Next Article in Journal
Modeling of Spray Combustion and Heat Transfer of MMH/N2O4 in a Small Rocket Engine Using Different Mechanisms
Previous Article in Journal
Flow Characteristics of Oil-Carrying by Water in Downward-Inclined and Horizontal Mobile Pipeline
Previous Article in Special Issue
Developing Expert Systems for Improving Energy Efficiency in Manufacturing: A Case Study on Parts Cleaning
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

A Systematic Review of Expert Systems for Improving Energy Efficiency in the Manufacturing Industry

by
Borys Ioshchikhes
*,
Michael Frank
and
Matthias Weigold
Institute for Production Management, Technology and Machine Tools (PTW), Technical University of Darmstadt, Otto-Berndt-Str. 2, 64287 Darmstadt, Germany
*
Author to whom correspondence should be addressed.
Energies 2024, 17(19), 4780; https://doi.org/10.3390/en17194780
Submission received: 23 August 2024 / Revised: 15 September 2024 / Accepted: 20 September 2024 / Published: 24 September 2024
(This article belongs to the Special Issue Energy-Efficient Manufacturing System Management)

Abstract

:
Against the backdrop of the European Union’s commitment to achieve climate neutrality by 2050, efforts to improve energy efficiency are being intensified. The manufacturing industry is a key focal point of these endeavors due to its high final electrical energy demand while simultaneously facing a growing shortage of skilled workers crucial for meeting established goals. Expert systems (ESs) offer the chance to overcome this challenge by automatically identifying potential energy efficiency improvements, thereby playing a significant role in reducing electricity consumption. This paper systematically reviews state-of-the-art ES approaches aimed at improving energy efficiency in industry with a focus on manufacturing. The literature search yields 1668 results, of which 62 articles published between 1987 and 2024 are analyzed in depth. These publications are classified according to the system boundary, manufacturing type, application perspective, application purpose, ES type, and industry. Furthermore, we examine the structure, implementation, utilization, and development of ESs in this context. Through this analysis, this review reveals research gaps, pointing toward promising topics for future research.

1. Introduction

Climate change is one of the most imperative topics in modern times. The European Climate Law addresses this threat by setting a net greenhouse gas emissions reduction target of at least −55% by 2030, compared to 1990 levels [1]. Energy use plays a critical role in the pursuit of this objective, as it was the source for almost three-quarters of global emissions in 2016. The industry sector is responsible for about 30% of emissions, with energy use accounting for 24.2%, making it the main source of emissions [2]. To achieve decarbonization, efforts are being made to increase the share of electricity in all sectors, and its share of final energy consumption is projected to rise from 20% today to over 50% by 2050. Since the industry sector was already the largest consumer of final electric energy in 2019, representing 41.9%, there is an imperative urgency to take action [3].
In response to this challenge, companies are prioritizing energy efficiency improvements as a way to simultaneously achieve affordability, supply security, and climate goals [4]. Energy efficiency is a measure that quantifies the utilization of energy in relation to the output or yield of services, goods, commodities, or energy [5]. Increasing energy efficiency therefore means making better use of existing resources. According to a 2022 global survey conducted among over 2200 industrial companies in 13 countries, 97% of them had either already invested or were planning to invest in energy efficiency. Moreover, 89% of these companies anticipated increasing their energy efficiency investments over the next five years, while 52% had the ambitious goal of achieving net zero within the same timeframe [4]. However, despite the increase in energy efficiency in the manufacturing sector in recent years, there is still considerable potential for further improvement [6]. Energy submetering of individual industrial processes is especially low, showing a deficiency of monitoring and targeting systems that are crucial for effective energy management and energy efficiency efforts. The survey also found that a lack of specialized contractors, a lack of digital skills in the workforce, and uncertainty about how to improve energy efficiency are significant barriers to investment in energy efficiency improvements [4].
The conclusion of a study undertaken by the International Energy Agency suggests that this situation could be improved with digitalization [7]. Expert systems (ESs) are one type of the advanced computer technology has have emerged from research in the field of artificial intelligence (AI) and are designed specifically to assist in decision making [8]. They pose a chance to overcome mentioned barriers by combining expert knowledge and analyzing energy data to automatically identify energy efficiency potentials [9]. In general, ESs, also referred to as knowledge-based systems or inference-based programs, can be described as computer programs that leverage expertise to solve problems and provide advice [8]. Unlike conventional applications, these systems emulate human reasoning by representing human knowledge and employ heuristic or approximate methods to solve problems [10]. Within organizational knowledge management, ESs can thereby serve as knowledge carriers and knowledge mediators, thus contributing to the qualification of personnel and reducing the shortage of skilled workers [11]. This further distinguishes ESs from the inherent complexity and limited explainability of other AI applications for decision support, such as machine learning and deep learning models [12]. The intelligent activity that characterizes ESs is the use of knowledge for processing and not just information. Information exists by itself without context, whereas knowledge has the added dimension that something is performed to process the information [8]. Energy data from a machine would be information. Reading measurement data, determining which energy efficiency potentials are present, and drawing conclusions on how to exploit existing energy efficiency potentials is the use of knowledge about energy data. This knowledge must be provided by human experts to develop the ES.
There are several literature reviews covering ESs in general and ESs or AI to optimize industrial production. Ref. [13] presented an overview of the development of ESs based on a literature review and a classification of articles from 1995 to 2004 into eleven methods. The respective applications of ESs for the associated methods are also given. This publication thus provides an overview of the various applications and techniques of ESs. However, ref. [13] did not address the use of ESs to increase energy efficiency and did not analyze any publications in this area. Ref. [14] examined literature from 1983 to 2015 dealing with decision support models for energy-efficient production planning. For this purpose, procedures for production systems, energy price laws, and energy efficiency criteria are considered. Although this literature analysis focused on improving energy efficiency in production, it did not include ESs to achieve this objective. Similar to refs. [13,15] provided a broad overview of ESs and their applications from 1984 to 2016. In comparison to refs. [13,15] introduced an indicator to estimate and compare the success of ES applications. Among many other applications of ESs, this publication also dealt with those in production. Nevertheless, applications for increasing energy efficiency were not discussed. Ref. [16] focused their literature review on applications of ESs in production planning regarding the handling of different products, process planning, tool selection, welding, and product development between 1981 and 2016 but without considering energy efficiency.
Unlike the previously cited reviews, this work focuses on ESs to overcome environmental and economic challenges by improving energy efficiency in manufacturing. Production describes the combination of goods and services to create new goods, while manufacturing is understood as a sub-area of production that describes the actual process of creating goods [17]. Other industrial sectors are also considered to ensure that the scope of the study is not overly restricted and to allow possible conclusions to be drawn about manufacturing. This is performed through a systematic literature review (SLR) with a comprehensive overview. Starting from a total of 1668 publications, the 62 most relevant publications are extracted and categorized according to system boundary, manufacturing type, application perspective, application purpose, ES type, and industry.
The remainder of this paper is organized as follows: Section 2 describes the methodological approach of the SLR, which is then applied in Section 3. This involves all steps, from narrowing down the topic to identifying research gaps. Finally, Section 4 draws conclusions and presents proposals for future research.

2. Review Methodology

To identify relevant publications on ESs for improving energy efficiency in manufacturing, an SLR is conducted. An SLR is a formal approach that aims to reduce bias due to overly restrictive selection of the available literature and to ensure a rigorous process [18]. Moreover, an SLR can clarify the state of research on a topic and highlight gaps and areas requiring further research [19]. The SLR in this paper follows the approach of [20] as well as [21], which is complemented with elements of [18] and consists of seven consecutive steps shown in Figure 1. This aligns with the criteria of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [22].
First, the topic is narrowed and conceptualized. When narrowing the topic (I), categories and their definitions are formulated, as consistent classifications help to ensure reliability when conducting content analyzes. In the conceptualization (II), research questions for the review itself as well as inclusion and exclusion criteria are formulated. Moreover, the keywords and the search query string are defined. Subsequently, databases suitable for the topic area are selected and searched (III). For this purpose, databases of previous literature analyses covering the same topic area or suggestions of university libraries for databases of individual research fields can be examined. For the literature search, we only consider metadata (title, abstract, keywords). Thereby it is prevented that the search terms are contained only in the bibliography of the found publications. After merging hits from different databases, literature is filtered (IV) by applying the defined inclusion and exclusion criteria. In the next step, the full texts of the remaining publications are thoroughly analyzed (V). This involves extracting and summarizing relevant information from the publications. By performing backward and forward snowballing, additional relevant publications are identified. Backward snowballing means using the reference list of selected literature to identify new publications to include. Forward snowballing, on the other hand, is carried out to find publications that cite the selected publication. The added publications pass through steps (III) to (V) again. During the synthesis (VI) of the literature, the essential characteristics are categorized. Finally, the findings are summarized, conclusions are derived with regard to the research questions, and research gaps are identified (VII).

3. Literature Review

3.1. Narrowing the Topic

To narrow the topic, relevant categories and associated dimensions are identified as shown in Figure 2. The classification scheme is based on [25] for the field of energy modeling in the manufacturing industry. In addition to the narrowing of the topic, the classification supports defining keywords for the search query string and the literature analysis.

