Next Article in Journal
Caregivers’ Knowledge and Food Accessibility Contributes to Childhood Malnutrition: A Case Study of Dora Nginza Hospital, South Africa
Previous Article in Journal
Particulate Matter Exposure and Allergic Rhinitis: The Role of Plasmatic Extracellular Vesicles and Bacterial Nasal Microbiome
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJERPH
Volume 18
Issue 20
10.3390/ijerph182010692
ijerph-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Systematic Knowledge Management of Construction Safety Standards Based on Knowledge Graphs: A Case Study in China

by
Yukun Jiang
,
Xin Gao
*,
Wenxin Su
and
Jinrong Li
School of Economics and Management, Tongji University, Shanghai 200092, China
*
Author to whom correspondence should be addressed.
Int. J. Environ. Res. Public Health 2021, 18(20), 10692; https://doi.org/10.3390/ijerph182010692
Submission received: 26 August 2021 / Revised: 1 October 2021 / Accepted: 3 October 2021 / Published: 12 October 2021
Browse Figures
Versions Notes

Abstract

:
Construction safety standards (CSS) have knowledge characteristics, but few studies have introduced knowledge graphs (KG) as a tool into CSS management. In order to improve CSS knowledge management, this paper first analyzed the knowledge structure of 218 standards and obtained three knowledge levels of CSS. Second, a concept layer was designed which consisted of five levels of concepts and eight types of relationships. Third, an entity layer containing 147 entities was constructed via entity identification, attribute extraction and entity extraction. Finally, 177 nodes and 11 types of attributes were collected and the construction of a knowledge graph of construction safety standard (KGCSS) was completed using knowledge storage. Furthermore, we implemented knowledge inference and obtained CSS planning, i.e., the list of standard work plans used to guide the development and revision of CSS. In addition, we conducted CSS knowledge retrieval; a process which supports interrogative input. The construction of KGCSS thus facilitates the analysis, querying, and sharing of safety standards knowledge.

1. Introduction

Construction is a risky industry [1,2]. Construction workers spend most of their time on site [3], and dynamic and complex site operations expose them to many unsafe factors [4,5]. According to previous studies, 22% of occupational fatalities in the United States, 27.2% in the United Kingdom, and 27.6% in South Korea occurred in the construction industry [6,7,8]. Extensive statistics show that the risk of fatal construction accidents is five times greater than in other industries, and the casualty rate is three times higher [9,10,11]. It has been shown that the violation of safety standards in high-risk industries is a significant cause of accidents [12], indicating that safety compliance is critical to maintaining occupational safety. In addition, the reason why workers are violating safety standards needs to be explained so that safety compliance can be improved.
Construction safety standards (CSSs) refer to the common use and reuse of normative documents that were developed by consensus and approved by a recognized authority in order to obtain the optimum degree of construction safety [13]. In practice, safety cannot be achieved without the implementation of a standard of behavior among contractors [14]. Therefore, safety standards are considered to be an important component of safety control. They are utilized to regulate the behavior of all parties engaged in a construction project. Effective management and implementation of standards are conducive to safety improvement [15].
The current state of CSS application has exposed a series of problems in standard management. In China, for example, the number of CSS is large and the integration is poor. Due to the lack of a unified approach to CSS [16], a significant amount of development in construction and science and technology in China has led to the proliferation of various standards [17]. According to incomplete statistics, there are more than 60 national CSSs, more than 40 industry standards, more than 90 local standards, and more than 10 enterprise standards. The large number of standards results in duplications, inconvenience of use, and a waste of compilation resources. Because contractors make comparisons and filtrations in the application of standards, the number of CSS with a high-frequency of use is less than 20. Second, there are contradictions between individual CSSs, leading to difficulties in governmental supervision and contractors’ production. For example, “Wire Rope Grips” stipulates that if a wire rope’s diameter is not greater than 9 mm, the rope card shall not be less than three groups. However, it is required that a rope card shall not be less than three groups if a wire rope’s diameter is not greater than 18 mm in “Steel wire ropes: Safety, Use and maintenance”, “Technical Specification for Safety Operation of Constructional Machinery”, and “Technical Specification for Safety Installation Operation and Dismantlement of Tower Crane in Construction”. Third, standards that are urgently needed in the market are missing, such as “Aluminum Formwork Safety Operation”. In summary, these problems diminish the function of CSS in safety control.
It is critical to address issues in CSS implementation. However, related research is far from complete. Although some research has proposed the implementation of standard management system reform in order to satisfy the objectives of economic and social development [18], this research generally concerns theoretical analysis, rather than concrete feasible measures or the domain of construction safety. Meanwhile, some studies have proposed integrating ISO 9001 and OHSAS 18001 in order to streamline management processes [19], and others have conducted standard knowledge transformation for contractors’ use [20]. Nonetheless, these studies are limited in their capacity to optimize the use of individual standards and are hardly generalizable to the standard management requirements of the construction industry as a whole.
It has been noted that security standards are a form of explicit knowledge. The term “knowledge” differs from “information” in that knowledge consists of facts, skills, and rules collected by individuals over a period [21]. Knowledge is classified into two categories: explicit knowledge and tacit knowledge [22]. Explicit knowledge is formal and can be easily encoded and transferred between individuals in documented and organized forms, such as databases, records, and regulations [21,23]. Safety standards belong to explicit knowledge, and a knowledge management approach can be applied for safety standard management.
From the perspective of knowledge management, the general approach is Semantic Web and knowledge graph (KG), which is extended from Semantic Web. Semantic Web is a graph-based data structure for storing knowledge [24]. KG is an emerging technology for massive knowledge management and intelligent services in the Big Data era [25,26]. Technically speaking, KG is a graph-based knowledge representation and organization method. It uses a set of “subject-predicate-object” triples to represent entities and their relationships [27]. KG can be seen as a huge network where nodes represent domain entities and contacts are regarded as semantic relationships between entities. KG can organize knowledge, establish connections, and solve the problem of “knowledge islands”. The term “knowledge islands” refers to isolated knowledge that lacks connections, and is difficult in terms of enabling knowledge sharing. In addition, KG can realize the functions of knowledge visualization, querying and reasoning [28].
In view of this, KG can establish connections between CSS and resolve issues of poor integration and contradiction. Moreover, KG can perform knowledge visualization, querying and inference, which provides practical measures for standard query and compilation. Therefore, it is feasible to integrate CSS by using KG.
However, few studies have introduced KG into CSS management. To bridge the research gap, KG is adopted to manage CSS knowledge. This paper first analyzes the structure of CSS knowledge, constructs a knowledge graph of construction safety standards (KGCSS), and realizes knowledge query and reasoning. Preliminary research of systematic CSS knowledge management is completed in this paper. The results of this paper not only solve the key problems of standard knowledge management, but also extend the results to other countries and industries.

