Industrial IoT refers to the application of IoT technology in industry to achieve value-added technical models unique to industrial settings.
All IoT aims to achieve interconnectivity, especially between objects, but Industrial IoT has its proprietary attributes. The reason is that the connectivity density, real-time connectivity, and heterogeneous requirements of consumer IoT, which is the counterpart of Industrial IoT, are not high, while the requirements of Industrial IoT mainly manifest in three aspects: connectivity density, real-time connectivity, and heterogeneous connectivity.
Thinking through all problems requires a refinement process from macro to micro, and Industrial IoT is no exception. I believe that in-depth thoughts on Industrial IoT need to be analyzed from the following five dimensions; otherwise, it may either lead to a narrow perspective or lofty aspirations.
The first question we need to consider is: what is the value, significance, and purpose of Industrial IoT? The second is what needs to be connected in Industrial IoT; this is a concept of scope. The third question we need to think about is the hierarchy of objects connected to the IoT, which is a question of depth. The fourth question we need to consider is the value-cost analysis of achieving IoT. The fifth question is how to build Industrial IoT.
First, what is the value, significance, and purpose of Industrial IoT?
The internet has achieved connections between computers or, in other words, connections between people. This connection has brought convenience to human interactions, and on this basis, many new and disruptive business models have emerged, such as e-commerce, instant messaging, social media, etc.; similarly, IoT will realize connections between people and objects, and between objects themselves. We also expect to bring about new and disruptive business models, and even further, hope to bring about completely disruptive modes of human life and production.
As the main battlefield of IoT, the expectations for Industrial IoT are to bring revolutionary changes in industrial design, manufacturing, and circulation, injecting new vitality into traditional industries, providing new potential, and driving industries to develop, innovate, and even transform at a higher dimension. With the improvement of computing and storage capabilities, especially with the development of big data and artificial intelligence, every industry has unprecedented demands for data acquisition methods. The demands for data acquisition mainly manifest in four characteristics: first, efficiency; second, accuracy; third, real-time; fourth, economy. Under current technological capabilities, Industrial IoT is the only solution that can meet all four characteristics. First, chip technology has developed to a point where an MCU with strong computing capability costs less than a dollar, and RFID chip prices have even reached the cent level, providing a material basis for Industrial IoT while meeting economic requirements. The development of communication technology over the past thirty years, from analog to digital, from simple modulation to the commercialization of complex modulation techniques, has made it possible for wireless communication to cover hundreds of meters or even kilometers at a very low cost, meeting the dense deployment requirements for data acquisition. Additionally, due to the permanent online nature of Industrial IoT, it satisfies the efficiency and real-time requirements for data acquisition. Recent advancements in microelectronics technology have also made significant breakthroughs in both price and progress, meeting the accuracy requirements for data acquisition.
In summary, the emergence of Industrial IoT is an irreversible trend that arises when the following conditions mature:
1. Rapidly changing markets require data support, generating urgent demands for data acquisition;
2. The development of MCU enhances computing capabilities rapidly;
3. Communication technology, centered on modulation technology, establishes the pipeline foundation for connectivity;
4. Sensing technology, particularly the development of microelectronics characterized by MEMS, provides assurance for perceiving the world.
Industrial IoT is not something that is planned; it is a product of the evolution of various technologies and demands, emerging naturally when life, production, and economic development reach a certain height. It is a natural product brought about by innovation across numerous industries driven by demand.
Through Industrial IoT, we can digitize things in the traditional economy that cannot be digitized, digitize behaviors that were traditionally non-digitizable, and turn what was once impossible into something possible, or even easily obtainable and solvable solutions.
Second, what needs to be connected in Industrial IoT?
This question continues from the first. If we do not consider economic feasibility, we could say Industrial IoT connects everything that can be connected. However, when we are creating a practical and valuable solution, we cannot ignore feasibility and economy. So what does Industrial IoT connect? We believe this is a question of origin and destination. From our analysis of value, significance, and purpose, we know we should reverse-engineer from purpose. Everything starts from purpose, and we must keep a close eye on the most critical links that enterprises need to address. For example, if there is a need to quantify OEE, we need to connect device status; if we want to reduce work-in-progress, we need to track work-in-progress; if energy consumption is a top priority for the enterprise, we need to connect energy efficiency, and so on. There are no two identical leaves in the world, just as there are no two identical enterprises. We can only conduct in-depth analysis of the enterprise itself, focusing closely on enterprise value, and determine the scope of Industrial IoT implementation based on economic feasibility. A core point of the scope of connectivity is the attributes of the connected objects, meaning we determine the breadth of Industrial IoT implementation by analyzing the coupling degree between the attributes of connected objects and the goals of building Industrial IoT.
Third, what data needs to be obtained from connected objects?