3.1.1. System Boundary

In the context of energy-related ES applications in industry, five dimensions are distinguished regarding the system boundary.
  • Factory: Distinct physical entity containing multiple devices [26].
  • Manufacturing cell/line: Logical organization of multiple machines to achieve a better division of labor [27].
  • Machine: Entity required to perform a specific production task [28].
  • Component: Individual parts or consumers of a machine which represent the lowest hierarchical level for energy metering [29].
  • Process: Value-adding and non-value-adding technical operations [27].

3.1.2. Manufacturing Type

Manufacturing systems can be characterized by their manufacturing type, which is determined by the frequency of repetition of a product or product group to be manufactured [27].
  • Job-shop manufacturing: Custom manufacturing, i.e., according to customer requirements, in which products are only manufactured once.
  • Repetitive manufacturing: Products are manufactured at irregular intervals. If orders are repeated, less preparation is required.
  • Variant manufacturing: Similar products of the same basic type, which generally involve similar manufacturing effort.
  • Serial manufacturing: Mostly contract manufacturing of standardized products in limited quantities.
  • Mass manufacturing: Manufacturing large quantities for an anonymous market. High initial investment costs, but low in relation to the sum of manufactured products.

3.1.3. Application Perspective

Adapted from [25], the application perspective subdivides the phases in which an ES is useful.
  • Engineering: ESs are used for energy-optimized design at a high level of abstraction, e.g., by supporting the selection of sustainable technologies.
  • Process planning: The objective of applying ESs in this phase is to plan and optimize manufacturing processes with regard to energy efficiency prior to actual operation, e.g., by optimizing parameters.
  • Operation: In the phase in which the actual manufacturing process takes place, ESs are utilized to improve the energy efficiency of the operation, e.g., by detecting inefficient operating points.

3.1.4. Application Focus

Two dimensions can be distinguished on which studies in the field of energy optimization focus.
  • Energy efficiency: Refers to the “relationship between the results achieved and the resources used, where resources are limited to energy” [30].
  • Energy flexibility: Describes the “ability of a production system to adapt quickly and in a process-efficient way to changes on the energy market” [31].

3.1.5. Expert System Type

ESs include a user interface, a knowledge base, and an inference engine. The interactions between humans and ESs happen via the user interface, e.g., to define a problem (input) or to read out the results (output). The knowledge base is the programmed knowledge of the expert and contains all pertinent facts and relationships as well as the rules or heuristics to solve problems. The inference engine is the “intelligent” part of an ES which can apply the knowledge from the knowledge base to solve the problem [8].
The way in which knowledge is represented within the knowledge base and how it is processed within the inference engine differs depending on the ES type. The typification of ESs in this paper is based on the categories in [13].
  • Rule-based expert system: A rule-based ES represents information in the form of IF–THEN rules. These rules are applied to perform operations on data in order to reach a conclusion [13].
  • Fuzzy expert system: Fuzzy ESs are characterized by dealing with uncertainties using fuzzy logic, while rule-based ESs only allow conditions or conclusions that are either true or false, fuzzy ESs also allow conditions or conclusions that are partly true or false. This approach is based on the premise that human experts often decide without precisely quantified information [13].
  • Machine learning (ML)-based expert system: This type of ES uses ML as its “intelligent” component to solve problems. Like ESs, ML belongs to the domain of AI and combines a collection of data-driven algorithms that can learn from data without being explicitly programmed. ML also includes deep learning and reinforcement learning [32].
  • Hybrid expert system: Hybrid ESs are a combination of several previously mentioned types or a previously mentioned type with a further approach. Further approaches can be mathematical optimization methods or physical simulation models.

3.1.6. Application Purpose

There are several reasons for the application of ESs that can contribute to energy efficiency improvements.
  • Transparency: To reduce the energy consumption in industry, stakeholders need a sufficient level of energy transparency to create a meaningful basis for decision-making [33].
  • Optimization: Optimization in the context of this work means improving energy efficiency as far as necessary and feasible.
  • Prediction: Prediction means determining unknown values from known inputs. For energy analysis, this means that the available observations at time t are used to predict the energy or energy efficiency at the same time t [25].
  • Forecasting: Statements are made about the future. In energy analyses, future values t + x for energy or energy efficiency are estimated based on current and/or past information at time t [34].

3.2. Conceptualization of the Topic

In the conceptualization, research questions are raised, inclusion and exclusion criteria are defined, and keywords as well as the search query string are formulated.

3.2.1. Research Questions

The following research questions (RQ) emerged for the SLR, which are intended to lead to the identification of research gaps:
RQ-1
Which industries deploy ESs to increase energy efficiency?
RQ-2
How have ESs been applied in the manufacturing industry to enhance energy efficiency?
RQ-3
How are ESs for improving energy efficiency in industry structured and implemented?
RQ-4
How are ESs for improving energy efficiency in industrial applications developed?

3.2.2. Inclusion and Exclusion Criteria

Inclusion (IC) and exclusion (EC) criteria are defined in advance to ensure consistent filtering within the literature search. Only those publications that pass this selection process are analyzed in depth.
IC-1
Studies written in English or German;
IC-2
Reviewed studies;
IC-3
Online full text availability;
IC-4
Empirical studies with a focus on ESs to improve energy efficiency in industry.
EC-1
Studies that do not meet the inclusion criteria;
EC-2
Duplicates;
EC-3
Excerpts from research results;
EC-4
Surveys or reviews (however, should they pertain to this review, they are integrated into Section 1 to address related work).

3.2.3. Keywords and Search Query String

The keywords and the search query string are determined based on the classification in Figure 2. We focus on industrial applications and do not restrict the search to a specific manufacturing type, application perspective, or application purpose. However, to find literature that is more relevant to the objective of this review, the search is limited by specifying the application focus and AI technique. For the search query, keywords and their synonyms are concatenated with “OR” operators to broaden the search. By using asterisks (*), words sharing a common root are included in the search as well. The search is restricted by using “AND” operators. The chosen keywords and their logical connections are presented in Table 1.

3.3. Search and Filter Literature

Literature databases were selected based on [23,35,36], which are literature reviews related to energy efficiency and flexibility in industry. Thus, the five electronic databases selected are Web of Science [37], ScienceDirect [38], IEEE Xplore [39], SpringerLink [40], and WorldCat [41]. The functionality and syntax of these databases differ slightly. ScienceDirect, for example, does not support wildcards such as asterisk (*), other databases including SpringerLink do not allow limiting the search to the abstract, so the search string has to be adjusted for a feasible filtering. The search queries used and the number of results found are listed in Table 1. Figure 3 provides an overview of the literature search and filtering process. A total of 1668 publications are found by entering the search query. When checking the titles, 1521 publications are filtered out according to the inclusion and exclusion criteria, so only 147 results remain. This number decreases to 69 articles after examining the abstract. By screening the full texts, another 18 publications are excluded. However, 11 publications are added through the snowball search. This results in a total of 62 publications for a detailed analysis.

3.4. Literature Analysis and Synthesis

Within the scope of the literature analysis and synthesis, initially descriptive information such as the publishing year, the document type, and the ES applications are extracted from the 62 selected publication. Each publication is assigned an identifier (ID) for subsequent analyses. The IDs are characterized by a number and an index, whereby the number is a consecutive sequence starting with the oldest and ending with the most recent publication. The index contains the relevance groups defined by the application areas of the ESs. Group A includes ESs within manufacturing and group B for other industries. Table 2 provides an overview of the analyzed publications.

3.4.1. Keyword Co-Occurrence Analysis

To explore the thematic structure of research based on our search queries defined in Table 1, we conduct a keyword co-occurrence analysis using VOSviewer [102]. This approach reveals relationships between keywords by analyzing their co-occurrence frequency in the literature. The keyword co-occurrence analysis thus helps to map the current state of research, highlighting the dominant areas of focus and uncovering potential gaps where further investigation may be needed.
For this analysis, results are exported from Web of Science [37], ScienceDirect [38], and IEEE Xplore [39] in Research Information System (RIS) format. SpringerLink [40] and WorldCat [41] are excluded due to the absence of export functions. Despite this limitation, the high number of included sources offers a comprehensive view of the research landscape. We apply the full counting method, giving equal weight to each co-occurrence with a threshold of five occurrences. Of the 4916 keywords, 242 meet this criterion. The association strength method is then used to normalize the strength of links between keywords. In the VOSviewer-generated graphic (see Figure 4), keywords are clustered based on their co-occurrence, grouping those that frequently appear together.
One of the largest and most significant clusters covers ESs and various other AI technologies (red). Other relevant clusters include sustainability (green), energy efficiency (yellow), and optimization (blue). The connections between clusters, depicted by lines, illustrate the interrelations among topics in the literature. Notably, the links between sustainability and ESs, as well as between energy efficiency and ESs, are less pronounced, indicating a potential gap in the literature. This suggests that further research should focus on the integration of ESs in these areas.

