A Bibliometric and Visualized Overview for the Evolution of Process Safety and Environmental Protection

This paper presents a bibliometric overview of the publications in the principal international journal Process Safety and Environmental Protection (PSEP) from 1990 to 2020 retrieved in the Web of Science (WoS) database to explore the evolution in safety and environmental engineering design and practice, as well as experimental or theoretical innovative research. Therefore, based on the WoS database and the visualization of similarities (VOS) viewer software, the bibliometric analysis and scientometric mapping of the literature have been performed from the perspectives of document types, publication and citation distribution over time, leading authors, countries (regions), institutions, the corresponding collaboration networks, most cited publications and references, focused research fields and topics, research trend evolution over time, etc. The paper provides a comprehensive and quantitative overview and significant picture representation for the journal’s leading and evolutionary trends by employing specific aforementioned bibliometric analysis factors. In addition, by reviewing the evolutionary trends of the journal and the proposed investigated factors, such as the influential works, main research topics, and the research frontiers, this paper reveals the scientific literature production’s main research objectives and directions that could be addressed and explored in future studies.


Introduction
Bibliometrics originated from library and information science [1]. A bibliographic analysis is mainly applied to characterize the structure and research trends of a specific field or journal by utilizing a quantitative methodology [2][3][4]. Additionally, it is a comprehensive visual analysis method augmented with network topology that could detect the influential authors, institutions, and countries in a specific research domain [5] and demonstrate a journal's influence and productivity [1].
Moreover, scientific literature mapping by utilizing bibliometric methods is an effective complement to the traditional structured literature review, as it is able to provide a The remainder of this paper is organized as follows: first, Section 2 presents the materials and bibliometric analysis methods utilized in the paper. Second, the statistical analysis and graphical analysis results are detailed, including the publication trends and citation distribution; leading authors, institutions, and countries/regions; influential works in PSEP; and identified research fields and research evolutionary trends in the perspective of keyword co-occurrence. Additionally, the accompanying discussions are conducted in Section 3. Finally, Section 4 concludes the main findings of the paper and delivers the recommendations for the readers.

Bibliographic Data
In this paper, a typical journal PSEP with a high reputation in the research fields of safety of industrial processes and protection of the environment and with a relatively rapid increase of impact factor, quick review speed, and online article publication time, etc., was selected as the candidate journal to be analyzed.
The data were retrieved on 9 January 2021 from the WoS Core Collection, which is owned by Clarivate Analytics. The advanced search module was employed, and the strategy for obtaining data was, "Publication Name: SO = (Process Safety and Environmental Protection), Indexes = SCI-EXPANDED, Timespan = 1990-2020". In total, 3152 publications were obtained from the Web of Science, and the PSEP publications had 13 different types (some of the papers were classified into more than one category). The proportion of each document type is shown in Figure 1. Note that articles and reviews are more essential document types in scientific outputs, and these two types have nearly 90.09% (2965 articles and reviews) in PSEP. The total number of citations was 44,879, and the average number of citations per publication was 14.24.

Bibliometric Methods and Analysis Tool
In the present paper, bibliometric methods were applied, and the bibliometric mapping tool VOSviewer was used to analyze the journal papers in a visual, user-friendly way. The bibliometric analysis originated from information and library science, which was first proposed by Otlet [23]. In the data science age, bibliometric methods were combined with network analysis and data visualization techniques, and then a new area named bibliometric mapping was produced. The bibliometric mapping was about quantitative methods (mathematics and statistics) for visually representing scientific literature based on bibliographic data.