2. Literature Review

2.1. CSS Management

The research on CSS management can be divided into two main areas. Firstly, the approach of integrating management of standards has been proposed. Some scholars have attempted to integrate ISO 9001 and OHSAS 18001 to optimize the process, avoid work duplication, and reduce resource input [19]. Secondly, studies have been carried out for the management of specific standards. The technical contents of “Code of Construction for Masonry Structure Engineering” have been analyzed. The influence of masonry mortar use time, the relative moisture content of blocks, and post-tensioning reinforcement on the quality and safety of projects has been derived, which facilitates the ease of comprehension for contractors [29].
As in the related research listed above, there is still a lack of a universal and operational integrated management methodology and tools in the field of CSS. It has been demonstrated that the essence of CSS is a technical system [18] and that knowledge characteristics are part of the technical system framework [30,31]. Previous studies have identified knowledge barriers to CSS application, which refer to the lack of standardization talents and low quality of standard operators [13]. On the other hand, strengthening technical standards knowledge management can reduce users’ time for selecting standard documents [32] and improve project performance. Therefore, it is necessary to organize and manage CSS knowledge in order to improve the sharing and reuse of CSS knowledge.

2.2. Knowledge Graph (KG)

KG technology originated from semantic networks in the 1960s [24]. In the 1990s, the idea of “ontology” was introduced into knowledge representation methods [33]. KG was first officially presented by Google in 2012. It refers to a graph-based method for representing and organizing knowledge. KG can associate a large amount of knowledge and information through an organization form and has knowledge reasoning ability. In recent years, the knowledge graph developed from the semantic web has attracted the attention of more and more researchers [34,35,36,37].
General knowledge graphs and domain knowledge graphs are two types of KG [38,39]. General knowledge graphs contain a wide range of general knowledge, and are also called open domain knowledge graphs. WordNet is the result of work to develop a dictionary database, in use since 1985 [40,41]. It implements the adding of simple facts to the knowledge network, but the division of concepts or entities is not very clear. Therefore, Cyc [42] and ConceptNet [43,44] were constructed. They relate concepts to common sense and incorporate complex relationships that exist in the real world. Currently, the most widely used are encyclopedic knowledge graphs, including DBpedia [45], YAGO [46], Freebase [47], Wikidata [48], and others. Wikimedia launched Wikidata in October 2012. Data in Wikidata are basically described by property-value pairs [38]. YAGO is a large semantic knowledge base conducted by the Max Planck Institute in Germany, with one million entities and more than five million facts. Freebase was launched in 2005. It was constructed based on Wikipedia and the idea of swarm intelligence. Encyclopedic knowledge graphs contain a large amount of semi-structured and unstructured data and occupy a central position in open-linked large-scale knowledge bases [49].
The general knowledge graph has a broad scope, a large scale, and a high degree of openness. However, due to its shallow research depth and relatively low accuracy, they cannot fully meet the requirements in some areas where there is a strong industry knowledge background.
Currently, there are few studies on KGCSS construction and application. However, there have been many results on domain knowledge graphs in other industries. GeoKG, which is represented with formal geographic knowledge, links discrete knowledge with the ability to query knowledge and solve geographic problems [50]. KG is used as a technical tool to link symptoms to diseases for efficient medical clinical decision making [51]. In the legal domain, KG is also used to manage case knowledge [52]. The maritime hazardous materials knowledge graph enables automatic determination of hazardous materials in isolation [53]. In addition, KG is used for knowledge management in education [54].
From the review above, KGCSS belongs to the domain of KG. Woo et al. [55] demonstrated the value and ability of KG to facilitate safe work practices. KG has been proven to improve security knowledge management because it helps to capture and prioritize knowledge through database searches and questionnaires, and to distribute important knowledge to specific working groups. Therefore, KGCSS should formally describe CSS and their interrelationships [56,57] and provide a systematical method of CSS knowledge management from the perspective of “classification” and “categorization” [58,59]. It is necessary and feasible to introduce KG into the field of CSS management in order to realize systematic knowledge management [60].

3. Materials and Methods

This study utilizes a top-down method to design the ontology of KGCSS. The top-down method refers to the design idea of defining the ontology of KG first and then adding entities to the knowledge base [61]. Ontologies are considered to be knowledge organization systems with elements that interact in a consistent conceptual structure [62]. Concept refers to the classification of a series of real objects. Entities are the most basic elements of KG and are the extrapolation of concepts which exist in the real world. Brainstorming is an effective way to utilize group wisdom and propose innovative ideas [63]. In this paper, we organized several seminars to discuss KGCSS construction. The participants included professionals from construction safety management authorities, standards management authorities and contractors. The research steps of KGCSS included knowledge structure analysis, conceptual layer design, entity layer construction and knowledge storage.

