Industry 4.0 solutions for preventive engine maintenance: a systematic review

Introduction

The integration of digital intelligence into products, machines, and equipment in the form of cyber-physical systems requires the adaptation and reconfiguration of production processes (BMI, 2020). Industry 4.0 is an umbrella term applied to the intelligent networking of machines and industrial processes with the aid of information and communication technology (BMI, 2021) and has been formulated as a catchword for industrial development programs of the German government and OECD (OECD, 2017). The autonomous, computer-assisted interaction and communication of industrial machinery, in brief, smart factory designs, promise to dynamize and modularize production, optimize logistics, monitor and control production processes, individualize outputs to customer requirements and define sustainable product lifecycles (BMI, 2021).

Predictive industrial maintenance refers to all processes of monitoring, analysis, and amendment applied before machinery or digital processes fail to retain and grant their full functionality (Zhong et al., 2017; Oztemel and Gursev, 2020). To integratedigital self-optimizing industry 4.0 elements in industrial production processe a paradigm shift of the maintenance and support process chain is required. At the same time industry 4.0 technologies promise to revolutionize maintenance and repair systems for industrial production (Reinhart, 2017, 409). To avoid downtime and bottlenecks in autonomous production networks, preventive maintenance is mandatory, intelligent digital technologies could enable industrial production systems to self-organize and self-implement their maintenance processes (Oztemel and Gursev, 2020).

Industry 4.0 is also, undoubtedly, a fashionable buzzword and a series of publications in preventive industrial maintenance use it (Sony & Naik, 2019; Bongomin et al., 2020). However, a comprehensive overview of relevant preventive maintenance approaches does not yet exist. Integration of the fragmented digitalized solutions for preventive maintenance as postulated by the industry 4.0 principle, is outstanding. Earlier reviews (Zhong et al., 2017; Oztemel and Gursev, 2020; Çınar et al., 2020; Bueno, Filho, and Frank, 2020) in industry 4.0 technologies are more general, i.e. not specialized in preventive maintenance and accordingly provide no in-depth data on the above points.

Methods

This study uses a systematic review of empirically proven technologies to develop an integrative model of preventive maintenance applicable and adjustable to diverse industries using industry 4.0 standards.

First eligibility criteria for study consideration have to be specified: The review includes publications in English language in academic peer-reviewed journals, as well as conference papers published since the initial public discussion of the term industry 4.0 in 2017. Studies published as working papers or final theses in a university context as well as books and book contributions are excluded since these are not peer-reviewed. Publications in languages beyond English and publications not available in full text have to be excluded, to ensure a detailed analysis of the content.

Second information sources the considered articles are retrieved from, have to be specified: to ensure the academic focus of results the study focuses on academic databases. Three major academic data bases are searched for relevant contributions: Web of Science, Scholar Google and Ebscohost. The search was done in June 2021 and the review thus considers publications until 2021.

A homogenous research strategy is applied across all three databases. It uses the following uniform keyword combination to retrieve appropriate studies: “industry 4.0” AND “preventive maintenance” OR “preventive repair” AND empirical. The research is limited to the period 2017 to 2021 to ensure topicality of results in the dynamically evolving field of research. Only studies discussing the issues of preventive maintenance or preventive repair in an industry 4.0 context and within the framework of an empirical study are retained for further evaluation. The titles of the initially retrieved 127 studies are screened manually for relevance to the review. After removal of 52 duplicate or not eligible studies 75 studies remain. Based on a scan of study abstracts 22 studies are deselected for content reasons. 53 studies remain to be sought for retrieval but only 25 studies are openly available in full text from the databases. Textual analysis is done for these studies. Due to lacking applicability four further studies are excluded, six evaluated studies contain no new information. A total of 15 studies were included in the final review. Figure 1 shows the PRISM flowchart (Fauska, 2022).

Figure 1. Identification of studies via databases (Concept of PRISMA Statement 2020).

The data interpretation process is based on a methodology suggested by Webster and Watson (2002). No software is required for this method. First the studies are catalogued in table form in alphabetical order of authors in the form of a content matrix. It summarizes discussed technologies, fields of application, potentials, limitations, and success factors of the respective technologies (Box 1). The textual evaluation (presented in section “Results”) is then organized by order of technologies and adopt a classification described in Deloitte (2017) and Çınar et al. (2020). Tt includes the sections “sensor equipment”, digital networks, technology integration and augmented behavior. Each section describes a technological subject field of predictive maintenance and repair. All four technological fields interact as shown in the section “discussion”.

The described review method avoids important biases of literature-based research, namely reporting, selection and interpretation biases by applying a coherent set of rules for study retrieval, selection and interpretation. However, some potential biases remain: Since there is only one coder in the person of the author, subjective biases and errors cannot fully be excluded. Further databases and key word combinations could deliver further results however cannot be considered here due to limitations in publication length.

Results

Predictive maintenance systems rely on predictive information to organize maintenance requirements (Jardine, Lin, and Banjevic, 2006), which is (1) gathered from decentral sensor equipment, (2) collected and integrated into digital networks, managed and evaluated in operation centred which rely on (3) technology integration and (4) augmented behaviour (Çınar et al., 2020):

Augmented behavior

(4) Augmented behavior refers to virtual computing and service applications administering the machine-human interface (Çınar et al., 2020). Preventive maintenance and repair information provides a reliable data basis for operative human repair and exchange activities (Saxby, Cano-Kourouklis, and Viza, 2020). Chen et al. (2017) see the key action field for preventive maintenance to be the level of data application. The analysis of data gathered from diverse sensor networks is integrated using a data mining strategy and enable active maintenance processes. Augmented reality applications enable humans to understand and control maintenance processes (Longo, Nicoletti, and Padovano, 2017; Bueno et al., 2020),

Integrating these points and referring to the Total Quality Management “Plan – Do- Check- Act” cycle (Johnson, 2002) a circular model for the integration and use of industry 4.0 technology to preventive maintenance processes results:

Figure 2. Smart preventative maintenance cycle.

Discussion

In the following discussion the opportunities, limitations and success factors of preventive maintenance using industry 4.0 technologies are discussed controversially.

Conclusions

Huge initial investments into a comprehensive ICT grid are required to comprehensively collect all necessary repair and maintenance data and evaluate these data (Sahal, Ali, and Breslin, 2020) based on joint information models (Jung et al., 2017).

Further inter-industry research is desirable to augment knowledge on the efficiency of smart maintenance in an industry 4.0 environment and to develop general industry standards.

Data availability

All data underlying the results are available as part of the article and no additional source data are required.

Reporting guidelines

FIGSHARE: PRISMA checklist and flowchart:

https://doi.org/10.6084/m9.figshare.20363727.v1 (Fauska, 2022)

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).