3.4.2. Data Sources and Publication Trend

The 62 publications include a total of 49 journal articles and 13 proceedings papers, whereby both document types are distributed rather evenly throughout the years. As presented in Figure 5, the publications are not concentrated on a small number of journals or conference proceedings but are spread across a total of 48 sources. The highest concentration can be observed in Expert Systems and Application, with six publications. Other sources listed in Figure 5 are represented by two publications each. The remaining 38 journals and conference proceedings are each represented by a single publication.
Among the publishers, Elsevier leads with 21 publications, followed by Springer with 19, IEEE with 10, MDPI with 5, Wiley with 2, and five other publishers each contributing individually.
Figure 6 shows the number of analyzed publications throughout the years, separated by both relevance groups. No time restriction was set for the literature search, which means that publications from 1987 to 2024 could be reviewed. The year 2024 is not fully covered since the search was carried out in August 2024. Until 2009, studies were published only occasionally. From 2012 onward, an increase in the average number of publications over the years can be observed for both groups. Furthermore, it is noticeable that the average number of publications in Group A has increased significantly more than in Group B over the last ten years.

3.4.3. Authors Country Distribution

The authors’ affiliations are an additional source of information that is extracted from the metadata of the selected publications. Figure 7 visualizes the authors’ distribution across individual countries. Most authors are affiliated with China (38). This is followed by a double-digit number of authors from Germany (29), India (28), the United States (14), and Spain (14). The continents where most authors are concentrated are Asia (110), Europe (86), and North America (18). The remaining authors originate from South America (9), Africa (9), and Australia (4). There is no observable trend of a particular change in the frequency of publication from individual countries or continents over certain time periods.

3.4.4. Literature Categorization

In this section, publications in relevance group A are categorized in accordance with Figure 2 by system boundary, manufacturing type, application perspective, application purpose and ES type. Since publications in relevance group B do not address manufacturing, they are only categorized by application perspective, application purpose, and ES type.
Regarding the system boundary (Table 3), all dimensions are covered, whereby the actual manufacturing process predominates with over one-third of the publications in group A. In terms of the manufacturing type (Table 4), it can be observed that the use of ESs is concentrated on serial and mass manufacturing, i.e., on larger quantities. Concerning the application perspective (Table 5), most of the studies address process planning and operation. The application purpose of ESs (Table 6) is primarily to create energy transparency and optimization, but there are also some publications that predict or forecast information in the context of energy efficiency. Among the ES types (Table 7), pure ML-based ESs are the least common. The other three dimensions—rule-based ES, fuzzy ES, and hybrid ES—are approximately evenly represented. Hybrid ESs are combinations of rule-based and fuzzy ESs (PA16); rule-based and ML-based ESs (PA11, PA35, PA38, PA52); fuzzy and ML-based ESs (PA4, PB9, PB13, PB18, PB54, PB56, PA59); rule-based or fuzzy ESs using mathematical optimization algorithms (PA25, PA45, PA48, PA49); ML-based ESs combined with simulation models (PA41, PA57, PA58); and a rule-based ES coupled with simulation models (PB10).

3.5. Identification of Research Gaps

Answers to the research questions raised in Section 3.2.1 are provided in the following and discussed to identify research gaps.

3.5.1. Classification of ESs by Industries

To answer RQ1, in which industries ESs are used to increase energy efficiency, the classification of economic sectors of the Federal Statistical Office in Germany is considered. The classification is hierarchically structured and is divided into sections, which are separated into divisions [103].
Table 8 summarizes the analyzed publications according to this classification scheme. We can classify the publications into seven industries. Some publications cannot be assigned clearly because they deal with cross-sectional technologies or do not provide any information about the exact industry. There is a noticeable concentration of the use of ESs in the manufacturing of metal goods. Furthermore, Table 8 shows again that the majority of ESs (66%) for improving energy efficiency are deployed in manufacturing (relevance group A).

3.5.2. Utilizations of ESs in Industry

Regarding the question of how ESs are applied to enhance energy efficiency in manufacturing (RQ2), it can be observed that ESs with a wider system boundary are more transferable than others with a more restricted system boundary. Thus, Refs. [47,48,63,75,76,95] describe ESs with system boundaries covering entire factories but are rather utilized for preliminary or generic energy analyses. Some other publications with a more limited system boundary such as [9,52,57,62,79] are applicable for similar use cases with few or no adjustments. The remaining publications in relevance group A describe ESs that are largely limited to individual case studies and cannot be transferred to other machines, systems, or processes without further modification.
The classification in Table 5 also shows that the utilization of ESs is strongly focused on either planning manufacturing processes regarding energy efficiency before actual operation or optimization during operation. ESs are deployed less frequently for energy-efficient planning of new energy systems. Furthermore, according to Table 6, the primary purpose of ESs is to create energy transparency as a basis for subsequent improvement processes as well as to contribute directly to improving energy efficiency.

3.5.3. Structure and Implementations of ESs

RQ3 addresses the structure and implementation of ESs. All ESs described within the analyzed 62 publications feature an implicit or explicit knowledge base, an inference engine, and a user interface, which are interconnected as shown in Figure 8. Some of the publications, for instance [75,76], additionally include distinct knowledge acquisition and explanation modules.
The expertise stored in the knowledge base can be represented in various forms. These include mathematical formulas [42], historical measurement data [43], physical and technical relationships [45], rules [49], and facts [57]. Furthermore, as outlined by [75], the knowledge base can be divided into short-term and long-term memory. The short-term memory stores information for particular use cases for which ESs are applied, whereas the long-term memory stores more general relationships, which correspond to the heuristic knowledge of human experts. Knowledge can be acquired using different techniques. Literature surveys are explicitly mentioned by [57,93] as well as expert consultations by [57,75,93]. The optional knowledge acquisition module enables the knowledge base to be enriched with new content while the ES is already deployed. Depending on the ES type (Table 7), there are variations in how the knowledge is represented in the knowledge base and how the inference engine utilizes the knowledge stored in the knowledge base to gain insights or solve problems.
The communication between humans and the ES (human–machine interface) can be realized by different means, including interactive questionnaires [42], dialog boxes [47], measurement data inputs [9], text outputs, and graphics [61]. The process interface enables the connection to other technical systems for the uni- or bidirectional transfer of data [104]. Other technical systems can be, for example, sensors or programmable logic controllers [11,82]. At least one of the two interface types is required to input information about the present situation. The optional explanation module clarifies the reasoning process in order to make it comprehensible and thus promote acceptance of the findings presented by the ES [75].
As the analyzed publications reveal, the software implementation of ES can be highly diverse. Ref. [42] use Fortran and Assembler as programming languages. More recent publications implement their ESs in Excel [57], C language [60], CLIPS [75], Python [79], and LabVIEW [92]. Refs. [11,66,82,84,97,98] use combinations of different programming languages to integrate simulation models or databases as part of the ESs.

3.5.4. Development of ESs for Industrial Applications

Many analyzed publications follow similar sequential and sometimes iterative steps in the development of their ES. These steps are as follows:
  • Situation analysis: Create an overview of the situation and examine the problem to be solved [48].
  • Parameter identification: Identify important and interactive parameters and factors, i.e., those that are both controllable and influential [55,84].
  • Data generation and acquisition: Data are generated and collected using simulation models [42] or experiments [60]. Experimental data generation is more common in the analyzed publications.
  • Data cleaning and preprocessing: Data filtering [55], scaling [71], and further preprocessing such as discriminant analysis [55] or feature extraction [92] is performed.
  • Creation of a knowledge base and knowledge representation: Knowledge is gathered and represented as described in Section 3.5.3.
  • Modeling: Build mathematical models [84], fuzzy models, or data-driven models such as neural networks [45]. The development steps of the modeling process depend on the respective model type.
  • Application and validation: The ES can be qualitatively validated via user interviews and quantitatively through case studies [84].
Refs. [11,75,84] specified corresponding development methods. Ref. [75] followed an incremental development model exclusively for the computational implementation, without providing a detailed description of the individual phases. Ref. [84] based the entire development of their ES on a design science method. The method described by [84] comprises three phases: conceptual design, tool development, and application and validation. During the conceptual design phase, influential factors are selected, and calculation rules are defined and iteratively refined. Subsequently, the computational implementation follows during the tool development phase. Finally, the ES is validated both qualitatively and quantitatively. Ref. [11] adopted the approach from [84] to manufacturing. This involves identifying relevant consumers and controllable parameters, acquiring measurement data, developing ML models and simulation models, defining energy performance indicators, formulating a rule base, and finally integrating all subsystems. References to practical implementation are provided for each step. Following [84], both application and validation were carried out, after which further refinements were possible. Additionally, in [11], personas with necessary capabilities were described for each development step.
To summarize, various development approaches for ESs can be found with some similarities in the context of improving industrial energy efficiency. Most development approaches, except for those described by [11,75,84], cannot be associated with a superior methodological approach. In the field of manufacturing, solely [11] offers a process model with a detailed description of the individual steps for developing ESs to increase energy efficiency. However, it is limited exclusively to the consideration of a single machine and requires the time-consuming development of an analytical simulation model as a core element [11].

