1. Definition of intelligent manufacturing
Manufacturing is to turn raw materials into suitable products. It should be noted that the meaning of manufacturing here is not limited to processing and production. For a manufacturing enterprise, its manufacturing activities include all related activities of "turning raw materials into suitable products", such as product research and development, process design, equipment operation and maintenance, procurement, and sales.
The most popular understanding of intelligent manufacturing is "using intelligent technology in manufacturing." But what is intelligence? What is artificial intelligence ? Although it has been more than half a century since the concept of artificial intelligence was proposed, the definition of artificial intelligence is still controversial. It is generally believed that the current research directions of artificial intelligence mainly focus on six major directions, including natural language processing, machine learning, computer vision , automatic reasoning, knowledge representation and robotics . But obviously people do not believe that enterprises must apply all the above technologies when implementing intelligent manufacturing.
has many definitions about intelligent manufacturing.
US Wright and Bourne define intelligent manufacturing as "modeling the skills and expert knowledge of manufacturing technicians through integrated knowledge engineering, manufacturing software systems, robot vision and robot control to enable intelligent machines to produce small batches without human intervention." The intelligent technologies that can be used in manufacturing activities today are not only listed in the above definitions, but also intelligent manufacturing is obviously not limited to small-scale production. But there is no reason for people to underestimate its significance because of the limitations of this definition. It was undoubtedly a visionary and pioneering work to propose the concept of intelligent manufacturing at that time (the 1980s) when the development of related technologies was not yet mature.
Lu Yongxiang once defined intelligent manufacturing: "A human-machine integrated intelligent system composed of intelligent machines and human experts. It can carry out intelligent activities in the manufacturing process, such as analysis, reasoning, judgment, conception and decision-making. Through cooperation between humans and intelligent machines, it expands, extends and partially replaces the mental labor of human experts in the manufacturing process. It updates and expands the concept of manufacturing automation to flexibility, intelligence and highly integrated." The human-machine integration emphasized in it is profound insight.
In China's "HTM2 Intelligent Manufacturing Technology Development "12th Five-Year Plan" for the Development of Intelligent Manufacturing Technology, intelligent manufacturing is defined as "facing the entire life cycle of the product, realizing information manufacturing under ubiquitous perception conditions. It is based on advanced technologies such as modern sensing technology, network technology, automation technology, and anthropomorphic intelligent technology. Through intelligent perception, human-computer interaction, decision-making and execution technology, the design process is intelligent, manufacturing process and manufacturing equipment, etc.." In this statement, the intelligence of the design process, manufacturing process and manufacturing equipment is only a phenomenon of intelligent manufacturing. In other words, intelligent design, equipment, etc. are just means of manufacturing, not goals.
The Ministry of Industry and Information Technology clearly defines intelligent manufacturing in the "Intelligent Manufacturing Development Plan (2016-2020)" released in 2016: Intelligent manufacturing is a new production method based on the deep integration of new generation information and communication technology and advanced manufacturing technology, running through all aspects of manufacturing activities such as design, production, management, and service. It has functions such as self-perception, self-learning, self-decision, self-execution, and self-adaptation. This definition undoubtedly draws on the wisdom of many scholars and experts, points out the technical basis and application links of intelligent manufacturing, and reveals its functional representation, but fails to touch on the essence and connotation of intelligent manufacturing.
SM (smart manufacturing) that is valued in the United States, EU , South Korea, etc. can be seen as a higher stage of the development of smart manufacturing. SM is the result of the rapid development of some cutting-edge technologies in recent years, such as the Internet of Things, big data, VR (virtual reality)/AR ( augmented reality ), intelligent sensing, cloud technology, new generation of artificial intelligence, etc. The National Bureau of Standards and Technology believes that SM is a fully integrated collaborative manufacturing system that responds to changes in demand and conditions in enterprises, supply chains, and customers in real time.This definition is quite simple, and does not directly point out the specific functions of the technology and system involved, but it more clearly reveals the goal of intelligent manufacturing.
Here is a minimalist definition of intelligent manufacturing and systems. The reason for this is precisely because intelligent manufacturing is still developing. A simple definition may include a wider range of functional and technical elements, whether existing or future; a simple definition may have a deeper meaning, whether superficial or internal; whether explicit or implicit.
Machine intelligence includes computing, perception, recognition, storage, memory, presentation, simulation, learning, reasoning, etc., including both traditional intelligent technologies (such as sensing, knowledge-based system KBS, etc.), and a new generation of artificial intelligence technologies (such as deep learning based on big data). Generally speaking, artificial intelligence is divided into three stages: computing intelligence, perceptual intelligence and cognitive intelligence.
The first stage is computing intelligence, that is, fast computing and memory storage capabilities. The second stage is perceptual intelligence, that is, perception abilities such as vision, hearing, and touch. The third stage is cognitive intelligence, which means being able to understand and think. Cognitive intelligence is the area with the biggest gap between machines and humans at present, making it extremely difficult for machines to learn reasoning and decision-making.