3.1. Knowledge Structure Analysis of CSS

CSS are divided into mandatory standards and recommended standards. Mandatory clauses are the most binding operational or regulatory requirements of mandatory standards. From the viewpoint of CSS release, CSS can be divided into national standards, industry standards, local standards, and corporate standards. In general, national standards have the largest scope of application, and enterprise standards have the smallest scope of application.
Knowledge structure analysis of CSS refers to the process of collecting standard documents, analyzing standard content, and dividing standard content levels. Knowledge structure reveals the components and levels of knowledge, which helps to understand CSS knowledge comprehensively and lay the foundation for the design of KGCSS. In order to analyze a CSS knowledge structure, we needed to obtain related materials from digital repositories.
Because official websites do not provide unified aggregation of CSSs, we aggregated related standards on websites, such as standard building library, national standard full text public system, and national standard information public service platform by searching for “construction management”, “construction technology”, and “construction safety”. There were 218 CSSs collected, including 65 national standards, 47 industry standards, 96 local standards, and 10 corporate standards. CSS knowledge can be divided into different granularity levels. Fine-grained knowledge is indivisible knowledge, and CSS fine-grained knowledge refers to the basic components of standard clauses. Coarse-grained knowledge is general and rough knowledge, and CSS coarse-grained knowledge refers to the full text of standards. Accordingly, the knowledge of CSS can be divided into the following three levels, and the knowledge structure is shown in Figure 1.
Terminology and definitions are basic components of standards and the basis of standard compilation. They include professional expression on objects and distinctive definitions of things, measures, status, and conditions in the construction industry.
Standard clauses are divided into mandatory and recommended clauses, including clauses that must be done, must not be done, recommended to be done, and should be done, etc. According to different objects, they can be divided into clauses for specialized construction matters, such as safety checks and firefighting, as well as specific targets like scaffolding and temporarily installed suspended access equipment.
The full texts of standards contain all the information of CSSs, including covers, tables of contents, prefaces, main contents, appendix, and explanation of clauses.

3.2. Construction of KGCSS

KGCSS is constructed in three steps: conceptual layer design, entity layer design, and knowledge storage.

3.2.1. Conceptual Layer Design

The concept of KGCSS refers to the categories to which CSS belong through a rational division. The concept layer of KGCSS is a normative and structured representation of CSS knowledge. There are five levels of CSS concepts. The first level is CSS. The second level is divided into basic, general, and specialized standards according to their applicability. The third level is a subdivision according to different usage scenarios. The fourth level is a further division for management standards. The fifth level is a division for specific objects or matters. In each level, concepts are further divided into different categories to suit different types of construction tasks. Accordingly, the conceptual layer framework is shown in Figure 2.
From the second level of CSS concepts, CSS is divided into three categories.
Basic Standards serve as the foundation for the formulation of other CSSs in the system. They serve to unify concepts covered by other standards and to minimize disputes in the compilation and application process caused by different interpretations of industrial terminology in the field of construction. Terminology Standards provide a unified language in the compilation process. Classification Standards put forward a criterion for reasonable division. Mark Standards specify the setting and maintenance of safety signs on the construction site. These three are defined as Basic Standards.
General Standards are unified standards that aim to regulate specific construction stages (such as transformation) and matters (such as electricity consumption), including general standards for construction safety technology and management. Construction Safety Technology General Standards make specific and specialized requirements for the operation of tools and the implementation of critical construction methods, including machinery, scaffolding, formwork, electricity, temporary construction, working at height, and construction supplies. Among the Construction Safety Management General Standards are management standards for subjects that provide guidance to safety management organizations, such as construction enterprises and supervision departments. Management standards for processes make provisions to regulate construction processes, including risk control, security check and other processes. Management standards for matters aim to guarantee important construction safety matters, such as hazardous major works. Management standards for things ensure the safe utilization of apparatuses in construction.
Compared to General Standards, Specialized Standards have a comparable structure to General Standards but relate to more specific management and technical objects.
Correspondingly, relationships between concepts in the KGCSS are organized into eight categories which are shown in Table 1.

3.2.2. Entity Layer Design

The entities of KGCSS are current CSS. KGCSS entity layer is a collection of clearly classified CSS corresponding to KGCSS concepts. The steps of KGCSS entity layer construction include: entity identification, attribute extraction, and entity extraction.
The 218 current standards collected in Section 3.1 are pending entities to be identified and categorized in accordance with CSS concepts. Furthermore, the names, numbers and mandatory clauses of current standards are extracted as properties of entities. Based on the structure of the conceptual layer, the collected 218 CSS were sorted and filtered to form an entity layer, which finally contains 147 standard entities. The relationships from the entity to the corresponding concept are designed with the name “is an instance of”. The construction logic of the entity layer is shown in Figure 3.
Entity extraction is the process of matching, classifying, and filtering entities with concepts one by one. It is a unique step in the establishment of KGCSS. For instance, the name “Safety Rules for Lifting Appliances” contains two main parts: “working at height” and “appliances”, which correspond to the concepts of “working at height” and “machinery” respectively (in Figure 2). Therefore, it is necessary to identify a conceptual classification for the standard according to its main content. The “Safety Rules for Lifting Appliances” applies to aerial work platforms and aerial work vehicles. It aims to guide the design, manufacture, use, maintenance, and management of aerial work machinery products. It also contains several mechanical terms, such as leveling mechanisms, wire ropes and hydraulic systems. In addition, the requirements for users are to maintain normal operation of the machinery, rather than guiding workers on how to work at height. Therefore, “Safety Rules for Lifting Appliances” should be classified as general standards for machinery technology. The materialization process is shown in Figure 3. In addition, standards will be removed from the entities if they do not meet any concept, as in the example of the general safety technology standard; the results of extracting entities are shown in Table 2.

3.2.3. Knowledge Storage

Knowledge storage refers to importing knowledge into a database to serve upper-layer applications, such as knowledge presentation, knowledge reasoning and intelligent Q&A. The knowledge of CSSs is stored in the Neo4j database in this research. Neo4j supports operational functions, including add, delete, and modify, making it convenient to adjust the design of KGCSS [64]. Organized CSS knowledge is stored in a Neo4j database via py2neo module in Pycharm. The knowledge graph contains 177 nodes, 263 relations and 11 properties. The result of knowledge storage is shown in Table 3.