4. Discussion and Conclusions

This paper presents an SLR that identifies, summarizes, and analyzes ESs aimed at improving energy efficiency in industry, with a particular focus on manufacturing. By following a formal six-step approach, we sought to minimize bias and enhance the reliability of the selected literature, resulting in a thorough analysis of 62 publications. The findings were organized by classifying the proposed ESs according to system boundary, manufacturing type, application perspective, application purpose, and ES type. The research questions raised were addressed, and several research gaps were identified.
Regarding existing literature concerning ESs, the main contribution of this work lies in the special focus on improving energy efficiency in manufacturing, although other industries were also considered, providing a comprehensive assessment. This study offers a structured classification of ESs across various dimensions, industries, and contexts, examining their structure, implementation, utilization, and development within the scope of energy efficiency. However, it is important to acknowledge the limitations of this study, including its reliance on digital sources and the exclusion of older publications, which may introduce a bias toward recent findings. Additionally, language restriction could have led to the exclusion of relevant studies published in other languages.
Despite these limitations, this study provides a detailed review of the current state of the art in the domain of ESs for energy efficiency in manufacturing. The analysis led to several noteworthy observations, while Figure 6 illustrates an increase in the number of publications, it remains unclear whether this trend reflects a growing interest in the topic or a broader expansion in research output. However, the comparison between relevance groups reveals a more pronounced growth in group A publications over the past decade, likely due to heightened attention to sustainability in manufacturing. Additionally, the analysis of authors’ country affiliations in Figure 7 suggests that countries with large populations, strong secondary economic sectors, and substantial research investments have contributed more significantly to this area of research. Moreover, the answer given to RQ1 (Table 8) combined with the categorization by manufacturing type (Table 4) indicates that ESs for improving energy efficiency are of particular interest to companies with high energy consumption and large output quantities. This could be related to the initial development effort arising from the lack of established process models (RQ4) and the low transferability of ESs to other use cases (RQ2).
The answers provided to the research questions revealed research gaps that can be addressed in future research. These include understanding why ESs are more prevalent in certain industries (RQ1) and the factors, such as energy consumption, output quantities, or digitalization, which influence this trend. Additionally, the low transferability of ESs (RQ2) suggests a need for research on developing more adaptable systems. Regarding the structure and implementation of ESs (RQ3), it has been revealed that although all described ESs have a knowledge base, an inference engine and a user interface, their implementation is highly diverse. Therefore, a tailored software tool (expert system shell [105]) could simplify the development process considerably. For the development of ESs to increase energy efficiency in manufacturing (RQ4), it was also pointed out that there is still no universally applicable methodology, which can lead to higher initial costs and may act as a barrier to the broader adoption of ESs. Hence, future research could aim to design a more broadly applicable methodology and streamlined software framework for the development of such ESs.

Author Contributions

Conceptualization, B.I.; methodology, B.I.; formal analysis, B.I.; investigation, B.I.; resources, M.W.; data curation, B.I.; writing—original draft preparation, B.I.; writing—review and editing, B.I., M.F. and M.W.; visualization, B.I.; supervision, M.W.; project administration, B.I.; funding acquisition, M.W. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the German Federal Ministry of Economic Affairs and Climate Action (BMWK) within the project KI4ETA (grant number 03EN2053A).

Data Availability Statement

No data were used for the research described in the article.

Acknowledgments

The authors gratefully acknowledge financial support from the German Federal Ministry of Economic Affairs and Climate Action (BMWK) and project supervision by the project management Jülich (PtJ).

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AIArtificial intelligence
ESExpert system
ECExclusion criteria
ICInclusion criteria
IDIdentifier
MLMachine learning
PPublication
RISResearch Information System
RQResearch question
SLRSystematic literature review