Bibliometric Methods and Analysis Tool
In the present paper, bibliometric methods were applied, and the bibliometric mapping tool VOSviewer was used to analyze the journal papers in a visual, user-friendly way. The bibliometric analysis originated from information and library science, which was first proposed by Otlet [23]. In the data science age, bibliometric methods were combined with network analysis and data visualization techniques, and then a new area named bibliometric mapping was produced. The bibliometric mapping was about quantitative methods (mathematics and statistics) for visually representing scientific literature based on bibliographic data.
Additionally, in bibliometrics, a threshold is used to select the minimum frequency of occurrence of the knowledge unit included in the network node. In the analysis of different knowledge units, the set of the specific threshold will have individual differences. Its pri-

Publication Trend and Citation Distribution
The publication trend (including the number of papers) is the mirror and indicator for reflecting and measuring the scientific activities and attention to a specific domain. Figure 2 and Table 1 show the annual increase trend of PSEP publications. The increase of the annual outputs shows the increased attention to the topic scope of the PSEP from scientific communities. PSEP, as one of the leading journals in industrial process safety and environmental protection, has released a total of 16 publications in 1990, according to its earliest record in WoS. Moreover, the number of publications has increased slowly before 2013, and the average value for the number of publications per year before 2013 was around 48. After 2013, the publication trend increased rapidly, and the outputs reached more than 100 papers per year, with the number of papers in 2019 being 432 (exceeding 400 the first time). Additionally, the cumulative percentage of the number of publications showed that nearly 50% of cumulative publications from PSEP were published after 2016, which means that the most recent five years (2016-2020) have contributed roughly half of all of the papers that have been published in PSEP from 1990 to 2020.    Note: NP = number of publications, CNP = cumulative number of publications, TC = total citations, CPP = citations per paper = TC/NP. The colors range from green to red in the related column indicate the smaller the number and the closer the color is to green in that column.

Leading Authors and Collaborations
Authors are the knowledge producers of PSEP, and an author's production and collaboration analysis can easily show the leading researchers and the author's social connectedness in PSEP. The whole author's collaboration network is illustrated in Figure 3, and the author who had a minimum number of publications of 10 was regarded as the leading author in PSEP. Table 2 lists the leading authors of PSEP, and indicates that there are 69.70% (23/33) of them within the giant connected component (GCC) of the authors' collaboration network (cluster 1, the red group ( Figure 8. Note that the larger the nodes and character fonts, the more often the key are used. Obviously, the keyword "adsorption" is the most used author's keyword w occurrences, followed by "kinetics" (80), "response surface methodology" (74), " sessment" (73), "optimization" (67), "heavy metals" (56), "safety" (56), etc. There w clusters of keywords, separated by different colors and representing the following t The blue group ( ) was mainly concentrated on waste and pollutants remed This cluster included keywords such as "adsorption" (122), "kinetics" (80), "heav als" (56), "wastewater" (43), "activated carbon" (40), "isotherm" (34), "biosorption "composting" (22), "thermodynamic" (22), "volatile organic compound" (22), etc.
The purple group ( ) was mainly about waste management and sustainable opment. This cluster was represented by keywords such as "anaerobic digestion "pyrolysis" (28), "recycling" (27), "lifecycle assessment" (24), "biomass" (21), "en ment" (20), "biogas" (19), "mathematical modeling" (18), "combustion" (17), "mas fer" (15), "energy" (14), "sustainable development" (12), "emissions" (12), etc. Figure 3). Considering the sparsely connected network structure of other groups, only the GCC of the authors' collaboration network was selected to analyze the author's social connection in PSEP. Note that authors with only the publication type of editorial material are not included in Table 2. The node size is proportionate to the number of publications of an author; the node color represents the clusters of authors in the same group. Additionally, the links between authors present the collaboration relations between authors, and the wideness of the link shows the authors' collaboration strength. Note: NP = number of publications, CNP = cumulative number of publications, TC = total citations, CPP = citations per paper = TC/NP. The colors range from green to red in the related column indicate the smaller the number and the closer the color is to green in that column.

