A Tale of Three Words: Knowledge, Safety, and Graphs

Simone, Francesco; Montaruli, Andrea; Fandino, Kristopher Hernandez; Patriarca, Riccardo

doi:10.3390/info17060599

Open AccessSystematic Review

A Tale of Three Words: Knowledge, Safety, and Graphs

Department of Mechanical and Aerospace Engineering, Sapienza University of Rome, Via Eudossiana 18, 00184 Rome, Italy

^*

Authors to whom correspondence should be addressed.

Information 2026, 17(6), 599; https://doi.org/10.3390/info17060599 (registering DOI)

Submission received: 9 April 2026 / Revised: 1 June 2026 / Accepted: 8 June 2026 / Published: 15 June 2026

(This article belongs to the Special Issue Knowledge Graph Technology and Its Applications, 3rd Edition)

Download

Browse Figures

Versions Notes

Abstract

The growing complexity of modern systems has pushed safety science beyond tradition-al analysis methods. In a world where the unknown matters as much as the known, knowledge graphs emerge as a powerful means for representing, connecting, and extending knowledge. However, the intersection between safety science and knowledge graphs remains largely unexplored. Which communities of researchers are leveraging knowledge graphs for safety? Is there any common pattern in how they are being used? This paper addresses these questions by presenting a systematic review of the literature on the use of knowledge graphs in the context of safety. Based on 173 eligible documents, we propose a classification framework structured around three dimensions: the originality of knowledge characterization, the originality of knowledge extraction, and the maturity of safety analysis. The framework identifies three archetypes of knowledge graph users: Assemblers, who rely on existing models and tools; Alchemists, who adapt available knowledge structures or extraction procedures; and Shapers, who develop novel ontologies, extraction methods, or both. The obtained results show how the latter represents the largest group among the reviewed studies, suggesting a tension between analytical maturity and the need for customized solutions. More broadly, the classification framework presented in this review may support researchers from both the safety and the artificial intelligence communities in fostering a shared path for the scientific development of these disciplines.

Keywords:

artificial intelligence; safety management; knowledge graphs; operations management; complex systems

1. Introduction

This paper presents a literature review that centers around three key concepts: knowledge, safety, and graphs. The following subsections explain the significance of each term in the context of the research and show how they are intricately interconnected.

1.1. Knowledge

Science has long pursued the understanding of reality [1], grounding this quest for knowledge in the scientific method, thus relying on observation, hypothesis, experimentation, and validation to develop knowledge [2]. This widespread accepted process boosts our confidence in how reality works, leading us to believe we can control and reproduce it. Yet even well-established knowledge continuously turns out to be flawed or incomplete [3]. As Socrates suggested [4], recognizing the boundaries of what we know is the first step to deepening and leveraging knowledge more effectively.

1.2. Safety

This problem of (not) knowing is particularly relevant to safety. Traditionally, safety has been associated with the absence of failures [5], relying on risk assessments and analyses to predict and avoid accidents, thus keeping systems safe. These approaches have promoted a vast body of knowledge, enough to consider safety as a science [6]. However, in real-world contexts, especially within complex socio-technical systems, traditional safety science falls short. Interactions among human, technology, and organizational factors push operations towards unpredictable unknowns. This complexity prompted a shift in safety science [7], especially with the introduction of Resilience Engineering, emphasizing the importance of adaptation and recovery, rather than the “simple” avoidance of failures [8]. In this new view of safety, it is acknowledged that we cannot fully know all possible failure modes [9]. Rather, safety deeply depends on the system’s adaptive capacity in response to both known and unknown challenges. This brings us to a key question: to what extent is the future of safety science shaped not only by the generation of new knowledge, but also by the need to more effectively organize and structure what is already known?

1.3. And Graphs

If we believe the answer is “yes”, then the way experts represent safety knowledge becomes crucial. In artificial intelligence (AI), knowledge representation refers to the structured ways to enable machines to simulate reasoning. To this end, human knowledge is formalized and transformed into models that machines can understand and use. Among the various techniques for doing so, knowledge graphs (KGs) have gained increasing attention, especially because of the few constraints they pose on representing knowledge, and their reliance on well-established mathematical formalisms [10]. KGs represent real-world entities as nodes and the relationships between them as edges, forming a semantic network whose elements are characterized following specific organizing principles. Based on these premises, do KGs have the potential to be the missing link in building a safety science that evolves with the growing complexity of our world?

1.4. Aim of This Research

Although KGs show great promise, their application in safety science is still evolving, with no formal investigation to characterize their role and utility within this field. This paper aims to fill this gap by conducting a systematic review to answer two research questions:

RQ1. What is the current state (number and typology of documents) of research leveraging KGs for safety-related purposes and who are the communities and industrial sectors mainly involved in it?

RQ2. Which common patterns can be identified in the way KGs are constructed and used to support safety analysis?

To address these questions, the paper makes three main contributions. First, it provides a systematic overview of the literature on KGs in safety contexts, highlighting publication trends, document types, and application domains. Second, it proposes a classification framework based on the originality of knowledge characterization, the originality of knowledge extraction, and the maturity of safety analysis, also identifying three archetypes of users with respect to the level of customization they employ in the KG construction, i.e., Assemblers, Alchemists, and Shapers. The proposed framework aims to bridge the gap between safety and AI domains, providing researchers with a structured tool to make sense of existing works, and to guide their future studies.

Third, the framework is applied to the pool of eligible documents, eventually discussing the trade-off between customization and analytical maturity in existing research.

2. Method

This systematic review was conducted in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [11]. The study selection process is summarized in the PRISMA flow diagram in Figure 1, and the main methodological steps are detailed as follows:

Identification. The Scopus database was selected as the source for collecting research contributions on the topic. This choice was driven by the database’s broad and comprehensive coverage of major venues in engineering and AI, ensuring a wide access to the relevant literature while minimizing the retrieval of duplicates. The following query was used: “TITLE-ABS-KEY (“*knowledge graph*”) AND (“*safety*”)”. The definition of the query was intentionally designed to ensure conceptual coherence and to focus exclusively on studies explicitly framed within this research stream.

Regarding the first keyword (i.e., “*knowledge graph*”), related terms such as “ontology”, “semantic network” and “RDF” were also considered. Nevertheless, control searches showed that their inclusion would either largely overlap with the retrieved corpus (i.e., 92%), or extend it toward adjacent but conceptually distinct research areas not consistently aligned with the KG technology. In particular, the term “ontology” substantially increased the retrieval volume (i.e., +245%) but introduced a different and broader research stream.

A similar rationale guided the use of “*safety*” as the second keyword. Alternative terms such as “hazard” and “resilience” were tested but did not significantly change the results yielding a 91% overlap, whereas “risk” produced a substantial increase in retrieved documents (i.e., +130%). However, this latter term is highly polysemous and frequently used in fields such as finance, project management, cybersecurity, and decision theory without a direct link to safety science. To reduce the conceptual noise, only “safety” was retained, as it typically denotes work focused on harm prevention or safe system behavior, including contexts where safety is interpreted as the absence of risk. These choices prioritize precision over recall, but also represent a limitation, as a broader query might have yielded a larger and more inclusive corpus.

Additional filters were then applied to refine the results to: (i) final-stage publications, (ii) written in English, and (iii) falling within the subject areas of computer science, engineering, mathematics, decision sciences, and environmental sciences. The document types were further filtered to exclude reviews, conference reviews, and surveys. This process yielded an initial pool of 523 records retrieved among those indexed in Scopus as of 20 June 2025.

Screening. Among the 523 documents, 273 were classified as out of scope after title, abstract, and keyword reading. In this phase, to promote initial consistency in the dataset while maintaining efficiency in the analysis process, all the documents were evaluated by a single author. Nevertheless, clear inclusion and exclusion criteria were jointly defined by all the authors, thereby minimizing the potential for subjective interpretation. Specifically: (i) documents that did not mention issues related to operational or occupational safety had to be excluded; (ii) documents that did not mention approaches nor solutions based on knowledge graphs had to be excluded; (iii) reviews and survey papers had to be excluded. Accordingly, among the resulting 273 exclusions: (i) 245 documents did not address safety science by means of operational or occupational safety as a central theme, but rather were related to, e.g., food and drug safety or medical care, or failed to mention safety-related aspects altogether; (ii) 12 documents were excluded because the use of KGs was not the primary focus of the research; (iii) 15 documents were non-empirical studies, consisting of reviews or surveys; and (iv) 1 document was identified as a duplicate entry. As a result, 250 documents were moved to the eligibility step.
Eligibility. The full text of the 250 documents was read to identify documents to be included in the analysis. At this stage, to ensure methodological rigor and minimize potential subjective bias, the dataset was evaluated independently by three raters, each assessing all documents’ eligibility. The inter-rater reliability was evaluated using Fleiss’ Kappa among the raters, which yielded a score of 0.773, indicating almost perfect agreement. The corresponding test statistic was z = 27.2, with a statistically significant result (p < 0.001), confirming that the observed level of agreement was substantially higher than would be expected by chance alone. Disagreements were discussed among the three raters in order to reach a common consensus. Based on this assessment, 76 documents were finally excluded for the following reasons: (i) 1 document was a short demo abstract, erroneously indexed as a conference paper, that lacked sufficient detail to be included; (ii) 23 documents did not explain the data source (either to construct or import the KG), or they provided only vague explanation in this regard, substantially lacking technical transparency; (iii) 43 documents were excluded due to the impossibility of accessing the full text under the authors’ institutional license; (iv) 4 documents were excluded because they treated the KGs as passive input data structures without any functional exploitation of the graph’s features, failing to demonstrate how the KG provided a specific advantage over traditional “flat” data formats; (v) 5 documents did not address safety-related topics explicitly; (vi) 1 document was later labeled as “retracted” by the publisher. As a result, 173 documents were deemed eligible.

A post hoc analysis was additionally conducted to provide transparency about the 43 documents excluded in (iii). They revealed a consistent distribution over time (average 8.3% per year, starting from 2019), indicating no significant temporal bias. However, the document type was highly skewed toward conference papers (33 documents out of 43, i.e., 77%), compared to journal articles (7 documents out of 43, i.e., 16%) and book chapters (3 documents out of 43, i.e., 7%), highlighting a potential bias against proceedings from smaller publishers or specialized venues.

Analysis. We recorded relevant metadata for each of the 173 eligible documents, including, e.g., the year of publication, the document type, the authors’ affiliation, the citation count, and the publication source-related metrics. An inductive coding process was then conducted to extract themes across the dataset. The codes reflected both application-specific and methodological aspects, capturing details such as, e.g., the algorithm and the software employed, the obtained results, and the input data structure. Rather than refining the code set for uniformity across the corpus, the codes were used as a foundational baseline for developing a broader classification framework, in which each document was subsequently categorized (see Section 3.2).

Figure 1. PRISMA flow diagram.

The complete set of documents included in the analysis is reported in Appendix A Table A1, and it is publicly available for transparency and reproducibility in [12].

3. Results

This section presents the documents from two perspectives: (i) a bibliometric overview of the eligible literature, followed by (ii) the introduction of the framework adopted for their classification.

3.1. Bibliometric Analysis

The bibliometric analysis considered: (i) the year of publication of the document, (ii) its type, and (iii) the industrial sector domain investigated. The year of publication shows that the use of KGs for safety is a recent trend starting from 2018. Figure 2 reports the temporal distribution of documents from 2018 to 2025. The observed numbers show a rapidly increasing trend from 2018 to 2024 that is well approximated by an exponential model, as confirmed by a high goodness-of-fit in log-linear space (i.e.,

R^{2} = 0.986

), suggesting near-exponential growth over the analyzed period (i.e., the number of publications doubled every 1.037 years). A downward trend can be observed after 2024, but this is justified by the documents’ collection taking place on 20 June 2025; that is, we did not consider documents published in the second half of 2025.

Figure 3 illustrates the type of documents through a pie chart. The majority (64.2%) are journal articles, suggesting that the KGs’ usage requires detailed explanation and discussion. This advocates the topic to be under active development, probably driven by the recent surge of interest in machine learning and large language models (LLMs). The other types of documents are conference papers (35.3%), which are usually shorter. In conference papers, there is less focus on the rationale that supports the KG usage, and more emphasis on the results obtained. Only one book chapter was included (0.5%).

The eligible documents were classified according to sector, industry group, and industry, through the Global Industry Classification Standard taxonomy [13]. Figure 4 includes such results, showing that the “Industrials” sector is the most investigated (61.3%), followed by the “Utilities” sector (19.0%). In general, the most covered cases relate to: (i) Construction and Engineering (19.0%), (ii) Electric Utilities (9.8%), and (iii) Aerospace and Defense (9.2%).

3.2. A Framework for Characterizing Safety Research Leveraging Knowledge Graphs

The in-depth review of the 173 documents permitted the deductive development of a framework to classify the use of KGs in safety-related research. The framework is structured around two core dimensions: (i) the way knowledge is characterized (i.e., the need for an organizing principle); and (ii) the way knowledge is extracted (i.e., the need to apply the organizing principle), each evaluated in terms of its originality. A third dimension, the maturity of the safety analysis, captures how effectively the KGs support safety insights.

The proposed dimensions were designed taking inspiration from existing maturity-oriented frameworks. In particular, the OC dimension relates to ontology-engineering maturity models such as the NeOn methodology, distinguishing between the reuse, adaptation, and development of novel conceptual structures [14]. Similarly, the OE dimension reflects increasing levels of procedural originality, conceptually aligned with technology-readiness level (TRL) and ML-pipeline maturity approaches, where the transition occurs from the adoption of established procedures to the development of fully novel extraction methods [15]. Finally, the MSA dimension parallels recognized reasoning-capability hierarchies in semantic systems, ranging from descriptive storage and retrieval functionalities to inference-oriented and operational knowledge exploitation [16,17]. On this path, the framework does not aim to replace prior maturity scales, but rather to integrate and contextualize them specifically for the analysis of KG-supported safety research. Accordingly, the three dimensions were assessed on four-level scales, as detailed hereafter. Please note that such scales represent discrete ordinal levels within a qualitative classification scheme, and the numbers serve as categorical markers rather than being the output of any calculation.

The originality of knowledge characterization (OC) refers to how knowledge is organized within the graph. Its scale includes:

(OC = 1) Unavailability of characterization. No organizing principle is mentioned, making it impossible to (re-)build a KG. This level represent a “zero point” in which a characterization cannot be present. As such, it serves as a theoretical baseline, designed to ensure symmetry with the extraction scale. In a review of KG applications to a specific field, no documents are expected to fall into this category, as its inclusion identifies the point where a study ceases to be KG-based research and reverts to traditional data analysis.
(OC = 2) Use of existing characterization. It denotes the use of an existing characterization without any modification by, e.g., adopting a pre-defined ontology. A clear reference to the used characterization shall be present.
(OC = 3) Adaptation of existing characterization. It involves adapting an existing model to better suit a specific application by, e.g., extending a taxonomy or redefining relationships. A clear reference to the adapted characterization shall be present.
(OC = 4) Development of characterization. Authors develop a fully customized knowledge characterization model from scratch, defining labels for nodes, edges, and attributes in detail. No reference to an existing characterization shall be present.