After analyzing what Industrial IoT connects, we have identified the content of connected objects. Next, we need to decide which properties of each/each type of connected object we should digitize. Here we encounter a unique obstacle of Industrial IoT: the connectivity of objects that need to be connected to the Industrial IoT, especially when devices are interconnected. Connectivity issues are particularly prominent. For instance, some devices have open communication protocols and usable communication interfaces, while others do not have open protocols, etc. Therefore, connectivity becomes a significant test for solution providers. Our experience suggests four options:
1. Use the device’s open protocol;
2. Use the sensors that come with the device;
3. Add new sensors;
4. Change the perspective and dimensions of observation to use a completely new collection mode.
Among these, the fourth option, changing the perspective and dimensions of observation to use a completely new connection method, employs first principles to bypass the obstacles of non-open protocols or interfaces, allowing us to obtain data from the essence. For example, we can obtain the operational status of a device through energy efficiency detection, analyze component failures through vibration sensing, or even measure speed, as long as we approach the necessary information from the perspective of first principles, rather than passively relying on the data provided by the device to offer IoT solutions. We directly set our information needs as the target and explore how else we can obtain the required information besides directly connecting to the device, because only the data we obtain can be ‘isomorphic’ with the data provided by the device. For instance, we can install a vibration sensor on our IoT device, and from the data obtained from the sensor, we can determine whether the device is powered on, whether it has started working, and even its speed. If we do not use first principles and insist on connecting with the device, we would need to collect at least three data points, which the device may not necessarily provide. This is a typical case of edge computing. The rules of edge computing must have customization capabilities; it can be said that edge computing is a knowledge container that conveniently integrates knowledge from customers, manufacturers, and even third parties. The devices we have developed that support scripting already have preliminary edge computing functions, and we need to continue to enhance support in this area.
Therefore, by analyzing enterprise value and the connectivity of objects, we can clearly define the hierarchy of connected objects, which also clarifies the depth of connectivity.
An important concept in the hierarchy of connected objects in the IoT is management granularity. For the manufacturing industry, the management granularity of connected objects can be roughly divided into the following levels:
1. Sensor level;
2. Device level;
3. Production line level;
4. Workshop level;
5. Enterprise level;
This means we need to define the granularity of data acquisition based on economic feasibility. Theoretically, finer granularity is always better and more valuable, but when cost analysis is involved, it may not necessarily mean that finer granularity is better. We need to find a balance point based on various constraints.
Fourth, value-cost analysis of Industrial IoT construction
Value-cost always holds the highest weight in a company’s decision-making process, whether in agreement or disagreement. After the first three analyses, we are left with one final unresolved question, which is key to value-cost: the management granularity issue. What level of granularity do we need to manage? This raises a philosophical question: does the world need black boxes? What does this mean? Once we determine a management granularity, information finer than the management granularity will be hidden in the black box, which will become a limiting factor and constraint for our depth of analysis or understanding. We can find this balance point through value-cost analysis, thereby clarifying the size of the black box and ultimately determining the characteristics of the objects connected to the Industrial IoT.
Fifth, how to build Industrial IoT?
Our expectation is the value view of Industrial IoT construction; everything else is methodology. First, when planning IoT, we should embody both a visionary outlook and a pragmatic, feasible spirit. When considering the size of the black box, we must aim high and design solutions to minimize the size of the black box, while the implementation plan should choose an appropriate size of the black box based on value-cost analysis, which means selecting an appropriate management granularity to ensure a balance between investment and returns. We can even define a larger black box to verify the feasibility of Industrial IoT, thereby minimizing the risks associated with its implementation.
In summary, we should determine the principles of Industrial IoT construction based on several plans:
1. What results do we expect to obtain?
2. What methods do we expect to use to achieve the desired results?
3. What information foundation is needed?
4. Can Industrial IoT obtain this information?
5. How can Industrial IoT obtain this information?
6. What is the cost-effectiveness of obtaining this information?
7. Return analysis, evaluate whether the expected results align with economic interests?
8. Implementation.
【Related Reading】
What books to read? How to read?
Work is a form of practice
Smart manufacturing extends the time dimension
The US crackdown on Huawei is a competition for product innovation factors
The future economy belongs to the integration of electromechanical and artificial intelligence
In-Depth Thoughts on Industrial IoT
1969, a year that cannot be bypassed in the history of technology
【Copyright Statement】
This article is original by Mr. Yin Jinguo, researcher at the Institute of Intelligent Development and CEO of Shenzhen Yuchentai Technology Co., Ltd., who holds all copyrights. Reposting is welcome, but please indicate the author’s name “Yin Jinguo” and the author’s public account “Aizhi Zatan”. Aizhi Zatan adheres to originality; please follow the public account by long-pressing the image below to recognize the QR code.
