Application Areas and Directions of the Industrial Internet of Things

Application Areas and Directions of the Industrial Internet of Things

To some extent, the Internet of Things can be described as a concept of “new wine in old bottles.” Proposing the idea of IoT (Internet of Things) was indeed innovative, but there is no sharply defined line separating it from traditional M2M—machine-to-machine communication. The greater difference lies in a change in mindset. In M2M, machines are the main actors, and networks are used to connect machines with sensors. As internet technology has advanced, M2M has also reached toward the internet, with web-based publishing and remote access making it feel increasingly similar to IoT. But M2M is ultimately not the same as IoT. As a shift in thinking, IoT works in the opposite direction: it starts from the network side and extends down toward devices and sensors. The difference between the two is a bit like the chicken-and-egg question.

Take a factory monitoring system as an example. Under the traditional way of thinking, there are first the machines, then the necessary control systems, and then distributed controllers such as HMIs or industrial computers. As computer and network technologies advance, the system can be upgraded by centralizing the controllers into one central control system and adding more sensors to collect more information. This is the M2M way of thinking. On top of that, network publishing functions can be added so that users can remotely view the actual operating status of the system. In this way, the equipment is connected to the internet, but the mindset is still M2M: the architecture is fixed, the objects are defined, the industry of application is fixed, and the scale of system expansion is predictable. These are all characteristics of M2M.

The IoT mindset is the opposite. It begins with a general-purpose or domain-specific internet-based interface. For users, this may still appear as a website, a standalone app, or something similar—functionally not too different from the network-facing side of M2M. But behind this network portal lies a series of complex applications, data analysis systems, machine learning algorithms, and supporting architectures designed to analyze and process large volumes of collected data. Data acquisition becomes another layer beneath this. The system does not necessarily need to centralize the previously distributed control systems, and it does not even have to be directly linked to the original control system. Through embedded sensors and distributed microprocessors, together with technologies such as 4G, Zigbee, or Bluetooth, devices and systems that need to be connected can be brought onto the internet. This connection is not built specifically for one workshop or one system through a separately written program. Instead, it is achieved through standardized sensor terminals. For different types of equipment, it may only be necessary to change the type and configuration of the terminal sensors.

IoT may be applied to a single workshop or system, or it may be deployed on a large scale across part of a power grid. Its growth is, to some extent, difficult to predict accurately. Therefore, it is essential from the outset to adopt scalable distributed architectures, cloud computing, and big data technologies, distributing the system across large computer clusters. This is also precisely the direction that modern internet architecture must consider from the very beginning.

At present, the industrial IoT is still far from where it should be. According to estimates, the output value of the power IoT sector is expected to exceed 24 billion USD by 2024, while the current figure has only just barely surpassed 1 billion. If we continue to use the traditional M2M mindset, the gap between industrial IoT development and internet technology will only continue to widen. Only by adopting an IoT way of thinking is it possible to dramatically narrow that gap. This may take 10–20 years, but it is more likely to take 5–10 years. Given the explosive progress of network technology, this timeframe could even shrink rapidly. For professionals in the field of industrial automation, if we want to keep up with this trend, the room and time left to us are extremely limited.

The foreseeable development areas for industrial IoT are concentrated in energy, manufacturing, the automotive industry, and smart cities.

In the energy sector, the construction of distributed power generation systems, microgrids, and energy storage networks all leave broad room for the application of IoT technologies. As electric vehicles increasingly replace fuel-powered cars in countries around the world, upgrading the power grid has become urgent. To cope with the enormous impact that EV charging will place on the power supply network, buffered energy storage systems and intelligent planning of energy distribution are indispensable. In manufacturing, rising labor costs and intensifying competition have created an increasingly urgent need for production tracking, production management, and product quality traceability. This, in turn, is also promoting the rapid development and adoption of industrial IoT.

In the automotive sector, Tesla and Apple are two major representatives driving IoT technology forward. Replacing outdated traditional vehicle operating systems with widely adopted smart interfaces is the most basic expression of automotive IoT, while its most advanced development is undoubtedly autonomous driving. Compared with the previous sectors, the development of smart cities remains more at the conceptual stage than the practical one. This is because, compared with any single specific industry, smart cities require far more technologies and much more complex implementation. In a sense, they are the culmination of all the earlier fields. But given that IoT technologies in various sectors are still not yet mature, smart cities are clearly unable to achieve any fundamental breakthrough or progress at present.