Leading Authors and Collaborations
Authors are the knowledge producers of PSEP, and an author's production and collaboration analysis can easily show the leading researchers and the author's social connectedness in PSEP. The whole author's collaboration network is illustrated in Figure 3, and the author who had a minimum number of publications of 10 was regarded as the leading author in PSEP. Table 2 lists the leading authors of PSEP, and indicates that there are 69.70% (23/33) of them within the giant connected component (GCC) of the authors' collaboration network (cluster 1, the red group ( ) in Figure 3). Considering the sparsely connected network structure of other groups, only the GCC of the authors' collaboration network was selected to analyze the author's social connection in PSEP. Note that authors with only the publication type of editorial material are not included in Table 2. The node size is proportionate to the number of publications of an author; the node color represents the clusters of authors in the same group. Additionally, the links between authors present the collaboration relations between authors, and the wideness of the link shows the authors' collaboration strength.   As shown in Table 2, Faisal Khan (Mem. Univ. Newfoundland, Canada) is the most productive author in PSEP with 62 publications. He is also the only author with more than 50 papers in PSEP, followed by Paul Amyotte (Dalhousie University, Canada), P.J. Thomas (193.13). The productivity distribution of the (leading) authors was not balanced; there were just a few authors who have published a large number of papers, resulting in the uneven distribution of the total number of citations of leading authors. In addition, the highly productive editorial board members of PSEP are also highlighted in Table 2. Among the leading authors, 24.24% (8/33) are editorial board members, which indicates that the editorial board members play a relatively important role among the leading authors, as well as within the process safety and environmental protection research domain. Additionally, as shown in Table 2, 27.27% (9/33) of the leading authors originate from the United Kingdom.

Leading Countries/Regions and Collaboration
The countries/regions cooperation relation in the explored field was also visualized and analyzed by utilizing collaboration networks analysis to investigate affiliated countries and institutes through the VOSviewer software. As shown in Figure 4, the minimum number of publications was 10, and 44 countries/regions were included in the network. It should be noted that a node was apportioned to each co-author of a publication in the networks. The node's color presents the average time for the publications of each country [41]. The node's size denotes the related publication number, and the thickness of the links indicates the international collaboration degree [38,42], i.e., the larger the node is, the more critical the country/region is, and the thicker the line is, the closer the cooperation relationship between countries/regions. The countries/regions cooperation relation in the explored field was also visualized and analyzed by utilizing collaboration networks analysis to investigate affiliated countries and institutes through the VOSviewer software. As shown in Figure 4, the minimum number of publications was 10, and 44 countries/regions were included in the network. It should be noted that a node was apportioned to each co-author of a publication in the networks. The node's color presents the average time for the publications of each country [41]. The node's size denotes the related publication number, and the thickness of the links indicates the international collaboration degree [38,42], i.e., the larger the node is, the more critical the country/region is, and the thicker the line is, the closer the cooperation relationship between countries/regions.     (25.36) are the only three countries whose average citations exceeded 25. Generally, more international collaboration needs to be promoted and enhanced to share knowledge globally in the future. Note: C/R = country/region, NP = number of publications, TC = total citations, APY = average publication year, CPP = citations per paper (average citations) = TC/NP. Figure 5 illustrates the institution collaboration network for PSEP during the entire explored timespan. The node's size presents the number of publications (the bigger the note is, the more publications the institution has), and the links between two nodes indicate the collaboration relationship between two institutions (the thicker the link is, the closer the cooperation they have). In addition, each institution in the network has a minimum of 10 publications in PSEP, and 77 institutions meet the threshold. As shown in Table 4