From the perspective of ontology-engineering maturity, OC = 2 corresponds to ontology reuse scenarios discussed in NeOn-like approaches, OC = 3 reflects ontology reengineering and extension practices, while OC = 4 captures fully original ontology development processes.

The originality of the knowledge extraction procedure (OE) assesses how the data is transformed into graph elements. Its scale includes:

(OE = 1) Availability of information. It indicates that all information is already available in a structured ready-to-graph manner (e.g., RDF format) and no extraction process is needed. Manual extraction is included here. Similarly to the “Unavailability of characterization” level, this one represent a “zero point” for the OE dimension, in which the already available information led to the non-necessity of extraction procedures.
(OE = 2) Use of existing procedure. It corresponds to using a known extraction method without any modification. A reference to the approach shall be present.
(OE = 3) Adaptation of existing procedure. It represents the modification or the combination of existing algorithms to better fit the problem context. A reference to the adapted procedure(s) shall be present.
(OE = 4) Development of procedure. It is assigned when a completely novel extraction method is developed and described, including, e.g., its logics or pseudocodes.

In this sense, the OE scale mirrors progression patterns commonly found in TRL-like and ML maturity models, where increasing levels correspond to greater methodological customization, integration complexity, and procedural innovation.

Combining the OC and OE dimensions allowed for the identification of three distinct archetypes of KGs users, namely:

(OC = 2 AND OE ≤ 2) Assemblers. They use pre-existing models and tools without the need for modifications. They employ KGs in a straightforward, utilitarian way, relying on established resources.
[(OC = 3 AND OE ≤ 3) OR (2 ≤ OC ≤ 3 AND OE = 3)] Alchemists. They transform or modify existing knowledge structures or processes to suit their needs. They innovate by adapting what is available rather than building anew.
[(OC = 4 AND OE ≤ 4) OR (2 ≤ OC ≤ 4 AND OE = 4)] Shapers. They are the most original users, creating entirely new models or extraction methods, or both. They treat the KG as a tailored solution, to be built specifically for the domain and purpose they investigate.

The partitioning logic of the OC and OE dimensions is structured to ensure that the archetypes are mutually exclusive. Consequently, each document is uniquely assigned to a single category, excluding any logical overlap between the profiles.

Beyond their construction, the framework also evaluates how KGs are leveraged for safety analysis, through a scale indicating the maturity of safety analysis (MSA):

(MSA = 1) Showcase of safety data. The graph is used to simply store and show safety-related data, without any analysis of them.
(MSA = 2) Retrieval of safety information. It involves basic information retrieval via simple queries which did not exploit the knowledge structure at its best (relationships are not navigated).
(MSA = 3) Reveal of safety knowledge. It represents a deeper use of the graph’s structure to reveal insights and relationships among seemingly disconnected information.
(MSA = 4) Inference of safety knowledge. Knowledge from the KG is not only derived but operationalized to solve practical safety problems.

The MSA scale is conceptually aligned with semantic reasoning hierarchies proposed in the knowledge-engineering literature, progressing from descriptive and retrieval-oriented uses toward inference-driven and decision-support-oriented exploitation of graph knowledge.

The conceptual space defined by the OC, OE, and MSA dimensions is represented in Figure 5, with archetypes of users highlighted by areas on the OC-OE plane. An area for “Uncharacterized concepts” identifies data with no organizing principle applied, thus stressing those occurrences in which representing knowledge through a KG is not feasible. Accordingly, given the scope of this review, no documents fell in this area from the eligible corpus. However, it would have included 27 documents (i.e., the 23 excluded in (ii), and the four excluded in (iv); see Eligibility phase in Section 2) if articles in which the KG employment was unclear had also been considered eligible.

The 173 documents were positioned within the framework’s space, with full classification details publicly available in [12]. Such classification was performed through a blind independent coding process involving three authors. Fleiss’ Kappa was calculated for each dimension, yielding values of 0.850 for OC, 0.821 for OE, and 0.908 for MSA, respectively. This indicated an almost perfect agreement among the raters, eventually supporting the reliability of the classification framework. In the few instances of disagreement, the final scores were determined through consensus-based discussions.

Figure 6 includes a bubble chart with the reviewed documents’ distribution in the OC-OE plane, showing the empirical distribution of the reviewed corpus in relation to the proposed archetype regions.

3.2.1. About Assemblers

Within the reviewed corpus, 28 documents (16%) were classified as Assemblers. All these studies scored 2 in the OC dimension, indicating a consistent reliance on pre-existing characterizations. In terms of the OE dimension, eight documents (29%) scored 1, while 20 (71%) scored 2, indicating a propension to specify the extraction procedure used rather than presenting the input data only. A variety of extraction models were employed in the 20 documents with OE = 2, with the most common including: natural language processing (NLP)-based techniques (four documents, 20%), BERT (six documents, 30%), BiLSTM (three documents, 15%), and prompt-based interactions with LLMs (three documents, 15%). The remaining four documents (20%) relied on other solutions. Among the eight documents with OE = 1, five out of eight (63%) did not specify how entities and relationships were extracted at all, possibly implying that the graph elements were readily available, while the remaining three (37%) were explicitly declared to perform manual extraction. Looking at MSA, 11 documents (39%) scored 4, five (18%) scored 3, and 12 (43%) scored 2. Among them, pure Assemblers (i.e., those scoring 2 in both OC and OE) account for 20 documents (71%). Within this subgroup, seven documents (35%) reached MSA = 4, four (20%) achieved MSA = 3, and nine (45%) scored MSA = 2. These findings suggest that the maturity of the safety analysis does not benefit from automated extraction. From a different perspective, this is further confirmed by the uncorrelation between MSA and OC and OE scores, verified by a Chi-square test of independence (χ² = 0.668, p = 0.716), and Fisher’s exact test (p = 0.778) due to the small sample size. Both tests indicate no statistically significant difference between the pure Assemblers and the rest of documents in the archetype. In other words, the comparison among the Assemblers’ documents suggests that among the 20 pure Assemblers, there is no clear trend linking OC and OE scores to higher MSA levels; while some studies achieved mature safety analyses despite relying solely on pre-existing ontologies and extraction methods, these cases appear isolated rather than systematic.

3.2.2. About Alchemists

The Alchemist archetype includes 67 documents (39%), more than double the number of Assemblers. This group comprises studies in which at least one of OC or OE scored 3. Among them, 42 documents (63%) rely on at least one pre-existing solution from the literature, applied without modification. The remaining 25 documents (37%) are classified as pure Alchemists, applying adaptations in both OC and OE. In the OE dimension, a recurring pattern emerged around the customization of BERT-like models: 18 documents (26%) propose modified extraction approaches based on variations in the encoder architecture. On the OC side, most Alchemists (48 out of 67, 71%) adapt a pre-existing ontology after conducting a preliminary screening, often guided by the Stanford 7-step [18] or similar frameworks [19]. This widespread practice suggests that formal methods of ontology construction serve as a key driver in adapting existing models. Turning to the MSA dimension, 26 documents (39%) scored 2, 23 (34%) scored 3, and 18 (27%) scored 4. This descending trend in MSA holds almost even when focusing exclusively on pure Alchemists, where 11 documents (44%) scored 2, eight (32%) scored 3, and six (24%) scored 4. These findings suggest that there is no clear correlation between being a pure Alchemist and achieving a higher level of analytical maturity. This hypothesis is confirmed, as a result of performing the Chi-square test of independence (χ2 = 0.462; p = 0.794), and Fisher’s exact test (p = 0.850), which suggest there is no statistically significant difference between the pure Alchemists and the rest of documents within the archetype. Despite the technical sophistication in many of the Alchemists’ studies, the analysis of the literature reveals a recurring challenge: the additional effort required to adapt existing characterization and extraction techniques often limits the scope and depth of the resulting safety analyses.

3.2.3. About Shapers

Within the eligible documents, 78 (45%) were categorized as Shapers. Among them, regarding OC, 63 documents (81%) scored a value of 4, six (7%) scored 3, and nine (12%) scored 2, suggesting a trend where authors prefer to develop a customized ontology, rather than use or adapt a ready-to-use one. As for Alchemists, among the documents exhibiting OC = 2 or OC = 3, most were guided by formalized building frameworks [18,19] (11 out of 15, 73%). Regarding OE, a broader distribution was observed in Shapers: 35 documents (45%) scored 4, 12 (15%) scored 3, 10 (13%) scored 2, while the remaining 21 (27%) scored 1, relying on readily available or manually extracted information. This distribution suggest that the characterization activity is a bit more critical when it comes to the KG construction. Overall, the analysis of documents showed that Shapers typically innovate in one dimension. Indeed, only 20 out of 78 documents (26%) were classified as pure Shapers, with both the OC and OE dimensions were equal to 4. Shapers follow two distinct patterns: some prioritize the adaptation or customization of extraction techniques, while others place emphasis on ontology development, mostly relying on readily available extraction algorithms or manually curated data. When looking at MSA, 37 documents (47%) scored 4, 21 (27%) scored 3, and 20 (26%) scored 2. When focusing on the pure Shapers subgroup, eight documents (40%) scored 4, six (30%) scored 3, and six (30%) scored 2. Once again, from the analysis emerged no significant association between being a pure Shaper and achieving a higher MSA level with the Chi-square test yielding χ2 = 1.143 and a p-value of 0.565, and Fisher’s exact test resulting in a p-value of 0.610. However, the MSA distributions qualitatively suggest that, compared to other archetypes, Shapers are more likely to reach higher levels of MSA, highlighting a perceived need to rely on ad hoc solutions in order to pursue more ambitious goals.

4. Discussion

When examining the review results from a broader perspective including all three archetypes together, some additional insights emerge. A cross-cutting look at the MSA levels reveals a general trend toward more advanced analytical approaches: 66 documents (38%) are categorized as MSA = 4, 49 (28%) as MSA = 3, and 58 (34%) as MSA = 2. In other words, documents at MSA = 3 and MSA = 4 together are nearly double in number to those at MSA = 2. However, this push toward higher levels of maturity seems to come at a cost. The literature shows a widespread need to develop tailored, often ad hoc, solutions to enable inference over safety-related knowledge. Indeed, the comparison between the three archetypes’ MSA distributions through the Chi-square test returns a χ² = 10.227 and a p-value of 0.037 confirming a statistically significant dependence between the archetypes and the MSA value. To further validate this result, we also performed Fisher’s exact test (Freeman–Halton extension), which yielded a p-value of 0.036. This confirms the statistical significance of the association and reinforces the robustness of the observed dependency, even in the presence of relatively small counts in some categories. Indeed, as previously highlighted, most of the reviewed documents fall under the Shapers’ archetype, i.e., 78 documents (45%), against 67 for Alchemists (39%), and only 28 for Assemblers (16%). This result is illustrated in Figure 7, which shows the distribution of documents by archetype and MSA level, in a sort of projection of the space depicted in Figure 5. Only a small number of MSA = 4 documents manage to keep the Assembler or Alchemist roles (11 and 18 out of 66, respectively 17% and 27%). In contrast, the majority (37 out of 66, 56%) are associated with the Shapers archetype. The pattern persists even when documents with MSA = 3 are included in the count: among the 115 documents at MSA levels 3 and 4 combined, only 16 (14%) are associated with Assemblers, 41 (36%) with Alchemists, and 58 (50%) with Shapers.

Thus, performing advanced safety analyses using KGs is believed to require a deep level of customization, including building domain-specific ontologies or algorithms from scratch. This conclusion is supported by the computation of standardized residuals, which allow the identification of local deviations between observed and expected frequencies across archetypes and MSA levels. Table 1 reports such results. Accordingly, Shapers show the largest positive deviation at MSA = 4 (i.e., z = 2.20), while Assemblers are at MSA = 2 (i.e., z = 1.05), confirming a higher-than-expected concentration of Shaper-associated studies at higher maturity levels. In addition, Shapers are under-represented at MSA = 2 (z = −1.85), while Alchemists are under-represented at MSA = 4 (z = −2.28), but, given the modest magnitude of most residuals often yielding |z| < 2, and the multiple comparisons involved, these effects should be interpreted as indicative rather than strongly inferential.

Nevertheless, this situation highlights a potential distance between the field of safety science and the available KG technologies: a gap caused by siloed approaches over collaboration and reuse. Ideally, this gap will close when safety scientists can act as Assemblers, relying on mature, reusable KG solutions that allow them to perform advanced analyses without needing to develop overly customized tools. Indeed, from the perspective of safety science, KG technology will be fully democratized only when it becomes accessible to all researchers and practitioners, without requiring substantial additional effort to implement their ideas. While this scenario remains aspirational, we deeply believe it offers a clear direction for future collaboration between the safety and KG research communities.

But are we alone in this quest for democratization? In an attempt to provide such an answer, we conducted an additional analysis of each document, re-applying our archetype and MSA tagging. To make the comparison feasible, we assumed that possible future research directions (if any) would have been proposed in the conclusion section.

This analysis is a simplified perspective, and it is meant to be explorative rather than evidentiary. In fact, this “prospective” tagging was not possible for all documents. Specifically: (i) 48 out of 173 (28%, mostly short conference papers) did not suggest any future research directions; and (ii) 63 out of 173 (36%) proposed future developments that would not change the original archetype and MSA classification. As a result, this extended analysis was performed on 63 documents only. The findings are presented in Figure 8, which displays only these 63 documents using the same visualization logic adopted in Figure 7, showing both their current classification (left) and the projected one based on future research directions (right). While the sample is limited, under the assumptions and limitations already discussed, the results align with our perception: there is a clear collective orientation towards more advanced MSA levels, accompanied by a shift away from the highly customized Shaper archetype. All re-tagged documents move beyond MSA = 2, indicating a growing interest in more mature safety analyses. At the same time, the number of Shapers decreases, while the number of Alchemists increases, suggesting a desire to balance the analytical improvement with more reusable and adaptable KG solutions. However, the number of Assemblers at MSA levels 3 or 4 remains largely unchanged, with only one additional document classified as an Assembler at MSA = 3, reinforcing the idea that, although progress is foreseen, the path is still widely open.

While this paper presents a classification of research into methodological archetypes, results suggest that structural and functional characteristics of a KG are often shaped by the domain. A cross-cutting observation of the corpus, Figure 4, reveals that safety research leveraging KGs is heavily concentrated in specific industrial sectors, notably Construction and Engineering (33 papers, 19%), Electric Utilities (18, ~10%), and Aerospace and Defense (16, ~9%). For instance, construction-safety KGs frequently emerge from established ontology traditions, whereas Aerospace and Defense applications are often driven by stringent regulatory codification and the availability of standardized incident reports. Currently, this study treats industrial sectors and user archetypes as distinct dimensions. However, we believe that a future direction should integrate these two aspects, better exploring how specific sectoral requirements could push to a Shaper or Assembler perspective. To support future investigations, the dataset made publicly available in [12] also includes the publication year, document type, and industrial sector associated with each framework classification. Interested readers may therefore derive additional insights by directly exploring how these contextual dimensions relate to the proposed archetypes.

