Energy system models: a review of concepts and recent advances using bibliometrics

ABSTRACT The present bibliometric review highlights the qualitative and quantitative study on Energy System Models by scanning the Scopus database between the years 2000 and 2023. Utilising co-citation analysis and bibliographic coupling from 1288 documents, the VOSviewer analysis presents meaningfully contributing authors, author keywords, nations, and cited references. Of 7930 keywords reviewed, the top keywords as per total link strength were ‘energy system model’, ‘energy systems’, ‘carbon dioxide’, and ‘energy policy’. From 2014 till 2020, research areas like energy systems, energy storage, and energy transitions were studied. Results from citations of sources, organisations, authors, and countries reveal that out of 2491 organisations, 189 have a minimum of two documents for analysis. European nations including the United Kingdom, Ireland and Germany lead in research on energy systems and power demand modelling. Future research will likely focus on ‘electric vehicles’, ‘long-term energy planning scenarios’, ‘use of alternative fuels’, ‘power demand modelling’ and ‘energy security’.


Introduction
Energy demand is growing in all the world countries (Wolfram, Shelef, and Gertler 2012).Many authors propound that various significant factors, such as the increasing use of energy in residential and industrial sector, and electric vehicles, have led to this increasing energy demand (Mairet and Decellas 2009).However, it is essential to plan for the required energy by mitigating energy generation scenarios (Sugiyama et al. 2021).Cornelius, Van de Putte, and Romani (2005) mention that planning and managing energy has become increasingly important as nations formulate social, economic and environmental policies for long-term scenarios.Numerous cross-disciplinary connections between the economy, environment and society occur due to the vital role that energy systems play in the socioeconomic systems of societies.Cajot et al. (2017) assert that the only way to address the numerous interconnected energy-related concerns is through long-term planning and collective effort by various stakeholders.Energy modelling or power demand modelling can help with long-term energy planning.
Future energy systems are examined using energy system models that are frequently used for addressing issues with climate change policy (Van Beeck 2000).Regarding the type, design, application, programming, amount of detail, scope, sophistication and flaws, the models themselves differ critically (Sovacool, Axsen, and Sorrell 2018).Numerous models use some mathematical optimisation to guide the solution procedure.Generating computer-based models of energy systems to study them is known as energy modelling or energy system modelling (Hilpert et al. 2018).These models often use scenario analysis to examine several hypotheses on the concerned technological and economic conditions.The system under examination's energy efficiency, use of natural resources, total cost of ownership, and viability may all be included as outputs.Many different strategies are used, from economic, environmental, and social to engineering in nature.The scope of a model may be worldwide, provincial, nationwide, local, or stand-alone.To develop energy policies, governments work on national energy models, whereas states work on regional energy models.
Although the world over, continuous studies are being carried out in the energy policies through energy systems modelling to meet energy requirements for varied sectoral needs, incorporation of Variable Renewable Energy Sources (VRES) in the energy systems planning for estimation of power demand is required for future energy security (Connolly et al. 2010;Gielen et al. 2019;Lund et al. 2015).In this light, European countries have made a considerable contribution to the research and development of energy system models, as is seen from the literature available in the energy systems modelling domain.However, many countries worldwide have ambitious targets of energy transition from fossil fuels to renewable resources for their national energy plans.
This paper highlights the key components and present dynamics of studies done in the energy systems modelling realm while outlining potential future research options.We pinpoint the publication trends and conceptual framework in this field using bibliometric analysis based on the Scopus database search.Bibliometric techniques have their roots in research using many bibliographic materials in the library and information sciences.Bibliometric studies categorise and analyse bibliographic content by developing illustrative synopses of the existing literature.Previous studies have used this method to examine journals, universities, nations, research areas, authors, and co-authorship, among other things.In their work, Merigó and Yang (2017) mention that bibliometrics can be called a statistical approach that uses mathematical techniques to quantitatively analyse research papers concerned with a particular topic.Basis their shared sources and referencing patterns, scientific works display intellectual convergence, according to Kessler (1963).On the other hand, Small (1973) argued that recurrent citation of two or more than two references in a third research paper shows that the citing and cited papers have similar concepts or ideas.Co-authorship and co-occurrence are other prevalent thoughts that are present in the bibliometrics.The co-occurrence of keywords illustrates the theoretical building of the body of literature, whereas co-authorship indicates the manner of authorship and association among the cooperating writers (Cheng et al. 2018;Koseoglu et al. 2016;Peters and Van Raan 1991).Additionally, it may evaluate the major research fields, assess the quality of the conducted studies, and forecast the course of future research (Bonilla, Merigó, and Torres-Abad 2015).Nearly all significant research papers are included in the Scopus online database, which also offers built-in tools for analysis to provide representative figures.
The present bibliometric analysis is likely to help understand the global positioning of countries in the work of energy system models.A thorough bibliographic evaluation has yet to be conducted in the literature for energy system models using VOS viewer as an analysis tool, making it challenging to assess how broadly dispersed or up-to-date the research efforts are.A graphic description and representation of the bibliometric analysis for energy systems modelling through word clouds have yet to be done.This review aims to identify deficits in geographic coverage by documenting the historical and geographic distribution of research in energy systems modelling as a noteworthy field of research globally.Geographic regions have been analysed to conclude the dominance of Europe in the energy systems modelling field.

Objectives of the study
The chief objective of this article is to offer the current state of knowledge on energy system models, with the subsequent inquiries defining its purview: Research Question 1: Concerning time, disciplines, journals, affiliated nations, authors and institutions, what are the contemporary trends in publishing in the arena of energy system models at urban, regional and national levels?
Research Question 2: What influential studies and research themes exist in this field?
The essay is formatted further: The data search and analysis techniques are outlined in Section 3, and findings on publication patterns are covered in Section 4. A keyword analysis is offered in Section 5. Section 6 covers the Citation analysis.Bibliographic coupling and Co-Citation analysis are covered in Section 7. Section 8 describes the evolution of clusters for the keywords and cited references and discusses the content of the significant papers cluster-wise.It outlines the theoretical underpinnings in the area of energy systems modelling.Section 9 has a discussion on the topic with future study directions, and Section 10 concludes the research paper.