4. Results

Based on the KGCSS constructed in Section 3, this subsection implements KGCSS in practice, including automatic reasoning of CSS planning knowledge retrieval based on natural language processing.

4.1. Automatic Reasoning of CSS Planning

Standard planning can be achieved from the current CSS system via knowledge reasoning. CSS planning involves a list of standard work plans that serve as a roadmap for CSS development and revision. Planning standards are derived from “Missing Standards”, “Standards to be Integrated” and “Standards That Require Complementary Knowledge”. First, standards that are absent from the list of basic and general standards are “Missing Standards”, such as “Construction Safety Terminology Standard”. In addition, under the same concept, standards with similar themes should be regarded as “Standards to be Integrated” [19]. For instance, “Temporarily Installed Suspended Access Equipment”, “Suspended Powered Platforms for Work at Height-safety Regulation”, and “Technical Specification for Installation Dismantlement and Operation of Temporarily Installed Suspended Access Equipment” are all intended to regulate the implementation of temporarily installed suspended access equipment, and as such they should be integrated and unified as “Safety Technical Regulations for Temporarily Installed Suspended Access Equipment”. Additionally, “Standards That Require Complementary Knowledge” refers to standards that are present in General Standards but not in Specialized Standards, such as “Safety Code for Tower Cranes”.
Moreover, planning standards are divided into four levels of priorities to indicate the importance and compilation urgency of each standard. Priorities are determined based on the planning standards’ attributes. Missing basic or general standards are the first priority. “Standards to be Integrated” are determined as the second priority. The third priority is set for “Standards That Require Complementary Knowledge”. Missing specialized standards are the fourth priority. Results of the CSS plan is shown in Table 4.

4.2. Knowledge Retrieval Based on Natural Language Processing

Natural language is the language used in daily human communication. The CSS knowledge Q&A function relies on domain entities and conducts reasoning based on KGCSS to answer users’ questions in depth [53]. Knowledge Q&A is better at addressing knowledge-based questions and is more direct and intuitive in providing replies than template-based chatbots. This subsection implements an automatic function of Q&A in the KGCSS via question classification, rule-based matching, query statement generation and answer template invocation [32].
First, types of questions are defined and classified. Common problems of standard users have been identified according to the results of expert seminars. Questions are divided into five categories, including: “latest version of the standard”, “mandatory clauses of the standard”, “standards involved in a certain field”, “mandatory clauses involved in a certain field”, and “standard planning in a certain field”.
Second, matching rules are defined for each question. Matching rules specify what conditions need to be met for a question sentence to be classified as a type of question we designed. The matching rules include a feature word lexicon and a question word bank. Feature words include concepts and entities of KGCSS. Question words refer to the objects of question sentences. When both feature words and question words contained in a question sentence match the rules of a question type, the sentence is classified as that type of question. For example, when the sentence “What are the mandatory standard clauses in the field of machinery management?” is typed in, the question type is obtained by matching as “mandatory clauses involved in a certain field”, as shown in Figure 4.
Third, query statements are designed for the five types of questions in order to activate the knowledge in KGCSS and answer templates. In other words, after the question is classified, the query statement will invoke its feature terms, search information in KGCSS and present the answer. The process and outcomes of the question “What are the mandatory standard clauses in the field of machinery management?” are shown in Figure 4.

5. Discussion

5.1. Innovation

CSS contains rich domain knowledge, which provides important materials for ensuring construction safety. The KGCSS constructed in this paper realizes CSS knowledge management, fills the research gap, solves the problems existing in the application of CSS, and has strong practical significance. The main innovations of this paper are as follows.
(1)
KGCSS is a universal and operational result for CSS management. It realizes systematical integration of CSS knowledge and better standard management.
(2)
The standard planning is obtained based on KGCSS reasoning, which is the most practical result of this paper. CSS planning is oriented to users’ needs and standard content, clarifying the future work tasks of CSS, and making CSS development and revision more in line with market needs.
(3)
Q&A function of KGCSS is designed. Compared with existing standard query websites, the CSS Q&A function realizes a knowledge search with the precision of clauses. Systematically integrated CSS knowledge is conducive to guiding workers’ behavior and supporting users’ decision making.
However, the exploration of CSS knowledge management is far from complete. Although KGCSS realizes deeper standard knowledge management, what KGCSS currently achieves is merely to communicate knowledge to users rather than assisting them how to make use of it, i.e., there is no guarantee of knowledge dissemination. In previous studies, technical requirements were translated into actionable instructions for workers through construction practice surveys [29]. However, the issue of how to achieve efficient dissemination of all standards in KGCSS remains to be addressed. Additionally, in other domains, machine learning methods and KG are combined for knowledge management, enabling more powerful applications, such as optimizing methods of knowledge representation of GeoKG [50] and distinguishing reliable relation paths among large amounts of meaningless relation paths [65]. Nevertheless, the knowledge in KGCSS is only accurate to the “clause” level, and query results of mandatory clauses are output in the form of paragraphs. The exploitability of KGCSS-based applications depends on the availability of a more formalized CSS knowledge structure, which also needs to be investigated in depth in the future.

5.2. Research Prospects

As a knowledge management tool, knowledge graph has broad research prospects. Based on this paper, future research of KGCSS can be focused on the following aspects.
Knowledge representation: the knowledge representation of CSSs solely reaches the accuracy of “clauses”, and mandatory clauses are only output as a whole paragraph. Thus, an event extraction algorithm can be introduced to create a structured representation of standard clause knowledge and to aid in more advanced knowledge reasoning.
Platform construction: the KGCSS can be used to establish an information platform accessible to the government and construction enterprises. With the accumulation of knowledge on the platform via standard users uploading their safety production experiences, experience feedback will serve as industrial knowledge to enrich current construction safety knowledge, as well as function as references to solve problems on various projects [66].