Another avenue for future research concerns the refinement of the proposed classification framework itself. The impact of introducing more fine-grained scales for OC, OE, and MSA, taking, e.g., additional inspiration from the already existing classifications such as the ones cited in Section 3.2, should be further investigated to assess whether or not it would enhance the discriminative power and overall classification performance.

5. Conclusions

This paper presented a literature review on the use of knowledge graphs in the safety domain, analyzing 173 documents retrieved through a systematic PRISMA process. The review was guided by two research questions. In response to RQ1, the results show that KG-based safety research saw an exponential growth in recent years, and it is mainly concentrated in industrial and technical domains, with the largest share of studies belonging to the Industrials sector, followed by Utilities. The most represented application areas include Construction and Engineering, Electric Utilities, and Aerospace and Defense, indicating that KGs are primarily being adopted in safety-critical domains characterized by complex operational knowledge and structured technical information. In response to RQ2, the review identified common ways in which KGs are constructed and used for safety analysis, formalizing them into a framework to classify contributions based on the originality of knowledge characterization and extraction, and the maturity of the safety analysis proposed. By applying this framework, we defined three archetypes of KG users, Assemblers, Alchemists, and Shapers, each reflecting distinct levels of customization and analytical maturity. Most studies are classified as Shapers, suggesting that high maturity levels of safety analysis require custom ontologies or dedicated extraction methods.

Nonetheless, the analysis presented in this review has some limitations. First, a subset of documents could not be included due to a lack of full-text access. Since many of these works are technical conference papers, this may have led to a partial under-representation of specific implementation details. Additionally, although the inter-rater reliability scores, represented by Fleiss’ Kappa, support the consistency of the classification process, they do not constitute an external validation of the proposed framework. Therefore, the validation of the framework remains an open task for future research, requiring application to independent datasets and/or comparison with external expert judgments.

Looking ahead, the proposed framework aims to represent a solid foundation to link the two disciplines. Further refinement may include the introduction of additional subcategories within each dimension to enable a more granular and objective assessment. Moreover, the proposed framework does not account for technical dimensions such as KG scalability, reasoning capabilities, or detailed extraction performance metrics, which are expected to be critical in practical applications. Future research should extend the framework, eventually accounting for such technical aspects, too. In the long run, it is recommended to revisit the scores over time to explore the field’s evolution, particularly as the KG technology and its widespread use mature. The preliminary evidence discussed in this study may suggest a gradual movement of Shapers toward more Alchemist-like positions, as methods, tools, and workflows become increasingly established and reusable over time. While this trend should be interpreted cautiously and requires further empirical validation, it may point toward a gradual increase in the accessibility and democratization of KG-based safety analysis, potentially leading to a scenario in which today’s Shapers will become tomorrow’s Assemblers.

Author Contributions

Conceptualization, F.S. and R.P.; methodology, F.S.; validation, F.S. and A.M.; formal analysis, F.S., A.M. and K.H.F.; investigation, F.S., A.M. and K.H.F.; data curation, F.S., A.M. and K.H.F.; writing—original draft preparation, F.S., A.M. and K.H.F.; writing—review and editing, F.S., A.M. and R.P.; visualization, F.S.; supervision, F.S. and R.P.; project administration, A.M. and R.P.; funding acquisition, R.P. All authors have read and agreed to the published version of the manuscript.

Funding

The contribution of Andrea Montaruli was carried out within the scope of the PhD project CRESCO (Cyber Resilience for Complex AAM Operations), co-funded by Engage 2-SESAR 3 Knowledge Transfer Network (GRANT ID 101114648; CUP B53C24010700006).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The full dataset related to the literature review conducted in this paper is available at https://zenodo.org/records/20645998 (accessed on 7 June 2026).

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:

AI	Artificial intelligence
BERT	Bidirectional encoder representations from transformers
BiLSTM	Bidirectional long short-term memory
KG	Knowledge graph
LLM	Large language model
MSA	Maturity of safety analysis
NLP	Natural language processing
OC	Originality of knowledge characterization
OE	Originality of knowledge extraction procedure
PRISMA	Preferred Reporting Items for Systematic Reviews and Meta-Analyses
RDF	Resource description framework
TRL	Technology readiness level

Appendix A

Table A1. List of studies included in the review (n = 173). Each study is linked to its corresponding reference number in the bibliography.