Techniques, data search and analysis
We offer a bibliometric overview of the literature on energy system models through network and descriptive studies.We consider total publications, citations, and citations for each publication in our descriptive analysis.Our network investigations also look at co-citations, co-occurrences, and bibliographic couplings in addition to descriptive analysis.Most mapping analyses in this study are performed using the VOSviewer programme.To graphically represent the nodal network, VOSviewer employs two defined weights: the quantity and overall strength of the links.The big or small size that relates to the nodes and the connecting lines between the nodes indicate the importance and force of the links.The relationship between items can be deduced from their proximity on a map provided by VOSviewer.The items are closely related when their distance is smaller (Van Eck and Waltman 2010).Thus, built on the 'visualisation of similarities' (VOS), VOSviewer helps in citation, co-citation, and keyword analysis.The analysis technique used in the paper is presented in the following Figure 1.
An extensive screening process produced the body of papers used in the bibliometric investigation.The Scopus database, the biggest multidisciplinary record of peer-reviewed research works, was accessed for the bibliographic information utilised in this work (Bartol et al. 2014).For quantitative analysis, Scopus is well known and regularly used (DuránSánchez et al. 2019).The papers were searched for in the title, abstract or keywords using a string of acceptable search terms ('power demand model*') OR ('energy system model*').To obtain all texts containing words with various root forms, the search query included an asterisk (*).The search was done for the years between 2000 till 06 January 2023, returning 2269 documents.The documents were limited to English language usage scaling them down to 2204.The search was then focused on selected subject areas, as shown in Table 1.As energy modelling is closely related to engineering, energy, environmental science, and management while being multidisciplinary, the subject areas have been thoughtfully chosen, as shown in Table 1.After carefully choosing the subject areas, the searches returned 1288 documents.The final search string is shown in Table 2.
It was observed that many documents were based on building energy simulations, heating, cooling and thermal comfort in buildings.While bibliometric analysis has been widely used in all subject areas, its use in the energy field is also broad.Masoumzadeh, Rasekhi, and Fathi (2016) created a complete model that analysed the techno-economic factors of an office building's energy supply.Zhang, Ling, and Lin (2023b) gave a complete analysis of risk management in East Asia from 1998 to 2021 utilising bibliometric methods.Using bibliometric analysis and data mining, Ghoshchi et al. (2022) studied machine learning methods for building energy modelling.Their study aimed to provide insight into the current state of machine learning applications for energy systems in the building and construction industry.The interplay of energy-related fields with Machine Learning, as well as how Artificial Intelligence can be applied in the energy field, was explored by Entezari et al. (2023), through a bibliographic review of Artificial Intelligence, Machine Learning, and the most commonly used energy-related terminology.Luo and Lin (2021) conducted a bibliometric analysis on Flash Translation Layer as a significant part of computing and cybernetics.According to Zhang, Ling, and Lin (2022), artificial intelligence-related technologies may successfully tackle difficulties linked to integrating renewable energy with power systems, 'such as solar and wind forecasting, power system frequency analysis and control, and transient stability assessment'.Maghzian,   2022) used bibliometric analysis and data mining to investigate the capture of CO 2 by greenhouse from combined cycle power plants.Zhang, Ling, and Lin (2023a) studied 633 publications on carbon neutrality to provide a complete review of this aspect through bibliometric analysis.For the present research, we aim to study the research for energy systems modelling in more comprehensive urban, regional and national settings.Thus, the documents were further filtered to omit the results associated with building energy simulations, heating and cooling in buildings, thermal comfort and air-conditioning in built structures and Data mining, Machine Learning and Artificial Intelligence domain.The results were shortlisted to 1288 documents that were then taken up for the bibliometric analysis.The date of retrieval of data was 6 January 2023.The co-occurrence, co-authorship, bibliographic coupling, citation, co-citation, and themes were examined using VOSviewer (version 1.6.18).Selected results from the Scopus built-in tool for analysis were also used along with the VOSviewer results for analysis.

Trends and analysis of publication patterns
The collected data was run through the VOSviewer software in Excel format for analysis of research publications by type, year of publication, source of the documents, keywords, country of origin, author, co-authorship, affiliation, co-citation, and bibliographic coupling.This was done to address Research Question 1.The select papers were analysed for their content to appreciate the current research advances in energy systems modelling.Discussion on trends in recent research, themes of the said research fields and content analysis of the clusters was done to perform the gap analysis in the field of research under consideration.This was done to address Research Question 2. The details of the bibliometric examination are presented in the succeeding sub-sections.

Analysis of the publication documents by type
1288 results were recognised from the Scopus online database from 2000 till 6 January 2023.874 articles contribute to 68% of the literature, followed by 240 conference papers, that are 19%.However, review, book chapter, data paper, book, conference review, note, editorial and erratum contribute less than 10% of the documents, as seen in Table 3.Thus, research articles are being published more in this field as against review or book chapters.

Analysis of the annual publication trends
The trend in energy system models research has been constantly growing since the year 2000, albeit slowly till 2018.Post-2018, the research in this field is seen increasing, but not at an ambitious pace, as there is a drop in the publications for 2019.The documents in 2018 were 116, compared to 107 in 2019.This trend has spiked since 2020, perhaps the key reason being the mandate of the United Nations' Sustainable Development Goals and the various climate policy changes worldwide.In this regard, the research was significant in 2020, 2021 and 2022, with 128, 156 and 175 publications, respectively, in each year.The details are seen in Figure 2.

Analysis of the documents by source, affiliation to organisations, countries and authors
The selected keyword search yielded 93 results for the source of the publications in the Scopus database, across which the selected 1288 articles have been spread.The 10 prominent sources out of the 93 yielded that have more than 20 publications are presented in Table 4. Applied Energy (183 publications) and Energy Policy (110 publications) have published the most in the field of energy system models, followed by Renewable And Sustainable Energy Reviews (88 publications), Energy Strategy Reviews (66 publications), Energy (65 publications), Journal of Cleaner Production (37 publications), Energy Procedia (34 publications), Renewable Energy (33 publications), Energy Conversion and Management (31 publications) and Energy Economics (21 publications) are a few prominent sources that have publications in the area of energy systems modelling since 2000 till 6 January 2023.The rest of the sources have less than 20 publications each.
160 organisations have publications on the subject of energy system models.However, only 55 organisations have 10 or more than 10 publications.The top 10 organisations have published 452 documents.The details of these top 10 organisations and their publication numbers are presented in Table 4.It is interesting to observe that all the top 10 organisations belong to European Countries, suggesting that the European countries are doing considerable research in energy systems modelling.
The analysis conducted for documents from different countries shows that 275 documents come from Germany, followed by the United Kingdom at 179 and the United States at 151.It is interesting to note that China and Japan are the only countries among Asian countries to make a name in the top 10 countries with 75 and 60 documents, respectively.This indicates that the European nations are leading in energy systems modelling.The details of the top 10 countries with the highest documents on energy systems modelling as an area of study are seen in Table 4.
26 authors have more than 10 publications in energy systems modelling, with author Gargiulo Maurizio with 20 documents having the highest number of publications.Alessandro Chiodi follows the list by 16 documents.Steve Pye has published 16 articles, whereas Christian Breyer and Diego García-Gusano have published 15 articles each.Diego Iribarren, Filip Johnsson and Neil Strachan have each contributed 14 articles.Table 4 details the number of documents per author.