6. Conclusions

Given that much attention has been paid to CSS management, the purpose of this paper is to improve CSS application and management. Knowledge characteristics of CSS make them suitable as objects for knowledge management [21]. In view of this, this paper employs KG techniques for systematic management of CSS. Firstly, a three-level CSS structure is obtained by analyzing the knowledge of 218 CSSs. Secondly, based on the expert opinions, a conceptual layer design is carried out, and a five-level CSS concept layer is obtained. The 218 standards are reasonably correlated to the conceptual layer to form an entity layer. Finally, knowledge storage is used to construct KGCSS, which contains 147 entities and 177 nodes.
This paper achieves significant improvements in methodological applications and results of CSS management. First, this paper proposes a novel standard planning based on KGCSS.A CSS revision plan is deduced, including twenty-one “Missing Standards”, seven “Standards to be Integrated” and eight “Standards That Require Complementary Knowledge”. The standards plan is a valuable tool for meeting the most urgent needs of CSS usage, such as government’s supervision and contractors’ production. The plan provides recommendations for standard revisions and improves the completeness of CSS knowledge. In addition, a CSS query function has also been devised and implemented. Unlike the general standard query platform, it offers CSS knowledge at a granularity of clauses rather than the entire standard documents. However, the existing CSS knowledge presentation is less structured, which should be further addressed in future studies.
Nevertheless, there are still some limitations in our research. First, the KGCSS does not collect all of local standards in China due to resource limitation. A KGCSS of rich local standards applicable to different regions can be improved in the future. Moreover, this study solely considers standards as KGCSS entities. More detailed knowledge representation, such as regarding terminology as entities and developing their connections, should be explored in future research.

Author Contributions

Conceptualization, Y.J., W.S. and X.G.; methodology, Y.J. and X.G.; software, Y.J.; formal analysis, W.S. and Y.J.; resources, X.G.; writing—original draft preparation, Y.J.; writing—review and editing, Y.J. and J.L.; supervision, X.G. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China, grant number 72174146.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data sharing not applicable.