Year	Title	Source	Reference
2024	PageRank Algorithm-Based Recommendation System for Construction Safety Guidelines	Buildings	[20]
2024	A new paradigm for construction safety management in China: Introducing knowledge graph and accident database into the early-stage of BIM	Journal of Cleaner Production	[21]
2024	Revealing the coupled evolution process of construction risks in mega hydropower engineering through textual semantics	Advanced Engineering Informatics	[22]
2024	Knowledge Reasoning- and Progressive Distillation-Integrated Detection of Electrical Construction Violations	Sensors	[23]
2024	Research on knowledge graph construction method for mine hoist fault field	2024 7th International Conference on Computer Information Science and Application Technology, CISAT 2024	[24]
2024	A data-driven and knowledge graph-based analysis of the risk hazard coupling mechanism in subway construction accidents	Reliability Engineering and System Safety	[25]
2024	A knowledge graph-based inspection items recommendation method for port state control inspection of LNG carriers	Ocean Engineering	[26]
2024	An Information Integration Technology for Safety Assessment on Civil Airborne System	Aerospace	[27]
2024	Multimodal knowledge graph construction for risk identification in water diversion projects	Journal of Hydrology	[28]
2024	Knowledge Management Model for Urban Flood Emergency Response Based on Multimodal Knowledge Graphs	Water (Switzerland)	[29]
2024	Road Traffic Accident Data Management and Application Analysis Based on Knowledge Graph Technology	ACM International Conference Proceeding Series	[30]
2024	Fault Diagnosis Method for On-Board Interface Equipment of CTCS-3 Based on Temporal Knowledge Graph Completion	ICNSC 2024-21st International Conference on Networking, Sensing and Control: Artificial Intelligence for the Next Industrial Revolution	[31]
2024	Information Integration of Regulation Texts and Tables for Automated Construction Safety Knowledge Mapping	Journal of Construction Engineering and Management	[32]
2024	Sequencial Event Graph Mining in Power Grid Accident Tracing Based on RED-GNN Algorithm	ACM International Conference Proceeding Series	[33]
2024	MAKG: A maritime accident knowledge graph for intelligent accident analysis and management	Ocean Engineering	[34]
2024	Evolutionary Game Strategy Research on PSC Inspection Based on Knowledge Graphs	Journal of Marine Science and Engineering	[35]
2024	Research on Large Model Text-to-SQL Optimization Method for Intelligent Interaction in the Field of Construction Safety	2024 5th International Symposium on Computer Engineering and Intelligent Communications, ISCEIC 2024	[36]
2024	Operational Fault Diagnosis and Mixed Reality Inspection for Building Fire Protection Facilities	Proceedings of 2024 IEEE 25th China Conference on System Simulation Technology and its Application, CCSSTA 2024	[37]
2024	Knowledge graph for safety management standards of water conservancy construction engineering	Automation in Construction	[38]
2024	An ADAS with better driver satisfaction under rear-end near-crash scenarios: A spatio-temporal graph transformer-based prediction framework of evasive behavior and collision risk	Transportation Research Part C: Emerging Technologies	[39]
2024	Graph-based intelligent accident hazard ontology using natural language processing for tracking, prediction, and learning	Automation in Construction	[40]
2024	Knowledge Graph Construction Method for Commercial Aircraft Fault Diagnosis Based on Logic Diagram Model	Aerospace	[41]
2024	A knowledge graph-based method for intelligent risk assessment of power grid	Journal of Physics: Conference Series	[42]
2024	Research on Joint Extraction Method of Elevator Safety Risk Control Knowledge Based on Multi-Perspective Learning	IEEE Access	[43]
2024	Thermal Fault Detection of High-Voltage Isolating Switches based on Hybrid Data and BERT	Arabian Journal for Science and Engineering	[44]
2024	Information Extraction of Aviation Accident Causation Knowledge Graph: An LLM-Based Approach	Electronics (Switzerland)	[45]
2024	Performance comparison of retrieval-augmented generation and fine-tuned large language models for construction safety management knowledge retrieval	Automation in Construction	[46]
2024	Constructing a Coal Mine Safety Knowledge Graph to Promote the Association and Reuse of Risk Management Empirical Knowledge	Sustainability (Switzerland)	[47]
2024	Development of a Knowledge Base for Construction Risk Assessments Using BERT and Graph Models	Buildings	[48]
2024	Causation Correlation Analysis of Aviation Accidents: A Knowledge Graph-Based Approach	Applied Sciences (Switzerland)	[49]
2024	Multi-domain fusion for cargo UAV fault diagnosis knowledge graph construction	Autonomous Intelligent Systems	[50]
2024	Enhancing aviation safety and mitigating accidents: A study on aviation safety hazard identification	Advanced Engineering Informatics	[51]
2024	A Knowledge-Driven Approach to Automate Job Hazard Analysis Process	Journal of Engineering, Project, and Production Management	[52]
2024	Named Entity Recognition Study for Distribution Network Operation	Advances in Transdisciplinary Engineering	[53]
2024	Urban flood vulnerability Knowledge-Graph based on remote sensing and textual bimodal data fusion	Journal of Hydrology	[54]
2024	Early-warning of unsafe hoisting operations: An integration of digital twin and knowledge graph	Developments in the Built Environment	[55]
2023	Dam Safety Monitoring and Early Warning Method Based on Knowledge Graph	Advances in Transdisciplinary Engineering	[56]
2024	Construction of Event Graph for Ship Collision Accident Analysis to Improve Maritime Traffic Safety	Complexity	[57]
2024	Knowledge Graph Generation and Application for Unstructured Data Using Data Processing Pipeline	IEEE Access	[58]
2024	Construction of Knowledge Graph of the Elevator Safety Accidents and Analysis of Key Risk Factors Based on KG-DEMATEL-ISM-MICMAC Method	IEEE Access	[59]
2024	Expanding Aviation Knowledge Graph using Deep Learning for Safety Analysis	AIAA Aviation Forum and ASCEND, 2024	[60]
2024	Research on the Construction and Application of Knowledge Graph in the Field of Coal Mine Safety Monitoring System	IMCEC 2024-IEEE 6th Advanced Information Management, Communicates, Electronic and Automation Control Conference	[61]
2024	Construction and Application of Knowledge Graph for Oil and Gas Pipeline Accidents Based on Graph Database	2024 5th International Conference on Information Science, Parallel and Distributed Systems, ISPDS 2024	[62]
2023	Emergency entity relationship extraction for water diversion project based on pre-trained model and multi-featured graph convolutional network	PLoS ONE	[63]
2023	Knowledge Graph for Identifying Geological Disasters by Integrating Computer Vision with Ontology	Journal of Earth Science	[64]
2023	Chinese Few-Shot Named Entity Recognition and Knowledge Graph Construction in Managed Pressure Drilling Domain	Entropy	[65]
2023	Knowledge-driven intelligent recommendation method for emergency plans in water diversion projects	Journal of Hydroinformatics	[66]
2023	Analyzing Long-Term and High Instantaneous Power Consumption of Buildings from Smart Meter Big Data with Deep Learning and Knowledge Graph Techniques	Energies	[67]
2024	Construction of an Event Knowledge Graph Based on a Dynamic Resource Scheduling Optimization Algorithm and Semantic Graph Convolutional Neural Networks	Electronics (Switzerland)	[68]
2024	Deepening Application Research on Substation Auxiliary Equipment Monitoring System Enabled by Advanced Digital Technology	2024 3rd International Conference on Energy, Power and Electrical Technology, ICEPET 2024	[69]
2024	Resilience Assessment of Multi-Layered Cyber-Physical Systems	2024 IFIP Networking Conference, IFIP Networking 2024	[70]
2023	A dynamic community gas risk-prediction method based on temporal knowledge graphs	Process Safety and Environmental Protection	[71]
2023	Extraction and analysis of risk factors from Chinese chemical accident reports	Chinese Journal of Chemical Engineering	[72]
2024	CPBA-CLIM: An entity-relation extraction model for ontology-based knowledge graph construction in hazardous chemical incident management	Science Progress	[73]
2023	Knowledge Graph Construction to Facilitate Indoor Fire Emergency Evacuation	ISPRS International Journal of Geo-Information	[74]
2023	Multi-Modal Spatio-Temporal Knowledge Graph of Ship Management	Applied Sciences (Switzerland)	[75]
2023	Knowledge Graph Engineering Based on Semantic Annotation of Tables	Computation	[76]
2024	A Semantic Approach to Dynamic Path Planning for Fire Evacuation through BIM and IoT Data Integration	Advances in Civil Engineering	[77]
2024	Enhancing Named Entity Recognition in Safety Hazard Analysis through GBD and LLMs	Proceedings-2024 7th International Conference on Information and Computer Technologies, ICICT 2024	[78]
2024	Named entity recognition technology improvements for Hazard and Operability analysis report	Chinese Control Conference, CCC	[79]
2024	Root Cause Analysis for Industrial Process Anomalies through the Integration of Knowledge Graph and Large Language Model	Chinese Control Conference, CCC	[80]
2023	Hazards correlation analysis of railway accidents: A real-world case study based on the decade-long UK railway accident data	Safety Science	[81]
2024	Earthquake event knowledge graph construction and reasoning	Geomatics, Natural Hazards and Risk	[82]
2023	Knowledge in graphs: investigating the completeness of industrial near miss reports	Safety Science	[83]
2023	Intelligent Exploration of Construction Accidents Based on Knowledge Graph	E3S Web of Conferences	[84]
2023	Building a knowledge graph for operational hazard management of utility tunnels	Expert Systems with Applications	[85]
2023	Architecture and Application of Traffic Safety Management Knowledge Graph Based on Neo4j	Sustainability (Switzerland)	[86]
2024	Research on the Construction Method and Application of Knowledge Graph of Power Operation Risk Pre-control	2024 3rd International Conference on Energy, Power and Electrical Technology, ICEPET 2024	[87]
2023	Dynamic data-driven railway bridge construction knowledge graph update method	Transactions in GIS	[88]
2023	Knowledge graph construction based on ship collision accident reports to improve maritime traffic safety	Ocean and Coastal Management	[89]
2023	Power Grid Fault Diagnosis Based on Knowledge Graph and Bayesian Inference	ACM International Conference Proceeding Series	[90]
2023	Application of Knowledge Graph Technology with Integrated Feature Data in Spacecraft Anomaly Detection	Applied Sciences (Switzerland)	[91]
2023	Construction Safety Knowledge Graph Integrating Text and Image Information	ACM International Conference Proceeding Series	[92]
2023	Knowledge Graph Improved Dynamic Risk Analysis Method for Behavior-Based Safety Management on a Construction Site	Journal of Management in Engineering	[93]
2023	Graph Structure-Based Implicit Risk Reasoning for Long-Tail Scenarios of Automated Driving	2023 4th International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering, ICBAIE 2023	[94]
2022	Construction of Knowledge Graph Based on Traffic Violations in Beijing	Proceedings-2022 4th International Conference on Intelligent Information Processing, IIP 2022	[95]
2022	A novel knowledge graph development for industry design: A case study on indirect coal liquefaction process	Computers in Industry	[96]
2023	A text mining-based approach for understanding Chinese railway incidents caused by electromagnetic interference	Engineering Applications of Artificial Intelligence	[97]
2023	Development of a Knowledge Graph for Automatic Job Hazard Analysis: The Schema	Sensors	[98]
2023	Analysing the Safety and Security of a UV-C Disinfection Robot	Proceedings-IEEE International Conference on Robotics and Automation	[99]
2023	Research on Knowledge Graph Construction for Operational Safety of Cryogenic Loading System	Proceedings of 2023 CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes, SAFEPROCESS 2023	[100]
2022	Hazardous Entity Recommendation for Safety Production Inspection Based on Multi-task Learning	2022 IEEE 8th International Conference on Computer and Communications, ICCC 2022	[101]
2023	Research on Personnel Safety Risk Early Warning Technology Based on Power Infrastructure Samples	2023 IEEE International Conference on Electrical, Automation and Computer Engineering, ICEACE 2023	[102]
2022	Using text mining to establish knowledge graph from accident/incident reports in risk assessment	Expert Systems with Applications	[103]
2023	Industrial safety management in the digital era: Constructing a knowledge graph from near misses	Computers in Industry	[104]
2023	Temporal Knowledge Graph Informer Network for Remaining Useful Life Prediction	IEEE Transactions on Instrumentation and Measurement	[105]
2022	Construction of petrochemical knowledge graph based on deep learning	Journal of Loss Prevention in the Process Industries	[106]
2023	Optical Cable Fault Diagnosis and Auxiliary Decision-making Based on Knowledge Graph	Journal of Physics: Conference Series	[107]
2023	Situation-aware system based on knowledge graphs derived from R-Map analysis of accident situational big data	Procedia Computer Science	[108]
2022	ROADSCENE2VEC: A tool for extracting and embedding road scene-graphs	Knowledge-Based Systems	[109]
2022	A Novel Method for Constructing Knowledge Graph of Railway Safety Risk	ACM International Conference Proceeding Series	[110]
2022	Towards Domain-Specific Knowledge Graph Construction for Flight Control Aided Maintenance	Applied Sciences (Switzerland)	[111]
2023	SailGenie: SAiling expertIse to knowLedge Graph through opEN Information Extraction	Procedia Computer Science	[112]
2022	Knowledge-driven recognition methodology for electricity safety hazard scenarios	Energy Reports	[113]
2023	Deep learning-based relation extraction and knowledge graph-based representation of construction safety requirements	Automation in Construction	[114]
2023	Analysis of Electricity Safety in Scientific Research and Production Sites: A Novel HMM-VA-based Knowledge Graph Approach	2023 IEEE International Conference on Sensors, Electronics and Computer Engineering, ICSECE 2023	[115]
2023	A BN-based risk assessment model of natural gas pipelines integrating knowledge graph and DEMATEL	Process Safety and Environmental Protection	[116]
2023	Explainable Recommendation for Hazard Inspection Reasoning Through Knowledge Graph	2023 IEEE 11th International Conference on Computer Science and Network Technology, ICCSNT 2023	[117]
2022	A temporal knowledge graphs prediction method for community gas risk	Proceedings-2022 4th International Conference on Intelligent Information Processing, IIP 2022	[118]
2023	A Study on a Knowledge Graph Construction Method of Safety Reports for Process Industries	Processes	[119]
2022	MLRP-KG: Mine Landslide Risk Prediction Based on Knowledge Graph	IEEE Transactions on Artificial Intelligence	[120]
2022	Construction of Knowledge Graph for Flag State Control (FSC) Inspection for Ships: A Case Study from China	Journal of Marine Science and Engineering	[121]
2023	Unstructured Transportation Safety Board Findings Categorization Using the Knowledge Graph Pipeline	Proceedings-2023 IEEE International Conference on Big Data, BigData 2023	[122]
2023	Construction Method of Equipment Defect Knowledge Graph in IoT	Intelligent Automation and Soft Computing	[123]
2022	Knowledge graph embedding and reasoning for real-time analytics support of chemical diagnosis from exposure symptoms	Process Safety and Environmental Protection	[124]
2021	Research on airspace security risk assessment technology based on knowledge Graph	Proceedings-2021 21st International Conference on Software Quality, Reliability and Security Companion, QRS-C 2021	[125]
2022	Analysis of Traffic Accident Based on Knowledge Graph	Journal of Advanced Transportation	[126]
2021	A Knowledge Management Framework for Vehicle Hazard Analysis	Proceedings-2021 IEEE International Conference on e-Business Engineering, ICEBE 2021	[127]
2022	Data-Driven Construction Safety Information Sharing System Based on Linked Data, Ontologies, and Knowledge Graph Technologies	International Journal of Environmental Research and Public Health	[128]
2021	Research on Ship Relation Graph Analysis Driven by Multi-source Data	6th International Conference on Transportation Information and Safety: New Infrastructure Construction for Better Transportation, ICTIS 2021	[129]
2020	Construction of public safety knowledge graphs	Proceedings of the 2020 International Conference on Computer, Information and Telecommunication Systems, CITS 2020	[130]
2021	Safety Analysis of Cryogenic Loading System Based on Knowledge Graph	Chinese Control Conference, CCC	[131]
2022	A Knowledge Graph for Automated Construction Workers’ Safety Violation Identification	Proceedings of the International Symposium on Automation and Robotics in Construction	[132]
2021	Research on Domain Entity Extraction in Civil Aviation Safety	Proceedings of 2021 IEEE 3rd International Conference on Civil Aviation Safety and Information Technology, ICCASIT 2021	[133]
2021	Remote supervision relation extraction method of power safety regulations knowledge graph based on ResPCNN-ATT	Proceedings of 2021 IEEE International Conference on Power Electronics, Computer Applications, ICPECA 2021	[134]
2020	Chinese Named Entity Recognition for Hazard and Operability Analysis Text	Proceedings of the 32nd Chinese Control and Decision Conference, CCDC 2020	[135]
2020	Knowledge graph for identifying hazards on construction sites: Integrating computer vision with ontology	Automation in Construction	[136]
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	[137]
2021	A fault diagnosis and visualization method for high-speed train based on edge and cloud collaboration	Applied Sciences (Switzerland)	[138]
2021	A knowledge graph-based approach for exploring railway operational accidents	Reliability Engineering and System Safety	[139]
2022	Integrating Knowledge Graph, Complex Network and Bayesian Network for Data-driven Risk Assessment	Chemical Engineering Transactions	[140]
2022	A Knowledge Graph to Digitalise Functional Resonance Analyses in the Safety Area	Contributions to Management Science	[141]
2020	Railway Train Device Fault Causality Model Based on Knowledge Graph	Proceedings of 2020 International Conference on Sensing, Diagnostics, Prognostics, and Control, SDPC 2020	[142]
2022	Research on Intelligent Question Answering System for Chemical Safety Based on Knowledge Graph	2022 International Conference on Artificial Intelligence and Computer Information Technology, AICIT 2022	[143]
2021	A knowledge graph-based method for modeling and analyzing the disaster risks of railway construction in the mountainous area of Southwest China	2021 4th International Symposium on Traffic Transportation and Civil Architecture, ISTTCA 2021	[144]
2020	Research on the classification of aviation safety reports based on text and knowledge graph	Journal of Physics: Conference Series	[145]
2020	Development of process safety knowledge graph: A Case study on delayed coking process	Computers and Chemical Engineering	[146]
2022	Safety-Critical Components Analysis Using Knowledge Graph For CNC Machine	IEEE International Conference on Automation Science and Engineering	[147]
2021	A knowledge graph method for hazardous chemical management: Ontology design and entity identification	Neurocomputing	[148]
2019	Construction of knowledge graphs for maritime dangerous goods	Sustainability (Switzerland)	[149]
2019	Research on construction method of knowledge graph in the civil aviation security field	Proceedings of 2019 IEEE 1st International Conference on Civil Aviation Safety and Information Technology, ICCASIT 2019	[150]
2018	Construction and application research of knowledge graph in aviation risk field	MATEC Web of Conferences	[151]
2019	Deep learning-based reasoning with multi-ontology for IoT applications	IEEE Access	[152]
2025	Risk factors extraction and analysis of Chinese ship collision accidents based on knowledge graph	Ocean Engineering	[153]
2025	Design and realization of compressor data abnormality safety monitoring and inducement traceability expert system	PLoS ONE	[154]
2025	Integrating machine learning and a large language model to construct a domain knowledge graph for reducing the risk of fall-from-height accidents	Accident Analysis and Prevention	[155]
2025	Construction of Knowledge Graph for Marine Diesel Engine Faults Based on Deep Learning Methods	Journal of Marine Science and Engineering	[156]
2025	A data-driven and knowledge graph-based research on safety risk-coupled evolution analysis and assessment in shield tunneling	Tunnelling and Underground Space Technology	[157]
2025	A knowledge graph for the vulnerability of construction safety system in megaprojects based on accident inversion	Engineering Applications of Artificial Intelligence	[158]
2025	Architecture and Application of Mine Ventilation System Safety Knowledge Graph Based on Neo4j	Sustainability (Switzerland)	[159]
2025	Ontology-Based Customisation Management System for Driver-Vehicle Interfaces: A Preventive Approach to Incident Reduction and Legal Accountability in Highly Automated Vehicles	Applied Sciences (Switzerland)	[160]
2025	Prediction of Equipment Remaining Useful Life Based on Graph Learning and Spatiotemporal Knowledge Graph	Quality and Reliability Engineering International	[161]
2025	How to realize the knowledge reuse and sharing from accident reports? A knowledge-driven modeling method combining ontology and deep learning	Journal of Loss Prevention in the Process Industries	[162]
2025	Identification and precise control of disaster-causing hazards in metro operation and maintenance: A new method for improving metro operation safety based on data mining	Computers and Industrial Engineering	[163]
2025	An automatic machine fault identification method using the knowledge graph–embedded large language model	International Journal of Advanced Manufacturing Technology	[164]
2025	Knowledge Graph-Augmented ERNIE-CNN Method for Risk Assessment in Secondary Power System Operations	Energies	[165]
2025	From surveys to simulations: Integrating Notre-Dame de Paris’ buttressing system diagnosis with knowledge graphs	Automation in Construction	[166]
2025	A Construction and Representation Learning Method for a Traffic Accident Knowledge Graph Based on the Enhanced TransD Model	Applied Sciences (Switzerland)	[167]
2025	HAZOPCT: A HAZOP analysis completeness tool based on knowledge graph reasoning	Process Safety and Environmental Protection	[168]
2025	A Domain Ontology For Safety of Road Users-SafeOn: Overview & Design	Transportation Research Procedia	[169]
2025	Beyond the images: Comprehensible unsafe behaviour recognition boosted by joint inference graph with multi-hop reasoning	Advanced Engineering Informatics	[170]
2025	A knowledge graph-enhanced large language model for question answering of hydraulic structure safety management	Advanced Engineering Informatics	[171]
2025	TH-RotatE: A Hybrid Knowledge Graph Embedding Framework for Fault Diagnosis in Railway Operational Equipment	Electronics (Switzerland)	[172]
2025	Secondary Operation Risk Assessment Method Integrating Graph Convolutional Networks and Semantic Embeddings	Sensors	[173]
2025	Knowledge graph exploitation to enhance the usability of risk assessment in construction safety planning	Advanced Engineering Informatics	[174]
2025	RCA Analysis of Multi-Source Faults in Autonomous Driving	International Journal of Information System Modeling and Design	[175]
2025	A multi-model approach to construction site safety: Fault trees, Bayesian networks, and ontology reasoning	Expert Systems with Applications	[176]
2025	Risk Evolution Analysis of Cross-Regional Water Diversion Projects Based on Spatio-Temporal Knowledge Graphs	Journal of Hydrology	[177]
2025	Risk propagation mechanisms in railway systems under extreme weather: A knowledge graph-based unsupervised causation chain approach	Reliability Engineering and System Safety	[178]
2025	Leveraging large language models for Human-Machine collaborative troubleshooting of complex industrial equipment faults	Advanced Engineering Informatics	[179]
2025	Question-Answering System Powered by Knowledge Graph and Generative Pretrained Transformer to Support Risk Identification in Tunnel Projects	Journal of Construction Engineering and Management	[180]
2025	Risk Assessment of Typhoon Disaster Chain Based on Knowledge Graph and Bayesian Network	Sustainability (Switzerland)	[181]
2025	Ontology-driven knowledge graph for decision-making in resilience enhancement of underground structures: Framework and application	Tunnelling and Underground Space Technology	[182]
2025	Intelligent emergency assisted decision-making method based on standard digitalization: Hazardous chemical accidents in industrial parks	Journal of Safety Science and Resilience	[183]
2024	Construction of a Multimodal Knowledge Graph for Power Grid Construction Safety Based on Large Language Models	Proceedings-2024 International Conference on New Power System and Power Electronics, NPSPE 2024	[184]
2024	A Research on Gas Safety Knowledge Graph and Retrieval-Augmented Generation Mechanism Based on Large Language Model	Proceedings of the IEEE International Conference on Computer and Communications, ICCC	[185]
2024	Fault Early Warning and Judgment System of Low-Voltage Substation Based on Deep Learning and Knowledge Map	Proceedings-2024 International Conference on Power, Electrical Engineering, Electronics and Control, PEEEC 2024	[186]
2024	Safety Risk Assessment in Fluid Catalytic Cracking Units Based on HAZOP and Knowledge Graph	Proceedings-2024 China Automation Congress, CAC 2024	[187]
2024	Ontology Construction of Fault Diagnosis Knowledge Graph for Civil Aircraft Maintenance	15th Global Reliability and Prognostics and Health Management Conference, PHM-Beijing 2024	[188]
2024	Construction of Safety Knowledge Graph for Near Electric Work Based on Graph Visualization	2024 5th International Conference on Clean Energy and Electric Power Engineering, ICCEPE 2024	[189]
2024	Emergency Disposal Decision Generation Method for Flight Test Based on Knowledge Graph	2024 4th International Conference on Communication Technology and Information Technology, ICCTIT 2024	[190]
2024	Research on railway operational accidents analysis method based on knowledge graph	Proceedings-2024 China Automation Congress, CAC 2024	[191]
2024	Distribution Transformer Fault Data Based on One-hot Coded Word Vector Knowledge Graph Construction Study	2024 5th International Conference on Clean Energy and Electric Power Engineering, ICCEPE 2024	[192]