Co-occurrence analysis for keywords
The authors' keywords denote the themes of research publications keywords (Comerio and Strozzi 2019).A keyword analysis was done with the use of VOSviewer to discover the most popular topics in energy systems modelling.The final analysis included keywords that were submitted by the authors and appeared more than 10 times in the Scopus database.A total of 7930 keywords were acknowledged in 1288 papers, of which 377 meet the threshold.Therefore, these 377 keywords that appeared at least 5 times or more are analysed.VOSviewer analysis showed that out of the 377 keywords, the most appearing keyword was 'energy system model' with a total link strength of 3134 and 294 occurrences, followed by 'energy systems' with a total link strength of 2725 and 240 occurrences.This shows that the 'energy system model' is not the only significant keyword  in the literature.It is closely followed by 'energy systems' and can be used interchangeably.The keyword 'carbon dioxide' has a link strength of 2614 and shows only 195 occurrences.Similarly, 'energy policy' as a keyword has a total link strength of 2511 but only 197 occurrences.It, therefore, means that even though the occurrence as a keyword is less for 'energy policy' and 'carbon dioxide', their relation with the literature is strong.The top 10 keywords that prominently appear in the analysis are presented in Table 5.
Additionally, a word cloud displayed the recurrence of the keywords that appeared more than 10 times.It should be noted that 'investments' and 'optimisation' are the other keywords that appear prominently.The word cloud created using VOSviewer is seen in Figure 3(a).The overlay visualisation of keywords, as shown in Figure 3(b), indicates that major research in 2014-2016 is associated with carbon dioxide, passenger transport, developing world, technology, and cost-benefit analysis among a few.In 2016-2018, research areas focused on energy system models, renewable energy resources, climate change, emission controls, and greenhouse gases, among a few others.From 2018 to 2020, the research areas revolved around energy systems models, energy transitions, variable renewable energy resources, decision making, investments, energy storage, hydrogen, modelling approach, sector coupling, integrated energy systems and open software, among many others.The trend of research is, thus, seen changing since 2014.Thus, emerging themes in this domain are seen with distinct new keywords.

Citation network analysis
Finding the publications that have had the most impact on the 'energy system models' is the goal of our second research question.We scrutinised the citation networks of 1288 papers to address Research Question 2. A document's citation count depicts how many citations it has accumulated.A document receiving more citations is supposed to be more influential and valuable than those receiving less.The most actual way to gauge the influence of a research document is by citation analysis (Ding and Cronin 2011;Tsay 2009).We can create intellectual connections using citations and references (Appio, Cesaroni, and Di Minin 2014).The amount of citations an article receives from other publications is a critical factor in determining its influence on citation analysis.

Citation of documents
The least number of citations per document was set at 5. With this constraint, 807 documents were selected.When VOSviewer ran the analysis, some of the 807 documents were disconnected.The 572 connected documents were further considered for the citation analysis.The top 20 most cited articles are shown in Table 6. Figure 4 shows a VOSviewer word cloud in this context.

Citation of sources, organisations, authors, and countries on the basis of total link strength and citations
The citation of sources was analysed for 276 sources.The minimum number of documents per source was considered at a threshold of 5, and the lowest number of citations of a source was fixed at 3. It is observed that 'Applied Energy' as a source has the greatest number of documents at 183 with 7424 citations.This is followed by 'Renewable and Sustainable Energy Reviews' having 4521 citations but only 88 documents.Although the journal 'Energy Policy' has 110 documents, the citations are 4452.No other source has documents that are more than 100.However, 'Energy Strategy Reviews' and 'Energy' are journals that have more than 50 Publications each.Considering the topmost 20 cited articles as per Table 7, it is detected that 5 publications are in 'Renewable and Sustainable Energy Reviews', 4 articles have been published in 'Applied Energy', and 3 articles in 'Energy Policy' and 'Energy' each.Considering the total link strength, 'Applied Energy' has the maximum total link strength of 673.'Renewable and Sustainable Energy Reviews' with a link strength of 537, 'Energy Policy', 'Energy Strategy Reviews' and 'Energy' with link strengths of 431, 352, 273, respectively follow closely.
The VOSviewer word cloud for the citation of sources is seen in Figure 5.As 'Applied Energy' and 'Renewable and Sustainable Energy Reviews' have very close scores, the word cloud of 'Renewable and Sustainable Energy Reviews' is overshadowed by that of 'Applied Energy'.The scale represents the average citation score for each of the journals.
When analysing the citation of authors, the lowest number of papers and citations of an author was fixed for a threshold of 2. Out of the 3081 authors, 789 met the threshold, of which 675 were considered for analysis.The analysis shows that author Lund H. with      Gusano D. and Chiodi A. have 15 documents each.However, their citation scores vary from 718, 645, and 484 to 402, respectively.A VOSviewer word cloud representing the citation of authors is revealed in Figure 6.Considering the total link strength, Strachan N., with 14 documents and a citation score of 1275 is the top author with a link strength of 801 followed by Gargiulo M., Lund H., Pye S., and Pfenninger S. as the top authors with link strengths of 699, 589, 569 and 522 respectively.A few other prominent names are Østergaard P.A., Deane J.P., Howells M., Hawkes A., Decarolis J., and Nijs W., who have contributed to energy systems modelling.
2491 organisations were identified, out of which 234 meet the threshold criteria of having minimum 2 numbers of documents, and considering the connectivity of the items, 189 items were selected for future analysis.
Table 7 shows that the United Kingdom and Ireland are the prominent names working in the energy transition and low carbon research with energy policies and energy modelling as a focus area.It is interesting to observe that except Columbia University, New York, United States, the top 20 list of organisations includes all the European Universities.Figure 7 shows the visualisation of citation scores based on the organisations where the organisations from European countries dominate the landscape of energy systems modelling.This strengthens the idea that European countries are doing prominent and prestigious research in energy systems and power demand modelling studies.It was observed that 82 countries were listed for the citation of documents.58 countries have at least 2 documents published.55 countries were connected, and VOSviewer analysis was run on these 55 countries for the citation of documents by countries, as seen in Table 7.
The total links strength of the documents related to the United Kingdom is 1434 as seen in Table 7, which shows the citations of the top 20 countries.The documents are 179, and the citations are 6085.275 documents from Germany have a total link strength of 1235, and citations are 5942.Besides, 151 documents from the United States have a total link strength of 918 and 3843 citations.The link strength of the United Kingdom is high, and the citation score is also high for these