Although machine intelligence is developed by humans, the strength of many unit intelligence (such as computing and memory) is far beyond that of human capabilities. The integration of machine intelligence into various manufacturing activities to realize intelligent manufacturing usually has the following benefits:
- The computing intelligence of intelligent machines is higher than that of humans. In some places where there are fixed mathematical optimization models, a large amount of calculations are required, but no knowledge reasoning is required, such as engineering analysis of design results, advanced planning production , pattern recognition, etc., compared with people's judgment based on experience, machines can give better solutions faster. Therefore, intelligent optimization technology helps improve design and production efficiency, reduce costs, and improve energy utilization.
- intelligent machines have higher active perception and automatic control capabilities of manufacturing conditions than humans. Taking the CNC machining process as an example, the vibration and temperature changes of the "machine tool/workpiece/cutter" system have an important impact on product quality, and it is necessary to adaptively adjust the process parameters , but it is obviously difficult for humans to perceive and analyze these changes in a timely manner. Therefore, the application of intelligent sensing and control technology to realize the "perception-analysis-decision-execution" closed-loop control can significantly improve the manufacturing quality. Similarly, there are many dynamic and changing environments in the manufacturing process of an enterprise. Some elements in the manufacturing system (equipment, detection mechanism, material delivery and storage system, etc.) must be able to respond dynamically and automatically to system changes, which also depends on the independent and intelligent decision-making of the manufacturing system. The product life cycle data owned by
- manufacturing companies may be massive. The development of technologies such as industrial Internet and big data analysis brings faster response speed, higher efficiency and far-reaching insight to enterprises. This is incomparable to traditional methods of judgment based on human experience and intuitiveness.
Enterprise manufacturing activities include R&D, design, processing, assembly, equipment operation and maintenance, procurement, sales, finance, etc.; integration means that it does not completely subvert the previous manufacturing methods, and further improve the efficiency of manufacturing by integrating machine intelligence. The definition points out that the purpose of intelligent manufacturing is to meet the corresponding goals of the enterprise. Although the specific goals are not specified, readers can easily understand that improving efficiency, reducing costs, and greening are all implicit.
In addition to the keywords in intelligent manufacturing, the keywords here include: system, people, resources, needs, environmental changes, dynamic adaptation, and optimization goals. Resources include raw materials, energy, equipment, tools, data... The demand can be external (not only considering customers, but also socially), or internal to the enterprise; the environment includes equipment working environment, workshop environment, market environment... In this definition, the system is a relative concept, as shown in the figure below.
means that the system can be a processing unit or production line, a workshop, a company, and an enterprise ecosystem composed of enterprises, their suppliers and customers; dynamic adaptation means being able to respond in real time to environmental changes (such as temperature changes, tool wear, market fluctuations); optimization goals involve the goals of enterprise operations, such as efficiency, cost, energy conservation and consumption reduction, etc. As for the various means required by the system, they are all implicit.
It is particularly important to note that the above definition implies:
Korean scholar Kang and others pointed out that intelligent manufacturing (SM) cannot only focus on economic indicators that increase efficiency and reduce costs, but should also be able to create new value for society for a long time. Lack of considerations about human and social issues may raise some problems. Intelligent manufacturing cannot be regarded simply as the application of cutting-edge IT technologies. It should be based on a manufacturing engine that can lead to sustainable growth that is oriented towards people and society.
2. The basic connotation of intelligent manufacturing
The introduction outlines the necessity of development from automation to digitalization, networking and then to intelligentization. After more than a hundred years of development, automation technology has been relatively mature. Take a little observation and a little abstraction to think about the problems that automation technology is suitable for solving.
is suitable for problems that automation technology can solve, basically all are certain. All automatic line , automatic machines, the process flow is determined, the motion trajectory is determined, and the goal of the control object is determined. Of course, there may be errors in the actual motion of the machine, which is reflected in the quality of the manufactured items, that is, uncertainty does not exist at all. However, as far as the design considerations of an automatic system are concerned, the system's input and output working methods, paths, goals, etc. are determined, and only the errors generated need to be ensured are within the allowable range.
The classic automation technology faces basically structured problems. Problems that can be described by classical control theory are structured, such as automatic adjustment problems, PID (proportional integral differential) control, etc. The development of electronic and computer technology has accelerated the application of program control and logic control in automated systems, and the problems it targets are also structured. In modern control systems, people use knowledge-based systems in some occasions, similar to IF-THEN, which is itself a structure, and the problems they deal with are structured.
The problems handled by traditional automation technology have their own fixed patterns, such as automatic processing, flow-through production, automatic material transportation, etc.
The problems targeted by traditional automation technology are relatively local, and there are few problems at the enterprise system level, such as supply chain issues, customer relationships, strategic responses, etc.