References

Chalmers, A.F. Science and Its Fabrication; Univ of Minnesota Press: Minneapolis, MN, USA, 1990. [Google Scholar]
Gauch, H.G. Scientific Method in Practice; Cambridge University Press: Cambridge, UK, 2015. [Google Scholar]
Kuhn, T.S. The Structure of Scientific Revolutions; University of Chicago Press: Chicago, IL, USA, 1962. [Google Scholar]
Schein, S.L.; West, T.G. Plato’s Apology of Socrates. An Interpretation, with a New Translation. Class. World 1980, 74, 39. [Google Scholar] [CrossRef]
Aven, T. Safety Is the Antonym of Risk for Some Perspectives of Risk. Saf. Sci. 2009, 47, 925–930. [Google Scholar] [CrossRef]
Hollnagel, E. Is Safety a Subject for Science? Saf. Sci. 2014, 67, 21–24. [Google Scholar] [CrossRef]
Hollnagel, E.; Woods, D.D. Cognitive Systems Engineering: New Wine in New Bottles. Int. J. Hum. Comput. Stud. 1999, 51, 339–356. [Google Scholar] [CrossRef] [PubMed]
Patriarca, R.; Bergström, J.; Gravio, G.D.; Costantino, F. Resilience Engineering: Current Status of the Research and Future Challenges. Saf. Sci. 2018, 102, 79–100. [Google Scholar] [CrossRef]
Le Coze, J.C. The ‘New View’ of Human Error. Origins, Ambiguities, Successes and Critiques. Saf. Sci. 2022, 154, 105853. [Google Scholar]
Peng, C.; Xia, F.; Naseriparsa, M.; Osborne, F. Knowledge Graphs: Opportunities and Challenges. Artif. Intell. Rev. 2023, 56, 13071–13102. [Google Scholar] [CrossRef]
Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G.; The PRISMA Group. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. J. Clin. Epidemiol. 2009, 62, 1006–1012. [Google Scholar] [CrossRef]
Montaruli, A.; Simone, F.; Fandino, K.H.; Patriarca, R. A Tale of Three Words: Knowledge, Safety, and Graphs—Dataset 2025. Available online: https://zenodo.org/records/20645998 (accessed on 7 June 2026).
MSCI Global Industry Classification Standard (Gics^®) Methodology. 2020. Available online: https://www.msci.com/gics (accessed on 20 December 2025).
Suárez-Figueroa, M.C.; Gómez-Pérez, A.; Fernández-López, M. The NeOn Methodology Framework: A Scenario-Based Methodology for Ontology Development. Appl. Ontol. 2015, 10, 107–145. [Google Scholar] [CrossRef]
Lavin, A.; Gilligan-Lee, C.M.; Visnjic, A.; Ganju, S.; Newman, D.; Ganguly, S.; Lange, D.; Baydin, A.G.; Sharma, A.; Gibson, A.; et al. Technology Readiness Levels for Machine Learning Systems. Nat. Commun. 2022, 13, 6039. [Google Scholar] [CrossRef]
Li, W.; Qi, G.; Ji, Q. Hybrid Reasoning in Knowledge Graphs: Combing Symbolic Reasoning and Statistical Reasoning. Semant. Web 2020, 11, 53–62. [Google Scholar] [CrossRef]
Zhang, D.; Wang, H.; Ding, Y. A Review of Inference Methods Based on Knowledge Graph. In Fuzzy Systems and Data Mining VI; Tallón-Ballesteros, A.J., Ed.; SAGE Publications: Thousand Oaks, CA, USA, 2020. [Google Scholar]
Noy, N.; Mcguinness, D. Ontology Development 101: A Guide to Creating Your First Ontology. Available online: https://protege.stanford.edu/publications/ontology_development/ontology101.pdf (accessed on 7 June 2026).
Fernández-López, M.; Gomez-Perez, A.; Juristo, N. METHONTOLOGY: From Ontological Art towards Ontological Engineering. In Engineering Workshop on Ontological Engineering (AAAI97); AAAI Press: Washington, DC, USA, 1997. [Google Scholar]
Lee, J.; Ahn, S. PageRank Algorithm-Based Recommendation System for Construction Safety Guidelines. Buildings 2024, 14, 3041. [Google Scholar] [CrossRef]
Li, W.; Wu, P.; Huang, J.; Xu, Y. A New Paradigm for Construction Safety Management in China: Introducing Knowledge Graph and Accident Database into the Early-Stage of BIM. J. Clean. Prod. 2024, 470, 143367. [Google Scholar] [CrossRef]
Cao, K.; Chen, S.; Yang, C.; Li, Z.; Luo, L.; Ren, Z. Revealing the Coupled Evolution Process of Construction Risks in Mega Hydropower Engineering through Textual Semantics. Adv. Eng. Inform. 2024, 62, 102713. [Google Scholar] [CrossRef]
Ma, B.; Liang, G.; Rao, Y.; Guo, W.; Zheng, W.; Wang, Q. Knowledge Reasoning- and Progressive Distillation-Integrated Detection of Electrical Construction Violations. Sensors 2024, 24, 8216. [Google Scholar] [CrossRef] [PubMed]
Luo, J.; Zhu, Z.; Zhu, H.; Dong, X. Research on Knowledge Graph Construction Method for Mine Hoist Fault Field; IEEE: New York, NY, USA, 2024; p. 145. [Google Scholar]
Huo, X.; Yin, Y.; Jiao, L.; Zhang, Y. A Data-Driven and Knowledge Graph-Based Analysis of the Risk Hazard Coupling Mechanism in Subway Construction Accidents. Reliab. Eng. Syst. Saf. 2024, 250, 110254. [Google Scholar] [CrossRef]
Zhang, X.; Liu, C.; Xu, Y.; Ye, B.; Gan, L.; Shu, Y. A Knowledge Graph-Based Inspection Items Recommendation Method for Port State Control Inspection of LNG Carriers. Ocean Eng. 2024, 313, 119434. [Google Scholar] [CrossRef]
Chen, X.; Zou, Q.; Bai, J.; Dong, L. An Information Integration Technology for Safety Assessment on Civil Airborne System. Aerospace 2024, 11, 459. [Google Scholar] [CrossRef]
Wang, L.; Liu, X.; Liu, Y.; Li, H.; Liu, J.; Yang, L. Multimodal Knowledge Graph Construction for Risk Identification in Water Diversion Projects. J. Hydrol. 2024, 635, 131155. [Google Scholar] [CrossRef]
Li, M.; Yuan, C.; Li, K.; Gao, M.; Zhang, Y.; Lv, H. Knowledge Management Model for Urban Flood Emergency Response Based on Multimodal Knowledge Graphs. Water 2024, 16, 1676. [Google Scholar] [CrossRef]
Yu, D.; Peng, W.; Chen, Y.; Yang, Y.; Chen, H. Road Traffic Accident Data Management and Application Analysis Based on Knowledge Graph Technology. In Proceedings of the 2024 International Conference on Digital Society and Artificial Intelligence; Association for Computing Machinery: New York, NY, USA, 2024; pp. 297–302. [Google Scholar]
Zhou, B.; Zhang, D.; Lv, J.; Liu, H.; Zhao, Z.; Luo, Z. Fault Diagnosis Method for On-Board Interface Equipment of CTCS-3 Based on Temporal Knowledge Graph Completion. In Proceedings of the 2024 International Conference on Networking, Sensing and Control (ICNSC); IEEE: New York, NY, USA, 2024; pp. 1–6. [Google Scholar] [CrossRef]
Wang, H.; Xu, S.; Cui, D.; Xu, H.; Luo, H. Information Integration of Regulation Texts and Tables for Automated Construction Safety Knowledge Mapping. J. Constr. Eng. Manag. 2024, 150, 04024034. [Google Scholar] [CrossRef]
Chen, X.; Zhou, X.; Pan, J.; Qiu, X. Sequencial Event Graph Mining in Power Grid Accident Tracing Based on RED-GNN Algorithm. In Proceedings of the International Conference on Computing, Machine Learning and Data Science; Association for Computing Machinery: New York, NY, USA, 2024; pp. 1–7. [Google Scholar]
Liu, D.; Cheng, L. MAKG: A Maritime Accident Knowledge Graph for Intelligent Accident Analysis and Management. Ocean Eng. 2024, 312, 119280. [Google Scholar] [CrossRef]
Liu, C.; Wang, Q.; Xiang, B.; Xu, Y.; Gan, L. Evolutionary Game Strategy Research on PSC Inspection Based on Knowledge Graphs. J. Mar. Sci. Eng. 2024, 12, 1449. [Google Scholar] [CrossRef]
Zhao, L.; Jiang, X.; Tao, F.; Liu, W.; Wang, X.; Wu, Y.H. Research on Large Model Text-to-SQL Optimization Method for Intelligent Interaction in the Field of Construction Safety. In Proceedings of the 2024 5th International Symposium on Computer Engineering and Intelligent Communications (ISCEIC); IEEE: New York, NY, USA, 2024; pp. 201–209. [Google Scholar] [CrossRef]
Feng, X.; Liu, J.; Yi, C. Operational Fault Diagnosis and Mixed Reality Inspection for Building Fire Protection Facilities. In Proceedings of the 2024 IEEE 25th China Conference on System Simulation Technology and its Application (CCSSTA); IEEE: New York, NY, USA, 2024; pp. 122–126. [Google Scholar] [CrossRef]
Chen, Y.; Lu, G.; Wang, K.; Chen, S.; Duan, C. Knowledge Graph for Safety Management Standards of Water Conservancy Construction Engineering. Autom. Constr. 2024, 168, 105873. [Google Scholar] [CrossRef]
Gao, J.; Yu, B.; Chen, Y.; Bao, S.; Gao, K.; Zhang, L. An ADAS with Better Driver Satisfaction under Rear-End near-Crash Scenarios: A Spatio-Temporal Graph Transformer-Based Prediction Framework of Evasive Behavior and Collision Risk. Transp. Res. Part C Emerg. Technol. 2024, 159, 104491. [Google Scholar] [CrossRef]
Hong, E.; Lee, S.; Kim, H.; Park, J.; Seo, M.B.; Yi, J.-S. Graph-Based Intelligent Accident Hazard Ontology Using Natural Language Processing for Tracking, Prediction, and Learning. Autom. Constr. 2024, 168, 105800. [Google Scholar] [CrossRef]
Peng, H.; Yang, W. Knowledge Graph Construction Method for Commercial Aircraft Fault Diagnosis Based on Logic Diagram Model. Aerospace 2024, 11, 773. [Google Scholar] [CrossRef]
Zhang, J.; Zhang, C.; Han, X.; Liu, X.; Zhu, X.; Chen, L.; Jin, X. A Knowledge Graph-Based Method for Intelligent Risk Assessment of Power Grid. J. Phys. Conf. Ser. 2024, 2914, 012016. [Google Scholar] [CrossRef]
Hao, S.; Shi, F. Research on Joint Extraction Method of Elevator Safety Risk Control Knowledge Based on Multi-Perspective Learning. IEEE Access 2024, 12, 159488–159502. [Google Scholar] [CrossRef]
Gong, Z.; Cao, Z.; Zhou, S.; Yang, F.; Shuai, C.; Ouyang, X.; Luo, Z. Thermal Fault Detection of High-Voltage Isolating Switches Based on Hybrid Data and BERT. Arab. J. Sci. Eng. 2024, 49, 6429–6443. [Google Scholar] [CrossRef]
Chen, L.; Xu, J.; Wu, T.; Liu, J. Information Extraction of Aviation Accident Causation Knowledge Graph: An LLM-Based Approach. Electronics 2024, 13, 3936. [Google Scholar] [CrossRef]
Lee, J.; Ahn, S.; Kim, D.; Kim, D. Performance Comparison of Retrieval-Augmented Generation and Fine-Tuned Large Language Models for Construction Safety Management Knowledge Retrieval. Autom. Constr. 2024, 168, 105846. [Google Scholar] [CrossRef]
Zhang, J.; Li, Y.; Wu, J.; Ren, X.; Wang, Y.; Jia, H.; Xie, M. Constructing a Coal Mine Safety Knowledge Graph to Promote the Association and Reuse of Risk Management Empirical Knowledge. Sustainability 2024, 16, 8848. [Google Scholar] [CrossRef]
Lee, W.; Lee, S. Development of a Knowledge Base for Construction Risk Assessments Using BERT and Graph Models. Buildings 2024, 14, 3359. [Google Scholar] [CrossRef]
Xu, J.; Chen, L.; Xing, H.; Tian, W. Causation Correlation Analysis of Aviation Accidents: A Knowledge Graph-Based Approach. Appl. Sci. 2024, 14, 6887. [Google Scholar] [CrossRef]
Xiao, A.; Yan, W.; Zhang, X.; Liu, Y.; Zhang, H.; Liu, Q. Multi-Domain Fusion for Cargo UAV Fault Diagnosis Knowledge Graph Construction. Auton. Intell. Syst. 2024, 4, 10. [Google Scholar] [CrossRef]
Xiong, M.; Wang, H.; Wong, Y.D.; Hou, Z. Enhancing Aviation Safety and Mitigating Accidents: A Study on Aviation Safety Hazard Identification. Adv. Eng. Inform. 2024, 62, 102732. [Google Scholar] [CrossRef]
Pandithawatta, S.; Rameezdeen, R.; Ahn, S.; Chow, C.W.K.; Gorjian, N. A Knowledge-Driven Approach to Automate Job Hazard Analysis Process. J. Eng. Proj. Prod. Manag. 2024, 14, 1. [Google Scholar] [CrossRef]
Zhang, P.; Yang, Z.; Dong, X.; Li, J.; Chen, S. Named Entity Recognition Study for Distribution Network Operation. In Mechatronics and Automation Technology; Xu, J., Ed.; SAGE Publications: Thousand Oaks, CA, USA, 2024. [Google Scholar]
Duan, C.; Zheng, X.; Li, R.; Wu, Z. Urban Flood Vulnerability Knowledge-Graph Based on Remote Sensing and Textual Bimodal Data Fusion. J. Hydrol. 2024, 633, 131010. [Google Scholar] [CrossRef]
Jiang, W.; Liu, Y.; Chen, K.; Liu, Y.; Ding, L. Early-Warning of Unsafe Hoisting Operations: An Integration of Digital Twin and Knowledge Graph. Dev. Built Environ. 2024, 19, 100490. [Google Scholar] [CrossRef]
Gong, S.; Sun, F.; Chen, K. Dam Safety Monitoring and Early Warning Method Based on Knowledge Graph. In Hydraulic and Civil Engineering Technology VIII; IOS Press: Amsterdam, The Netherlands, 2023; pp. 842–850. [Google Scholar]
Ma, J.; Wang, Y.; Wang, L.; Xu, L.; Zhao, J. Construction of Event Graph for Ship Collision Accident Analysis to Improve Maritime Traffic Safety. Complexity 2024, 2024, 4998195. [Google Scholar] [CrossRef]
Thushara Sukumar, S.; Lung, C.-H.; Zaman, M.; Panday, R. Knowledge Graph Generation and Application for Unstructured Data Using Data Processing Pipeline. IEEE Access 2024, 12, 136759–136770. [Google Scholar] [CrossRef]
Shi, F.; Wu, Z. Construction of Knowledge Graph of the Elevator Safety Accidents and Analysis of Key Risk Factors Based on KG-DEMATEL-ISM-MICMAC Method. IEEE Access 2024, 12, 43615–43631. [Google Scholar] [CrossRef]
Jing, X.; Sawant, K.; Bendarkar, M.V.; Elias, L.R.; Mavris, D. Expanding Aviation Knowledge Graph Using Deep Learning for Safety Analysis. In AIAA Aviation Forum and Ascend 2024; AIAA Aviation Forum and ASCEND Co-Located Conference Proceedings; American Institute of Aeronautics and Astronautics: Reston, VA, USA, 2024. [Google Scholar]
Wu, K. Research on the Construction and Application of Knowledge Graph in the Field of Coal Mine Safety Monitoring System. In Proceedings of the 2024 IEEE 6th Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC); IEEE: New York, NY, USA, 2024; pp. 612–617. [Google Scholar] [CrossRef]
Chen, J.; Ye, Y.; Jiang, L.; Jiang, Y. Construction and Application of Knowledge Graph for Oil and Gas Pipeline Accidents Based on Graph Database. In Proceedings of the 2024 5th International Conference on Information Science, Parallel and Distributed Systems (ISPDS), Guangzhou, China, 31 May–2 June 2024; pp. 295–299. [Google Scholar]
Wang, L.H.; Liu, X.M.; Liu, Y.; Li, H.R.; Liu, J.Q.; Yang, L.B. Emergency Entity Relationship Extraction for Water Diversion Project Based on Pre-Trained Model and Multi-Featured Graph Convolutional Network. PLoS ONE 2023, 18, e0292004. [Google Scholar] [CrossRef] [PubMed]
Qiu, Q.; Xie, Z.; Zhang, D.; Ma, K.; Tao, L.; Tan, Y.; Zhang, Z.; Jiang, B. Knowledge Graph for Identifying Geological Disasters by Integrating Computer Vision with Ontology. J. Earth Sci. 2023, 34, 1418–1432. [Google Scholar] [CrossRef]
Wei, S.; Liang, Y.; Li, X.; Weng, X.; Fu, J.; Han, X. Chinese Few-Shot Named Entity Recognition and Knowledge Graph Construction in Managed Pressure Drilling Domain. Entropy 2023, 25, 1097. [Google Scholar] [CrossRef]
Wang, L.; Liu, X.; Liu, Y.; Li, H.; Liu, J. Knowledge-Driven Intelligent Recommendation Method for Emergency Plans in Water Diversion Projects. J. Hydroinformatics 2023, 25, 2522–2540. [Google Scholar] [CrossRef]
Wang, R.-G.; Ho, W.-J.; Chiang, K.-C.; Hung, Y.-C.; Tai, J.-K.; Tan, J.-C.; Chuang, M.-L.; Ke, C.-Y.; Chien, Y.-F.; Jeng, A.-P.; et al. Analyzing Long-Term and High Instantaneous Power Consumption of Buildings from Smart Meter Big Data with Deep Learning and Knowledge Graph Techniques. Energies 2023, 16, 6893. [Google Scholar] [CrossRef]
Liu, X.; Zhang, L.; Zheng, Q.; Wei, F.; Wang, K.; Zhang, Z.; Chen, Z.; Niu, L.; Liu, J. Construction of an Event Knowledge Graph Based on a Dynamic Resource Scheduling Optimization Algorithm and Semantic Graph Convolutional Neural Networks. Electronics 2024, 13, 11. [Google Scholar] [CrossRef]