Acknowledgments

The authors thank the anonymous referees for their constructive suggestions.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Chi, C.F.; Chang, T.C.; Ting, H.I. Accident patterns and prevention measures for fatal occupational falls in the construction industry. Appl. Ergon. 2005, 36, 391–400. [Google Scholar] [CrossRef] [PubMed]
  2. Chan, A.P.C.; Wong, F.K.W.; Chan, D.W.M.; Yam, M.C.H.; Kwok, A.W.K.; Lam, E.W.M.; Cheung, E. Work at height fatalities in the repair, maintenance, alteration, and addition works. J. Constr. Eng. M. 2008, 134, 527–535. [Google Scholar] [CrossRef]
  3. Zhang, S.J.; Teizer, J.; Pradhananga, N.; Eastman, C.M. Workforce location tracking to model, visualize and analyze workspace requirements in building information models for construction safety planning. Automat. Constr. 2015, 60, 74–86. [Google Scholar] [CrossRef]
  4. Kines, P. Occupational injury risk assessment using injury severity odds ratios: Male falls from heights in the Danish construction industry, 1993-1999. Hum. Ecol. Risk Assess. 2001, 7, 1929–1943. [Google Scholar] [CrossRef]
  5. Kines, P.; Mikkelsen, K.L. Effects of firm size on risks and reporting of elevation fall injury in construction trades. J. Occup. Environ. Med. 2003, 45, 1074–1078. [Google Scholar] [CrossRef] [Green Version]
  6. Dong, X.S.; Fujimoto, A.; Ringen, K.; Stafford, E.; Platner, J.W.; Gittleman, J.L.; Wang, X. Injury underreporting among small establishments in the construction industry. Am. J. Ind. Med. 2011, 54, 339–349. [Google Scholar] [CrossRef]
  7. Health and Safety Executive of UK. Fatal Injuries in Great Britain. Available online: https://www.hse.gov.uk/statistics/fatals.htm (accessed on 15 August 2020).
  8. Im, H.-J.; Kwon, Y.-J.; Kim, S.-G.; Kim, Y.-K.; Ju, Y.-S.; Lee, H.-P. The characteristics of fatal occupational injuries in Korea’s construction industry, 1997–2004. Saf. Sci. 2009, 47, 1159–1162. [Google Scholar] [CrossRef]
  9. Sousa, V.; Almeida, N.M.; Dias, L.A. Risk-based management of occupational safety and health in the construction industry—part 1: Background knowledge. Saf. Sci. 2014, 66, 75–86. [Google Scholar] [CrossRef]
  10. Sousa, V.; Almeida, N.M.; Dias, L.A. Risk-based management of occupational safety and health in the construction industry—part 2: Quantitative model. Saf. Sci. 2014, 74, 184–194. [Google Scholar] [CrossRef]
  11. Sorock, G.S.; Smith, E.O.; Goldoft, M. Fatal occupational injuries in the New Jersey construction industry: 1983–1989. J. Occup. Med. 1993, 35, 916–921. [Google Scholar] [CrossRef] [PubMed]
  12. Mason, S. Procedural violations -causes, costs and cures. In Human Factors in Safety-Critical Systems; Redmill, F., Rajan, J., Eds.; Butterworth Heinemann: Oxford, UK, 1997; pp. 287–318. [Google Scholar]
  13. Wang, W.; Zhang, S.J.; King, A.P. Research on the adoption barriers of the engineering construction standards in China. Struct. Surv. 2016, 34, 367–378. [Google Scholar] [CrossRef] [Green Version]
  14. Hale, A.; Borys, D. Working to rule, or working safely? Part 1: A state of the art review. Saf. Sci. 2013, 55, 207–221. [Google Scholar] [CrossRef] [Green Version]
  15. Zhou, Z.P.; Goh, Y.M.; Li, Q.M. Overview and analysis of safety management studies in the construction industry. Saf. Sci. 2015, 72, 337–350. [Google Scholar] [CrossRef]
  16. Zhu, H.L.; Zhang, J. Analysis of the current situation of the engineering construction standards and its management system in China. Stand. Sci. 2011, 6, 10–14. [Google Scholar]
  17. Zhu, H.l.; Cui, J.J. Study on the relationship between engineering construction standards and laws of interaction. Sci. Technol. Prog. Policy 2009, 26, 99–102. [Google Scholar]
  18. Liu, S.J.; Liu, H. Considerations and Path for the Reform of China’s Standardization System. Strategy Decis. Mak. 2015, 7, 1–12. [Google Scholar]
  19. Zeng, S.X.; Tam, V.; Tam, C.M. Towards occupational health and safety systems in the construction industry of China. Saf. Sci. 2008, 46, 1155–1168. [Google Scholar] [CrossRef]
  20. Zhang, Y.L.; Gu, C.L.; Ding, Y.J. The Application of Functional Safety Standards on the Safety Management of a Chemical Process Tank. Appl. Mech. Mater. 2012, 130–134, 1590–1595. [Google Scholar] [CrossRef]
  21. Sherehly, B.; Karwowski, W. Knowledge management for occupational safety, health, and ergonomics. Hum. Factors Ergon. Manuf. 2006, 16, 309–319. [Google Scholar] [CrossRef]
  22. Hadikusumo, B.; Rowlinson, S. Capturing Safety Knowledge Using Design-for-Safety-Process Tool. J. Constr. Eng. Manag. 2004, 130, 281–289. [Google Scholar] [CrossRef]
  23. Nonaka, I.; Takeuchi, H. The Knowledge Creating Company. Harv. Bus. Rev. 1995, 1, 995. [Google Scholar]
  24. Sowa, J.F. Principles of Semantic Networks: Exploration in the Representation of Knowledge; Morgan Kaufmann Publishers: San Mateo, CA, USA, 1991; pp. 135–157. [Google Scholar]
  25. Wang, H.F. Large scale knowledge graph technology. Chin. Soc. Comput. Sci. 2014, 10, 64–68. [Google Scholar]
  26. Yu, T.; Liu, J.; Jia, L.R.; Zhang, Z.L.; Yang, S.; Liu, L.H. Study on the construction of large scale knowledge graph of traditional Chinese medicine. Chin. J. Digit. Med. 2010, 10, 80–82. [Google Scholar]
  27. Yan, J.; Lv, T.; Yu, Y. Construction and Recommendation of a Water Affair Knowledge Graph. Sustainability 2018, 10, 3429. [Google Scholar] [CrossRef] [Green Version]
  28. Yu, T.; Li, J.; Yu, Q.; Tian, Y.; Shun, X.; Xu, L.; Zhu, L.; Gao, H. Knowledge graph for TCM health preservation: Design, construction, and applications. Artif. Intell. Med. 2017, 77, 48–52. [Google Scholar] [CrossRef] [PubMed]
  29. Sun, Y.M.; Zhang, C.X. Establishment and Research of National Standard Code of Construction of Masonry Structures Engineering. Eng. J. Wuhan Univ. 2015, 48, 387–390+396. [Google Scholar]
  30. Malerba, F.; Orsenigo, L. Technological Regimes and Firm Behavior. Industrial and Corporate Change. Organization and Strategy in the Evolution of the Enterprise; Palgrave Macmillan: London, UK, 1996; pp. 42–71. [Google Scholar]
  31. Malerba, F.; Orsenigo, L. Schumpeterian patterns of innovation are technology-specific. Res. Policy 1996, 25, 451–478. [Google Scholar] [CrossRef]
  32. Xiao, Z.W.; Zhang, C.X. Construction of Meteorological Simulation Knowledge Graph Based on Deep Learning Method. Sustainabilityn 2021, 13, 1311. [Google Scholar] [CrossRef]
