We conducted a search on the Web of Science core collection with the following subject line: (“web-based energy management system” OR “web-based energy information systems” OR “web-based energy monitoring system” OR “web-based energy scheduling”) and a publication time range of (2003-01-01 to 2022-12-31). The search retrieved a total of 1451 papers. To identify the main topics and subtopics discussed in the literature, we conducted a content analysis of the retrieved papers. Based on this analysis, we categorized the topics into two groups: “core knowledge domains” and “peripheral knowledge domains.” The “core knowledge domains” refer to the main topics that were most frequently discussed in the literature, while the “peripheral knowledge domains” refer to the subtopics that were discussed less frequently.

To measure the impact of new papers or scholars on changes in the structure of existing citation networks, we employed the Structural Variation Analysis (SVA) model proposed by Prof. Chaomei Chen in 2012. The SVA model uses three main measures to quantify the degree of structural transformation: the rate of change of patternedness (ΔM), the rate of change of links between clusters (ΔCLw), and the degree of centrality dispersion (ΔCkl).

The rate of pattern change (ΔM) measures the increase in network links among cited documents that may appear in the same or different academic groups. The rate of linkage change between clusters (ΔCLw) focuses on the impact of linkage between different clusters before and after the introduction of new literature into the network. Centrality dispersion (ΔCkl) measures the changes in the network structure caused by new literature before and after considering the introduction of new literature. The equations for these measures are as follows.

(1) The rate of pattern change (△M) mainly considers the increase of knowledge base network links among the cited documents, and these increased links may appear in the same academic group or different academic groups. The more significant the increase of network links caused by the cited documents, the larger the value of the pattern change rate index, and the more significant the potential impact of the corresponding new documents on the whole network.

The basic computational procedure of the structural transformation model is as follows: assume that the co-citation network G is divided into k clusters by partition C such that G = c ₁ + c ₂ + … + c _k . First define Q (G,C) as follows: Q G , C = 1 2 m ∑ i j = 0 n δ c i , c j ⋅ A i j − deg ⁡ ⁡ n i ⋅ deg ⁡ ⁡ n j 2 m (1)

Where m is the total number of edges in the network G and n is the number of nodes in G. Known as Kronecker’s delta, it is 1 if nodes n _i and n _j belong to the same cluster and 0 otherwise. In addition, deg (n _i ) is the degree of node n _i .

Network patterns are a measure of the overall network structure, which ranges between −1 and 1. The rate of change of patterns of scientific papers is mainly referred to the relative structural change of the network baseline of published papers. For each article (scholar)a and a base network G _baseline , the rate of change in patterns is defined as follows. M C R ⁡ a = Q G b a s e l i n e , C − Q G b a s e l i n e ⊕ G a , C Q G b a s e l i n e , C ⋅ 100 (2) where G _baseline and G _a are network baselines updated based on the information in article a. For example, assuming that reference nodes n _i and n _i are not connected to the network baseline of co-cited references but are co-cited by article a, new links between n _i and n _i will be added to the network baseline such that the article changes the structure of the network base, and the remaining two metrics inter-cluster link change rate and centrality dispersion are measured similarly.

In mapping the knowledge evolution of the research area “Web-based Energy Management Information System,” this study divides the data from 2003 to 2022 into two phases (Figure 1). The first stage is from 2003 to 2012 (Figure 1A), and the second is from 2013 to 2022 (Figure 1B). The core knowledge areas are distributed in “1. MATHEMATICS, SYSTEMS, MATHEMATICAL” and “3. ECOLOGY, EARTH, MARINE” in 2003–2012. The core knowledge base is “1. SYSTEM, COMPUTING, COMPUTER,” “2. ENVIRONMENT, TOXICOLOGY, NUTRITION,” and “10. PLANT, ECOLOGY, ZOOLOGY.” The marginal knowledge bases are “5. HEALTH, NURSING, MEDICINE,” “7. PSYCHOLOGY, EDUCATION, SOCIAL”, “8. MOLECULAR, BIOLOGY, GENETICS,” “12. ECONOMICS, ECONOMIC, POLITICAL”. In 2013-2022, the core knowledge areas were distributed in “1. MATHEMATICS, SYSTEMS, MATHEMATICAL,” “3. ECOLOGY, EARTH, MARINE” and “7. VETERINARY, ANIMAL, SCIENCE.” The core knowledge base is “1. SYSTEM, COMPUTING, COMPUTER,” “2. ENVIRONMENT, TOXICOLOGY, NUTRITION,” and “10. PLANT, ECOLOGY, ZOOLOGY.” The marginal knowledge bases are “5. HEALTH, NURSING, MEDICINE,” “7. PSYCHOLOGY, EDUCATION, SOCIAL,” “8. MOLECULAR, BIOLOGY, GENETICS,” “9. SPORTS, REHABILITATION, SPORT,” “12. ECONOMICS, ECONOMIC, POLITICAL.”