Influential Works Published by PSEP
Publications with a large number of citations often indicate the influence of the publication in a specific research domain, i.e., the number of publications exceeding a certain citation threshold allows the identification of the number of publications that have a certain level of influence [43,44]. In this paper, the publications with more than 100 citations are identified as influential works in PSEP. Therefore, 25 publications are listed in Table 5. The paper by Ho and McKay [45] held the highest number of citations of 1530 and the biggest average number of citations per year of 66.52. Moreover, there were seven publications with a total number of citations of more than 200. Additionally, there were two papers among the first five most cited papers (2/5 = 40.00%), and 24.00% (6/25) of the top 25 most cited papers were review papers, while, as shown in Figure 1, only 2.89% of all publications were the review papers. The statistical fact that a relatively small number of publications accomplished with a relatively high total number of citations indicated that the document type of review paper was more likely to get more citations.  Furthermore, Figure 6 demonstrates the citation distribution of PSEP publications from 1990 to 2020. Overall, according to the increasing of the number of citations, the number of publications gradually decreased. In addition, there were 1924 publications, the highest number of publications among various intervals, that had no more than ten citations. Note that, among the publications cited no more than ten times, there were 447 publications with no citations. Considering that a paper's publication requires a certain period, the citations cannot be counted in time. However, except for the 183 publications of 2020, 59.06% (264/447) of the publications had zero citations. Note that, since most of the influential works' research topics were cross-fused with the research hotspots in Section 3.4, the related literature productions are not discussed and analyzed in detail in this section.

Influential Works Published by PSEP
Publications with a large number of citations often indicate the influence of the publication in a specific research domain, i.e., the number of publications exceeding a certain citation threshold allows the identification of the number of publications that have a certain level of influence [43,44]. In this paper, the publications with more than 100 citations are identified as influential works in PSEP. Therefore, 25 publications are listed in Table  5. The paper by Ho and McKay [45] held the highest number of citations of 1530 and the biggest average number of citations per year of 66.52. Moreover, there were seven publications with a total number of citations of more than 200. Additionally, there were two papers among the first five most cited papers (2/5 = 40.00%), and 24.00% (6/25) of the top 25 most cited papers were review papers, while, as shown in Figure 1, only 2.89% of all publications were the review papers. The statistical fact that a relatively small number of publications accomplished with a relatively high total number of citations indicated that the document type of review paper was more likely to get more citations.
Furthermore, Figure 6 demonstrates the citation distribution of PSEP publications from 1990 to 2020. Overall, according to the increasing of the number of citations, the number of publications gradually decreased. In addition, there were 1924 publications, the highest number of publications among various intervals, that had no more than ten citations. Note that, among the publications cited no more than ten times, there were 447 publications with no citations. Considering that a paper's publication requires a certain period, the citations cannot be counted in time. However, except for the 183 publications of 2020, 59.06% (264/447) of the publications had zero citations. Note that, since most of the influential works' research topics were cross-fused with the research hotspots in Section 3.4, the related literature productions are not discussed and analyzed in detail in this section.

Influential Works Cited by PSEP
Highly cited works cited by PSEP papers in our local dataset can be considered the intellectual bases of PSEP. The co-citation network of highly cited references (the minimum number of citations of a paper was 15) was constructed. In total, 43 highly cited references were identified and obtained from the 89,287 references of PSEP. The co-citation network among these 43 papers is displayed in Figure 7. The node stands for a highly cited reference, and the size is proportional to the number of cites from the PSEP papers. The label here just shows the first author or first two authors and the publication year of a paper. In addition, links between each node present the co-citation relations of highly cited references. Link wideness indicates the co-citation strength between these references. The color shows the different groups of these references, which was clustered based on the co-citation strength of these references by using the bibliometric data analysis based on the clustering method included in the widely used VOSviewer software, as introduced in Section 2.2. Note that the investigated references can only be included in one cluster, and their position in the overall network and the connections to the references in other clusters show how closely related it is, both within its own cluster and with other clusters.