References

  1. European Parliament; Council of the European Union. Regulation (EU) 2021/1119. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32021R1119 (accessed on 29 July 2024).
  2. Ritchie, H.; Roser, M.; Rosado, P. CO2 and Greenhouse Gas Emissions. Available online: https://ourworldindata.org/co2-and-greenhouse-gas-emissions (accessed on 27 July 2024).
  3. International Energy Agency. Key World Energy Statistics 2021. Available online: https://www.iea.org/reports/key-world-energy-statistics-2021/final-consumption (accessed on 27 July 2024).
  4. ABB Ltd. Accelerating Ambition: How Global Industry is Speeding up Investment in Energy Efficiency. Available online: https://www.energyefficiencymovement.com/wp-content/uploads/2022/04/ABB-Energy-Efficiency-Survey-Report-2022.pdf (accessed on 16 August 2024).
  5. EN ISO 50001:2018; Energy Management Systems—Requirements with Guidance for Use. ISO: Geneva, Switzerland, 2018.
  6. International Energy Agency. Energy Efficiency 2022. Available online: https://www.iea.org/reports/energy-efficiency-2022 (accessed on 31 July 2024).
  7. International Energy Agency. Digitalization and Energy. Available online: https://www.iea.org/reports/digitalisation-and-energy (accessed on 6 November 2023).
  8. DeTore, A.W. An introduction to expert systems. J. Insur. Med. 1989, 21, 233–236. [Google Scholar]
  9. Ioshchikhes, B.; Elserafi, G.; Weigold, M. An Expert System-Based Approach For Improving Energy Efficiency Of Chamber Cleaning Machines. In Proceedings of the Conference on Production Systems and Logistics (CPSL 2023), Querétaro, Mexico, 28 February–2 March 2023. [Google Scholar] [CrossRef]
  10. Jackson, P. Introduction to Expert Systems, 3rd ed.; Addison-Wesley Longman Publishing Co., Inc.: Boston, MA, USA, 1998. [Google Scholar]
  11. Ioshchikhes, B.; Frank, M.; Elserafi, G.; Magin, J.; Weigold, M. Developing Expert Systems for Improving Energy Efficiency in Manufacturing: A Case Study on Parts Cleaning. Energies 2024, 17, 3417. [Google Scholar] [CrossRef]
  12. Hassija, V.; Chamola, V.; Mahapatra, A.; Singal, A.; Goel, D.; Huang, K.; Scardapane, S.; Spinelli, I.; Mahmud, M.; Hussain, A. Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence. Cogn. Comput. 2024, 16, 45–74. [Google Scholar] [CrossRef]
  13. Liao, S.H. Expert system methodologies and applications—A decade review from 1995 to 2004. Expert Syst. Appl. 2005, 28, 93–103. [Google Scholar] [CrossRef]
  14. Biel, K.; Glock, C.H. Systematic literature review of decision support models for energy-efficient production planning. Comput. Ind. Eng. 2016, 101, 243–259. [Google Scholar] [CrossRef]
  15. Wagner, W.P. Trends in expert system development: A longitudinal content analysis of over thirty years of expert system case studies. Expert Syst. Appl. 2017, 76, 85–96. [Google Scholar] [CrossRef]
  16. Kumar, L.S.P. Knowledge-based expert system in manufacturing planning: State-of-the-art review. Int. J. Prod. Res. 2019, 57, 4766–4790. [Google Scholar] [CrossRef]
  17. Piller, F.T. Einführung: Informationsrevolution und industrielle Produktion. In Mass Customization; Picot, A., Reichwald, R., Franck, E., Eds.; Springer Fachmedien Wiesbaden GmbH: Wiesbaden, Germany, 2006; pp. 1–9. [Google Scholar] [CrossRef]
  18. Tranfield, D.; Denyer, D.; Smart, P. Towards a Methodology for Developing Evidence-Informed Management Knowledge by Means of Systematic Review. Br. J. Manag. 2003, 14, 207–222. [Google Scholar] [CrossRef]
  19. Feak, C.B.; Swales, J.M. Telling a Research Story: Writing a Literature Review, 2nd ed.; The University of Michigan Press: Ann Arbor, MI, USA, 2009. [Google Scholar] [CrossRef]
  20. Reynolds, N.; Simintiras, A.; Vlachou, E. International business negotiations. Int. Mark. Rev. 2003, 20, 236–261. [Google Scholar] [CrossRef]
  21. Glock, C.H.; Hochrein, S. Purchasing Organization and Design: A Literature Review. Bus. Res. 2011, 4, 149–191. [Google Scholar] [CrossRef]
  22. Tricco, A.; Lillie, E.; Zarin, W.; O’Brien, K.; Colquhoun, H.; Levac, D.; Moher, D.; Peters, M.; Horsley, T.; Weeks, L.; et al. PRISMA extension for scoping reviews (PRISMA-ScR): Checklist and explanation. Ann. Intern. Med. 2018, 169, 467–473. [Google Scholar] [CrossRef] [PubMed]
  23. Panten, N. Deep Reinforcement Learning zur Betriebsoptimierung Hybrider Industrieller Energienetze; Schriftenreihe des PTW: Innovation Fertigungstechnik; Shaker Verlag: Düren, Germany, 2019. [Google Scholar]
  24. vom Brocke, J.; Simons, A.; Niehaves, B.; Reimer, K.; Plattfaut, R.; Cleven, A. Reconstructing the Giant: On the Importance of Rigour in Documenting the Literature Search Process. In Proceedings of the European Conference on Information Systems. 2009. Available online: https://aisel.aisnet.org/ecis2009/161 (accessed on 29 July 2024).
  25. Walther, J.; Weigold, M. A Systematic Review on Predicting and Forecasting the Electrical Energy Consumption in the Manufacturing Industry. Energies 2021, 14, 968. [Google Scholar] [CrossRef]
  26. Duflou, J.R.; Sutherland, J.W.; Dornfeld, D.; Herrmann, C.; Jeswiet, J.; Kara, S.; Hauschild, M.; Kellens, K. Towards energy and resource efficient manufacturing: A processes and systems approach. CIRP Ann. 2012, 61, 587–609. [Google Scholar] [CrossRef]
  27. Westkämper, E. Einführung in Die Organisation der Produktion; Springer: Berlin/Heidelberg, Germany, 2006. [Google Scholar]
  28. Weck, M. Werkzeugmaschinen 4; Springer: Berlin/Heidelberg, Germany, 2006. [Google Scholar] [CrossRef]
  29. Kara, S.; Bogdanski, G.; Li, W. Electricity Metering and Monitoring in Manufacturing Systems. In Glocalized Solutions for Sustainability in Manufacturing; Hesselbach, J., Herrmann, C., Eds.; Springer: Berlin/Heidelberg, Germany, 2011; pp. 1–10. [Google Scholar] [CrossRef]
  30. ISO 14955-1:2017; Machine Tools—Environmental Evaluation of Machine Tools—Part 1: Design Methodology for Energy-Efficient Machine Tools. ISO: Geneva, Switzerland, 2017.
  31. VDI 5207 Blatt 1; Energy-Flexible Factory–Fundamentals. VDI: Düsseldorf, Germany, 2020.
  32. Ongsulee, P. Artificial intelligence, machine learning and deep learning. In Proceedings of the 2017 15th International Conference on ICT and Knowledge Engineering (ICT&KE), Bangkok, Thailand, 22–24 November 2017; pp. 1–6. [Google Scholar] [CrossRef]
  33. Posselt, G. Towards Energy Transparent Factories; Springer International Publishing: Cham, Switzerland, 2016. [Google Scholar] [CrossRef]
  34. Box, G.E.P.; Jenkins, G.M.; Reinsel, G.C.; Ljung, G.M. Time Series Analysis: Forecasting and Control, 5th ed.; John Wiley and Sons Inc.: Hoboken, NJ, USA, 2015. [Google Scholar]
  35. dos Santos, S.B.A.; Soares, J.M.; Barroso, G.C.; Prata, B.d.A. Demand response application in industrial scenarios: A systematic mapping of practical implementation. Expert Syst. Appl. 2023, 215, 119393. [Google Scholar] [CrossRef]
  36. Solnørdal, M.; Foss, L. Closing the Energy Efficiency Gap—A Systematic Review of Empirical Articles on Drivers to Energy Efficiency in Manufacturing Firms. Energies 2018, 11, 518. [Google Scholar] [CrossRef]
  37. Clarivate. Web of Science. Available online: https://www.webofscience.com/wos/woscc/basic-search (accessed on 8 August 2024).
  38. Elsevier. ScienceDirect. Available online: https://www.sciencedirect.com/ (accessed on 8 August 2024).
  39. IEEE. IEEE Xplore. Available online: https://ieeexplore.ieee.org (accessed on 8 August 2024).
  40. SpringerNature. SpringerLink. Available online: https://link.springer.com/ (accessed on 8 August 2024).
  41. OCLC. WorldCat. Available online: https://www.worldcat.org (accessed on 8 August 2024).
  42. Keviczky, L.; Bányász, C.; Haber, R.; Hilger, M. Cemexpert: An Expert System for the Cement Grinding Process. IFAC Proc. Vol. 1987, 20, 157–162. [Google Scholar] [CrossRef]
  43. Miyayama, T.; Tanaka, S.; Miyatake, T.; Umeki, T.; Miyamoto, Y.; Nishino, K.; Harada, E. A combustion control support expert system for a coal-fired boiler. In Proceedings of the IECON ’91: 1991 International Conference on Industrial Electronics, Control and Instrumentation, Kobe, Japan, 28 October–1 November 1991; Volume 2, pp. 1513–1516. [Google Scholar] [CrossRef]
  44. Zhu, M.S.; Wang, B.X.; Xiao, Y.H. A multi-criteria decision making procedure for the analysis of an energy system. J. Therm. Sci. 1992, 1, 221–225. [Google Scholar] [CrossRef]
  45. Tuma, A.; Haasis, H.D.; Rentz, O. Development of emission orientated production control strategies using fuzzy expert systems, neural networks and neuro-fuzzy approaches. Fuzzy Sets Syst. 1996, 77, 255–264. [Google Scholar] [CrossRef]
  46. Kontopoulos, A.; Krallis, K.; Koukourakis, E.; Denaxas, N.; Kostis, N.; Broussaud, A.; Guyot, O. A hybrid, knowledge-based system as a process control ‘tool’ for improved energy efficiency in alumina calcining furnaces. Appl. Therm. Eng. 1997, 17, 935–945. [Google Scholar] [CrossRef]
  47. Dunning, S.; Segee, B.; Allen, V. A Self-Assessment Software Application for Industrial Manufacturers. Energy Eng. 1999, 96, 70–77. [Google Scholar] [CrossRef]
  48. Lau, H.; Cheng, E.; Lee, C.; Ho, G. A fuzzy logic approach to forecast energy consumption change in a manufacturing system. Expert Syst. Appl. 2008, 34, 1813–1824. [Google Scholar] [CrossRef]
  49. Lin, C.H.; Chen, C.S.; Chuang, H.J.; Huang, M.Y.; Huang, C.W. An Expert System for Three-Phase Balancing of Distribution Feeders. IEEE Trans. Power Syst. 2008, 23, 1488–1496. [Google Scholar] [CrossRef]
  50. Soyguder, S.; Alli, H. Predicting of fan speed for energy saving in HVAC system based on adaptive network based fuzzy inference system. Expert Syst. Appl. 2009, 36, 8631–8638. [Google Scholar] [CrossRef]