  33. Gruber, T.R. A Translation Approach to Portable Ontology Specifications. Knowl. Acquis. 1993, 5, 199–220. [Google Scholar] [CrossRef]
  34. Chau, K.W. An ontology-based knowledge management system for flow and water quality modeling. Adv. Eng. Softw. 2007, 38, 172–181. [Google Scholar] [CrossRef] [Green Version]
  35. Ding, L.Y.; Zhong, B.T.; Wu, S.; Luo, H.B. Construction risk knowledge management in BIM using ontology and semantic web technology. Saf. Sci. 2016, 87, 202–213. [Google Scholar] [CrossRef] [Green Version]
  36. Zhang, C.; Romagnoli, A.; Zhou, L.; Kraft, M. Knowledge management of eco-industrial park for efficient energy utilization through ontology-based approach. Appl. Energy 2017, 204, 1412–1421. [Google Scholar] [CrossRef]
  37. Antezana, E.; Kuiper, M.; Mironov, V. Biological knowledge management: The emerging role of the Semantic Web technologies. Brief. Bioinform. 2009, 10, 392–407. [Google Scholar] [CrossRef] [PubMed]
  38. Lin, J.J.; Zhao, Y.Z.; Huang, W.Y.; Liu, C.F.; Pu, H.T. Domain knowledge graph-based research progress of knowledge representation. Neural. Comput. Appl. 2021, 33, 681–690. [Google Scholar] [CrossRef]
  39. Yang, Y.; Xu, B.; Hu, J.; Tong, M.; Zhang, P.; Zheng, L. Accurate and Efficient Method for Constructing Domain Knowledge Graph. J. Softw. 2018, 29, 2931–2947. [Google Scholar]
  40. Miller, G.A. Wordnet—A Lexical Database for English. Commun. ACM 1995, 38, 39–41. [Google Scholar] [CrossRef]
  41. Miller, G.A.; Beckwith, R.; Fellbaum, C.; Gross, D.; Miller, K.J. Introduction to WordNet: An On-line Lexical Database. Int. J. Lexicogr. 1990, 3, 235–244. [Google Scholar] [CrossRef] [Green Version]
  42. Lenat, D.B. Cyc—A Large-Scale Investment in Knowledge Infrastructure. Commun. ACM 1995, 38, 33–38. [Google Scholar] [CrossRef]
  43. Speer, R.; Havasi, C. Representing General Relational Knowledge in ConceptNet 5. In Lrec 2012—Eighth International Conference on Language Resources and Evaluation; European Language Resources Association: Istanbul, Turkey, 2012; pp. 3679–3686. [Google Scholar]
  44. Liu, H.; Singh, P. ConceptNet—A practical commonsense reasoning tool-kit. Bt. Technol. J. 2004, 22, 211–226. [Google Scholar] [CrossRef]
  45. Lehmann, J.; Isele, R.; Jakob, M.; Jentzsch, A.; Kontokostas, D.; Mendes, P.N.; Hellmann, S.; Morsey, M.; van Kleef, P.; Auer, S.; et al. DBpedia—A large-scale, multilingual knowledge base extracted from Wikipedia. Semant. Web 2015, 6, 167–195. [Google Scholar] [CrossRef] [Green Version]
  46. Rebele, T.; Suchanek, F.; Hoffart, J.; Biega, J.; Kuzey, E.; Weikum, G. YAGO: A Multilingual Knowledge Base from Wikipedia, Wordnet, and Geonames. Lect. Notes Comput. Sci. 2016, 9982, 177–185. [Google Scholar] [CrossRef]
  47. Bollacker, K. Freebase: A collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD International Conference on Management of Data, Vancouver, BC, Canada, 10–12 June 2008. [Google Scholar]
  48. Tanon, T.P.; Vrandecic, D.; Schaffert, S.; Steiner, T.; Pintscher, L. From Freebase to Wikidata: The Great Migration. In Proceedings of the 25th International Conference on World Wide Web (Www’16), Montreal, WI, Canada, 11–15 May 2016; pp. 1419–1428. [Google Scholar] [CrossRef] [Green Version]
  49. Bizer, C.; Heath, T.; Berners-Lee, T. Linked Data—The Story So Far. Int. J. Semant. Web Inf. 2009, 5, 1–22. [Google Scholar] [CrossRef] [Green Version]
  50. Wang, S.; Zhang, X.Y.; Ye, P.; Du, M.; Lu, Y.X.; Xue, H.N. Geographic Knowledge Graph (GeoKG): A Formalized Geographic Knowledge Representation. ISPRS Int. J. Geo-Inf. 2019, 8, 184. [Google Scholar] [CrossRef] [Green Version]
  51. Rotmensch, M.; Halpern, Y.; Tlimat, A.; Horng, S.; Sontag, D. Learning a Health Knowledge Graph from Electronic Medical Records. Sci. Rep. 2017, 7, 1–11. [Google Scholar] [CrossRef]
  52. Dong, B.; Yu, H.Z.; Li, H.S. A Knowledge Graph Construction Approach for Legal Domain. Teh. Vjesn. 2021, 28, 357–362. [Google Scholar] [CrossRef]
  53. Zhang, Q.; Wen, Y.Q.; Zhou, C.H.; Long, H.; Han, D.; Zhang, F.; Xiao, C.S. Construction of Knowledge Graphs for Maritime Dangerous Goods. Sustainability 2019, 11, 2849. [Google Scholar] [CrossRef] [Green Version]
  54. Chen, P.; Lu, Y.; Zheng, V.; Chen, X.; Yang, B. KnowEdu: A System to Construct Knowledge Graph for Education. IEEE Access 2018, 6, 31553–31563. [Google Scholar] [CrossRef]
  55. Woo, J.H.; Clayton, M.J.; Johnson, R.E.; Flores, B.E.; Ellis, C. Dynamic Knowledge Map: Reusing experts’ tacit knowledge in the AEC industry. Automat. Constr. 2004, 13, 203–207. [Google Scholar] [CrossRef]
  56. Nickel, M.; Murphy, K.; Tresp, V.; Gabrilovich, E. A Review of Relational Machine Learning for Knowledge Graphs. Proc. IEEE 2016, 104, 11–33. [Google Scholar] [CrossRef]
  57. Paulheim, H. Knowledge Graph Refinement: A Survey of Approaches and Evaluation Methods. Semant. Web 2017, 8, 489–508. [Google Scholar] [CrossRef] [Green Version]
  58. Banko, M.; Cafarella, M.J.; Soderland, S.; Broadhead, M.; Etzioni, O. Open Information Extraction from the Web. In Proceedings of the 20th International Joint Conference on Artificial Intelligence, Yokohama, Japan, 7–15 January 2007; pp. 2670–2676. [Google Scholar]
  59. Carlson, A.; Betteridge, J.; Kisiel, B.; Settles, B.; Hruschka, E.R.; Mitchell, T.M. Toward an Architecture for Never-Ending Language Learning. In Proceedings of the Twenty-Fourth Aaai Conference on Artificial Intelligence (Aaai-10), Atlanta, GA, USA, 11–15 July 2010; pp. 1306–1313. [Google Scholar]
  60. Leivestad, S.; Mehus, J. Standards: A system of tools for the construction industry—interface and interaction between standards. Struct. Concr. 2010, 11, 57–62. [Google Scholar] [CrossRef]
  61. Qiao, L.; Yang, L.; Hong, D.; Yao, L.; Zhiguang, Q. Knowledge graph construction techniques. J. Comput. Res. Dev. 2016, 53, 582–600. [Google Scholar]
  62. Gruber, T. Ontology. In Encyclopedia of Database Systems; Liu, L., Özsu, M.T., Eds.; Springer: New York, NY, USA, 2009; ISBN 978-0-387-35544-3. [Google Scholar]
  63. Geoffrey, J.R. Creative Thinking and Brainstorming; Routledge: London, UK, 1981. [Google Scholar]
  64. Miller, J. Graph Database Applications and Concepts with Neo4j. In Proceedings of the Southern Association for Information Systems Conference, Atlanta, GA, USA, 23–24 March 2013; pp. 141–147. [Google Scholar]
  65. Seo, S.; Oh, B.; Lee, K.H. Reliable Knowledge Graph Path Representation Learning. IEEE Access 2020, 8, 32816–32825. [Google Scholar] [CrossRef]
