This section presents the results divided into two parts. In the first part, a bibliometric analysis is carried out to answer the questions, Q1, Q2, Q3, and Q4. In the second part, a content analysis was conducted in order to answer the questions Q5, Q6, Q7, and Q8.
4.2. Content Analysis
The number of articles where it was possible to see each reference to I4.0 technologies was considered, so that the most relevant I4.0 technologies could be established. It should be noted that only I4.0 technologies with a frequency equal to or greater than 10% were considered, which corresponds to a minimum of 5 articles where the technology was referred.
Table 4 summarizes these same technologies. The top 3 most referred technologies were Internet of Things (with 47%), Big Data Analytics (with 43%) and Cloud Computing (with 35%). These three technologies are indicated as the most applied by lean, which allows for the elevation of fragility more than conscious of lean. According to the literature, it is said that lean demonstrates a weakness with data processing, since it brings together several tools for its extraction; however, they are only used to monitor and not improve processes. Thus, the application of IoT in lean tools improves the capture of this data, Big Data Analytics comes to treat and give meaning to these data that are in their most primitive state, and cloud computing offers a place to store this created information.
As in RQ5, the number of articles that referred in their content to each of the technologies belonging to the I4.0 paradigm was analyzed; here, the same was carried out; however, the scope of the scrutiny was based on lean tools. The ones that demonstrated just one article where the tool was referred were not considered (corresponding this to about 2%).
Table 5 summarizes these tools. Here, the top 3 most referred technologies were just-in-time (with 18%), value stream mapping (with 18%) and heijunka (with 16%). The existence of more references in these tools may constitute the conclusion that they are seen in the academia as being more viable and easier to integrate with the I4.0 paradigm. The 5S (Sort, Set in Order, Shine, Standardize, Sustain), for example, which appear at the end of the table, can nevertheless be considered a very mechanical tool, where the application of technology does not bring added direct benefits.
Based on content analysis, this section will highlight the impact of 4.0 technologies on lean tools. Every lean tool was evaluated by studying the combination with I4.0 technologies.
In visual management 4.0, the automated acquisition of data (using IoT) saves time for managers and employees since boards can be automatically updated with information (with pre-processed data) [
45]. Smart visualization abilities emerge from this combination, and entire processes and activities can be visualized across the supply chain, allowing a better risk management, predicting future incidents [
36]. This is possible to use for not just IoT, but also the cloud which makes information available to all the right people. Big Data Analytics can be used to extract and process the data collected, convert it in information and the augmented reality (AR) provides and presents the visual information to managers and/or employees [
8].
In Wagner et al. [
30], just-in-time is integrated in an IT-system to support a lean Just-in-time materials flow process (cyber-physical just-in-time delivery). In this project, Kanban cards were switched by a vertical integrated solution, creating a gapless information flow between manufacturing order, material supply, material stock, and material consumption, not forgetting the computerized purchase order to the supplier. Sensors detect every material movement (IoT tracks products in real-time) [
8], displaying the information into a basic big data architecture, that in connection with the material consumption of the manufacturing machines, can send an automatic order to the supplier, each time the minimum inventory level is hit. An analytics service on statistical data was aggregated here in order to have a prognostic of material requirement, comparing the available information with a digital model of the complete process [
30]. Robotics, more specifically collaborative or even autonomous robots, are able to adjust the productive flow and act promptly, ensuring that production runs smoothly [
8]. With 3D Printing, since it can be mounted near the customer’s location, it is possible to reduce distance and delivery cost, which enhances the JIT principle, decreasing lead times and augmenting logistics performance [
1]. CPS-based devices will be able to provide information about cycle times to operators using it for that augmented reality (AR), which will support the performance of JIT tasks [
1]. Therefore, JIT can be of benefit alongside I4.0 in a way that provides visualization of the entire supply chain, improved demand forecast and accuracy, responsiveness to changes, and superior inventory management and control [
36].
Several studies have been already investigated the digitization of conventional Kanban cards, thus emergence of the e-Kanban system. With this system, missing or empty bins can be exposed, and replenishment can be triggered automatically [
42]. In Bittencourt et al. [
20], a case study is cited, carried out by the Wurth Company, which introduced an order replenishment system based on Kanban baskets. The new program can send orders automatically to suppliers, decreasing, in that way, stock and consequently space clearance on the shop floors occurs. Above that, orders are concise with demand. Another study in the Wittenstein Company uses automated guided vehicles (AGVs—included in the Robotics group) which provide and establish the milk-run system-based interval via real-time demand [
20], possibly changing the e-Kanban system. Additionally, Cloud Computing can be of superior interest when it is necessary to have an exchange platform to facilitate JIT supply between the producer and the supplier [
8]. With the implementation of a system like the ones mentioned, lost Kanban will not cause problems anymore, and modifications in kanbans due to shifts in batch sizes, work plans, or cycle times will be easier [
42,
43].
A CPS-based jidoka system was already planned and implemented by Ma et al. [
29]. It is considered a distributed system self-possessed of analogue and digital parts, actuators, controllers, ICT, software, and jidoka rules. Pereira et al. [
1] also cited an integrated and standardized approach to implement and design a CPS-based smart jidoka system, which was a system mainly based on CPS technology, including other technologies such as Cloud Computing and IoT, capable of allowing data collection of resources and flexible configuration of the system itself. Rosin et al. [
8] also cited the use of autonomous robots capable of detecting and correcting production errors, which makes part of the jidoka principle. Thus, the defect detecting process is carried out with more accuracy, the identification of any errors is supported in real-time, preventing them from moving to the next process, and it is easier to manage the identification of the causes of any errors occurring [
36]. Together and closely linked to the principle of jidoka, there is the poka yoke lean tool, which is improved and more effective, using the technologies mentioned for jidoka. Haddud et al. [
36] analyses this tool together with jidoka, and the benefits encountered were the same as the mentioned above. Although, some authors show some reluctance with the relation between simulation and poka yoke, since the first one can foresee potential difficulties and mitigate failures in the production process; however, it does not avoid errors (which is poka yoke’s goal) [
2]. A particular complementarity between these two methods can be achieved, even though practical application is needed [
2].
In the Andon principle, IoT provides products to connect with equipment and send a warning once the incorrect product is being produced, offering the capability of the equipment to react to errors, discontinuing the work or changing products [
8]. In Pereira et al. [
1] and Kolberg et al. [
43] it is mentioned that the use of CPS-based smart devices (smart watches, SMS or even email, for example) by operators provides the reception of error messages in real time, alerting the operator in case of failure, prompting repair actions, and reducing delay times due to failure incidences. With this approach, recognizing failures will not depend on location of employees. In a more standardized environment, CPS will be capable of automatically trigger fault-repair actions on another CPS [
43]. Alert notifications becoming more frequent can occur, notifying a possible failure for the system. In that way, it will be important for PLC’s to be programmed to generate an alarm whenever any sensor value captured is outside of the tolerance or even when the rate of the recurrence of the number of alerts goes beyond a certain limit [
24]. For this to happen, the combination with artificial intelligence would be important, in a way that it could be confirmed, based on historical data, if the problem would be on the product or machine.
For the heijunka principle, Ante et al. [
49] reveal some projects in the I4.0 context. One of them is related with the construction of a digital heijunka board. It is intended that the system automatically creates the production Kanban cards and places them in the build to order slot of the Digital heijunka board, which has been developed following the standard leveling rules for assembly. The logistics department sends the assembly program automatically to the system board and leveling performance is calculated automatically. In Kolberg et al. [
42], another project is revealed by the Wittenstein Company which digitized their heijunka board. Graphical user interfaces connected to the production line and MES are displayed, which contribute to diminish information flows and efforts for updating the board. Pekarčíková et al. [
35] exposes some relations between heijunka and I4.0 technologies, such as augmented reality (AR), virtual reality (VR), horizontal and vertical integration, Cloud Computing, Big Data, data analytics and IoT. The last five can be assumed by the examples listed above, although the first two still need practical application. However, it is appropriate to extrapolate the use of AR and VR to display the heijunka board, eliminating the one that is physical. Pereira et al. [
1] sums up all the seven I4.0 technologies as artificial intelligence, arguing that this technology is indicated to provide analytical support in the decision-making process and apply intelligence environment approaches that allows complex analysis and learning.
Value stream mapping (VSM) is the lean tool with more practical applications, and because of that has more evidence of the capabilities gathered through the connection with I4.0 technologies. Phuong et al. [
50] developed sustainable value stream mapping, which involves three dimensions above the traditional one; they are economy, societal factors, and environment. I4.0 technologies were integrated in this tool, such as RFID, providing real-time tracking, allowing, at the same time, employees to be more quickly reactive to potential incidences. Additionally, Big Data collected by the real-time tracking of SVSM can be used for forecasting reasons, possibly preventing waste in resources consumption and any damage to workers. Molenda et al. [
15], inspired from value stream mapping 4.0, suggested a new methodology for the visualization, analysis, and assessment of information processes in manufacturing companies—The VAAIP mapping. For the visualization, quantitative measurements and a qualitative analysis were carried out. Ramadan et al. [
32] presented a real-time scheduling and dispatching module (RT-DSM) that traces the flow of products and detects the incompatibilities and inconsistencies between the physical and virtual world that are caused by lean waste. This module runs on dynamic value stream mapping (DVSM) to prevent a frozen production schedule, producing appropriate reactions and directives to be executed both by machines or a human to relieve the impact of incidents and try to match up the virtual value stream mapping with the actual value stream mapping. Huang et al. [
33] also designs a DVSM version that is included in a cyber-physical multi-agent system that real-time and virtual attributes make visible the conditions of material, workforce, and machine. The DVSM is considered by authors capable of providing valuable information for the decision-making process. Therefore, DVSM or VSM 4.0 provides real-time data which allows appropriate action in the right time, overcoming the static behavior of VSM and, above that, the current value stream can be constantly displayed and bottlenecks as well as improvements continuously ascertained, which facilitates the implementation and concretization of Kaizen activities [
7,
23]. Balaji et al. [
23] also refers the enhancement of the team’s morale as an advantage, since they are able to see the results of their kaizen activity very quickly and suggest the standardization of measurement methods across the organization.
Gambhire et al. [
24], Pagliosa et al. [
2], Pekarčíková et al. [
35], and Wagner et al. [
30] considered that the 5S lean tool can usufruct the I4.0 technologies’ integration, believing that Virtual Reality and Augmented Reality are the ones capable of having a greater impact. This can be explained by the fact that this lean tool is still some kind of mechanical, which just can be solved by I4.0 tools capable of representing it in a virtual world, in order to facilitate the shop floor disposition.
Kaizen strategies relay profoundly on well-timed detection of errors and abnormalities all over the processes and supply chain operations [
36]. Because of this, some tools mentioned above, such as jidoka, VSM, heijunka, kanban, and andon, properly integrated with I4.0 technologies can be a source to provide insights to adopt kaizen strategies or even can be the kaizen strategies themselves. In that way, a totally integrated production system will actually increase the value chain’s performance and the responsiveness of the whole system [
12]; make it easy to catch, process, and distribute information to the right people, permit suppliers, and customers participation; and make timely improvements, since identification of errors is easier and promptly [
36].
Key performance indicators (KPIs) are metrics often used by lean. Pereira et al. [
1] makes reference to a case in automotive electronics production where the main problem was the missing traceability for the shop floor KPI reporting process. To apply a solution, data analytics and a cloud solution were essential to process the live data collected from all lines in the production network. Abd Rahman et al. [
38] also applied simulation in their work, but the integration was more specific, since the overall equipment effectiveness was the KPI metric chosen. This metric is the most used in companies to measure their efficiency.
Ayabakan et al. [
13], as already mentioned, analyzed a digitalization of a Kanban system, but above that, an automatic change over system was the focus of attention too (based on single-minute exchange-of-die—SMED—lean principle). The system uses RFID (radio frequency identification) to recognize each die and know their storage address. It was concluded that, with this system, an increase of the line’s productivity was felt, as well as its capacity, reducing, at the same time, the number of production people. Pagliosa et al. [
2] supports the idea that IoT can have a crucial impact in the execution of adjustments and setup of workstations, but states too that the integration of robotics with SMED can cause conflicting efforts for operational improvement. This can happen because elevated levels of robotization and automation can create less flexible production lines, limiting the customization of products and weakening changeover time. However, on the other hand, the authors assume the capacity of carry complex activities with the utilization of advanced robotization. Therefore, more study and practical applications are needed in this aspect, to converge results.
The total productive maintenance (TPM) is another lean tool that has been recently attracting attention for integration with I4.0 technologies. Big data analytics, cloud-based systems, and IoT enable real-time information and data that can support productive and preventive maintenance [
36]. Sensors produce data which are then contrasted to the information from the machine and the specific workpiece being processed, allowing to continuously keep in check and predict the incidence of failures, as there are multiple signs and tendencies that the component demonstrates “symptoms” of forthcoming failure or degradation in performance [
24,
30]. The timely information sharing, and real-time data provides better inventory management and shorter downtimes [
24,
36]. Smarter maintenance is capable of guarantee better processing equipment performance and fewer defects, which aids an increase in the product’s quality [
41]. With prompt information about the equipment’s state and properly triggered repair actions, a smart planner can be easily updated on reconfiguring production lines and updating kanbans in real-time, based on changes [
1]. Pagliosa et al. [
2] refers to the connection between TPM and AR as being explained by the support in performing maintenance remotely through knowledge sharing and technical guidance. Marcello et al. [
51] carried out a construction of an ensemble-learning model that combines prediction results from multiple algorithms, using big data analytics, to estimate failure rates of equipment subject to distinct operating conditions (reached an accuracy value of 96.15%). On the other side, Passath et al. [
52] created a standard criticality analysis as a foundation of an agile, smart, and value-oriented asset management system to dynamically adjust the maintenance strategy. It was concluded that the more complex and disparate the assets are, the more essential it was to have a guideline to dynamically adapt the maintenance approach due to the environmental variations as well as production circumstances.
After analyzing each of the lean principles and their correlation with I4.0 technologies, it can be concluded that lean and digital technologies support organizations in becoming faster, more efficient, and more economically sustainable [
11,
53].
Since waste reduction is the lean’s main goal, an overall analysis about the impact between the seven wastes and the I4.0 introduction should be done. Overproduction can be reduced in the I4.0 context, since a better order management is provided and information is communicated through the shop floor directly and constantly [
41]. The waiting time is able to be decreased too as smarter decisions are made on site and feedback from related stakeholders can be received by vertical or horizontal levels [
41]. Here, the horizontal and vertical integrations will have a huge impact in identifying waste, as well as big data which can be used to detect, in real time, unusual situations in the production system and identify the root causes of these conditions [
8]. IoT is already assumed as an important tool to reduce transportation, since it takes advantage of real-time product tracking to see unnecessary transportation [
8], and robotics and infrastructures brought by I4.0 support the transport by itself and even the calculation of best routes [
41]. Simulation allied with augmented reality or virtual reality is a possible resolution to over-processing, because it allows the replication of scenarios for testing ideas, providing managers space to choose the most promising ones. Additionally, the defects and unnecessary processes can be minimized with this tool, as a copy of the production system can be constructed, and several scenarios can be provided to solve production problems [
8]. Additionally, the digitalization of value streams provides real-time feedback, allowing the control of processes’ efficiency [
41]. A better control of production and raw materials is given by the connectivity between customer, supply chain, and individual processing equipment, which offers the ability to decrease the inventory [
41].
Besides, since lean puts people at the center of almost every strategy, the I4.0 impact in the way collaborators do their work is essential to be understood. Augmented reality, for example, is considered by Rosin et al. [
8] and Dutta et al. [
28] as a useful tool to learn new processes, perform material audits, and, further down, to execute on-site maintenance tasks, allowing the sharing of knowledge with other employees. Simulation is another tool considered to be able to validate human operations and train new employees (in conjunction with AR and VR) [
2,
8,
10,
28]. I4.0 pretends to change the workers’ role from machine operator to augmented operator, whereby the main position is supervising the work (while it is being performed by the machine) [
12].
Although, before introducing I4.0 technologies, companies should appropriately weigh the maturity of their organization [
21], paying special attention to structure, jobs, and competences. Experimentation is essential to recognizing the interventions at both technological and organizational levels and companies must never misjudge the time required [
17].
The matrix presented in
Table 6 arose from a unification of the content of articles subject to the systematic review. In this way, and with a view to a consistent follow-up, the vertical axis refers to I4.0 technologies identified as being the most relevant and most referred in the articles. The vertical axis refers to lean practices, which were also presented as being the most significant. Every lean tool was evaluated by considering the combination with I4.0 technologies, which resulted in the introduction of an “x” in case of existing some kind of impact between the two (I4.0 technology and lean practices).
The two lean practices that can most benefit from the contribution of I4.0 technologies are just-in-time and standardization, as can be seen from the
Table 6.
In the previous chapter, I4.0’s contribution to lean was analyzed. Therefore, it is time to analyze the reverse.
Architectures for establishing dynamical and self-controlled Industry 4.0 productions centered on CPSs exist, but they are predominantly high-level methodologies focus barely on technology approaches. It has been said in academia that smart factories must consider the technology perspective, as well as the organizational and human point of view [
42].
According to Tortorella et al. [
4], Rauch et al. [
54], Tortorella et al. [
5], Erro-Garcés et al. [
31], Rossini et al. [
3], Bittencourt et al. [
20], and Rossini et al. [
27], lean practices implementation shows a great potential of a higher adoption of I4.0 technologies. This is assumed because of the solid behavioral and processes foundation that lean can provide. However, Tortorella et al. [
4] considered that, to implement I4.0 technologies, it is necessary to have a minimum maturity level of lean practices. If this lean maturity level is not adequate, results of the implementation of industry 4.0 practices will be below expectations, producing management frustration and financial waste. Tortorella et al. [
5] discuss the association regardless the companies’ size and the conclusion is that size is not a barrier for implementation of I4.0 technologies. Nonetheless, Rossini et al. [
3] establish the independence of lean practices adoption from the presence of I4.0 technologies. Lean practices adoption effects still prevail over the impact of I4.0, and this happens since companies’ perception and implementation maturity with respect to lean are considerably larger than I4.0.
There are already some more concrete contributions from lean to I4.0. It is the case for Rosienkiewicz et al. [
19], whose contribution settles in the conception of a lean production control system that uses the Glenday Sieve lean tool (“states that a small percentage of procedures, processes, units or activities account for a large portion of sales, and includes a color-coding system for labelling processes by output volume” [
19]) and I4.0 technologies, such as artificial neural networks (ANNs). The lean tool was used to establish which type of products were able to be predicted by ANNs.
Bittencourt et al. [
20] recommend having a framework for the implementation of I4.0 technologies in a production system and that it must adopt tools such as process standardization and production flow, intrinsic to lean, which will ensure transparency of the process and gain of productivity. Qu et al. [
55] creates a framework to overpass gaps between requirements for traditional manufacturing systems and smart manufacturing systems. Some of the design requirements settle in lean principles, such as standardization, in which it pretends to establish a data dictionary, uniform document format and sheet design, and standardizes the database.
Constantinescu et al. [
56] suggest another approach, which the authors called just-in-time information retrieval (JITIR), founded on using, as an input, the users’ environment and activity, and delivering, as an output to the user, information reclaimed, proposing documents which potentially match the users’ concern. Furthermore, JITIR agents are capable of being performed automatically and; therefore, decrease considerably the cost of search, behaving as a time-saver search. Other authors carried out the analysis of the importance of using the JIT principle to display information, since they agree about the power of having information displayable, reusable, and provided to the right person, in the right format, at the right moment [
14,
57]. Lean information management (LIM) is a management practice enhanced by Teixeira et al. [
57] as a benefit for systems such as enterprise resource planning (ERP) and manufacturing execution system (MES), because of its capacity to eliminate all waste in terms of avoidable data and processes.
Bittencourt el al. [
20] also conclude that, on one side, lean thinking focuses on waste reduction, while, on the other side, I4.0 concentrates on the use of new technologies driven by IoT. However, with different tactics, the concepts can and ought to be complementary, since the implementation of lean will inspire a company to stimulate thinkers who will be necessary in implementing the changes needed by I4.0.