Analysis of the documents by bibliographic coupling and co-citation method
Whenever two papers, cite the same third article, it is known as bibliographic coupling (Kessler 1963).It is widely used to gauge the document similarity (Small 1973).Small (1973) also mentions that co-citation and bibliographic coupling are frequently misunderstood or incorrectly associated.Co-citation happens when two papers are cited jointly by a shared third article (Garfield 1988).While co-citation examines the cited papers, bibliographic coupling examines the citing documents (Small 1973).It should be observed that as time passes, more papers get bibliographically associated with a given document.One of the reasons bibliographic coupling is helpful for document clustering is due to this dynamic feature of it.When two documents quote the same source, this is known as 'bibliographic coupling', which concerns those two texts.For example, Figure 9 demonstrates that there exists bibliographic coupling between publications A and B if both of them cite document C. The arrow denotes a straight citation from the current text (arrow end) to the earlier document that is being referenced (arrowhead).
In VOSviewer, a researcher is represented by each circle.Large circles represent numerous publications by researchers.Small circles signify authors of few publications or researchers.The closer two scholars are to each other in the visualisation, the more closely they are frequently related to one another, according to bibliographic coupling.In other words, researchers closely located to one another often cite the same papers, whereas researchers who are far apart rarely do so.The use of colour is done to identify groups of researchers closely related to one another.These groups are called clusters.7.1.Bibliographic coupling of documents, sources, authors, organisations and countries Papers were analysed for bibliographic coupling.177 documents with minimum citations of 50 were selected out of 1288 papers; some of which were not connected in the network.The most extensive set of connected items was 166, and the analysis was run on them.Figure 10 The unit of analysis for bibliographic coupling is sources.Of the 276 sources, 40 had at least 5 publications each.The VOSviewer analysis shows that most authors have cited publications from 'Applied Energy', 'Energy Policy', 'Energy' and 'Renewable and Sustainable Energy Reviews', indicating that these sources are closely linked and have solid bibliographic coupling.The word cloud for the same is shown in Figure 11.
3081 authors were identified for the bibliographic coupling of authors.Authors with the minimum number of 5 documents were selected, thereby filtering the authors to 144 numbers.143 authors were linked to each other.The VOSviewer analysis was run on the 143 selected authors.There exists a strong coupling between the authors Lund H., Mathiesen B.V., Thellufsen J.Z., and Østergaard P.A. whereas, Prina M.G., Sparber W., Manzolini G., and Breyer C. are closely linked.Gargiulo M., Pye S., Strachan N., Ó Gallachóir B.P., and Deane J.P. are coupled.Figure 12 shows the VOSviewer density visualisation for the bibliographic coupling of authors.2491 organisations were identified for analysing the bibliographic coupling of organisations.The lowest number of documents for an organisation was fixed at 5, and 18 met this threshold.The VOSviewer word cloud shown in Figure 13

Analysis by co-citation for cited references, cited sources, and cited authors
When a third publication mentions two other publications, those two publications are said to be cocited (Small 1973).The relationship between any two publications is robust when more publications mention both of them.To evaluate and depict relationships between publications, Griffith et al. (1974) recommended using co-citations.Thus, when two papers are frequently cited in combination, this is known as co-citation.Two papers are more comparable in terms of the entire research area the more often they are co-cited (Culnan 1987).
Too recent documents fail to sufficiently impact the study (Pilkington and Fitzgerald 2006).We chose a co-citation criterion of 15 papers for our research area to focus on the most authoritative publications.Pilkington and Fitzgerald (2006) recommend a co-citation criterion of 15 papers over  a 10-year period.Similarly, Özmen Uysal ( 2007) applied a 20-year threshold of 10 co-citations.Based on these parameters, co-citation analysis was performed.60,740 cited references were identified, and the lowest number of cited references of a document was secured to 10, shortening the list of cited references to 37. Table 8 details 8 presents the details of the top 10 co-citations for cited sources.51,246 cited authors were identified, with 1054 having a minimum of 20 cited authors for a document.Figure 16 represents the word cloud for the most co-cited authors, and the details are seen in Table 8.Lund H. and Mathiesen B.V. remain the top 2 cocited authors, followed by Breyer C., Østergaard P.A. and Duic N., among others.