Li, M.; Lang, Q.; Wu, Q. Deepening Application Research on Substation Auxiliary Equipment Monitoring System Enabled by Advanced Digital Technology. In Proceedings of the 2024 3rd International Conference on Energy, Power and Electrical Technology (ICEPET); IEEE: New York, NY, USA, 2024; pp. 1835–1840. [Google Scholar] [CrossRef]
Dagnas, R.; Barbeau, M.; Garcia-Alfaro, J.; Yaich, R. Resilience Assessment of Multi-Layered Cyber-Physical Systems: 23rd International Federation for Information Processing on Networking Conference, IFIP Networking 2024. In Proceedings of the 2024 IFIP Networking Conference, IFIP Networking; IEEE: New York, NY, USA, 2024; Volume 2024, pp. 634–639. [Google Scholar] [CrossRef]
Shi, Y.; Zhao, L.; Zhou, M.; Yin, X.; Guo, W.; Li, C. A Dynamic Community Gas Risk-Prediction Method Based on Temporal Knowledge Graphs. Process Saf. Environ. Prot. 2023, 177, 436–445. [Google Scholar] [CrossRef]
Luo, X.; Feng, X.; Ji, X.; Dang, Y.; Zhou, L.; Bi, K.; Dai, Y. Extraction and Analysis of Risk Factors from Chinese Chemical Accident Reports. Chin. J. Chem. Eng. 2023, 61, 68–81. [Google Scholar] [CrossRef]
Du, W.; Wang, X.; Zhu, Q.; Jing, X.; Liu, X. CPBA-CLIM: An Entity-Relation Extraction Model for Ontology-Based Knowledge Graph Construction in Hazardous Chemical Incident Management. Sci. Prog. 2024, 107, 00368504241235510. [Google Scholar] [CrossRef]
Da, M.; Zhong, T.; Huang, J. Knowledge Graph Construction to Facilitate Indoor Fire Emergency Evacuation. ISPRS Int. J. Geo-Inf. 2023, 12, 403. [Google Scholar] [CrossRef]
Zhang, Y.; Xu, R.; Lu, W.; Mayer, W.; Ning, D.; Duan, Y.; Zeng, X.; Feng, Z. Multi-Modal Spatio-Temporal Knowledge Graph of Ship Management. Appl. Sci. 2023, 13, 9393. [Google Scholar] [CrossRef]
Dorodnykh, N.; Yurin, A. Knowledge Graph Engineering Based on Semantic Annotation of Tables. Computation 2023, 11, 175. [Google Scholar] [CrossRef]
Pang, B.; Shi, J.; Jiang, L.; Pan, Z. A Semantic Approach to Dynamic Path Planning for Fire Evacuation through BIM and IoT Data Integration. Adv. Civ. Eng. 2024, 2024, 8839865. [Google Scholar] [CrossRef]
Dong, H.; Wu, B. Enhancing Named Entity Recognition in Safety Hazard Analysis through GBD and LLMs. In Proceedings of the 2024 7th International Conference on Information and Computer Technologies (ICICT), Honolulu, HI, USA, 15–17 March 2024; pp. 13–19. [Google Scholar]
Yang, H.; Lu, D.; Wang, Z.; Gao, D.; Li, D. Named Entity Recognition Technology Improvements for Hazard and Operability Analysis Report. In Proceedings of the 2024 43rd Chinese Control Conference (CCC); IEEE: New York, NY, USA, 2024; pp. 8200–8205. [Google Scholar] [CrossRef]
Sun, Q.; Li, Y.; Zhou, C.; Tian, Y.-C. Root Cause Analysis for Industrial Process Anomalies through the Integration of Knowledge Graph and Large Language Model. In Proceedings of the 2024 43rd Chinese Control Conference (CCC); IEEE: New York, NY, USA; pp. 6855–6860. [CrossRef]
Wang, N.; Yang, X.; Chen, J.; Wang, H.; Wu, J. Hazards Correlation Analysis of Railway Accidents: A Real-World Case Study Based on the Decade-Long UK Railway Accident Data. Saf. Sci. 2023, 166, 106238. [Google Scholar] [CrossRef]
Qiu, P.; Pang, L.; Luo, Y.; Liu, Y.; Xing, H.; Liu, K.; Zhuang, G. Earthquake Event Knowledge Graph Construction and Reasoning. Geomat. Nat. Hazards Risk 2024, 15, 2383768. [Google Scholar] [CrossRef]
Simone, F.; Ansaldi, S.M.; Agnello, P.; Di Gravio, G.; Patriarca, R. Knowledge in Graphs: Investigating the Completeness of Industrial near Miss Reports. Saf. Sci. 2023, 168, 106305. [Google Scholar] [CrossRef]
Xu, H.; Liao, H.; Tan, Y.; Xing, B.; Hou, B. Intelligent Exploration of Construction Accidents Based on Knowledge Graph. E3S Web Conf. 2023, 409, 04002. [Google Scholar] [CrossRef]
Peng, F.-L.; Qiao, Y.-K.; Yang, C. Building a Knowledge Graph for Operational Hazard Management of Utility Tunnels. Expert Syst. Appl. 2023, 223, 119901. [Google Scholar] [CrossRef]
Yuan, D.; Zhou, K.; Yang, C. Architecture and Application of Traffic Safety Management Knowledge Graph Based on Neo4j. Sustainability 2023, 15, 9786. [Google Scholar] [CrossRef]
Minfu, A.; Guo, K.; Yin, K. Research on the Construction Method and Application of Knowledge Graph of Power Operation Risk Pre-Control. In Proceedings of the 2024 3rd International Conference on Energy, Power and Electrical Technology (ICEPET); IEEE: New York, NY, USA, 2024; pp. 1360–1366. [Google Scholar] [CrossRef]
Lai, J.; Zhu, J.; Guo, Y.; You, J.; Xie, Y.; Wu, J.; Hu, Y. Dynamic Data-Driven Railway Bridge Construction Knowledge Graph Update Method. Trans. GIS 2023, 27, 2099–2117. [Google Scholar] [CrossRef]
Gan, L.; Ye, B.; Huang, Z.; Xu, Y.; Chen, Q.; Shu, Y. Knowledge Graph Construction Based on Ship Collision Accident Reports to Improve Maritime Traffic Safety. Ocean Coast. Manag. 2023, 240, 106660. [Google Scholar] [CrossRef]
Su, F.; Zhang, J.; Zhang, C.; Zhu, X.; Shen, S.; Liu, K.; Liu, H.; Han, Y. Power Grid Fault Diagnosis Based on Knowledge Graph and Bayesian Inference. In Proceedings of the 2023 4th International Conference on Computer Science and Management Technology; Association for Computing Machinery: New York, NY, USA, 2024; pp. 657–662. [Google Scholar]
Yi, X.; Huang, P.; Che, S. Application of Knowledge Graph Technology with Integrated Feature Data in Spacecraft Anomaly Detection. Appl. Sci. 2023, 13, 10905. [Google Scholar] [CrossRef]
Wu, W.; Yuan, Q.; Chen, Q.; Cao, Y. Construction Safety Knowledge Graph Integrating Text and Image Information. In Proceedings of the 2023 6th International Conference on Information Management and Management Science, Chengdu, China, 25–27 August 2023; pp. 26–32. [Google Scholar] [CrossRef]
Chen, Q.; Long, D.; Yang, C.; Xu, H. Knowledge Graph Improved Dynamic Risk Analysis Method for Behavior-Based Safety Management on a Construction Site. J. Manag. Eng. 2023, 39, 04023023. [Google Scholar] [CrossRef]
Li, X.; Liu, J.; Li, J.; Yu, W.; Cao, Z.; Qiu, S.; Hu, J.; Wang, H.; Jiao, X. Graph Structure-Based Implicit Risk Reasoning for Long-Tail Scenarios of Automated Driving; IEEE: New York, NY, USA, 2023; p. 85. [Google Scholar]
Li, J.; Shi, Y.; Li, S.; Wang, Q. Construction of Knowledge Graph Based on Traffic Violations in Beijing; IEEE: New York, NY, USA, 2022; p. 30. [Google Scholar]
Wang, Z.; Zhang, B.; Gao, D. A Novel Knowledge Graph Development for Industry Design: A Case Study on Indirect Coal Liquefaction Process. Comput. Ind. 2022, 139, 103647. [Google Scholar] [CrossRef]
Liu, C.; Yang, S. A Text Mining-Based Approach for Understanding Chinese Railway Incidents Caused by Electromagnetic Interference. Eng. Appl. Artif. Intell. 2023, 117, 105598. [Google Scholar] [CrossRef]
Pandithawatta, S.; Ahn, S.; Rameezdeen, R.; Chow, C.W.K.; Gorjian, N.; Kim, T.W. Development of a Knowledge Graph for Automatic Job Hazard Analysis: The Schema. Sensors 2023, 23, 3893. [Google Scholar] [CrossRef]
Nurchalifah, D.; Blumenthal, S.; Lo Iacono, L.; Hochgeschwender, N. Analysing the Safety and Security of a UV-C Disinfection Robot. In Proceedings of the 2023 IEEE International Conference on Robotics and Automation (ICRA); IEEE: New York, NY, USA, 2023; pp. 12729–12736. [Google Scholar] [CrossRef]
Li, S.; Chai, Y.; Mao, Y.; Liu, Y.; Wang, Y. Research on Knowledge Graph Construction for Operational Safety of Cryogenic Loading System. In Proceedings of the 2023 CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS), Yibin, China, 22–24 September 2023; pp. 1–5. [Google Scholar]
Wang, X.; Ai, X.; Guo, Y.; Chen, Z.; Zhang, Y. Hazardous Entity Recommendation for Safety Production Inspection Based on Multi-Task Learning; IEEE: New York, NY, USA, 2022; p. 398. [Google Scholar]
Zhang, L.; Guo, Q.; Qin, Y.; Wang, X.; Huang, X.; Nie, W. Research on Personnel Safety Risk Early Warning Technology Based on Power Infrastructure Samples. In Proceedings of the 2023 IEEE International Conference on Electrical, Automation and Computer Engineering (ICEACE), Changchun, China, 29–31 December 2023; pp. 1696–1701. [Google Scholar]
Liu, C.; Yang, S. Using Text Mining to Establish Knowledge Graph from Accident/Incident Reports in Risk Assessment. Expert Syst. Appl. 2022, 207, 117991. [Google Scholar] [CrossRef]
Simone, F.; Ansaldi, S.M.; Agnello, P.; Patriarca, R. Industrial Safety Management in the Digital Era: Constructing a Knowledge Graph from near Misses. Comput. Ind. 2023, 146, 103849. [Google Scholar] [CrossRef]
Zhang, Y.; Zhou, W.; Huang, J.; Jin, X.; Xiao, G. Temporal Knowledge Graph Informer Network for Remaining Useful Life Prediction. IEEE Trans. Instrum. Meas. 2023, 72, 3528610. [Google Scholar] [CrossRef]
Zhao, Y.; Zhang, B.; Gao, D. Construction of Petrochemical Knowledge Graph Based on Deep Learning. J. Loss Prev. Process Ind. 2022, 76, 104736. [Google Scholar] [CrossRef]
Li, Y.; Li, C.; Wang, Y.; Guo, X.; Li, L.; Ma, J. Optical Cable Fault Diagnosis and Auxiliary Decision-Making Based on Knowledge Graph. J. Phys. Conf. Ser. 2023, 2661, 012029. [Google Scholar] [CrossRef]
Srinivasan, S.B.T.; Ozaki, M.; Nishida, Y.; Oono, M.; Yamanaka, T. Situation-Aware System Based on Knowledge Graphs Derived from R-Map Analysis of Accident Situational Big Data. Procedia Comput. Sci. 2023, 220, 436–445. [Google Scholar] [CrossRef]
Malawade, A.V.; Yu, S.-Y.; Hsu, B.; Kaeley, H.; Karra, A.; Al Faruque, M.A. Roadscene2vec: A Tool for Extracting and Embedding Road Scene-Graphs. Knowl.-Based Syst. 2022, 242, 108245. [Google Scholar] [CrossRef]
Wang, Y.; Wang, Y.; Ye, Y. A Novel Method for Constructing Knowledge Graph of Railway Safety Risk. In Proceedings of the 6th International Conference on Computer Science and Application Engineering, Online, 21–22 October 2022; pp. 1–6. [Google Scholar] [CrossRef]
Li, C.; Yang, X.; Luo, S.; Song, M.; Li, W. Towards Domain-Specific Knowledge Graph Construction for Flight Control Aided Maintenance. Appl. Sci. 2022, 12, 12736. [Google Scholar] [CrossRef]
Carta, S.; Fariello, P.; Giuliani, A.; Piano, L.; Podda, A.S.; Tiddia, S.G. SailGenie: SAiling expertIse to knowLedge Graph through opEN Information Extraction. Procedia Comput. Sci. 2023, 225, 2224–2233. [Google Scholar] [CrossRef]
Qu, Z.; Zhang, Z.; Liu, S.; Cao, J.; Bo, X. Knowledge-Driven Recognition Methodology for Electricity Safety Hazard Scenarios. Energy Rep. 2022, 8, 10006–10016. [Google Scholar] [CrossRef]
Wang, X.; El-Gohary, N. Deep Learning-Based Relation Extraction and Knowledge Graph-Based Representation of Construction Safety Requirements. Autom. Constr. 2023, 147, 104696. [Google Scholar] [CrossRef]
Xu, Z.; Yuan, J.; Pan, M.; Tian, S.; Song, L. Analysis of Electricity Safety in Scientific Research and Production Sites: A Novel HMM-VA-Based Knowledge Graph Approach. In Proceedings of the 2023 IEEE International Conference on Sensors, Electronics and Computer Engineering (ICSECE); IEEE: New York, NY, USA, 2023; pp. 7–12. [Google Scholar] [CrossRef]
Bai, Y.; Wu, J.; Ren, Q.; Jiang, Y.; Cai, J. A BN-Based Risk Assessment Model of Natural Gas Pipelines Integrating Knowledge Graph and DEMATEL. Process Saf. Environ. Prot. 2023, 171, 640–654. [Google Scholar] [CrossRef]
Chen, Z.; Ai, X.; Guo, Y.; Huang, Y.; Yang, J. Explainable Recommendation for Hazard Inspection Reasoning Through Knowledge Graph. In Proceedings of the 2023 IEEE 11th International Conference on Computer Science and Network Technology (ICCSNT), Dalian, China, 21–22 October 2023; pp. 37–42. [Google Scholar]
Zhao, L.; Shi, Y.-T.; Zhang, T.; Zhou, M. A Temporal Knowledge Graphs Prediction Method for Community Gas Risk. In Proceedings of the 2022 4th International Conference on Intelligent Information Processing (IIP), Guangzhou, China, 14–16 October 2022; pp. 23–27. [Google Scholar]
Yin, Z.; Shi, L.; Yuan, Y.; Tan, X.; Xu, S. A Study on a Knowledge Graph Construction Method of Safety Reports for Process Industries. Processes 2023, 11, 146. [Google Scholar] [CrossRef]
Ma, L.; Wang, J.; Cheng, J.; Wang, X.; Zhu, W. MLRP-KG: Mine Landslide Risk Prediction Based on Knowledge Graph. IEEE Trans. Artif. Intell. 2022, 3, 78–87. [Google Scholar] [CrossRef]
Gan, L.; Chen, Q.; Zhang, D.; Zhang, X.; Zhang, L.; Liu, C.; Shu, Y. Construction of Knowledge Graph for Flag State Control (FSC) Inspection for Ships: A Case Study from China. J. Mar. Sci. Eng. 2022, 10, 1352. [Google Scholar] [CrossRef]
Panday, R.; Lung, C.-H. Unstructured Transportation Safety Board Findings Categorization Using the Knowledge Graph Pipeline. In Proceedings of the 2023 IEEE International Conference on Big Data (BigData); IEEE: New York, NY, USA, 2023; pp. 2624–2632. [Google Scholar] [CrossRef]
Yang, H.; Yang, W.; Zhang, N.; Wei, S.; Shang, Y. Construction Method of Equipment Defect Knowledge Graph in IoT. IASC 2023, 37, 2745–2765. [Google Scholar] [CrossRef]
Shin, E.; Yoo, S.; Ju, Y.; Shin, D. Knowledge Graph Embedding and Reasoning for Real-Time Analytics Support of Chemical Diagnosis from Exposure Symptoms. Process Saf. Environ. Prot. 2022, 157, 92–105. [Google Scholar] [CrossRef]
Yang, Y.; Huang, C.; Zhang, H.; Feng, C.; Wang, Z.; Cui, Z. Research on Airspace Security Risk Assessment Technology Based on Knowledge Graph. In Proceedings of the 2021 IEEE 21st International Conference on Software Quality, Reliability and Security Companion (QRS-C), Hainan, China, 6–10 December 2021; pp. 980–986. [Google Scholar]
Zhang, L.; Zhang, M.; Tang, J.; Ma, J.; Duan, X.; Sun, J.; Hu, X.; Xu, S. Analysis of Traffic Accident Based on Knowledge Graph. J. Adv. Transp. 2022, 2022, 3915467. [Google Scholar] [CrossRef]
Liu, X.; Zheng, R.; Wang, H.; Butala, M.D.; Liu, D.; Ren, X.; Hu, S. A Knowledge Management Framework for Vehicle Hazard Analysis. In Proceedings of the 2021 IEEE International Conference on e-Business Engineering (ICEBE), Guangzhou, China, 12–14 November 2021; pp. 165–169. [Google Scholar]
Pedro, A.; Pham-Hang, A.-T.; Nguyen, P.T.; Pham, H.C. Data-Driven Construction Safety Information Sharing System Based on Linked Data, Ontologies, and Knowledge Graph Technologies. Int. J. Environ. Res. Public Health 2022, 19, 794. [Google Scholar] [CrossRef]
Wan, H.; Xiao, Y.; Mo, J.; Yu, Y.; Li, H. Research on Ship Relation Graph Analysis Driven by Multi-Source Data. In Proceedings of the 2021 6th International Conference on Transportation Information and Safety (ICTIS), Wuhan, China, 22–24 October 2021; pp. 655–660. [Google Scholar]
Huang, Y.; Yin, P.; Zhou, G.; Liu, P.; Tang, Y.; Li, W. Construction of Public Safety Knowledge Graphs. In Proceedings of the 2020 International Conference on Computer, Information and Telecommunication Systems (CITS), Hangzhou, China, 5–7 October 2020; pp. 1–4. [Google Scholar]
Liu, Y.; Chai, Y.; Lin, W.; Li, X. Safety Analysis of Cryogenic Loading System Based on Knowledge Graph. In Proceedings of the 2021 40th Chinese Control Conference (CCC), Shanghai, China, 26–28 July 2021; pp. 4677–4682. [Google Scholar]
Zhu, Y.; Luo, X. A Knowledge Graph for Automated Construction Workers’ Safety Violation Identification. In ISARC Proceedings 2022, 2022 Proceedings of the 39th ISARC, Bogotá, Colombia, 13–15 July 2022; IAARC Publications: Singapore, 2022; pp. 312–319. [Google Scholar] [CrossRef]