Influential Works Cited by PSEP
Highly cited works cited by PSEP papers in our local dataset can be considered the intellectual bases of PSEP. The co-citation network of highly cited references (the minimum number of citations of a paper was 15) was constructed. In total, 43 highly cited references were identified and obtained from the 89,287 references of PSEP. The co-citation network among these 43 papers is displayed in Figure 7. The node stands for a highly cited reference, and the size is proportional to the number of cites from the PSEP papers. The label here just shows the first author or first two authors and the publication year of a paper. In addition, links between each node present the co-citation relations of highly cited references. Link wideness indicates the co-citation strength between these references. The color shows the different groups of these references, which was clustered based on the co-citation strength of these references by using the bibliometric data analysis based on the clustering method included in the widely used VOSviewer software, as introduced in Section 2.2. Note that the investigated references can only be included in one cluster, and their position in the overall network and the connections to the references in other clusters show how closely related it is, both within its own cluster and with other clusters. Furthermore, Table 6 lists the highly cited references of PSEP publications ranked by the number of citations. Most of the influential works cited by PSEP were journal articles, accounting for 81.40% (35/43). Additionally, the blue group ( ) was the biggest cluster with the most citations (277), and the red group ( ) was the biggest cluster with the most publications (12). As shown in Figure 7, there were six clusters (groups) for the highly cited references: The blue group ( ) was primarily concentrated on environmental protection theories and techniques (especially adsorption theory and application). The most influential works in this group were the theory for adsorption in solution [46][47][48] and adsorption of gases [49], modeling for the sorption processes [50], and isotherms systems [51]. Moreover, the review on methodologies and techniques for removing heavy metal ions from wastewaters by Fu and Wang [52] and the fundamental theory for the constitution and properties of solids and liquids by Langmuir [53], etc., are also impactful works in this group.
The red group ( ) mainly focused on the methodologies and models for process safety and risk management in chemical and process industries. Dynamic safety analysis and risk assessment theory and models, e.g., Bayesian theory [54,55], bow-tie approach [56], etc., are developed and widely used, and the aforementioned research was recognized as the influential works. Additionally, the other impactful original research was the Furthermore, Table 6 lists the highly cited references of PSEP publications ranked by the number of citations. Most of the influential works cited by PSEP were journal articles, accounting for 81.40% (35/43

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the show the paper's core content. Author keywords are imperative the topics/concepts/methods that are presented to deliver and co tific community by the authors. The author keyword co-occurr strates another perspective of themes in PSEP, and it can be obser main author keywords that frequently occur together in PSEP.

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the p show the paper's core content. Author keywords are imperative, since they are the topics/concepts/methods that are presented to deliver and communicate to th tific community by the authors. The author keyword co-occurrences network strates another perspective of themes in PSEP, and it can be observed that it illustr main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they ob have not had significant influences on the main themes of PSEP. Therefore, in the study, to focus on the main themes, only the keywords occurring at least five tim selected to construct the co-occurrence analysis map and indicate the research Thus, 407 keywords were extracted based on the threshold of the keyword's freq The keyword co-occurrence network for the clusters (groups) of PSEP is sh Figure 8. Note that the larger the nodes and character fonts, the more often the ke are used. Obviously, the keyword "adsorption" is the most used author's keyword w occurrences, followed by "kinetics" (80), "response surface methodology" (74), " sessment" (73), "optimization" (67), "heavy metals" (56), "safety" (56), etc. There w clusters of keywords, separated by different colors and representing the following The blue group ( ) was mainly concentrated on waste and pollutants reme This cluster included keywords such as "adsorption" (122), "kinetics" (80), "hea als" (56), "wastewater" (43), "activated carbon" (40), "isotherm" (34), "biosorptio "composting" (22), "thermodynamic" (22), "volatile organic compound" (22), etc.
) was the biggest cluster with the most publications (12). As shown in Figure 7, there were six clusters (groups) for the highly cited references: The blue group (

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.
) was primarily concentrated on environmental protection theories and techniques (especially adsorption theory and application). The most influential works in this group were the theory for adsorption in solution [46][47][48] and adsorption of gases [49], modeling for the sorption processes [50], and isotherms systems [51]. Moreover, the review on methodologies and techniques for removing heavy metal ions from wastewaters by Fu and Wang [52] and the fundamental theory for the constitution and properties of solids and liquids by Langmuir [53], etc., are also impactful works in this group.
The red group (