Mapping the trends in energy research for the last ten years by cluster analysis method
Clustering is a VOSviewer method where the network nodes are allocated to clusters.A powerfully tied-together collection of nodes is called a cluster.One cluster is specified for each node in a network.VOSviewer employs colours to denote the cluster where a node has been assigned when displaying a bibliometric network.A cluster analysis of keywords was done to understand the underlying theme in more connected keywords.For this purpose, the keywords were clustered into 5 clusters derived from the VOSviewer analysis.The top 10 keywords of each cluster are presented in Table 9.
Energy system model, energy systems, carbon dioxide and renewable energy resources were the keywords with the highest link strength, with energy system model as a keyword having the most occurrences.Modelling, as a keyword, did not exhibit a rising trend in usage with only 89      occurrences, inconsistent with the general rise in the number of research articles published in the subject.Due to multidisciplinary research trends, it may have been replaced with more prevalent keywords such as optimisation, investments, costs and numerical models.A closer study also demonstrated the significance of the energy systems modelling for a futuristic system that is sustainable and low in carbon.Therefore, keywords like renewable energy resources, greenhouse gases, carbon capture, carbon dioxide, and emission controls have found a place in the keyword search.Our analysis of the keywords clustering depicts that at the beginning of the past decade, research focused on fossil fuels, energy consumption, greenhouse gases and energy efficiency.Although the topics continue to be discussed and debated, more relevant topics with in-depth studies have emerged, which are associated with energy systems, modelling, optimisation, renewable energy resources, optimisation, investment and, more recently, hydrogen.Decarbonisation, electricity, and carbon emissions have an excellent total link strength but low occurrences.Keywords such as optimisation, alternative energy, and renewable energies are multidisciplinary and can be related to the subject areas of computer science, mathematics, engineering, and sustainability, among a few.Therefore, their occurrences, when narrowed to a single field, might have given restricted results, but they can be assumed to have a much broader premise.
After co-citation analysis, a complete content analysis of the 37 publications resulting from the co-citation analysis for cited references was done.All the 37 publications were organised into five clusters, and a common motif was found within each cluster after rigorous examination.The Table 10.Clustering of Articles, their year of publication and their presentation based on total link strength and co-citation of cited references.

Cluster 1
Cluster 2 Cluster 3 Cluster 4 Cluster 5 Brown, T., Schlachtberger, D., Kies, A., Schramm, S., Greiner, M., (2018) representation of the documents in clusters is seen in Table 10.Table 11 outlines the theme of each cluster and mentions major topics that authors under each theme have explored.
The underlying theme of Cluster 1 is integrating renewable energy systems in several energy system models concerning spatial and temporal resolutions for optimisation.Jebaraj and Iniyan (2006) examined and presented many models for energy supply-demand, energy planning, renewable energy, forecasting, optimisation, and emission reduction.Connolly et al. (2010) have reviewed 37 tools for their application with regard to single building to national energy systems, while Hall and Buckley (2016) discovered that there are references to about 100 models in academic writing, out of which 22 models were categorised using the schema after considering the United Kingdom's model market.Mathiesen et al. (2015) discussed integrating intelligent gas, thermal, and electric grids to enable 100% renewable energy while considering to suffice for transport solutions.Comparing a wide range of economically advantageous future power solutions for Great Britain, Pfenninger and Keirstead (2015) suggested that with only a slight cost increase, variable renewable energy capacity of up to 60% is feasible.Østergaard (2015) examined how the EnergyPLAN model was applied geographically and the several simulations and scenario evaluations that were run on the model, examined the several performance indicators used in the energy system simulations and proposed new advanced energy system performance indicators.Poncelet et al. (2016) mentioned that for large proportions of renewable energy sources, the degree of time-based precision has the most significant influence and that the temporal representation can be enhanced by establishing a separate time slice level for renewable energy sources.Connolly, Lund, and Mathiesen (2016) analysed the shift to 100% renewable energy in 2050 from a no-change (business-as-usual) scenario with the use of the Smart Energy System technique in Europe, while Gils et al. (2017) introduced and applied the Renewable Energy Mix (REMix) energy system model that allows for the evaluation of capacity growth and hourly dispatch at varying degrees of wind and solar power integration.Brown et al. (2018) mention that increasing cross-border transmission of energy lowers system costs.However, the value of transmission reinforcement diminishes, and the more rigidly the energy sectors are intertwined, when considering integration of renewables.
Cluster 2 discusses the broad area of energy storage and an energy system's optimal design.Nahmmacher et al. (2016) offer and test a brand-new, computationally effective time slice approach that can easily offer input information for several power system models.Staffell and Pfenninger (2016) analysed current and future wind energy for European nations, while Pfenninger and Staffell (2016) analysed Europe's long-term PV output trends to conclude that net power demand changes significantly as PV deployment increases.Pfenninger (2017b) mentioned that improved planning and modelling techniques are required to handle inter-year variability with high renewable sources.Kotzur et al. (2018) observed that aggregation level effects are very system-specific and cannot be generalised for an optimal energy system design.Ringkjøb, Haugan, and Solbrekke (2018) reviewed 75 energy models and presented an updated summary of the modelling tools and their capabilities to help modellers identify and select an acceptable Power sector, variable renewable sources and their impact on power sector, energy transition pathways, decarbonisation pathways 5 Trends in energy system modelling Power system, energy system, demand response and flexibility, challenges and trends in energy systems modelling model.Gabrielli et al. (2019) studied the impact of data input uncertainty on costs, emissions, and multi-energy systems' robustness.In contrast, in another study design, the requirement for seasonal energy storage were defined by extensive sensitivity analysis.(Gabrielli et al. 2018).Cluster 3 is themed around urban energy systems and energy system models.Allegrini et al. (2015) give a thorough analysis of the modelling techniques and related software resources for district-level energy systems.Best practices are discussed by DeCarolis et al. (2017) for energy system optimisation modelling.Hansen, Breyer, and Lund (2019) mention that holistic cross-sector analysis is becoming the norm and that there will be a need in the future to connect the local and global levels.Keirstead, Jennings, and Sivakumar (2012) look at ways to enhance the way that urban energy systems modelling is currently done, concentrating on the possibilities of cloud computing and sensitivity analysis, data collecting and integration methods, and the usage of activity-based modelling to arrive at an organising framework.According to the findings, there appears to be a massive opportunity for energy systems modelling in the urban energy domain to advance past traditional boundaries of disciplinary studies and toward a knowledgeable, integrated viewpoint that better reflects the theoretic complexity of urban energy systems.Mancarella (2014) gives the reader a thorough and critical outline of the most recent models and evaluation methods currently available to study multi-energy systems and specifically distributed multi-generation systems, including, for example, ideas like microgrids, virtual power plants and energy hubs.Pfenninger, Hawkes, and Keirstead (2014) look at four issues they deal with and the steps to solve them.They mention that time and space must be reconciled, uncertainty and transparency must be balanced, the energy system's increasing complexity must be addressed, and social risks and possibilities must be integrated.Pfenninger et al. (2017) have stressed the significance of open-source data and software availability for energy systems modelling.
Cluster 4 broadly studies power systems, energy models and energy transition.Deane et al. (2012) provide a 'soft-linking methodology' that uses thorough simulation results from a specific power systems model to obtain awareness of the outcomes of the portfolio of electricity generation plants for the power sector from a different energy systems model.For both long-term models of the power sector and energy system models, a novel modelling typology founded on a study of the literature is offered by Després et al. (2015).It is suggested to use new comparison criteria that concentrate on the grid and electricity storage, which are two essential parts of the power sector.Trutnevyte (2016) examines the justification for using cost optimisation and poses concerns about whether cost-optimal models are reliable representations of the energy transition in the UK, wherein ex-post modelling demonstrates that cost optimisation does not closely resemble the actual UK electrical system transformation.Pye, Sabio, and Strachan (2015) describe a global sensitivity analysis-related approach to uncertainty analysis, adding that uncertainties significantly impact the price and viability of necessary mitigation and, therefore, the process between analysts and policymakers must be iterative.Usher and Strachan (2012) demonstrate the intricate relationship between the UK's long-life infrastructure and generation technologies and the elasticity of the energy system in terms of reducing the costs associated with uncertainty.
Cluster 5 underlines the trends in energy system modelling.Collins et al. (2017) present modern approaches to overcoming the difficulties of integrating fluctuating renewable energy sources.An evaluation of the possibility for theoretical Demand Response in Europe is given where geographical distribution and temporal availability of flexible loads are given particular consideration (Gils 2014).Lopion et al. (2018) analysed historical patterns in the growth of energy system models mentioning that more flexibility for geographic or time resolution exists in recently developed models.They also observed that Python has emerged as the most common programming language in the simulation of energy systems.Lund et al. (2015) analyse various methods, tools and tactics for managing massive renewable energy projects that produce variable electricity, such as solar and wind power while taking both supply and demand side actions.
It is seen from Table 10 that major research in the area of energy systems modelling has been done in the past decade from 2012 to 2019, for which considerable citations can be seen.Recent research publications need more citations and therefore are not seen in the co-cited references.However, as the cited references are from 2012 to 2019, recent articles are putting in considerable research efforts on energy systems modelling.