Let us observe and think about the actual problems of the company again. There are a lot of uncertainty issues in enterprises, such as quality issues that any enterprise must pay attention to. For some pre-knowledgeable, certain problems that may cause quality defects, they can be solved by setting up corresponding processes and automation methods, which is within the reach of traditional automation technology. There are many random factors that affect quality, such as temperature, vibration, etc. Although these factors are known in advance to affect quality, they are only qualitative concepts and cannot set the control quantity in advance. This requires real-time monitoring of changes in relevant factors during the manufacturing process, and applying corresponding controls according to the changes, such as adjusting the ambient temperature, or automatically compensating the processing error . This is the initial intelligent control. Although random factors such as
that cause quality problems are uncertain, they are obvious and easy to be aware of. There is also a type of uncertainty factor that is hidden and difficult for engineers and managers to even realize. For example, how many related and combined factors affecting its performance of an advanced and complex engine system? To what extent does it affect? For example, what are the parameters that may exist in a new process that invisibly affect process performance? The degree of impact? For engineers, these may be uncertain.
In fact, some of the factors and their correlations affect the certain side, but people still lack understanding of their objective laws, resulting in subjective uncertainty.In addition, there are some original certainty issues, which lead to uncertainty in people's understanding of them due to failure to digitize. For example, the arrangement of various activities and processes in an enterprise is inherently certain. But because there are too many people involved and the time of occurrence is different, if there are no special means, it will be chaotic to people's understanding. This is also the subjective uncertainty or cognitive uncertainty of people. Why are subjective uncertainty also regarded as uncertainty in manufacturing systems? Because the manufacturing system should have included relevant people.
There is also a type of implicit influencing factors that are inherently uncertain. For example, the subtle inconsistency of raw material performance in the precision manufacturing process, energy instability, sudden environmental factors (such as sudden external vibrations), etc., lead to quality instability; quality problems caused by temporary changes in personnel positions in the workshop; quality problems caused by changes in work and rest time caused by special major social activities (such as football world cup ) during a certain period; the specific degree of impact on the enterprise after major public health safety occurs, these are related to various special characteristics such as the enterprise supply chain, regional location, abortion, and infection of enterprise personnel (each enterprise is different). At present, people can only have an abstract and qualitative understanding of such problems, and it is difficult to respond relatively carefully based on the specific degree of influence. For such problems, classic automatic control technology is naturally shelved, and even modern control technology with certain intelligent characteristics is powerless.
Note: Explicit and implicit uncertainty factors!
There are a lot of problems in enterprises that are unstructured. When people want to improve quality as much as possible, it is difficult to find that the construction of factors affecting quality problems is difficult; after major public health safety occurs, it is difficult to have quantitative analysis of the specific impact on the enterprise, let alone deal with it; these are all because the environment and the problems themselves are unstructured. There is a large amount of information and unconventional numerical data or structured data stored in the database that can be logically expressed by two-dimensional table structure, such as full text, images, sound, hypermedia, etc., which is unstructured data. These unstructured data are all useful information for enterprises, such as reports from R&D personnel, collected external information (text, images, etc.)... Traditional automation technology fails to effectively utilize this information and can only stop there.
Many problems in enterprises are non-fixed patterns. Nowadays, many companies implement personalized customization in order to better meet customer needs. Different types of enterprises implement personalized customization methods. Even for the same company, different models may be needed for different products and different types of customers. The collection and processing of data, data-driven personalized design and production methods are different. For example, energy saving in factories or workshops, different types of enterprises may have different ways of saving energy. Even if companies with similar products have different equipment, different regional environments, and different factory structures, they will lead to different energy-saving models. Technical workers engaged in traditional automatic control will naturally not care about such non-fixed mode problems.
Our ancestors had a good cultural tradition, that is, they focused on overall connections. The material view of in ancient China, metal, wood, water, fire and earth, are mutually generated and restrained. Although this statement is not scientific, its idea of focusing on overall connection has reasonable elements. Traditional Chinese medicine regards people as a whole, as meridians say, and actually emphasizes the overall connection of the human body. Although it has limitations from a scientific point of view, the reasonable components of its ideas can still be seen from the effectiveness of certain practices (such as acupuncture).
Enterprise is a large system, with many subsystems and subsystems, various activities (design, processing, assembly...), various resources (raw materials, tools, parts, equipment, manpower...), suppliers, customers... Are there so many factors in the large system related and influence each other? Affirmative influence – based on imagination and feeling. What is the specific impact on the overall effectiveness of large systems? Senior managers and engineers may not be clear.
Even if a device system has the mutual influence between its components, subsystems, operating parameters, environment and other elements, people can only qualitatively know some impacts, and it is difficult to clearly understand the degree of impact in all. In short, our understanding of the overall connection between large and subsystems of the enterprise is very limited. The reason for this is not only because the system is large and complex, but also because the system is full of the above-mentioned uncertainty, unstructured, and non-fixed modes.
does not mean that people have not realized the existence of problems such as overall connection and uncertainty, but are just suffering from the lack of tools and lack of brain power. Humans never stop pursuing the tool of "supernatural existence". Based on the desire to understand and even control issues such as overall connections, uncertainty, unstructured, and non-fixed models more clearly and more meticulously, human beings have finally created suitable tools, namely the Internet of Things, big data analysis, artificial intelligence (especially the new generation), etc. It is precisely with these tools and means that we cannot continue to plague problems such as overall connections and uncertainties, and the manufacturing field is no exception. At this point, we can understand the connotation of intelligent manufacturing more deeply:
Welcome real estate peers and industry peers to like + follow, and professional exchanges, and will continue to update and release professional information.