Digital twins-based smart manufacturing system design in Industry 4.0: A review

The general birth process of a manufacturing system can be described as follows: a designer generates an idea in his mind, then draws it in drawings (forms a model) and tries to make some physical prototypes for testing, and finally assembles the system. Table 1 provides an overview of the manufacturing system design in different stages of industry development from Industry 1.0–4.0.

In Industry 1.0 and 2.0, the information in the R&D stage exists in the drawings; and in the physical prototype stage, the information uses the physical system as the carrier. The conventional manufacturing system design method requires physical prototypes to be assembled to check and verify whether the virtual space model is accurately matched, including structure, ergonomics, and performance indexes. However, using physical prototypes to verify the design is very costly. To avoid unnecessary risks on physical prototypes, simulations of the performance and manufacturability of a model in the virtual space are critical [1].

In Industry 3.0, many Computer-Aided Design (CAD) software providers (e.g., PTC, Siemens, ANSYS, Dassault) proposed the concept of virtual prototype, digital prototype, active prototype, and others. A digital prototype is a kind of rehearsal in which the information model replaces physics. Among various digital prototype concepts, the Digital Mock-Up [2] concept is widely used, which emphasizes complex mappings and contextual relationships among 3D model simulations. The existence of digital prototypes greatly reduces the failure of physical prototypes (sometimes they simply directly replace real prototypes). Based on a virtual simulation of the physical manufacturing process, the SMS performance can be evaluated early to guide cutting down the reconfiguration costs/losses in establishing physical SMS prototypes (e.g., Plant Simulation). Using virtual reality could greatly enhance the convenience and efficiency of SMS design (e.g., FlexSim). Due to the consistency of information transmission in digital prototypes, the difficulty of the manufacturing system design is greatly reduced [3].

In Industry 4.0, smart manufacturing has become a development direction of the world's manufacturing industry. New national advanced manufacturing strategies around the world resulted in the increasing need of designing new smart manufacturing systems. A Smart Manufacturing System (SMS) is a multi-field physical system composed of intelligent machines, materials, products, and complex couplings among various elements. In the digital design process, an SMS can be broken down into digital models of various granularities in digital space, while the physical products and manufacturing processes exist in another physical space. In the SMS design process, high-fidelity cyber models mapping the real worlds of SMSs are critical to fulfilling the gap between its design domain and operation domain [5].

Digital twins (DT) technology is a solution. Nowadays, the digital twins trend is gaining momentum because of rapidly evolving simulation and modeling capabilities, better interoperability and IoT sensors, and more availability of tools and computing infrastructure (www2.deloitte.com/). Recent MarketsandMarkets research suggests that the global digital twin market size was valued at USD 3.1 billion in 2020 and is projected to reach USD 48.2 billion by 2026 (www.marketsandmarkets.com). A digital twin could realize the feedback of the digital model of cyberspace to the real physical system. In this way, it is possible to ensure the coordination of the digital and physical spaces within the scope of the entire life cycle. Zhang et al. [5] firstly applied the digital twins-based approach in the manufacturing system design sector. After a review of existing SMS development models, Mahmoud and Grace [4] suggested a digital twins-based simulation approach for SMS configuration. Because of the supremacy of digital continuity, realistic modeling, and interaction between disciplines against conventional simulation, the creation of a digital twin in the SMS designing phase will be highly valuable during all the lifecycle of the SMS.

However, the application of the digital twins concept in the SMS design remains vague [6]. Therefore, it is necessary to explore and sort out the research on the rational integration of digital twins technology into the SMS design. This article attempts to show how digital twins technology is integrated into the SMS design to truly promote the development of smart manufacturing. A literature search on digital twins-based SMS design is conducted in the Web of Science database. The collected literature is further refined by following three phases. The first phase is to screen related papers via keyword retrieving. Two combinations of keywords, namely, {“digital twin” and “manufacturing system design”} and {“digital twin” and “manufacturing system planning”}, are used for retrieving papers. The retrieving yields 202 papers and 54 papers, respectively (up to 31 December 2020). After the deletion of duplicated papers, this retrieving yield 220 papers for further screening. The second phase is a theoretical screening process to identify high-quality studies related to the concept, key enabling techniques, models, systems, frameworks, and case studies on SMS design. Research not related to the digital twins or SMS design domain and short papers less than four pages are excluded. In this phase, 159 papers are finally included to discover the key issues, new solutions, challenges, and promising directions in the research of the digital twins technology in the SMS design process. Additionally, 33 additional papers are referred to make this survey more comprehensive. Finally, this survey includes 192 references, and the statistics of collected literature are shown in Fig. 1.

The definitions, frameworks, major design steps, new blueprint models, key enabling technologies, design cases, and research directions of digital twins-based SMS design in industry 4.0 will be presented in this survey. The organization of this survey is outlined in Fig. 2.