Jiao, Y.; Han, J.; Xu, B.; Xiao, M.; Shen, B.; Sun, H. Research on Domain Entity Extraction in Civil Aviation Safety. In Proceedings of the 2021 IEEE 3rd International Conference on Civil Aviation Safety and Information Technology (ICCASIT), Changsha, China, 20–22 October 2021; pp. 384–388. [Google Scholar]
Sun, J.; Zhao, D.; Wang, L.; Chen, X.; Yi, M.; Xia, L. Remote Supervision Relation Extraction Method of Power Safety Regulations Knowledge Graph Based on ResPCNN-ATT. In Proceedings of the 2021 IEEE International Conference on Power Electronics, Computer Applications (ICPECA), Shenyang, China, 22–24 January 2021; pp. 54–58. [Google Scholar]
Li, F.; Zhang, B.; Gao, D. Chinese Named Entity Recognition for Hazard and Operability Analysis Text. In Proceedings of the 2020 Chinese Control and Decision Conference (CCDC), Hefei, China, 22–24 August 2020; pp. 374–378. [Google Scholar]
Fang, W.; Ma, L.; Love, P.E.D.; Luo, H.; Ding, L.; Zhou, A. Knowledge Graph for Identifying Hazards on Construction Sites: Integrating Computer Vision with Ontology. Autom. Constr. 2020, 119, 103310. [Google Scholar] [CrossRef]
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. [Google Scholar] [CrossRef]
Zhang, K.; Huang, W.; Hou, X.; Xu, J.; Su, R.; Xu, H. A Fault Diagnosis and Visualization Method for High-Speed Train Based on Edge and Cloud Collaboration. Appl. Sci. 2021, 11, 1251. [Google Scholar] [CrossRef]
Liu, J.; Schmid, F.; Li, K.; Zheng, W. A Knowledge Graph-Based Approach for Exploring Railway Operational Accidents. Reliab. Eng. Syst. Saf. 2021, 207, 107352. [Google Scholar] [CrossRef]
Bai, Y.; Xing, Y.; Wu, J. Integrating Knowledge Graph, Complex Network and Bayesian Network for Data-Driven Risk Assessment. Chem. Eng. Trans. 2022, 90, 31–36. [Google Scholar] [CrossRef]
De Nicola, A.; Villani, M.L.; Costantino, F.; Di Gravio, G.; Falegnami, A.; Patriarca, R. A Knowledge Graph to Digitalise Functional Resonance Analyses in the Safety Area. In Resilience in a Digital Age: Global Challenges in Organisations and Society; Matos, F., Selig, P.M., Henriqson, E., Eds.; Springer International Publishing: Cham, Switzerland, 2022; pp. 259–269. [Google Scholar]
Zeng, Y.; Qin, Y.; Liu, D.; Fu, Y.; Gong, M.; Zhang, X. Railway Train Device Fault Causality Model Based on Knowledge Graph. In Proceedings of the 2020 International Conference on Sensing, Diagnostics, Prognostics, and Control (SDPC), Beijing, China, 5–7 August 2020; pp. 385–390. [Google Scholar]
Liu, Z.; Zheng, S.; Shi, X. Research on Intelligent Question Answering System for Chemical Safety Based on Knowledge Graph. In Proceedings of the 2022 International Conference on Artificial Intelligence and Computer Information Technology (AICIT), Yichang, China, 16–18 September 2022; pp. 1–4. [Google Scholar]
Xu, Y.; Liang, Y.; Li, K. A Knowledge Graph-Based Method for Modeling and Analyzing the Disaster Risks of Railway Construction in the Mountainous Area of Southwest China. In Proceedings of the 2021 4th International Symposium on Traffic Transportation and Civil Architecture (ISTTCA), Suzhou, China, 12–14 November 2021; pp. 259–264. [Google Scholar]
Zhang, W.; Shi, H.; Yang, Y.; Luo, Y. Research on the Classification of Aviation Safety Reports Based on Text and Knowledge Graph; IOP: London, UK, 2020; Volume 1646, p. 012028. [Google Scholar]
Mao, S.; Zhao, Y.; Chen, J.; Wang, B.; Tang, Y. Development of Process Safety Knowledge Graph: A Case Study on Delayed Coking Process. Comput. Chem. Eng. 2020, 143, 107094. [Google Scholar] [CrossRef]
Duan, X.; Chen, Y.; Ji, Z.; Pei, Z.; Yi, W. Safety-Critical Components Analysis Using Knowledge Graph for CNC Machine. In Proceedings of the 2022 IEEE 18th International Conference on Automation Science and Engineering (CASE); IEEE: New York, NY, USA, 2022; pp. 1617–1621. [Google Scholar] [CrossRef]
Zheng, X.; Wang, B.; Zhao, Y.; Mao, S.; Tang, Y. A Knowledge Graph Method for Hazardous Chemical Management: Ontology Design and Entity Identification. Neurocomputing 2021, 430, 104–111. [Google Scholar] [CrossRef]
Zhang, Q.; Wen, Y.; Zhou, C.; Long, H.; Han, D.; Zhang, F.; Xiao, C. Construction of Knowledge Graphs for Maritime Dangerous Goods. Sustainability 2019, 11, 2849. [Google Scholar] [CrossRef]
Cheng, Y.; Jiao, Y.; Wei, W.; Wu, Z. Research on Construction Method of Knowledge Graph in the Civil Aviation Security Field. In Proceedings of the 2019 IEEE 1st International Conference on Civil Aviation Safety and Information Technology (ICCASIT), Kunming, China, 17–19 October 2019; pp. 556–559. [Google Scholar]
Zhao, Q.; Li, Q.; Wen, J. Construction and Application Research of Knowledge Graph in Aviation Risk Field. In Proceedings of the MATEC Web of Conferences; Adiguzel, O., McAndrew, I., Yokoi, Y., Koryanov, V., Eds.; EDP Sciences: Les Ulis, France, 2018; Volume 151, p. 05003. [Google Scholar]
Liu, J.; Zhang, X.; Li, Y.; Wang, J.; Kim, H.-J. Deep Learning-Based Reasoning With Multi-Ontology for IoT Applications. IEEE Access 2019, 7, 124688–124701. [Google Scholar] [CrossRef]
Chen, J.; Zhuang, C.; Shi, J.; Jiang, H.; Xu, J.; Liu, J. Risk Factors Extraction and Analysis of Chinese Ship Collision Accidents Based on Knowledge Graph. Ocean Eng. 2025, 322, 120536. [Google Scholar] [CrossRef]
Wang, Y.; Hu, S. Design and Realization of Compressor Data Abnormality Safety Monitoring and Inducement Traceability Expert System. PLoS ONE 2025, 20, e0315917. [Google Scholar] [CrossRef]
Zhou, Z.; Yu, X.; Magoua, J.J.; Cui, J.; Luan, H.; Lin, D. Integrating Machine Learning and a Large Language Model to Construct a Domain Knowledge Graph for Reducing the Risk of Fall-from-Height Accidents. Accid. Anal. Prev. 2025, 215, 108009. [Google Scholar] [CrossRef]
Tian, X.; Gan, H.; Liu, Y. Construction of Knowledge Graph for Marine Diesel Engine Faults Based on Deep Learning Methods. J. Mar. Sci. Eng. 2025, 13, 693. [Google Scholar] [CrossRef]
Li, X.; Li, S.; Yuan, J.; Wan, Z.; Liu, X. A Data-Driven and Knowledge Graph-Based Research on Safety Risk-Coupled Evolution Analysis and Assessment in Shield Tunneling. Tunn. Undergr. Space Technol. 2025, 162, 106657. [Google Scholar] [CrossRef]
Yang, Y.; Xiang, P. A Knowledge Graph for the Vulnerability of Construction Safety System in Megaprojects Based on Accident Inversion. Eng. Appl. Artif. Intell. 2025, 150, 110630. [Google Scholar] [CrossRef]
Zhou, K.; Lu, X.; Yang, C.; Chen, Z.; Liu, W.; Yan, H. Architecture and Application of Mine Ventilation System Safety Knowledge Graph Based on Neo4j. Sustainability 2025, 17, 3209. [Google Scholar] [CrossRef]
Cappelli, M.A.; Di Marzo Serugendo, G. Ontology-Based Customisation Management System for Driver-Vehicle Interfaces: A Preventive Approach to Incident Reduction and Legal Accountability in Highly Automated Vehicles. Appl. Sci. 2025, 15, 1043. [Google Scholar] [CrossRef]
Men, C.; Han, Y.; Huang, C.-G. Prediction of Equipment Remaining Useful Life Based on Graph Learning and Spatiotemporal Knowledge Graph. Qual. Reliab. Eng. Int. 2025, 41, 1209–1224. [Google Scholar] [CrossRef]
Xue, N.; Zhang, W.; Zhong, H.; Liao, W.; Zhao, T. How to Realize the Knowledge Reuse and Sharing from Accident Reports? A Knowledge-Driven Modeling Method Combining Ontology and Deep Learning. J. Loss Prev. Process Ind. 2025, 94, 105525. [Google Scholar] [CrossRef]
Ding, X.; Hu, H.; Hu, Y.; Jin, J.; Liu, Z.; Pan, H.; Shi, G. Identification and Precise Control of Disaster-Causing Hazards in Metro Operation and Maintenance: A New Method for Improving Metro Operation Safety Based on Data Mining. Comput. Ind. Eng. 2025, 201, 110899. [Google Scholar] [CrossRef]
Wu, P.; Mou, X.; Gong, L.; Tu, H.; Qiu, L.; Yang, B. An Automatic Machine Fault Identification Method Using the Knowledge Graph–Embedded Large Language Model. Int. J. Adv. Manuf. Technol. 2025, 138, 725–739. [Google Scholar] [CrossRef]
Huang, X.; Li, P.; Wang, Y.; Ren, X.; Zhao, Z.; Li, G. Knowledge Graph-Augmented ERNIE-CNN Method for Risk Assessment in Secondary Power System Operations. Energies 2025, 18, 2104. [Google Scholar] [CrossRef]
Gros, A.; De Luca, L.; Dubois, F.; Véron, P.; Jacquot, K. From Surveys to Simulations: Integrating Notre-Dame de Paris’ Buttressing System Diagnosis with Knowledge Graphs. Autom. Constr. 2025, 170, 105927. [Google Scholar] [CrossRef]
Liu, X.; Wu, H.; Yu, D.; Chen, Y.; Wu, H. A Construction and Representation Learning Method for a Traffic Accident Knowledge Graph Based on the Enhanced TransD Model. Appl. Sci. 2025, 15, 6031. [Google Scholar] [CrossRef]
Li, Z.; Zhao, J. HAZOPCT: A HAZOP Analysis Completeness Tool Based on Knowledge Graph Reasoning. Process Saf. Environ. Prot. 2025, 197, 107025. [Google Scholar] [CrossRef]
Jha, A.N.; Gupta, R.K.; Chatterjee, N.; Tiwari, G. A Domain Ontology for Safety of Road Users—SafeOn: Overview & Design. Transp. Res. Procedia 2025, 82, 1925–1948. [Google Scholar] [CrossRef]
Cui, D.; Xu, S.; Wang, S.; Zhang, K. Beyond the Images: Comprehensible Unsafe Behaviour Recognition Boosted by Joint Inference Graph with Multi-Hop Reasoning. Adv. Eng. Inform. 2025, 66, 103454. [Google Scholar] [CrossRef]
Zhang, D.; Ma, G.; Qu, T.; Wang, X.; Zhou, W.; Wang, X. A Knowledge Graph-Enhanced Large Language Model for Question Answering of Hydraulic Structure Safety Management. Adv. Eng. Inform. 2025, 66, 103468. [Google Scholar] [CrossRef]
Yang, X.; Li, H.; Yan, J.; He, R. TH-RotatE: A Hybrid Knowledge Graph Embedding Framework for Fault Diagnosis in Railway Operational Equipment. Electronics 2025, 14, 1656. [Google Scholar] [CrossRef]
Zhu, P.; Li, Y.; Xu, P.; Li, P.; Zhao, Z.; Li, G. Secondary Operation Risk Assessment Method Integrating Graph Convolutional Networks and Semantic Embeddings. Sensors 2025, 25, 1934. [Google Scholar] [CrossRef]
Johansen, K.W.; Schultz, C.; Teizer, J. Knowledge Graph Exploitation to Enhance the Usability of Risk Assessment in Construction Safety Planning. Adv. Eng. Inform. 2025, 65, 103305. [Google Scholar] [CrossRef]
Xu, X. RCA Analysis of Multi-Source Faults in Autonomous Driving. Int. J. Inf. Syst. Model. Des. 2025, 16, 1–28. [Google Scholar] [CrossRef]
Shi, D.; Gan, S.; Zurada, J.; Guan, J.; Wang, F.; Weichbroth, P. A Multi-Model Approach to Construction Site Safety: Fault Trees, Bayesian Networks, and Ontology Reasoning. Expert Syst. Appl. 2025, 288, 127817. [Google Scholar] [CrossRef]
Wang, L.; Liu, X.; Dong, Y.; Zhao, D.; Wang, Z.; Chen, X. Risk Evolution Analysis of Cross-Regional Water Diversion Projects Based on Spatio-Temporal Knowledge Graphs. J. Hydrol. 2025, 650, 132533. [Google Scholar] [CrossRef]
Huang, Y.; Zhang, Z.; Hu, H. Risk Propagation Mechanisms in Railway Systems under Extreme Weather: A Knowledge Graph-Based Unsupervised Causation Chain Approach. Reliab. Eng. Syst. Saf. 2025, 260, 110976. [Google Scholar] [CrossRef]
Wen, S.; Li, F.; Zhuang, W.; Pan, X.; Yu, W.; Bao, J.; Li, X. Leveraging Large Language Models for Human-Machine Collaborative Troubleshooting of Complex Industrial Equipment Faults. Adv. Eng. Inform. 2025, 65, 103235. [Google Scholar] [CrossRef]
Isah, M.A.; Kim, B.-S. Question-Answering System Powered by Knowledge Graph and Generative Pretrained Transformer to Support Risk Identification in Tunnel Projects. J. Constr. Eng. Manag. 2025, 151, 04024193. [Google Scholar] [CrossRef]
Lu, Y.; Qiao, S.; Yao, Y. Risk Assessment of Typhoon Disaster Chain Based on Knowledge Graph and Bayesian Network. Sustainability 2025, 17, 331. [Google Scholar] [CrossRef]
Gan, B.-L.; Zhang, D.-M.; Huang, Z.-K.; Zheng, F.-Y.; Zhu, R.; Zhang, W. Ontology-Driven Knowledge Graph for Decision-Making in Resilience Enhancement of Underground Structures: Framework and Application. Tunn. Undergr. Space Technol. 2025, 163, 106739. [Google Scholar] [CrossRef]
Tao, Z.; Liu, X.; Li, Y.; Hu, P.; Tang, W.; Luo, N.; Wu, J.; Yang, R. Intelligent Emergency Assisted Decision-Making Method Based on Standard Digitalization: Hazardous Chemical Accidents in Industrial Parks. J. Saf. Sci. Resil. 2025, 6, 79–92. [Google Scholar] [CrossRef]
Zhou, X.; Shi, J.; Dong, L.; Zhang, Y.; Pan, J.; Huang, H. Construction of a Multimodal Knowledge Graph for Power Grid Construction Safety Based on Large Language Models. In Proceedings of the 2024 International Conference on New Power System and Power Electronics (NPSPE), Dalian, China, 16–18 August 2024; pp. 21–28. [Google Scholar]
Zheng, Z.; Zhang, H.; Zhang, J.; Liu, P.; Pan, Y. A Research on Gas Safety Knowledge Graph and Retrieval-Augmented Generation Mechanism Based on Large Language Model. In Proceedings of the 2024 10th International Conference on Computer and Communications (ICCC), Chengdu, China, 13–16 December 2024; pp. 311–315. [Google Scholar]
Liu, Z.; Hong, H.; Deng, Y.; Li, Q.; Shang, M.; Tang, X. Fault Early Warning and Judgment System of Low-Voltage Substation Based on Deep Learning and Knowledge Map. In Proceedings of the 2024 International Conference on Power, Electrical Engineering, Electronics and Control (PEEEC), Athens, Greece, 14–16 August 2024; pp. 797–802. [Google Scholar]
Liu, Y.; Xu, P.; Wang, R.; Sun, Q.; Zhou, C. Safety Risk Assessment in Fluid Catalytic Cracking Units Based on HAZOP and Knowledge Graph. In Proceedings of the 2024 China Automation Congress (CAC), Qingdao, China, 1–3 November 2024; pp. 2102–2107. [Google Scholar]
Peng, H.; Sun, K.; Zhou, Q.; Zhang, Y.; Yang, W.; Li, Y. Ontology Construction of Fault Diagnosis Knowledge Graph for Civil Aircraft Maintenance. In Proceedings of the 2024 Global Reliability and Prognostics and Health Management Conference (PHM-Beijing), Beijing, China, 11–13 October 2024; pp. 1–7. [Google Scholar]
Zeng, D.; Zhao, Y.; Yang, Z. Construction of Safety Knowledge Graph for Near Electric Work Based on Graph Visualization. In Proceedings of the 2024 5th International Conference on Clean Energy and Electric Power Engineering (ICCEPE), Yangzhou, China, 9–11 August 2024; pp. 756–759. [Google Scholar]
Zhang, W.; Yang, M.; Nie, R.; Lai, R. Emergency Disposal Decision Generation Method for Flight Test Based on Knowledge Graph. In Proceedings of the 2024 4th International Conference on Communication Technology and Information Technology (ICCTIT), Guangzhou, China, 27–29 December 2024; pp. 636–642. [Google Scholar]
Ji, L.; Liu, J.; Chen, K. Research on Railway Operational Accidents Analysis Method Based on Knowledge Graph. In Proceedings of the 2024 China Automation Congress (CAC), Qingdao, China, 1–3 November 2024; pp. 4370–4375. [Google Scholar]
Luo, Z.; Lu, C.; Wang, Y.; Lv, M.; Liu, X.; Xu, C. Distribution Transformer Fault Data Based on One-Hot Coded Word Vector Knowledge Graph Construction Study. In Proceedings of the 2024 5th International Conference on Clean Energy and Electric Power Engineering (ICCEPE), Yangzhou, China, 9–11 August 2024; pp. 1073–1076. [Google Scholar] [CrossRef]