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics.
Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.
The keyword co-occurrence network for the clusters (groups) of PSEP is shown in Figure 8. Note that the larger the nodes and character fonts, the more often the keywords are used. Obviously, the keyword "adsorption" is the most used author's keyword with 122 occurrences, followed by "kinetics" (80), "response surface methodology" (74), "risk assessment" (73), "optimization" (67), "heavy metals" (56), "safety" (56), etc. There were six clusters of keywords, separated by different colors and representing the following themes: The blue group ( ) was mainly concentrated on waste and pollutants remediation.
The red group ( ) mainly focused on environmental protection methodologies and technologies. This cluster contained keywords such as "response surface methodology" ) mainly focused on the methodologies and models for process safety and risk management in chemical and process industries. Dynamic safety analysis and risk assessment theory and models, e.g., Bayesian theory [54,55], bow-tie approach [56], etc., are developed and widely used, and the aforementioned research was recognized as the influential works. Additionally, the other impactful original research was the reviews of the available techniques and methodologies for risk analysis in chemical process industries by Khan and Abbasi [57] and the methods and models in process safety and risk management by Khan et al. [58], the research for fuzzy sets by Zadeh [59], the predictive accident model for system hazard identification, prediction, and prevention by Rathnayaka et al. [60], and the utilization of Bayesian network and fault tree approaches for the safety analysis in process facilities [61] and dependable systems [62], etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. ) mainly focused on the inherent safety and hazard identification and assessment in chemical and process industries. The group contained the early influential works by Edwards and Lawrence [63] on the exploration for the relation between plant costs and the inherent safety of chemical process routes, and the multivariate system hazard identification and ranking methods for fire and explosion and toxic chemical release hazards by Khan and Abbasi [64]. In addition, Gupta and Edwards [65] proposed a graphical methodology for inherent safety measurement, Khan and Amyotte [66] detailed the cost and system design model for integrated inherent safety index, Koller et al. [67] presented a safety, health, and environmental impact assessment methodology for selecting the most reliable data from a variety of substance databases or estimation method, and Khan et al. [68] developed a safety-weighted hazard index for chemical process industry hazard identification and risk assessment. The research mentioned above was widely cited by the publications in PSEP and identified as influential works as well.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. ) concentrated on loss prevention in process industries. The influential works were primarily focused on major accident hazards, accident causes, and consequences analysis, and some specific original contributions on managing the risk of the domino effect of chemical accidents. The most cited works in this group were the book on loss prevention in the process industry by Lees [69], the book on chemical process safety by Crowl and Louvar [70], the book on the guidelines for chemical process quantitative risk analysis by CCPS [71], and the book on the identification, assessment, and prevention of hazards for dust explosion accidents, which presented the evaluation of prevalent activities, testing methods, and design measures for safe operation techniques by Eckhoff [72]. The significant and influential articles included the research for the domino effect for chemical accident features, sequences analysis, and accidental event escalation threshold assessment, as well as the research by Khan and Abbasi [73] on the common causes or errors and consequences of major accidents in chemical process industries, etc.
The purple group ( Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.
) was primarily about waste treatment (especially wastewater treatment). The most significant original research contributions in this group concerned the work on the examination of water and wastewater standard methods by APHA [74][75][76], and the significant overview or review articles included the analyzing and summarizing of the recent developments for the technologies of photocatalytic water treatment [77] and landfill leachate treatment [78], etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period.
) mainly focused on coal mine safety; it primarily contained compendia works, for instance, Cheng et al. [79] designed an intelligent gel for fire prevention and extinguishing to control coal spontaneous combustion (CSC), Wang et al. [80] established a model of airflow dust migration for underground mine tunnels and achieved a better air curtain dust suppression effect, and Karacan et al. [81] delivered an overview for the capture and utilization of coal mine methane in the perspective of mining safety and greenhouse gas reduction, etc. The aforementioned works were highly influential in this group.  Note: Cluster matches the color for each cluster group in Figure 7, DT = document type, JA = journal articles.

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies. The keyword co-occurrence network for the clusters (groups) of PSEP is shown in Figure 8. Note that the larger the nodes and character fonts, the more often the keywords are used.

