16

2018

-

03

The direction of chemical intelligent manufacturing|Artificial intelligence is not reliable

Author:


Finance Dad 17-08-2115: 24
  
In the past two years, due to the outstanding performance of artificial intelligence in some fields (not the industrial field), some people have begun to be optimistic, feeling that China is overtaking on the curve of manufacturing and industrialization through "Internet +" and artificial intelligence. Is there really such optimism?   When I wrote my doctoral dissertation in 2007, the first sentence of the introduction to the first chapter was, "The Report of the 16th National Congress (2002) clearly proposed that 'industrialization is driven by informatization, and informatization is promoted by industrialization'". Later, the 17th National Congress (2007) put forward: "integration of industrialization and informatization"; In order to achieve this two modernizations, the state also established the "Ministry of Industry and Information Technology" in 2008, which shows that the state has a deep understanding and attention to this development strategy issue. Later, the 18th National Congress (2012) put forward: "deep integration of industrialization and industrialization". Until now, in the boom of intelligent manufacturing, the integration of industrialization and industrialization is still the "foundation of the ministry" of the Ministry of Industry and Information Technology. More than a decade has passed, and the two modernizations are still being mentioned, which shows that this matter is not easy, and the progress is not as smooth as the government thinks.  
  
At present, there is a lot of talk about intelligent manufacturing discrete industries, but there is very little talk about intelligent manufacturing in the chemical industry. So, chemical intelligent manufacturing, in which direction is it developing? The chemical industry has long realized the primary intelligent system - automatic control. Due to the continuity of the chemical process and the large-scale of the plant, and the huge investment of the plant (billions of billions of investment), the chemical industry (including oil refining, petrochemical) has long put forward very high requirements for process automation, and began to use DCS for process control in the 70s of last century. Automation improves the stability and safety of chemical production, and it is also easy to increase the profit margin of the plant (improving profits is the direct driving force for enterprises to adopt new technologies in the market economy environment).
  
The current level of technology can make more than 80% of the production workshop and operation of chemical production unmanned, mainly in some solids processing and transportation to achieve automation is still difficult. It is normal to rely on pumps and compressors to realize the flow of materials in the closed pipeline system, and rely on various temperature, pressure, liquid level and flow control to realize the automatic operation of materials and energy in each operating unit. Technically feasible or optimal does not mean optimal economic efficiency. Especially some small devices, completely using automatic control system system system unit cost is high; When the labor cost is low, it is preferable to use manual operation. Therefore, the automation rate of the real world chemical industry is determined by the level of technology and economic benefits (investment costs, labor costs).
  
Traditional AI is not suitable for the chemical industry The core of traditional artificial intelligence (big data, machine learning) is to summarize the rules of historical data extraction to predict the future. The theoretical basis is that the operating data contains all the important hidden information of the system, and there is no need to study the problem mechanism, and the laws and knowledge of the system can be directly mined from the data. This artificial intelligence is not suitable for the chemical industry, and its role in intelligent production in the chemical industry is extremely limited.
  
Based on three reasons: 1. The operation mechanism and mathematical model of the chemical plant are relatively complete. As an engineering discipline that has developed for more than 100 years, chemical engineering has a relatively complete knowledge system. As a manual design system, the designer has clear the intrinsic characteristics and mechanism of the device and knows the mathematical model of the device. So there is no need to use artificial intelligence to dig and discover knowledge. Even when the mechanism is unclear or the boundaries are uncertain, some conventional, traditional data analysis methods are sufficient to deal with the problems in the chemical industry. 2. As a strictly controlled system, the chemical plant has a lot of data but monotonous, and the amount of information is too low to mine knowledge. Because the chemical process is strictly controlled by various control systems and the production is stable, the data generated is large but narrowly distributed, and artificial intelligence cannot be used to extract laws or knowledge from such big data with less information. 100, 10,000 identical data contains the same amount of information as 1 data. 3. The reliability and safety requirements of chemical plants do not accept the black box knowledge generated by artificial intelligence systems. Chemical production has extremely strict requirements for safety and reliability, and in the event of an accident, it is catastrophic, and the damage to the environment and employee lives is irreparable. Artificial intelligence relies entirely on the input and output data of the system to produce a black-box model. When this black-box model is applied, firstly, the cause of the failure or problem cannot be found according to the model, and secondly, it is difficult to evaluate the reliability of the model.
  
Traditional artificial intelligence is more suitable for human intellectual activities with extremely complex systems (making it difficult to study the mechanism) and do not have strict requirements for system causality and reliability, such as finance, business, medicine, artificial intelligence will revolutionize these fields, and these changes are really happening around us. The field of science and technology is essentially the pursuit of causality and reliability, scientists and engineers have long attached importance to and applied data, and the degree of change of artificial intelligence to science and technology will be limited from the perspective of knowledge discovery and extraction. Knowledge automation is the main direction As a typical engineering discipline, chemical engineering is characterized by semi-theoretical and semi-experimental. Because some phenomena are too complex, involving machinery, materials, physics, chemistry, thermodynamics, kinetics and transfer, a variety of factors are associated and coupled together, and the causes or conclusions of certain phenomena cannot be derived through pure theoretical logic, and it is necessary to separate various factors in a laboratory environment for independent research (thorough research methods), or synthesize them and only study the influence of the main factors on the results. That is, most of the theoretical knowledge of chemical engineering comes from laboratory research.
  