Discussion and future trends
This study will help academic scholars, policy-makers, and regulators by educating them on the fundamentals of energy systems and power demand modelling and pointing out areas that warrant further research.From a decade ago, studies have considerably increased, and the area is now fast developing.This review is an early attempt to identify the knowledgeable links among the significant works of the last 22 years, trace back the convergent tendencies in the area of energy systems modelling and identify future research directions.
As evident from the Scopus database, the study in this field is done with due diligence in the European nations.Germany is the leading country researching the area of energy system models.Most research papers are published in Germany, making it a global focus of energy-related research.The vast differences in policy details between nations make it difficult for the authors to work together.Most of the literature on the topic concentrates on the collaboration of individual authors.A positive note would be when the ideas are exchanged and interchanged within and outside the affiliations and in the open domain by the collaborating authors and their affiliations.If the most critical scientific issues are only discussed inside specific affiliations, it may reduce the exposure of scientific debate to the general public.
While there already exist bibliometric assessments that have considered some of the studies done in this research, this document thoroughly reviews the prevailing peer-reviewed publications with an emphasis on energy systems modelling.Numerous other bibliometric evaluations place most of the highly prolific journals in the present article, including Energy, Applied Energy, Energy Policy, Energies, and Renewable Energy, in the top positions.The result is evident in the present study and as mentioned in the analysis of Weinand (2020), who also claims that 37% of the retrieved papers have been published in the topmost journals (Energy, Applied Energy, Energy Policy, Energies, Renewable Energy) and backed by Dominković et al. (2022).
However, despite their significance in energy systems modelling, extremely well-known papers have yet to appear on the list of works that have been mentioned the most.For instance, the energy systems field has come under fire for its lack of modelling transparency (DeCarolis et al. 2017;Strachan, Fais, and Daly 2016), and there are growing demands to make study hypotheses, information sources, as well as applied models more visible and available to the public as mentioned by Abramo, D'Angelo, and Di Costa 2009 and asserted by Pfenninger (2017).More articles that discuss other allied fields are by Bhattacharyya and Timilsina (2010), who observe that the energy system models are concentrated on projecting long-term energy scenarios for various nations and spatial spread and reducing carbon emissions and endorsing sustainability as has been observed by Liu (2014).Jebaraj and Iniyan (2006) have seconded sustainability as one of the most discussed topics associated with the energy system models.Malla and Timilsina (2019) observed that despite the predicted drop in the country's population, Romania's demand for electricity is expected to rise due to rising household income and an enlarged services industry, which is rather electricity heavy.They, therefore, came up with an end-use energy model for Romania.The studies have often propounded the use of renewable energy resources in power generation systems whilst also mentioning their inevitable role in mitigating climate change.Mohammed, Mustafa, and Bashir (2014) have observed the same in their work.Discussions about how the integration of variable renewable energy resources brings instability to the power systems have often been observed (Medina, Ana, and González 2022) by many authors (Banjar-Nahor et al. 2018;Pourarshad et al. 2021;Zaman 2018).There have been suggestions for using smart grids and microgrids in related articles to overcome this issue (Worighi et al. 2019).
There are no citations for the most recent research on the subject.However, it is clear from a thorough content analysis that inventive research on energy systems, energy models, power demand, and the function of governments, organisations, and energy modellers can further pave the way for future advancement (Moustakas et al. 2020).Finally, we have noted a few restrictions on this research.Despite our best efforts to guarantee that the search phrases used accurately reflected the field, a few studies may have been overlooked because the search parameters needed to include any terms that were closely connected to them.For example, the indication of the research area's inclination to place more emphasis on the study of renewable energy systems is the fact that no terms that could be linked to fossil fuels made it to the 10 most popular keywords.However, it does reflect in keyword clustering.
Results demonstrating regional collaboration more prominently within the European countries were evident from the findings.However, that was not anticipated, and they point to the necessity for more aggressive efforts to increase it, especially on a global scale (Bodin 2017).The advantages of wider and more intense collaboration are obvious in the energy systems modelling sector, where study frequently combines various disciplines including engineering, environmental science, and economics.Additionally, as has been welcomed by the 'open' modelling community in current years, the creation and use of many energy system models provide a compelling case for closer cooperation.Several projects and platforms are needed for model and data sharing in a collaborative workspace (Connolly et al. 2010).More intense international collaboration should result if curiosity in and use of the resources continues to soar.However, thus far, there are limitations to this collaboration's worldwide component.These barriers could be related to a lack of funding options and insufficient networks, among other issues (Cholewa, Mammadov, and Nowaczek 2022).
Another observation is the supremacy of the North European (Nordic) nations in the outcomes.This could also be connected to the energy practices and ideologies in these nations, which have established themselves as leaders in sustainable energy.The inference is that a significant portion of the energy models and energy systems modelling literature contains robust features of wasteto-energy, district heating, and wind energy, all of which are important parts of the Nordic energy system (Persson 2015).However, regarding the application, this can be a drawback for areas with drastically diverse climates.There is no generalised way to work on energy system models or sustainable energy systems; thus, research must be customised to local conditions.A monopoly from a few authors, organisations, or nations only sometimes ensures this as geographical considerations must be taken for customising energy modelling studies.Considerations of social sciences and humanities as allied area needs to be done for energy modelling studies.Socio-technical and socio-commercial considerations drive any nation's energy transition and therefore form an essential ingredient for studies related to national policies or financial and political decision-making.Also, decision sciences need to be considered as an allied study area.Besides, as we might have to deal with cross-border energy trade or alliances for energy security and sustainability, international relations and political sciences (Falkner 2014) should be aligned to the study of energy systems modelling.
The authors have identified a few research areas that willing researchers can further study.Investigations can be done to study the benefits of Artificial Intelligence and Machine Learning to discover new opportunities (Ghoshchi et al. 2022;Khare et al. 2021).Other topics that Entezari et al. (2023) have mentioned include energy storage, uncertainty analysis, wastewater treatment, pollution control, manufacturing of biofuel, supply chain in the energy domain, renewable energies, assessment of risks, and demand response as well as heat storage, environment management, thermal energy, carbon capture techniques, economic and investment, renewable energies and internet of things wherein detailed studies and analyses can be done.Chen, Lin, and Zhuang (2022) have mentioned similar views in their study on wastewater treatment.Zhang, Ling, and Lin (2023a) have mentioned that it is vital to decrease carbon emissions by fully utilising renewable energy and increasing the utilising carbon sinks and carbon collection, utilisation, and storage; and therefore, studying carbon neutrality can be a potential research area which is similarly mentioned by Uche, Okere, and Das (2023).As uncertainty and risk assessment are significant areas of research in the energy field, the study conducted by Zhang, Ling, and Lin (2023b) feels apt in giving future directions for research in this area.Zhang, Ling, and Lin (2022) observe that AI-related technologies are mainly used to forecast short-term solar radiation and wind speed and rarely analyse the impact of power electronics on frequency response while failing to investigate the geographical and timedependent characteristics of frequency response.Offsetting this lacuna can be a potential research zone.Zhong and Lin (2022) studied the economic aspect globally and felt that interventions in policy formulations, development of the economy and financial markets need to be studied.Therefore, studying the economic and financial markets is essential for the energy field as well.Nielsen and Karlsson (2007) had very early propounded that energy scenario studies should be enhanced by focusing more on outlining short-term policies and activities that could lead to future states while considering market conditions, lifestyle, and societal developments.Collins et al. (2017) suggest integrating short-term variations of the power system into integrated energy system models to increase the penetration of renewable energy into the energy system.A multi-disciplinary approach towards an integrated system perspective in energy systems can significantly better the modelling of energy systems (Keirstead, Jennings, and Sivakumar 2012).Hybrid renewable energy systems should be optimised using artificial intelligence and computing intelligence (Ajiboye et al. 2023).Relationships between modellers, empirical researchers, policymakers, and diverse academic disciplines must be created to enhance model development and theory maturation (Fodstad et al. 2022).Behavioural aspects should be considered while planning energy systems and therefore, the models also should have integration across disciplines (Huckebrink and Bertsch 2021).Wiese et al. (2018) suggest to bring more focus on model generators and frameworks can be developed for the energy system models.
A significant portion of the articles that made it past the Scopus screening were reviews.Therefore, a few review papers were included in this analysis's total of 1288 articles in this analysis.This degree of analysis, which combines bibliometric and manual analysis, is still reasonably high.While broad keywords, for instance, were possible to study, it needed to be more practical to go into the specifics of the many models and approaches.However, the study makes up for its need for more depth with its breadth.