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.
The purple group (

Research Fields Identification and Research Trends Evolution
Keywords are one of the essential elements supplied by the authors of the paper to show the paper's core content. Author keywords are imperative, since they are used as the topics/concepts/methods that are presented to deliver and communicate to the scientific community by the authors. The author keyword co-occurrences network demonstrates another perspective of themes in PSEP, and it can be observed that it illustrates the main author keywords that frequently occur together in PSEP.
Considering the fact that many keywords only appeared a few times, they obviously have not had significant influences on the main themes of PSEP. Therefore, in the present study, to focus on the main themes, only the keywords occurring at least five times were selected to construct the co-occurrence analysis map and indicate the research topics. Thus, 407 keywords were extracted based on the threshold of the keyword's frequencies.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. ) mainly concentrated on accident prediction and hazard assessment methods and models (especially fire and explosion mitigation). This cluster was characterized by keywords such as "computational fluid dynamics" (49), "modeling" (46), "explosion" (43), "numerical simulation" (37), "dust explosion" (34), "consequence analysis" (23), "gas explosion" (19), "hydrogen" (19), etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. ) mainly focused on the process safety and risk assessment in chemical and process industries. This cluster was represented by keywords such as "risk assessment" (73), "process safety" (43), "Bayesian network" (34), "inherent safety" (34), "risk analysis" (23), "human factors" (22), "risk management" (18), "hazard identification" (17), "process design" (17), etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. ) mainly focused on safety and risk management strategy and cost-benefit analysis. This cluster was composed of keywords such as "safety" (56), "risk" (40), "J-value" (19), "health" (14), "reliability" (13), "flash point" (11), "radiation" (11), "nuclear" (10), etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period. The green group ( ) mainly concentrated on accident prediction and hazard assessment methods and models (especially fire and explosion mitigation). This cluster was characterized by keywords such as "computational fluid dynamics" (49), "modeling" (46), "explosion" (43), "numerical simulation" (37), "dust explosion" (34), "consequence analysis" (23), "gas explosion" (19), "hydrogen" (19), etc.
Additionally, to further understand and analyze the research topics, evolutionary process, and research frontiers for PSEP, the chronological evolution of research keywords based on the average publication date for the publications in which the keywords appeared in PSEP is demonstrated in Figure 9. Note that the VOSviewer software automatically settled the most suitable time interval and span for 407 keywords based on the publication time (i.e., average publication year) of each keyword. The main research topics in various periods could be recognized and summarized based on the occurrence frequency of the keywords in each period. Table 7 summarizes the top 20 keywords that appeared most frequently in each sub-period.   Overall, the PSEP was mainly focused on the topics related to the fields of safety of industrial processes and protection of the environment. In the present study, to explore and extract the evolutionary process and trends of the main research topics in PSEP, the special topics of publications in the separated four periods are summarized and analyzed.