To give a simple example, for example, a new set of binary systems, in the absence of vapor-liquid balance experimental data, which model dares to say that its prediction accuracy is within 5%? Although there are millions of sets of binary vapor-liquid balance experimental data in chemical literature and databases, chemical thermodynamicists have studied vapor-liquid balance prediction models for nearly 50 years, but once they encounter key applications, they still have to go to the laboratory to do experiments to obtain experimental data. Due to the complexity of chemical phenomena, some phenomena in industrial devices exhibit different characteristics than laboratory experimental devices or even phenomena that cannot be observed in laboratories, the so-called "amplification effect", the essence of which is still unclear investigation of certain factors and not correctly predicted. At this point, we can get feedback from industrial installations to expand our knowledge of chemical engineering. In addition, from the operation of industrial plants, a large amount of empirical knowledge in operation, maintenance, safety can be obtained, beyond the scope of laboratory research. Although the chemical industry is not a good image in China, and it is not a good major in universities, its subject knowledge structure and research methods are relatively complex, and chemical engineering is a high-income major in engineering disciplines in Europe and the United States. Turning experience into data, data into knowledge, and knowledge into automated systems is knowledge automation, which is the core of intelligent manufacturing. It can be seen that the knowledge of the process mechanism of a chemical plant has basically been integrated into the initial design and operation automation control, and more than 80% of the knowledge has been automated; The knowledge of device operation, mainly involving personnel management, asset and equipment management, operation, maintenance, supply chain knowledge still exists in various SOPs, and exists in the human brain as experience, which is still far from knowledge automation. Knowledge automation is the key direction of informatization and intelligence in the chemical industry in the future. So, what exactly can knowledge automation and smart manufacturing do in the chemical industry? Production process is the core of a chemical enterprise, the chemical industry production process digitalization, automation and intelligence in the forefront of the entire industrial system, basically has achieved unmanned production in the workshop, but the central control room is still sitting people, through the computer screen to observe and monitor the production process, ready to manually intervene remotely or even to the site intervention. In fact, the quality of our final products is not static, very stable, and the utilization rate of raw materials and energy in the production process is not optimal.
  
A chemical device is not automated, unmanned even if the goal of intelligent manufacturing is achieved, the new goal is no longer to meet the stability of the system in a state, but to let the device system automatically run under the constraints of device safety, product quality, raw materials, energy, asset utilization of the optimal state. Multi-product fine chemical plants also need to be agile and flexible to quickly respond to market changes and order requirements. The architecture of "intelligent manufacturing" in the petrochemical industry has long been determined, that is, to realize knowledge automation and intelligence at the three levels of process control, production management and operation management, corresponding to process control system (PCS), production execution system (MES), and enterprise resource planning (ERP).  Figure 1: Architecture of "intelligent manufacturing" in the petrochemical industry, source: Popular Science Chemical Process Control System (PCS): An automatic control system that characterizes the parameters of the production process as the controlled quantity to make it close to a given value or maintain it within a given range, represented by DCS and PLC, including advanced process control APC. Enterprise resource planning (ERP): is an effective and comprehensive planning and management of enterprise resources, including product ordering, raw material procurement, distribution, sales, accounting and a series of business flows, represented by SAP. Production Process Execution Management System (MES): As the interface between DCS and MES, it realizes the integration of production performance management and operation data, and the functional modules include short-term production planning, job scheduling and scheduling (APS), maintenance management, technical information management, online real-time optimization RTO, product quality management LIMS. No matter at which level, to let the computer solve the problem, it is necessary to digitize and abstract the problem into a mathematical model, so the core of intelligence is five modernizations: "digitalization, visualization, modeling, automation, integration". Only by digitizing experience and information through detection technology, finding laws through data visualization and analysis, and then converting the laws into mathematical models, and then automatically solving and making decisions on the mathematical models through computers, and automatically transmitting decisions or conclusions to the previous or next layer, so as to achieve integration. This complete process is knowledge automation, knowledge automation frees people from repetitive work and focuses on innovation and high value-added activities, obviously it has become more demanding of talent. A long way to go The monitoring and pre-maintenance of equipment is one of the few feasible applications of big data technology in the chemical industry.
  
The technology that needs to be developed: First, new sensing technology: vibration, sound, image, current and other signals are integrated into the monitoring model, and a key characteristic signal can replace dozens of signals with weak associations. The biggest achievement of artificial intelligence technology based on big data is the processing and recognition of sound and images, and there are relatively few applications based on these two types of signals in the industry, and it is time to apply them. The second is signal processing methods and algorithms: how to correlate signal phenomena with fault types requires the use of advanced pattern recognition techniques. But even so, similar industry replication is still very difficult and cannot be simply ported. First, because of the input and output ratio of various intelligent manufacturing systems. One of the main inputs of intelligent manufacturing is all levels of software, a feature of software investment is that it has almost nothing to do with the scale of the device, the price of a set of software will not change with the size of the device, but the benefits generated are basically proportional to the scale of the device, so it seems that the investment income of the software system also has an almost linear scale effect. SMEs must use low-cost software systems suitable for SMEs. Second, because the three-level optimization in addition to PCS direct pure and equipment communication, but ERP and MES have more interfaces with people, and involve the management culture of the enterprise, the system investment involves management, cultural changes, or the system is customized according to the management culture of the enterprise. Things that involve people cannot be simply copied.
  
Postscript: The reliable road is not in the sky under the feet Although the concept of artificial intelligence is very hot now, telling countless stories and attracting countless investments, the impact on the chemical industry (and even the process industry) can be considered negligible. When the chemical industry talks about intelligent manufacturing, it does not rely on artificial intelligence based on big data, but relies on the digitalization and automation of knowledge and experience. The framework of "intelligent manufacturing" in the petrochemical industry has long been determined, that is, in the process control, production management, operation management three levels through PCS, MES and ERP to achieve knowledge automation and intelligence, which is a reliable road.

 

 

 

Chemical Engineering, Knowledge, Systems, Devices, Automation, Data, Intelligence, Manufacturing, Artificial Intelligence