Conclusion
With the study in the energy sector picking up in the last decade more than its previous one, primary research is seen in the energy sector.Climate change, GHG emissions and renewable energy resources have caught researchers' attention in the energy sector.Globally, all nations are working towards achieving energy security and equilibrium in energy supply and demand.Finding useful information and assessing the calibre of such a large quantity of research articles is especially critical.
The research on collaboration networks also revealed a substantial convergence of published papers among nations and institutions.More of the leading institutions are concentrated in a select few nations.The author and institution collaboration networks reveal that they primarily work together in the same university and nation.Additionally, there is a close relationship between the best journals, writers, and organisations.
The method's benefit in a broader scope is equally its drawback given its detail.The main topics and publication patterns in energy systems modelling during the previous few decades may be recognised.However, exploring particular models and approaches intending to perform a more in-depth study of the research's findings was impossible.Such bibliometric study methods, which tend to place more emphasis on the results than their content, have this as an inherent weakness.
The terms that find relevance as a growing trend of research are electric vehicles, long-term energy planning scenarios, use of alternative fuels, 100% incorporation of Variable Renewable Energy Resources (VRES) in the power systems, power demand modelling and energy security.These are the areas in which the future scope of research can be envisioned.
The study, nevertheless, has many field-related implications.It gives a complete summary of the research output over the past two decades and helps concerned readers understand the significant movements in the industry.It also offers a strong basis for involved researchers to broaden their links with collaborators to engage with an added and varied range of individuals, thereby improving the applicability of results.The study's findings can also help academics fresh to the subject better understand global trends and important authors.

Figure 1 .
Figure1.Process of analysis used for the study.Source: Author's compilation.

Figure 2 .
Figure 2. Publication of documents by year.Source: Author's compilation from the Scopus analysis built in tool.

Figure 4 .
Figure 4. Overlay Visualisation of Citations based on score of cited articles per year and total link strength.

Figure 5 .
Figure 5. Overlay Visualisation of Citations by Sources based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.