  51. Beck, A.; Göhner, P. An approach for model-based energy cost analysis of industrial automation systems. Energy Effic. 2012, 5, 303–319. [Google Scholar] [CrossRef]
  52. Garcia, R.F. Improving heat exchanger supervision using neural networks and rule based techniques. Expert Syst. Appl. 2012, 39, 3012–3021. [Google Scholar] [CrossRef]
  53. Klein, G.M.; Schrooten, T.; Neuhaus, T. Reduction of Energy Demand for Industrial Dedusting Plants. Chem. Ing. Tech. 2012, 84, 1121–1129. [Google Scholar] [CrossRef]
  54. Roy, S.S.; Pratihar, D.K. Soft computing-based approaches to predict energy consumption and stability margin of six-legged robots moving on gradient terrains. Appl. Intell. 2012, 37, 31–46. [Google Scholar] [CrossRef]
  55. Monedero, I.; Biscarri, F.; León, C.; Guerrero, J.I.; González, R.; Pérez-Lombard, L. Decision system based on neural networks to optimize the energy efficiency of a petrochemical plant. Expert Syst. Appl. 2012, 39, 9860–9867. [Google Scholar] [CrossRef]
  56. Djatkov, D.; Effenberger, M.; Martinov, M. Method for assessing and improving the efficiency of agricultural biogas plants based on fuzzy logic and expert systems. Appl. Energy 2014, 134, 163–175. [Google Scholar] [CrossRef]
  57. Do, T.T.H.; Schnitzer, H.; Le, T.H. A decision support framework considering sustainability for the selection of thermal food processes. J. Clean. Prod. 2014, 78, 112–120. [Google Scholar] [CrossRef]
  58. Mitra, S.; Singh, R.K.; Mondal, A.K. An Expert System Based Process Control System for Silicon Steel Mill Furnace of Rourkela Steel Plant. In Proceedings of the 2014 Fourth International Conference of Emerging Applications of Information Technology, Kolkata, India, 19–21 December 2014; pp. 29–33. [Google Scholar] [CrossRef]
  59. Bagheri, N.; Tarabi, N.; Javadikia, H. Development and evaluation of an adaptive neuro-fuzzy interface models to predict performance of a solar dryer. Agric. Eng. Int. CIGR -J. 2015, 17, 112–121. [Google Scholar]
  60. Iqbal, A.; Zhang, H.C.; Kong, L.L.; Hussain, G. A rule-based system for trade-off among energy consumption, tool life, and productivity in machining process. J. Intell. Manuf. 2015, 26, 1217–1232. [Google Scholar] [CrossRef]
  61. Ochoa, C.E.; Capeluto, I.G. Decision methodology for the development of an expert system applied in an adaptable energy retrofit façade system for residential buildings. Renew. Energy 2015, 78, 498–508. [Google Scholar] [CrossRef]
  62. Singh, P.P.; Madan, J. A computer-aided system for sustainability assessment for the die-casting process planning. Int. J. Adv. Manuf. Technol. 2016, 87, 1283–1298. [Google Scholar] [CrossRef]
  63. Singh, S.; Olugu, E.U.; Musa, S.N. Development of Sustainable Manufacturing Performance Evaluation Expert System for Small and Medium Enterprises. Procedia CIRP 2016, 40, 608–613. [Google Scholar] [CrossRef]
  64. Teja, R.; Sridhar, P.; Guruprasat, M. Control and Optimization of a Triple String Rotary Cement Kiln using Model Predictive Control. IFAC-PapersOnLine 2016, 49, 748–753. [Google Scholar] [CrossRef]
  65. Cai, Y.; Shao, H. Energy Efficiency State Identification in Milling Processing Based on Improved HMM. In Proceedings of the ASME 2017 12th International Manufacturing Science and Engineering Conference collocated with the JSME/ASME 2017 6th International Conference on Materials and Processing, Volume 2: Additive Manufacturing; Materials. Los Angeles, CA, USA, 4–8 June 2017; pp. 1–11. [Google Scholar] [CrossRef]
  66. Deng, Z.; Zhang, H.; Yahui, F.; Wan, L.; Lv, L. Research on intelligent expert system of green cutting process and its application. J. Clean. Prod. 2018, 185, 904–911. [Google Scholar] [CrossRef]
  67. Cheng, C.C.; Liu, K.W. Optimizing energy savings of the injection molding process by using a cloud energy management system. Energy Effic. 2018, 11, 415–426. [Google Scholar] [CrossRef]
  68. Fernández-Cerero, D.; Fernández-Montes, A.; Ortega, J.A. Energy policies for data-center monolithic schedulers. Expert Syst. Appl. 2018, 110, 170–181. [Google Scholar] [CrossRef]
  69. Garofalo, P.; Campi, P.; Vonella, A.V.; Mastrorilli, M. Application of multi-metric analysis for the evaluation of energy performance and energy use efficiency of sweet sorghum in the bioethanol supply-chain: A fuzzy-based expert system approach. Appl. Energy 2018, 220, 313–324. [Google Scholar] [CrossRef]
  70. Burow, K.; Franke, M.; Deng, Q.; Hribernik, K.; Thoben, K.D. Sustainable Data Management for Manufacturing. In Proceedings of the 2019 IEEE International Conference on Engineering, Technology and Innovation (ICE/ITMC), Valbonne Sophia-Antipolis, France, 17–19 June 2019; pp. 1–5. [Google Scholar] [CrossRef]
  71. Debnath, S.; Reddy, J.; Jagadish; Das, B. An expert system-based modeling and optimization of corrugated plate solar air collector for North Eastern India. J. Braz. Soc. Mech. Sci. Eng. 2019, 41, 273. [Google Scholar] [CrossRef]
  72. Simmons, A.; Sarao, G.; Campain, D. A Cement Mill Upgrade Story Reboot. In Proceedings of the 2019 IEEE-IAS/PCA Cement Industry Conference (IAS/PCA), St. Louis, MO, USA, 28 April–2 May 2019; pp. 1–8. [Google Scholar] [CrossRef]
  73. Wang, J.; Fei, Z.; Chang, Q.; Li, S.; Fu, Y. Multi-state decision of unreliable machines for energy-efficient production considering work-in-process inventory. Int. J. Adv. Manuf. Technol. 2019, 102, 1009–1021. [Google Scholar] [CrossRef]
  74. Zhao, J.; Peng, S.; Li, T.; Lv, S.; Li, M.; Zhang, H. Energy-aware fuzzy job-shop scheduling for engine remanufacturing at the multi-machine level. Front. Mech. Eng. 2019, 14, 474–488. [Google Scholar] [CrossRef]
  75. Buccieri, G.P.; Balestieri, J.A.P.; Matelli, J.A. Energy efficiency in Brazilian industrial plants: Knowledge management and applications through an expert system. J. Braz. Soc. Mech. Sci. Eng. 2020, 42, 577. [Google Scholar] [CrossRef]
  76. Grigoras, G.; Neagu, B.C. An Advanced Decision Support Platform in Energy Management to Increase Energy Efficiency for Small and Medium Enterprises. Appl.-Sci.-Basel 2020, 10, 3505. [Google Scholar] [CrossRef]
  77. Khayum, N.; Rout, A.; Deepak, B.B.V.L.; Anbarasu, S.; Murugan, S. Application of Fuzzy Regression Analysis in Predicting the Performance of the Anaerobic Reactor Co-digesting Spent Tea Waste with Cow Manure. Waste Biomass Valorization 2020, 11, 5665–5678. [Google Scholar] [CrossRef]
  78. Kim, S.H.; Nam, J. Can Both the Economic Value and Energy Performance of Small- and Mid-Sized Buildings Be Satisfied? Development of a Design Expert System in the Context of Korea. Sustainability 2020, 12, 4946. [Google Scholar] [CrossRef]
  79. Petruschke, L.; Elserafi, G.; Ioshchikhes, B.; Weigold, M. Machine learning based identification of energy efficiency measures for machine tools using load profiles and machine specific meta data. MM Sci. J. 2021, 2021, 5061–5068. [Google Scholar] [CrossRef]
  80. Qu, N.; You, W. Design and fault diagnosis of DCS sintering furnace’s temperature control system for edge computing. PLoS ONE 2021, 16, e0253246. [Google Scholar] [CrossRef]
  81. Davydenko, L.; Davydenko, N.; Bosak, A.; Bosak, A.; Deja, A.; Dzhuguryan, T. Smart Sustainable Freight Transport for a City Multi-Floor Manufacturing Cluster: A Framework of the Energy Efficiency Monitoring of Electric Vehicle Fleet Charging. Energies 2022, 15, 3780. [Google Scholar] [CrossRef]
  82. Ioshchikhes, B.; Borst, F.; Weigold, M. Assessing Energy Efficiency Measures for Hydraulic Systems using a Digital Twin. Procedia CIRP 2022, 107, 1232–1237. [Google Scholar] [CrossRef]
  83. Kalayci, O.; Pehlivan, I.; Coskun, S. Improving the performance of industrial mixers that are used in agricultural technologies via chaotic systems and artificial intelligence techniques. Turk. J. Electr. Eng. Comput. Sci. 2022, 30, 2418–2432. [Google Scholar] [CrossRef]
  84. Li, P.; Lu, Y.; Qian, Y.; Wang, Y.; Liang, W. An explanatory parametric model to predict comprehensive post-commissioning building performances. Build. Environ. 2022, 213, 108897. [Google Scholar] [CrossRef]
  85. Mendia, I.; Gil-Lopez, S.; Grau, I.; Del Ser, J. A novel approach for the detection of anomalous energy consumption patterns in industrial cyber–physical systems. Expert Syst. 2022, 41, e12959. [Google Scholar] [CrossRef]
  86. Rahman, H.F.; Chakrabortty, R.K.; Elsawah, S.; Ryan, M.J. Energy-efficient project scheduling with supplier selection in manufacturing projects. Expert Syst. Appl. 2022, 193, 116446. [Google Scholar] [CrossRef]
  87. Rahmati, A.H.; Nikbakht, A.M. Fuzzy-based modeling of thermohydraulic aspect of solar air heater roughened with inclined broken roughness. Neural Comput. Appl. 2022, 34, 2393–2412. [Google Scholar] [CrossRef]
  88. Taleb, H.; Nasser, A.; Andrieux, G.; Charara, N.; Cruz, E.M. Energy Consumption Improvement of a Healthcare Monitoring System: Application to LoRaWAN. IEEE Sens. J. 2022, 22, 7288–7299. [Google Scholar] [CrossRef]
  89. Wang, X.l.; Lu, M.y.; Wei, S.m.; Xie, Y.f. Multi-objective optimization based optimal setting control for industrial double-stream alumina digestion process. J. Cent. South Univ. 2022, 29, 173–185. [Google Scholar] [CrossRef]
  90. Choudhury, B.; Chandrasekaran, M. Electron Beam Welding Investigation of Inconel 825 and Optimize Energy Consumption Using Integrated Fuzzy Logic-Particle Swarm Optimization Approach. Int. J. Fuzzy Syst. 2023, 25, 1377–1399. [Google Scholar] [CrossRef]
  91. Duman, E.; Seckin, D. Enhancing the efficiency of cabin heaters in emergency shelters after earthquakes through an optimized fuzzy controller. Build. Simul. 2023, 16, 1759–1776. [Google Scholar] [CrossRef]
