi.​org/​10.​1007/​s00170-​017-​1077-4 (2018).
23. Xue, Y. & Lai, Y. Integration of macro energy thinking and big data thinking part two applications and explorations. Autom. Electr.
Power Syst. 40(8), 1–13 (2016).
24. Farhan, U., Tolouei-Rad, M. & Osseiran, A. Indexing and retrieval using case-based reasoning in special purpose machine designs.
Int. J. Adv. Manuf. Technol. 92(5–8), 2689–2703. https://​doi.​org/​10.​1007/​s00170-​017-​0274-5 (2017).
25. Lin, S. W. & Chen, S. C. Parameter tuning, feature selection and weight assignment of features for case-based reasoning by artificial
immune system. Appl. Soft Comput. 11(8), 5042–5052. https://​doi.​org/​10.​1016/j.​asoc.​2011.​05.​054 (2011).
26. Shen, Qi. Optimization of CBR model utilizing genetic algorithm. Appl. Mech. Mater. 543, 1827–1831. https://​doi.​org/​10.​4028/​
www.​scien​tific.​net/​AMM.​543-​547.​1827 (2014).
27. Jia, R., Pan, W. & Liu, F. Virtual enterprise partner selection based on a rough set theory and adaptive genetic algorithm. J. Harbin
Eng. Univ. 33(6), 730–734 (2012).
28. Noori, B. Developing a CBR system for marketing mix planning and weighting method selection using fuzzy AHP. Appl. Artif.
Intell. 29(1), 1–32. https://​doi.​org/​10.​1080/​08839​514.​2014.​962282 (2015).
29. Cunningham, P. A taxonomy of similarity mechanisms for case-based reasoning. IEEE Trans. Knowl. Data Eng. 21(11), 1532–1543.
https://​doi.​org/​10.​1109/​TKDE.​2008.​227 (2009).
30. Meditskos, G. & Bassiliades, N. Structural and role-oriented web service discovery with taxonomies in OWL-S. IEEE Trans. Knowl.
Data Eng. 22(2), 278–290. https://​doi.​org/​10.​1109/​TKDE.​2009.​89 (2010).
31. Mawussi, K. B. & Tapie, L. A knowledge base model for complex forging die machining. Comput. Ind. Eng. 61(1), 84–97. https://​
doi.​org/​10.​1016/j.​cie.​2011.​02.​016 (2011).
32. Yang, B., Qiao, L., Na, C., Zhu, Z. & Wulan, M. Manufacturing process information modeling using a metamodeling approach.
Int. J. Adv. Manuf. Technol. 94(5), 1579–1596. https://​doi.​org/​10.​1007/​s00170-​016-​9979-0 (2018).
33. Katia, L., Franca, G., Marina, M. & Jean-Philippe, P. Content-based multi-criteria similarity assessment of CAD assembly models.
Comput. Ind. 112, 103111–103111. https://​doi.​org/​10.​1016/j.​compi​nd.​2019.​07.​001 (2019).

Acknowledgements
This work was supported in part by the National Defense Basic Scientific Research Program under Grant
JCKY2018414C002 and the National Natural Science Foundation of China (grants NO. 52075229), and The
Natural Science foundation of the Higher Education Institutions of Jiangsu Province (grants NO. 20KJA460009),
and the High-tech Ship Scientific Research Project of the Ministry of Industry and Information Technology
(NO. CJ04N20).

Author contributions
J.L.: methodology, writing, supervision, validation—original draft preparation. J.D.: Investigation and method-
ology. X.J.: Supervision, writing—reviewing and editing. X.C.: Validation. C.D.: Supervision. L.L.: Supervision.

Competing interests
The authors declare no competing interests.

Additional information
Correspondence and requests for materials should be addressed to J.L.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 13

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol.:(0123456789)
 www.nature.com/scientificreports/

Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

© The Author(s) 2022

Scientific Reports | (2022) 12:2911 | https://doi.org/10.1038/s41598-022-06940-y 14

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Vol:.(1234567890)
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]