Figure 6 .
Figure 6.Overlay Visualisation of Citations of Authors based on citation scores and total link strength.Source: Compiled by the authors using VOSviewer.

Figure 7 .
Figure 7. Overlay Visualisation of Citations Score of Organisations based on total link strength and citation scores.Source: Authors' contribution using VOSviewer.

Figure 8 .
Figure 8. Overlay Visualisation of Citations Score of Documents by Countries based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.
(a) shows the word cloud for the same.Figure 10(b) shows a scaled-up version of the same.The book 'Solar Energy Engineering Processes and Systems' (2009) by Soteris A. Kalogirou has the highest total link strength at 449 with a citation score of 479, followed by the research article 'Formalizing Best Practice for Energy System Optimization Modelling' by DeCarolis et al. (2017) with a total link strength of 275 and citation score of 171.This is followed by the article 'Trends in Tools and Approaches for Modelling the Energy Transition' by Chang et al. (2021), with total link strength of 239 and a citation score of 79.The highest citation score of 1167 with a total link strength 105 is seen for the article 'A Review of Computer Tools for Analysing the Integration of Renewable Energy into Various Energy Systems' by Connolly et al. (2010).This is followed by the article 'Electric Vehicles and Smart Grid Interaction: A Review on Vehicle to Grid and Renewable Energy Sources Integration' by Mwasilu et al. (2014) with a citation score of 637 and total link strength of 15 and then by the article 'Energy Systems Modeling for Twenty-First Century Energy Challenges' by Pfenninger, Hawkes, and Keirstead (2014) with a citation score of 586 and total link strength 232.

Figure 10 .
Figure 10.(a) Bibliographic coupling of Documents based on total link strength and citation scores.(b) Scaled up figure showing Bibliographic coupling of Documents based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.
depicts the details of linkages and couplings.The results show that the School of Engineering, Energy Policy and Modelling Group, MaREI Centre and Environmental Research Institute at University College Cork, Ireland, are very closely coupled as organisations.Institute for Renewable Energy, Italy; Technical University of Denmark; RWTH Aachen University and Europa-Universität Flensburg from Germany; Electric Power Research Institute and National Renewable Energy Laboratory from USA; Systems Analysis Unit, IMDEA Energy from Spain and Lappeenranta University of Technology from Finland are closely coupled.Other closely grouped organisations are UCL Energy Institute, UK, the Centre for Renewable Energy Sources and Saving from Greece and the Institute of Energy, Environment and Economy from China.82 countries were identified for the bibliographic coupling for analysing the strong linkages between countries.37 countries have published at least 5 documents each.The United Kingdom, Germany, Italy, Greece, Portugal and Ireland are closely coupled.Belgium, Croatia, the Netherlands and India are closely linked.Denmark, China, Norway, Finland and Estonia form a link.France, Japan, the USA, New Zealand and South Korea are closely connected.The VOS viewer word cloud shown in Figure 14 depicts these linkages and couplings.

Figure 11 .
Figure 11.Bibliographic coupling of Sources based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.

Figure 12 .
Figure 12.Bibliographic coupling of Authors based on total link strength.Source: Compiled by the authors using VOSviewer.

Figure 13 .
Figure 13.Bibliographic coupling of Organisations based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.
the top 10 cited references through co-citation analysis.'Energy Systems Modeling for Twenty-First Century Energy Challenges' by Pfenninger, Hawkes, and Keirstead (2014) is the top co-cited reference with 45 citations, followed by 'Dealing with Multiple Decades of Hourly Wind and PV Time Series in Energy Models: A Comparison of Methods to Reduce Time Resolution and the Planning Implications of Inter-Annual Variability' by Pfenninger (2017a ) at 22 citations.15,853 cited sources were identified, and considering a minimum of 20 cited sources of a document, 204 were selected.The VOSviewer word cloud, as seen in Figure 15 reinforces that the sources such as 'Energy', 'Applied Energy', 'Renewable and Sustainable Energy Reviews' and 'Energy Policy' are co-cited.Table

Figure 14 .
Figure 14.Bibliographic coupling of Countries based on total link strength and citation scores.Source: Compiled by the authors using VOSviewer.
Link strength, C = Citations.(Source: Compiled by the authors using VOSviewer).

Figure 15 .
Figure 15.Co-citation analysis for cited sources based on citation scores.Source: Compiled by the authors using VOSviewer.

Figure 16 .
Figure 16.Co-citation analysis for cited authors based on citation scores.Source: Authors' Contribution using VOSviewer.

Table 1 .
Subject areas chosen for the document search.

Table 2 .
Chen, Lin, and Zhuang (2022)a result of search queries in the Scopus literature database.Aslani, andZahedi (2022)used data-mining analysis based on the bibliometric method to find advances in research, trends, and existing gaps to look at microalgae as a carbon-capturing technology to provide insight into the present state of 'microalgae research and development as a Renewable Energy resource with carbon-capturing' potential.Chen, Lin, and Zhuang (2022)used bibliometric analysis to present a complete assessment of wastewater treatment and emerging pollutants research from 1998 to 2021.Zahedi et al. (

Table 3 .
Publication of documents by type.Compiled by the authors from the Scopus analysis built in tool.

Table 4 .
Analysis of the number of documents by source, affiliation, country and author.Compiled by the authors from the Scopus analysis built in tool.

Table 5 .
Topmost keywords of energy systems modelling built on the occurrences and link strength.
8 documents, has the highest citation score of 1903.Mathiesen B.V., with 6 documents has a citation score of 1793.Connolly D. has just 2 documents, yet the citation score is 1691.Strachan N., with 14 documents, has a citation score of 1275, and Keirstead J., with 4 documents, has a citation score of 1145.It is seen that Gargiulo M. has 19 documents, but the citation score is 603.It is seen that Pye S., Breyer C., García-

Table 6 .
Most often cited articles among the scientific publications on Energy System Models retrieved from VOS viewer.

Table 7 .
Top 20 sources authors, organisations and countries on the basis of total link strength and citations.
Source: Compiled by the authors using VOSviewer.

Table 8 .
Co-Citation analysis for top 10 cited authors, cited sources and cited references , cited sources and cited authors.

Table 9 .
Clustering of Keywords(K) and representing top 10 keywords from each cluster on the basis of Total Link Strength (TLS) and Occurrences (O).

Table 11 .
Cluster themes and major topics explored in each cluster.