  92. Wang, J.; Tian, Y.; Hu, X.; Fan, Z.; Han, J.; Liu, Y. Development of grinding intelligent monitoring and big data-driven decision making expert system towards high efficiency and low energy consumption: Experimental approach. J. Intell. Manuf. 2023, 35, 1013–1035. [Google Scholar] [CrossRef]
  93. Perera, J.C.; Gopalakrishnan, B.; Bisht, P.S.; Chaudhari, S.; Sundaramoorthy, S. A Sustainability-Based Expert System for Additive Manufacturing and CNC Machining. Sensors 2023, 23, 7770. [Google Scholar] [CrossRef] [PubMed]
  94. Reddy, J.; Jagadish.; Das, B.; Debnath, S. Experimental investigation and optimization of sand-coated solar air collector parameters by fuzzy-MCDM integrated decision approach. J. Therm. Anal. Calorim. 2023, 148, 5543–5556. [Google Scholar] [CrossRef]
  95. Banadkouki, M.R.Z. Selection of strategies to improve energy efficiency in industry: A hybrid approach using entropy weight method and fuzzy TOPSIS. Energy 2023, 279, 128070. [Google Scholar] [CrossRef]
  96. Ilić, D.; Žarković, M. Machine learning for power generator condition assessment. Electr. Eng. 2023, 106, 2691–2703. [Google Scholar] [CrossRef]
  97. Mohamed, E.G.; Ahmed, M.; Redouane, M. Smart Energy Consumption Control System While Respecting the Quality of the Injection Molded Product. In Proceedings of the 2023 3rd International Conference on Innovative Research in Applied Science, Engineering and Technology (IRASET), IEEE, Mohammedia, Morocco, 18–19 May 2023; pp. 1–6. [Google Scholar] [CrossRef]
  98. Nadim, K.; Ouali, M.S.; Ghezzaz, H.; Ragab, A. Learn-to-supervise: Causal reinforcement learning for high-level control in industrial processes. Eng. Appl. Artif. Intell. 2023, 126, 106853. [Google Scholar] [CrossRef]
  99. Lourenço, G.; de Arruda, G.T.; La Correia de Rosa, V.E.; Castoldi, M.F.; Goedtel, A.; de Souza, W.A. Three-Phase Induction Motor Electrical Parameter Estimation Using Artificial Neural Networks. In Proceedings of the 2024 International Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), IEEE, Napoli, Italy, 9–21 June 2024; pp. 883–888. [Google Scholar] [CrossRef]
  100. Pawuś, D.; Paszkiel, S. Identification and Expert Approach to Controlling the Cement Grinding Process Using Artificial Neural Networks and Other Non-Linear Models. IEEE Access 2024, 12, 26364–26383. [Google Scholar] [CrossRef]
  101. Sharma, M.; Pareek, S.; Singh, K. An efficient power extraction using artificial intelligence based machine learning model for SPV array reconfiguration in solar industries. Eng. Appl. Artif. Intell. 2024, 129, 107516. [Google Scholar] [CrossRef]
  102. van Eck, N.J.; Waltman, L. Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics 2010, 84, 523–538. [Google Scholar] [CrossRef]
  103. Statistisches Bundesamt. Klassifikation der Wirtschaftszweige 2008 (WZ 2008). Available online: https://www.destatis.de/DE/Methoden/Klassifikationen/Gueter-Wirtschaftsklassifikationen/Downloads/klassifikation-wz-2008-3100100089004-aktuell.pdf?__blob=publicationFile (accessed on 18 June 2024).
  104. Wehking, K.H.; Albrecht, W.; Becker, J.; Hager, H.J.; Kille, C.; Popp, J.; Scherner, T.; Yousefifar, R. Technisches Handbuch Logistik; Springer: Berlin/Heidelberg, Germany, 2020. [Google Scholar]
  105. Quinn, K. Expert system shells: What to look for. Ref. Serv. Rev. 1990, 18, 83–86. [Google Scholar] [CrossRef]
Figure 1. Methodological steps for a systematic literature review (illustration adapted from [23] based on [24]).
Figure 1. Methodological steps for a systematic literature review (illustration adapted from [23] based on [24]).
Energies 17 04780 g001
Figure 2. Categories and dimensions for classification based on [25].
Figure 2. Categories and dimensions for classification based on [25].
Energies 17 04780 g002
Figure 3. Overview of the literature search and filtering. Illustration based on [25].
Figure 3. Overview of the literature search and filtering. Illustration based on [25].
Energies 17 04780 g003
Figure 4. Keyword co-occurrence network generated using VOSviewer version 1.6.20 [102].
Figure 4. Keyword co-occurrence network generated using VOSviewer version 1.6.20 [102].
Energies 17 04780 g004
Figure 5. Journal, proceedings, and publisher distribution.
Figure 5. Journal, proceedings, and publisher distribution.
Energies 17 04780 g005
Figure 6. Publishing years of the analyzed publications.
Figure 6. Publishing years of the analyzed publications.
Energies 17 04780 g006
Figure 7. Authors’ country distribution.
Figure 7. Authors’ country distribution.
Energies 17 04780 g007
Figure 8. Simplified structure of ESs.
Figure 8. Simplified structure of ESs.
Energies 17 04780 g008
Table 1. Databases and applied search queries on 8 August 2024.
Table 1. Databases and applied search queries on 8 August 2024.
DatabaseSearch QueryNumber of Results
Web of Science((TS = (energ*) OR TS = (load) OR TS = (electri*) OR TS = (power)) AND
(TS = (industr*) OR TS = (manufactur*)) AND (TS = (efficien*))
AND (TS = (expert system)))
817
ScienceDirect(energy OR load OR electricity OR power) AND (industry OR manufacturing) AND (efficiency) AND (expert system)128
IEEE Xplore(“All Metadata”: energ* OR load OR electri* OR power) AND (“All Metadata”:
industr* OR manufactur*) AND (“All Metadata”: efficien*)
AND (“All Metadata”: expert system)
139
SpringerLink(“energy efficiency”) AND (industr* OR manufactur*) AND (“expert system”)270
WorldCat(energy efficiency) AND (industr* OR manufactur*) AND (“expert system”)314
Table 2. Overview of analyzed publications.
Table 2. Overview of analyzed publications.
IDReferenceDocument TypeExpert System Application
PA1[42]Proceedings paperAssists operators in setting the technological parameters for cement grinding to optimal working points
PB2[43]Proceedings paperSupports the control of coal-fired boiler combustion with the aim of high-efficiency operation
PA3[44]Journal articleSupports multi-criteria decision making for analyzing energy systems by simultaneously considering economic and energy-related criteria
PA4[45]Journal articleSelects control strategies to influence energy and material flows in interconnected production units and processes
PA5[46]Journal articleGuides furnace operators to reduce sensible heat losses by operating with a marginal excess air
PA6[47]Journal articleAssists manufacturers in assessing their facilities, minimizing waste, and improving energy efficiency
PA7[48]Journal articleForecast energy consumption change to reduce the uncertainty, inconvenience, and inefficiency resulting from variations in the production factors
PB8[49]Journal articleEnhances the three-phase balance of distribution systems and reduces the neutral current of distribution feeders
PB9[50]Journal articlePredicts the fan speed and the damper gap rates of a heating, ventilating, and air-conditioning system
PB10[51]Journal articleEnables a fully automated, user-centric energy cost analysis for industrial automation systems
PA11[52]Journal articleDiagnoses the fouling condition and heat transfer efficiency of a heat exchanger
PA12[53]Journal articleEnables identifying energy and operating cost reductions for new and existing filtering installations
PB13[54]Journal articlePredicts specific energy consumption and stability margin in the ascending and descending motions of a six-legged robot
PA14[55]Journal articleIncreases the energy efficiency of a petrochemical plant by helping the operator take decisions to optimize future operating points
PB15[56]Journal articleAssesses and supports improving the efficiency of agricultural biogas plants based on specific performance figures
PA16[57]Journal articleSupports selecting thermal process technologies in the food industry considering sustainability
PA17[58]Proceedings paperPredicts furnace heating cycles with the objective of achieving the desired product quality with minimal energy utilization
PB18[59]Journal articlePredicts and evaluates the energy efficiency of a forced-convection solar dryer
PA19[60]Journal articleSuggests suitable settings for cutting parameters leading to a trade-off between energy consumption, tool life, and productivity in a machining process
PB20[61]Journal articleDiagnoses suitable technologies to adapt energy façade systems for residential buildings
PA21[62]Journal articleAssesses the sustainability performance of a die-casting process plan by determining CO2 emissions, solid waste, and energy use
PA22[63]Proceedings paperAssesses the sustainable manufacturing performance in small- and medium-sized enterprises
PA23[64]Proceedings paperControls and optimizes a triple string rotary cement kiln to improve energy efficiency and increase production while ensuring good quality
PA24[65]Proceedings paperDetermines the initial probability of the energy efficiency state of milling process for further classification
PA25[66]Journal articleOptimizes cutting parameters for machine tools based on cutting process cases
PA26[67]Journal articleOptimizes parameters of an injection molding process and measures the energy savings
PB27[68]Journal articleApplies energy policies based on shutting machines off in order to reduce data-center energy consumption
PA28[69]Journal articleAssesses the impact of soil treatment and mineral nitrogen on the energy performance and efficiency of sweet sorghum in the bioethanol supply chain
PA29[70]Proceedings paperOptimizes the energy efficiency of a running production process by utilizing scattered data from multiple distributed sources