Figure 2. Temporal distribution of eligible documents. The graph includes contributions indexed up to 20 June 2025. The lower number in 2025 represents incomplete coverage.

Figure 3. Distribution of documents per type of document.

Figure 4. Number of documents per sector, industry group, and industry [13].

Figure 5. Framework’s space to classify safety research leveraging KGs.

Figure 6. Distribution of documents in the OC-OE plane of the framework. Different colors for areas identify different archetypes, as for Figure 5.

Figure 7. Number of documents per MSA grouped by archetype. Darker colors identify bigger bubbles in the presented chart.

Figure 8. Number of documents (future perspectives) per MSA grouped by archetype. Darker colors identify bigger bubbles in the presented chart.

Table 1. Standardized residuals for archetypes and MSA levels.

Archetype	MSA = 2	MSA = 3	MSA = 4
Assemblers	1.05	−1.49	0.18
Alchemists	1.05	1.25	−2.28
Shapers	−1.85	−0.09	2.20

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Simone, F.; Montaruli, A.; Fandino, K.H.; Patriarca, R. A Tale of Three Words: Knowledge, Safety, and Graphs. Information 2026, 17, 599. https://doi.org/10.3390/info17060599

AMA Style

Simone F, Montaruli A, Fandino KH, Patriarca R. A Tale of Three Words: Knowledge, Safety, and Graphs. Information. 2026; 17(6):599. https://doi.org/10.3390/info17060599

Chicago/Turabian Style

Simone, Francesco, Andrea Montaruli, Kristopher Hernandez Fandino, and Riccardo Patriarca. 2026. "A Tale of Three Words: Knowledge, Safety, and Graphs" Information 17, no. 6: 599. https://doi.org/10.3390/info17060599

APA Style

Simone, F., Montaruli, A., Fandino, K. H., & Patriarca, R. (2026). A Tale of Three Words: Knowledge, Safety, and Graphs. Information, 17(6), 599. https://doi.org/10.3390/info17060599

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

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.

Article Menu

A Tale of Three Words: Knowledge, Safety, and Graphs

Abstract

1. Introduction

1.1. Knowledge

1.2. Safety

1.3. And Graphs

1.4. Aim of This Research

2. Method

3. Results

3.1. Bibliometric Analysis

3.2. A Framework for Characterizing Safety Research Leveraging Knowledge Graphs

3.2.1. About Assemblers

3.2.2. About Alchemists

3.2.3. About Shapers

4. Discussion

5. Conclusions

Author Contributions

Funding

Institutional Review Board Statement

Informed Consent Statement

Data Availability Statement

Conflicts of Interest

Abbreviations

Appendix A

References

Share and Cite

Article Metrics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI