Recently, manufacturing is undergoing major changes in terms of technology development, environmental concern and digitalization with the existence of emerging technologies, wireless sensor networks, big data, Internet of Things (IoT) and embedded system ushering in the fourth industrial revolution. As a result, the term Industry 4.0 was introduced. It integrates emerging technical advancement to enhance industry so as to deal with some global challenges [1]. In this paper, we will compare the implementation of smart factory in Sweden and South Korea. Then, we will illustrate the main features of Industry 4.0 and explore the framework, technical features and benefits of the smart factory of Industry 4.0.
Implementation of smart factory in South Korea
In South Korea, the country uses ‘The Manufacturing Innovation Strategy 3.0 (Strategy 3.0)’ as an initiative to comprehensively modernize their manufacturing industry. In fact, the Strategy 3.0 is based on direct inspiration from ‘Industry 4.0’ plan. The objective of the Strategy 3.0 is to promote the growth of segments incorporating manufacturing with information technology, examples of which include the integration of information technology into energy management and industrial safety sectors [2].
In South Korea, the smart factory is defined as “a manufacturing system which all business processes of planning, design, production, distribution and sales are automated, connected and integrated by various information and communication technologies. It can produce personalized products by customers with satisfactory time, cost and quality” [3].
CPS (Cyber Physical System) is considered one of the core technologies for developing a smart factory based on the cyber model, DT (Digital Twin) [4]. It realizes an integrated system, monitoring and controlling changes in the shop floor autonomously by merging information and communication technologies with products, manufacturing processes & resources, workers, production cells and lines [5]. Figure 1 illustrates basic concepts and structure of a smart factory with CPS.
Fig. 1. Basic concepts and the structure of ‘smart factory’ with CPS[6]
The country applies smart factory in two directions: ‘Operation Excellences in Manufacturing’ and ‘Personalized Manufacturing for Servitization’ [6]. The first smart factory case is a company which produces engine pistons of autos and targets for ‘Operation Excellences in Manufacturing’. The major processes involve melting, casting, machining and assembly. Due to the characteristics of engine pistons, a permanent mold casting process with accurate dimension control is utilised for mass production. As a result, the company have developed and implemented a Digital Twin model which can analyse and predict errors in productivity, logistics and quality via the use of smart sensors, IoT and big data analytics. The Digital Twin model is shown in Figure 2. After using this model, the quality issue reduced by 26% and net profits increased by 14% in 2017 [7].
Fig. 2. Digital twin model
Another case of smart factory that aim for ‘Personalized Manufacturing for Servitization’ is a micro smart factory for FaaS (Factory-as-a-Service). Figure 3 illustrates a concept of FaaS for personalization. It is also developed and operated using the digital twin model. This type of factory is a completely automated manufacturing system to produce customized products in low volume for individual customers or start-up companies. Besides, it has been reported that the overall cost is reduced by 28% compared to traditional production [7].
Fig. 3. Concept of FaaS for mass personalization, including its CPS system (top right) [8].
Implementation of smart factory in Sweden
On the other hand, Swedish’s effort in implementing smart factory is discovered at Scania, a large truck, bus and engine manufacturer. Scania’s lean based production system principle indicates that the company should always operate with consistent enhancements of production system to increase effectivity and efficiency. An example of its initiative is the establishment of smart truck- and bus production lab. The purpose of the lab is to adapt, evaluate and implement production and logistics technologies at one geographical location [7].
Currently, Scania has developed and evaluated an event driven IT architecture. Collaborative robots, automotive guide vehicles, hand tools and various sensors are connected in order to control equipment and collect data in real time. So far, the company has emphasized on the three lower levels of the Scania’s smart factory pyramid, see Figure 4. Scania is currently targeting for the high levels in the pyramid to be able to analyse, predict and prescribe [7].
Fig. 4. Scania’s smart factory pyramid [7].
According to the overview of national efforts within the two countries, one reflection could be that the South Korea’s initiatives on Smart Factory appears to be more target-driven and emphasized on developing SME’s smart manufacturing capabilities, while Sweden national Smart Factory effort appears to be more diversified and decentralized in supporting smart capabilities to industry in general. Apparently, the degree of maturity varies between companies, starting from companies that are still exploring what digitalization means to them and how they could take some first actions, to companies that integrates digitalization into their business, model and product. Digitalization is considered crucial to achieve increased efficiency and cost reduction and as a way to product revenues via new products and product features, new businesses and business models[7].
Features of Industrie 4.0
As mentioned in previous research, Industry 4.0 has three basic features which comprises of horizontal integration through value networks to facilitate inter-corporation collaboration, vertical integration of hierarchical subsystems inside a factory to create flexible and reconfigurable manufacturing system, and end-to-end engineering integration across the whole value chain to support product customization.[9] Horizontal integration proposes the idea that a corporation, more specifically corporations that are related to each other should not just cooperate but compete with each other in a healthy manner to form an efficient ecosystem. The second feature, which is the vertical integration. It basically helps a smart machine self organise so that they can be reconfigured to adapt to different product types and the massive information is collected and processed to make the production process transparent. The third and last feature as can be seen in the diagram is the end-to-end engineering integration which is basically a chain of activities which are involved in a product-centric value creation. Through the integration process, any continuous and consistent product model can be reused at every stage. Figure 5 clearly shows an illustration of these three kinds of integrations and their relationship to one another.
Figure 5: Illustration of three main features of Industrie 4.0[9]
In light of all the emerging information technology innovations that exists in the world such as big data, cloud computing, IoT and artificial intelligence technologies, they indirectly become enabling factors of Industry 4.0. With cloud computing, almost all our information systems are deployed on the cloud, giving way to an unconventional world of IoT and services that will lay a strong base for the three kinds of integration mentioned. With all these Industrie 4.0 features mentioned, we can see how it expects to vertically integrate the hierarchical subsystems to transfer the traditional factory into the highly flexible and reconfigurable manufacturing system, that we call the smart factory.
Framework of the smart factory
This smart factory is said to be the main foundation that supports the other two kinds of integrations, namely, the horizontal integration through value networks and the end-to-end digital integration of engineering. Figure 6 clearly depicts the framework of a smart factory that has four tangible layers, which are the physical resource layer, industrial network layer, cloud layer and the supervision and control terminal layer[10].
Figure 6: Framework of the smart factory[9]
The physical resource layer contains all kinds of physical artifacts such as smart machines, smart products, and smart conveyors. These artifacts achieve communication through the industrial network, and through this they can collaborate to achieve a system-wide goal. The industrial network layer forms an important infrastructure that connects the physical resource layer to the cloud layer and aids communication among artifacts. The cloud layer is another type of significant infrastructure that supports the smart factory. Since cloud computing became the worldwide trend, even the Internet can be seen as a resource pool. The cloud enables big data application where both storage space and computing ability can be scaled on demand. Technology has become so advanced that sometimes we wonder if a connection between it and people still exists. This is where the supervision and control terminal layer comes in handy. It links people to the smart factory. Examples of components in this layer are PCs, tablets and mobile phones. People can access the information and statistics provided by the cloud, perform maintenance even remotely through the Internet.
Technical features of the smart factory
Through the eyes of a control engineer, this smart factory can be seen as a closed- loop system. The centre of the system is a network of smart artifacts which possess the 3C capabilities, autonomy and sociality[11]. When the artifacts are said to be autonomous, it means that they can make decisions for themselves, no other entities can control its behaviour. These smart artefacts can also understand a common set of knowledge and abide to rules for negotiation. So, imagine a society of these artifacts, the manufacturing system that can exist will be highly flexible, self-organised and reconfigurable, hence the term smart. This closed-loop system can clearly be visualised in Figure 7.
Figure 7: Smart factory in the eyes of a control engineer[9]
The traditional production line is quite different from the smart factory production system. The traditional production line only aims to produce one type of product while the smart factory production system aims to process multiple types of products. Figure 8 will further illustrate these differences.
Figure 8: differences between smart factory and traditional production
Benefits of the smart factory
After all is said and done about the smart factory, what are the benefits that it brings to us? Firstly, it is highly flexible. The smart artifacts can be configured automatically to produce multiple kinds of products. It is able to cope with the ever changing market and consumption demands. It is also robust due to its self-organisation and dynamic reconfiguration nature. As seen in the above figure, the smart factory can produce small-IoT products of different kinds much more efficiently than the traditional production line. This smart factory is also energy efficient. This is said because through big data analytics, people are able to predict the production process and guarantee system with a stable oricut quality level and the rate of finished products. With this knowledge, the amount of raw materials needed can be determined before production and product redundancy can be minimised.
Intelligent Manufacturing
Intelligent manufacturing or smart manufacturing is the following concept. With the advancement of technology, modern manufacturing systems is more complicated. The reason is because the elements are being integrated into common systems. Using human as analogy, intelligent capability makes up of three functions namely, sensing, decision-making and action. Blooming of technologies in areas such as big data analytics and cloud computing greatly improve manufacturing the intelligent capability of products[12].
Big data analytics
Big data analytics brings the meaning of extracting details and knowledge from big data to produce a clear and systematic trends that is easily recognized and more accurate decision can be made. Today’s manufacturing processes enormous amount of data from machines, production, logistics and feedbacks from users. Data like these will not exist during the traditional manufacturing environment. Relevant and core information are suitable to be used from terabytes of datasets in manufacturing with big data analytics. Also, right decision can be made effectively. A proactive decision-making is more prominent in controlling the manufacturing systems.
Predictive analytics for intelligent manufacturing
In the application of big data technology, predictive analytics can be one of the most successful in industrial applications today. It is undeniable that enormous of data is produce throughout manufacturing their systems and the data is not fully utilized else goes to a waste. To date, predictive analytics is the most promising counter to the waste of data. Companies are figuring out analysis algorithms for more effective yields in future predictions of machine. This is to reduce the equipment down time and maintenance can be carried out more efficiently[11].
Application IBM predictive analytics
IBM launched a specific predictive analytics service platform: SPSS Predictive Analytics Enterprise. To objective to provide clear descriptive and predictive analytics with a wide variety of applications for industry like manufacturing, administration management and health management. It is very welcoming to know how manufacturers benefit from IBM knowledge-building engine: IBM Watson Analytics.
IBM states that the SPPS Predictive Analytics Enterprise can achieve benefits as below [13] :
Statistical analysis in descriptive and inference manner;
In numerical, text and graphic data formats for predictive modelling and advanced algorithms is enabled;
Ability for interactive visual and plain-language presentations of data and information;
A secure and automatic data collection and management using framework;
Real-time scoring for the predictive control of company assets.
With cloud computing which allows scalable storage and computing ability, big data is more feasible. We should focus on special features of manufacturing on big data, despite cloud computing is good in this sense. Questions like what data should be collected, how data can be gathered, what the data mean and how to analyse should be answered. For example, as malfunction machines will greatly reduce product quality and also the finished product ratio. Thus, the state of machine and its operation history shall be monitored to forecast the problems in order for the moderator to respond in advance. As a machine moderator, we should take into consideration the processing time of each operation and time taken for each machine to operate. This help to improve efficiency, the bottleneck performance and load unbalance of machines.
Conclusion
To conclude this article, we can see that the smart factory approach the fourth industrial revolution is taking will somehow solve all the problems that innovators have been facing during the first three industrial revolutions. With the existence of cloud computing, and all these new emerging technologies, we can save more time and resources efficiently without harming the environment and spending too much money. We can also clearly see how well the countries that have implemented smart factory are doing in terms of production efficiency. The next move would be to further promote smart factory to benefit third world countries as the current rise in information technology have taken a toll on these countries the most. We should make the world aware of such methods to further benefit the world, not just modern countries.