Conclusions
This study presents a bibliometric journal analysis perspective to explore the evolution in safety or environmental engineering design and practice, as well as experimental or theoretical innovative research. Various bibliometric analyses, scientometric mapping, and statistical techniques are utilized to identify and unearth the evolution trends and detailed characteristics of the publications as indexed by the Web of Science. To validate the application of our purpose, the principal and influential international journal PSEP is taken as the case study journal to explore the influencing factors behind the rapid development of journals and reveal the main research trends of the safety of industrial processes and the protection of the environment.
The main findings for our case study journal could be summarized as follows: (a) until the statistical time of this paper, PSEP has published 3152 literature productions and drawn 44,879 citations, and the corresponding number keeps increasing over time. (b) Faisal Khan is the most productive author and the only author with more than 50 papers in PSEP, followed by Paul Amyotte and P. J. Thomas , and the document type of review paper is more likely to get more citations. However, 61.04% publications in PSEP have no more than ten citations, and there are still a certain number of publications that are not cited. Furthermore, there are 43 highly cited references (most of them are journal articles) that can be regarded as the core intellectual bases of PSEP. These works concern adsorption theory, process safety and risk management methods and models, inherent safety, loss prevention, domino effect, waste treatment, and coal mine safety. The related reviews for the highly cited references could deliver further detailed insights in the explored domain for scholars. (f) Several keywords/topics, such as "adsorption", "kinetics", "response surface methodology", "risk assessment", "optimization", "heavy metals", and "safety", are the most popular hotspots of PSEP. Additionally, the popular keywords in various periods reveal the chronological evolutionary process and trends of the emerging or long-lasting research hotspots of PSEP during the last 30 years.
The topics of the selected publications from the case study journal mainly concentrate on the research fields of protection of the environment and safety of industrial processes. For specific research areas, the mainstream research areas of our case study journal, including the waste and pollutants remediation, environmental protection methodologies and technologies, waste management and sustainable development, accident prediction and hazard assessment methods and models, process safety and risk assessment, and safety and risk management strategy and cost-benefit analysis, etc., are recognized. In addition, some new emerging topics in the research community could be highlighted and recommended for scholars and stakeholders. In respect to environmental protection, some emerging topics are recognized, such as catalytic pyrolysis of waste, sustainability of the production of fuels, analysis of kinetics and adsorption mechanisms, activated carbon adsorption, supercritical water processes, climate change mitigation technology (especially catalytic conversions of greenhouse gas), catalytic degradation of organics (especially utilizing microbial community), application of microwave for waste treatment, safety of drinking water and its treatment, mineralization technology for pollutant emission control, etc. In view of industrial process safety, the mainly focused topics are failure of complex systems, mine fires and coal spontaneous combustion prevention, inherent safety and cost evaluation, risk analysis and prevention considering synergistic effects and domino effects, optimization for hazardous materials transportation, J-value analysis and assessment of accidents, safety culture and process safety education, human and organizational factors of safety, etc. As for Industry 4.0 (especially smart manufacturing and production, big data, Internet of Things), accompanied with artificial intelligence (e.g., the application of artificial intelligence for wastewater treatment and deep learning-based forecast modeling for safety/risk-related topics), are widely used as the hotspots for the research fields of PSEP. In terms of techniques/methods, heterogeneous catalysis, photocatalysis degradation, advanced oxidation process, biochar for adsorption, machine learning (especially deep learning), response surface methodology, numerical simulation, HAZOP, bowtie, dynamic Bayesian network, dynamic and computer-aided fault tree analysis, hybrid artificial neural network, genetic algorithm approach, etc., are widely used and applied advanced emerging techniques, as well as traditional methods, among global scholars for solving the environment protection, safety, and risk analysis problems. The aforementioned new future trends contribute to offering safer and cleaner living and production environments for human beings, which will further promote the harmonious coexistence of man and nature and achieve social and economic development sustainability.
This paper provides a comprehensive and quantitative overview and significant picture representation for the journal's leading and evolutionary trends by employing specific bibliometric key impact factors, such as the document types, publication distribution and citation structure, most cited works, the trends of research topics, and prominent contributing authors, countries (regions), and institutions, etc. Additionally, by reviewing the evolution trends of the journal and the proposed investigated factors, such as the influential works, main research topics, and the research frontiers, this paper delivers various research objectives and directions that could be addressed and explored in future studies for related scholars worldwide, as well as for related journal editors, to position their journal to align with the focus topic of the safety of industrial processes and the protection of the environment. However, note that the results obtained from the present paper are dynamic, and may change over time along with the emergence of new research hotspots or mainstream subjects and some specific variables increasing/decreasing in the study.
In the follow-up research, some other data sources, e.g., Scopus, Google Scholar, and EconLit, etc., and some other document types, e.g., books, proceedings books, PhD theses, etc., could be contained as a supplement for scholars applying our proposed methodology for exploring the evolutionary trends for a particular knowledge domain. Additionally, some comparative analysis and inductive research for different journals in a specific research area could be conducted in further study. Nevertheless, some expert knowledge should be considered when selecting the most relevant journals and articles, especially for a very wide study field that may lead to a huge workload of a subsequent analysis and an overloading and time-consuming phenomena for the analysis tools.