The keywords are often a high summary of the research theme of literature, so the keyword analysis of the literature can clearly grasp the research theme of the literature and then grasp the research hotspots and frontiers of the literature. The basic principle is to count the co-occurrence of keywords or noun phrases in the literature, reflecting the strength of association between different keywords or noun phrases, and represent the research hotspots and research frontiers in the subject area according to the frequency of co-occurrence of these keywords or noun phrases. In order to reveal the research hotspots in the field of “Web-based energy management information system,” this study used CiteSpace software to analyze the keyword co-occurrences in the dataset from 2003 to 2022. The parameters were set as: time 2003-2022, Node Types = Keyword, and the threshold value was chosen as g index k = 15, and the network density was 0.0231. It was found that the cryptocurrency research domain had 375 nodes with 1621 connections in the keyword co-occurrence network from 2003 to 2022 (Figure 2). The top 20 keywords with the highest frequency of co-occurrence were listed as keyword information (Table 1). Based on this, the top 20 keywords with the highest frequency of co-occurrence were listed as keyword information (Table 1).

According to the TimZone view of the keyword co-occurrence network from 2003 to 2022 (Figure 2), it can be found that the number of new keywords in the research field of “Web-based energy management information system” increases gradually every year, indicating that the scholars’ attention to this field cannot be increased continuously. The system’s frequency is 133 with a centrality of 0.09, the frequency of management is 112 with a centrality of 0.2, and the frequency of the model is 86 with a centrality of 0.13. The frequency of the Internet of Things is 77, with a centrality of 0.04. The frequency of the keyword energy efficiency is 68; the centrality is 0.1. The frequency of the keyword energy is 67; the centrality is 0.09. The frequency of the keyword cloud computing is 53; the centrality is 0.03. The frequency of the keyword design is 52; the centrality is 0.04. The frequency of the keyword design is 52; the centrality is 0.04. The frequency of the internet is 49 times with a centrality of 0.02. The frequency of performance is 45 times with a centrality of 0.1. The frequency of energy consumption is 40 times with a centrality of 0.02 (Table 1).

From 2003 to 2022, there were six hot topics in the research area of “Web-based energy management information system” (Figure 3). The main keywords in the food web hotspot are management, food web, impact, climate change, and energy efficiency. The main keywords in food web hot topics are management, food web, impact, climate change, eco path, community, temperature, biodiversity, dynamics, and energy flow. Internet of things hot topics are the main keywords in the hot topics of the internet of things, cloud computing, internet, energy management, technology, wireless sensor network, and challenge. Photovoltaic system hot topics are model, performance, energy consumption, optimize. The main keywords in the hot topics of the photovoltaic system are model, performance, energy consumption, optimization, framework, consumption, efficiency, simulation, and neural network. Energy efficiency hot topics are energy efficiency, smart grid, web service, web, power, architecture, information system, and demand side management. The keywords in the semantic web hot topics are system, energy, renewable energy, decision support system, semantic web, and decision making. Circular economy hot topics are design, smart city, generation, life cycle assessment, management system, and management system: life cycle assessment, management system, bibliometric analysis, and circular economy.

Burst word analysis is a method of word relevance analysis using keyword co-occurrence data. Its basic idea is to discover the correlation between keywords by analyzing the number of co-occurrence between keywords, to uncover some potential hot knowledge. Specifically, burst word analysis is to find the correlation between keywords by comparing the number of co-occurrences between keywords, to discover the correlation between keywords, and dig out some potential hot knowledge. In this study, Burst analysis in CiteSpace software is used to explore the correlation between different keywords in the research field of “Web-based energy management information system” and the impact of emergent keywords on the research field. This study helps scholars to understand more deeply the development hotspots in the research field, adjust the research direction in time, and seize development opportunities.

From 2003 to 2022, we found a total of 27 Burst keywords in the research field of “Web-based Energy Management Information System,” which can be divided into three phases in general (Table 2). The first phase is from 2003 to 2009; the hotspots are information systems and dynamics. The second phase is from 2010 to 2017. The hotspots are the smart grid, food web, wireless sensor network, energy efficiency, eco path, eco slim, and energy management information system. The third phase is from 2018 to 2022. The research hotspots are the internet of things (IoT), internet, smart building, algorithm, framework, generation, real-time systems, and dynamics. Generation, real-time system, simulation, management system, bibliometric analysis.

The given passage presents a study that predicts potentially influential literature on “Web-based energy management information system” based on pattern change rates (△M), inter-cluster linkage change rates (△CLw), and centrality dispersion (△Ckl) indices. The study finds nine papers with a pattern change rate greater than 0, indicating a change in the research trend or direction. Out of these nine papers, five have an inter-cluster linkage change rate greater than 0, indicating a change in the relationship between clusters of research. However, there are no papers with a centrality dispersion index greater than 0, suggesting no significant change in the importance of authors’ roles in the network.

The passage lists the top five papers with high pattern change rates and inter-cluster link change rates, which are considered to be noteworthy. These papers are “A framework for integrating BIM and IoT through open standards,” “Designing Low-Res Lighting Displays as Ambient Gateways to Smart Devices,” “IoT big data and HPC based smart flood management framework,” “Energy Harvesting towards Self-Powered IoT Devices,” and “Toward Sustainable Energy-Independent Buildings Using Internet of Things.”

Furthermore, Table 3 provides the model change rate of the top 20 English literature, including the nine papers mentioned earlier. The table lists the pattern change rate (△M), inter-cluster linkage changes rate (△CLw), centrality dispersion (△Ckl), title, author, and publication date of each paper. The table shows that all the papers in the top 20 have a pattern change rate of 100, indicating a significant change in the research trend or direction. Additionally, five papers have an inter-cluster linkage change rate greater than 0, indicating a change in the relationship between clusters of research. However, there are no papers with a centrality dispersion index greater than 0, which is consistent with the findings presented in the passage.