PB30[71]Journal articleOptimized parameters of solar air collectors with corrugated plates under different climatic conditions
PA31[72]Journal articleImproves the stability of a cement mill and reduces energy consumption for a cement mill circuit
PA32[73]Journal articleSwitches machines in manufacturing systems with serial, disassembly, and assembly workstations into sleep state at an appropriate opportunity
PA33[74]Journal articleSchedules engine remanufacturing at a multi-machine level to minimize remanufacturing energy consumption
PA34[75]Journal articlePreliminarily diagnoses energy efficiency potential in Brazilian industrial plants and manages knowledge in organizational environments
PA35[76]Journal articleSupports energy consumption management decisions for small- and medium-sized enterprises
PB36[77]Journal articlePredicts the performance of an anaerobic reactor for the production of biogas using cow dung with spent tea waste in different proportions
PB37[78]Journal articlePresents multiple design decision paths for small- and mid-sized buildings that pursue a balance between economic value and energy performance
PA38[79]Proceedings paperSupports energy efficiency improvements of machine tools by providing energy efficiency measures and their evaluation
PA39[80]Journal articlePredicts the situation of sintering furnaces to diagnoses faults with the aim of reducing energy consumption during smelting
PB40[81]Journal articleSupports scheduling the charging of an electric vehicle fleet by assigning each electric vehicle a priority for connection to the charging station
PA41[82]Proceedings paperIdentifies system leakage and increased flow resistance in hydraulic systems and quantifies potential energy savings
PA42[83]Journal articleIncreases production performance of industrial mixers in terms of product quality, homogeneity, time, and energy savings
PB43[84]Journal articlePredicts existing building commissioning outcomes for various types of public buildings
PA44[85]Journal articleCharacterizes the nominal performance of a factory in terms of production and energy consumption
PB45[86]Journal articleProposes manufacturing projects an energy-efficient resource constrained project scheduling plan embedded with a supplier selection strategy
PB46[87]Journal articlePredicts the thermohydraulic performance of solar air heater roughened with inclined broken roughness
PB47[88]Journal articleDetermines the duration of the sleep mode and the sent data rate of a healthcare monitoring system for power consumption optimization
PA48[89]Journal articleFinds solutions to maintain the balance between resource consumption and process indices for a double-stream alumina digestion process
PA49[90]Journal articleOptimizes parameters during electron beam welding of Inconel 825 for minimizing net input energy without compromising product quality
PB50[91]Journal articleReduces the fuel consumption of cabin heaters in emergency shelters and enhances their efficiency
PA51[9]Proceedings paperAnalyzes the energy consumption of chamber cleaning machines and provides measures to improve energy efficiency
PA52[92]Journal articleControls and optimizes grinding process parameters based on monitored power data
PA53[93]Journal articleAids metal manufacturing facilities in selecting binder jetting, direct metal laser sintering, or CNC machining
PB54[94]Journal articleOptimizes parameters of sand-coated solar air collectors for different climatic conditions
PA55[95]Journal articleOffers industrial owners, developers, and planners strategies to enhance energy efficiency in their multi-criteria decision-making processes
PA56[96]Journal articleCreates maintenance schedules for power generators to fit assessments made by human experts
PA57[97]Proceedings paperRecommends optimal parameter values for controlling an injection molding process.
PA58[98]Journal articleSuggests control policies to reduce energy consumption for heat recovery systems in a pulp and paper mill
PA59[11]Journal articleIdentifies inefficient parameter settings, calculates potential energy savings, and prioritizes actions for a throughput parts-cleaning machine
PA60[99]Proceedings paperEstimates the electrical parameters of the equivalent motor circuit for three-phase induction motors to enhance system efficiency and reliability
PA61[100]Journal articleProvides control signals to operators for managing a multi-dimensional, non-linear, and non-stationary cement grinding process
PB62[101]Journal articlePredicts suitable configurations for solar photovoltaic arrays under partial shading conditions
Table 3. Categorization of group A publications by system boundaries.
Table 3. Categorization of group A publications by system boundaries.
FactoryManufacturing Cell/LineMachineComponentProcess
PublicationPA6; PA7; PA22; PA34; PA35; PA44; PA45; PA55PA4; PA16; PA21; PA29; PA31; PA32; PA33; PA53; PA55PA14; PA17; PA19; PA28; PA38; PA52; PA55PA3; PA11; PA17; PA38; PA41; PA55; PA60PA1; PA5; PA12; PA17; PA23; PA24; PA25; PA26; PA38; PA39; PA48; PA49; PA51; PA55; PA57; PA58; PA59
Table 4. Categorization of group A publications by manufacturing type.
Table 4. Categorization of group A publications by manufacturing type.
Job-Shop ManufacturingRepetitive ManufacturingSerial ManufacturingMass Manufacturing
PublicationPA53PA35; PA53PA4; PA7; PA16; PA19; PA21; PA24; PA25; PA26; PA32; PA33; PA38; PA39; PA44; PA51; PA52; PA53; PA57; PA59PA1; PA4; PA5; PA7; PA14; PA17; PA19; PA21; PA23; PA24; PA25; PA26; PA28; PA31; PA32; PA33; PA38; PA39; PA44; PA48; PA51; PA52; PA53; PA55; PA57; PA58; PA59; PA60; PA61
Table 5. Categorization of group A publications by application perspective.
Table 5. Categorization of group A publications by application perspective.
EngineeringProcess PlanningOperation
PublicationPA3; PA16; PA35; PA55; PA60PA1; PA4; PA6; PA12; PA16; PA19; PA21; PA22; PA25; PA26; PA28; PA29; PA33; PA34; PA45; PA48; PA52; PA53; PA57; PA58; PA59PA1; PA5; PA7; PA11; PA14; PA17; PA23; PA24; PA31; PA32; PA38; PA39; PA41; PA44; PA49; PA51; PA52; PA59; PA61
Table 6. Categorization of publications by application purpose.
Table 6. Categorization of publications by application purpose.
TransparencyOptimizationPredictionForecasting
PublicationPA1; PA5; PA6; PB10; PA11; PA12; PB13; PB15; PA16; PA21; PA22; PA24; PA26; PB30; PA31; PA34; PA35; PA38; PB40; PA41; PB43; PA44; PB46; PA51; PA52; PA53; PB54; PA57; PA59PA1; PB2; PA3; PA4; PB8; PB9; PB10; PA12; PA14; PB15; PA16; PA17; PA19; PB20; PA23; PA25; PA26; PB27; PA28; PA29; PB30; PA31; PA32; PA33; PA35; PB37; PA38; PA39; PB40; PA41; PB42; PB43; PA45; PB47; PA48; PA49; PB50; PA51; PA52; PB54; PA55; PA57; PA58; PA59; PA60; PA61; PB62PA12; PB13; PB18; PA19; PB27; PA28; PB36; PA41; PB43; PA57; PA59; PA61PA6; PA7; PA11; PA35; PA51; PB56; PA58
Table 7. Categorization of publications by expert system type.
Table 7. Categorization of publications by expert system type.
Rule-Based ESFuzzy ESML-Based ESHybrid ES
PublicationPA1; PA3; PA5; PA6; PB8; PA12; PA17; PB20; PA21; PA24; PA26; PB27; PA29; PA31; PA34; PB37; PB43; PA53PB2; PA7; PB15; PA19; PA22; PA23; PA28; PB30; PA32; PA33; PB36; PA39; PB40; PB42; PB46; PB47; PB50; PA51; PA55; PB62PA14; PA44; PA60; PA61PA4; PB9; PB10; PA11; PB13; PA16; PB18; PA25; PA35; PA38; PA41; PA45; PA48; PA49; PA52; PB54; PB56; PA57; PA59; PA58
Table 8. Classification of ESs by industries.
Table 8. Classification of ESs by industries.
SectionDivisionPublication
ManufacturingFood and feed manufacturingPA12; PA16
Textile manufacturingPA4; PA7
Chemical product manufacturingPA14; PA28
Plastic goods manufacturingPA26; PA57
Ceramic wall and floor tile manufacturingPA55
Cement manufacturingPA1; PA12; PA23; PA31; PA61
Iron and steel manufacturing and processingPA17; PA21; PA49
Aluminum manufacturing and initial processingPA4; PA48
Metal goods manufacturingPA19; PA24; PA25; PA38; PA39; PA41; PA51; PA52; PA53; PA59
Automotive and automotive parts manufacturingPA33; PA35; PA44
Pulp, paper, paperboard and cardboard manufacturingPA58
OthersPA3; PA6; PA11; PA22; PA29; PA32; PA34; PA45; PA60
Energy supplyElectricity generationPB56
Electric power distributionPB8
Gas productionPB15; PB36
Heating and cooling supplyPB2
ConstructionElectrical installationsPB62
Provision of professional, scientific and technical servicesArchitectural and engineering officesPB20; PB37; PB43
Information servicesData processing, hosting, and related activitiesPB27
Transport and storagePassenger transport by landPB40
HealthcareHospitalsPB47
Others PB9; PB10; PB13; PB18; PB30; PB42; PB46; PB50; PB54
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Ioshchikhes, B.; Frank, M.; Weigold, M. A Systematic Review of Expert Systems for Improving Energy Efficiency in the Manufacturing Industry. Energies 2024, 17, 4780. https://doi.org/10.3390/en17194780

AMA Style

Ioshchikhes B, Frank M, Weigold M. A Systematic Review of Expert Systems for Improving Energy Efficiency in the Manufacturing Industry. Energies. 2024; 17(19):4780. https://doi.org/10.3390/en17194780

Chicago/Turabian Style

Ioshchikhes, Borys, Michael Frank, and Matthias Weigold. 2024. "A Systematic Review of Expert Systems for Improving Energy Efficiency in the Manufacturing Industry" Energies 17, no. 19: 4780. https://doi.org/10.3390/en17194780

APA Style

Ioshchikhes, B., Frank, M., & Weigold, M. (2024). A Systematic Review of Expert Systems for Improving Energy Efficiency in the Manufacturing Industry. Energies, 17(19), 4780. https://doi.org/10.3390/en17194780

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

Article Metrics

Back to TopTop