  66. Kamsu-Foguem, B.; Abanda, F.H. Experience modeling with graphs encoded knowledge for construction industry. Comput. Ind. 2015, 70, 79–88. [Google Scholar] [CrossRef] [Green Version]
Ijerph 18 10692 g001 550
Figure 1. Structure analysis of construction safety standard knowledge.
Figure 1. Structure analysis of construction safety standard knowledge.
Ijerph 18 10692 g001
Ijerph 18 10692 g002 550
Figure 2. Conceptual layer framework of KGCSS.
Figure 2. Conceptual layer framework of KGCSS.
Ijerph 18 10692 g002
Ijerph 18 10692 g003 550
Figure 3. Construction logic of the entity layer of KGCSS.
Figure 3. Construction logic of the entity layer of KGCSS.
Ijerph 18 10692 g003
Ijerph 18 10692 g004 550
Figure 4. Q&A function of KGCSS.
Figure 4. Q&A function of KGCSS.
Ijerph 18 10692 g004
Table
Table 1. Concept relationships of construction safety standard knowledge graph.
Table 1. Concept relationships of construction safety standard knowledge graph.
Relation TypesExplanation
Basis of compilationIt is the relation between Basic Standards and Construction Safety Standards, denoting that standard compilation should comply with terminology and definitions in Basic Standards.
All categories applicableIt is the relation between General Standards and Construction Safety Standards, denoting applicable standards can be found in General Standards for all types of construction projects.
Specific objects applicableIt is the relation between Specialized Standards and Construction Safety Standards.
Kind ofIt is the classification relation between concepts.
Classified by subjectsIt is the relation between management standards for subjects and construction safety management standards.
Classified by processesIt is the relation between management standards for processes and construction safety management standards.
Classified by mattersIt is the relation between management standards for matters and construction safety management standards.
Classified by thingsIt is the relation between management standards for things and construction safety management standards.
Table
Table 2. Entities of construction safety standard knowledge graph (part).
Table 2. Entities of construction safety standard knowledge graph (part).
ConceptsEntities (Current Standards)
Machinery Technology General StandardSafety Rules for Lifting Appliances
Technical Specification for Safety of Installation, Use and Disassembly of Building Hoist in Construction
Safety of Machinery: Prevention of Unexpected Start-up
Safety of Machinery: Safety Distances to Prevent Hazard Zones being Reached by Upper and Lower Limbs
Safety Rules for Aerial Work Machinery
Mold Frame Technology General StandardTechnical Code for Safety of Forms in Construction
Technical Code for Temporary Support Structures in Construction
Scaffolding Technology General StandardUnified Standard for Safety of Scaffold in Construction
Electricity Use Technology General StandardGeneral Guide for Safety of Electric User
Code for Safety of Power Supply and Consumption for Construction Site
Technical Code for Safety of Temporary Electrification on Construction Site
Work-at-height Technology General StandardTechnical Code for Safety of Working at Height of Building Construction
…………
Table
Table 3. Aggregation of construction safety standard knowledge storage results.
Table 3. Aggregation of construction safety standard knowledge storage results.
Knowledge TypesQuantityExamples
ConceptsLevel 11Construction Safety Standard
Level 23Basic Standard, General Standard, Specialized Standard
Level 34Construction Safety Management General Standard, Construction Safety Technology General Standard, Construction Safety Management Specialized Standard, Construction Safety Technology Specialized Standard
Level 48Construction Safety Management General Standard for Subjects, Construction Safety Management General Standard for Processes, etc.
Level 551Enterprise Safety Management General Standard, Safety Plan General Standard, Safety Management General Standard for Dangerous and Large Projects, etc.
EntitiesCurrent standards111Technical Code for Safety of Lifting in Construction, Technical Code for Fire Safety of Construction Site, Safety Rules for Lifting Appliances—Part1: General, etc.
PropertiesStandard numbers111GB2893-2020, GB2894-2018, JGJ59-2011, etc.
Mandatory clauses58What should be done, what is forbidden, etc.
………………
RelationsRelations between concepts8See Table 1
Relations between entities1Instance of
Table
Table 4. Construction safety standard planning.
Table 4. Construction safety standard planning.
PriorityQuantityNames of Planning Standards
1st20Construction Safety Terminology Standard, Construction Material Classification Standard, Risk Level Classification Standard, Construction Project Safety Supervision Standard, Construction Project Team Safety Management Standard, etc.
2nd7Construction Machinery Safety Technology Unified Standard, Construction Site Temporary Electricity Safety Technology Standard, Construction Mold Frame Unified Safety Technology Standard, Construction Scaffolding Unified Safety Technology Standard, etc.
3rd8Fixed Scaffolding Safety Technical Standard, Mobile Scaffolding Safety Technical Standard, Tower Crane Safety Technical Standard, Prefabricated Building Safety Technical Standard, etc.
4th1New Technology, New Techniques, New Materials, New Equipment and Safety Contents involved
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Jiang, Y.; Gao, X.; Su, W.; Li, J. Systematic Knowledge Management of Construction Safety Standards Based on Knowledge Graphs: A Case Study in China. Int. J. Environ. Res. Public Health 2021, 18, 10692. https://doi.org/10.3390/ijerph182010692

AMA Style

Jiang Y, Gao X, Su W, Li J. Systematic Knowledge Management of Construction Safety Standards Based on Knowledge Graphs: A Case Study in China. International Journal of Environmental Research and Public Health. 2021; 18(20):10692. https://doi.org/10.3390/ijerph182010692

Chicago/Turabian Style

Jiang, Yukun, Xin Gao, Wenxin Su, and Jinrong Li. 2021. "Systematic Knowledge Management of Construction Safety Standards Based on Knowledge Graphs: A Case Study in China" International Journal of Environmental Research and Public Health 18, no. 20: 10692. https://doi.org/10.3390/ijerph182010692

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop