1 Background on Industrial process optimization
In modern factories, no matter how advanced the process technology is, they are pre-designed solutions after all. During the production and operation of the factory, it is difficult to make the production equipment in Optimal economic operation state; complex chemical production process characteristics also determine the non-linearity of its production process, such as raw material consumption, product output and energy consumption. How to optimize the production process, improve production efficiency, reduce raw material consumption, improve product yield and quality, and reduce energy consumption have become urgent problems for enterprises to solve.
As one of the four dimensions of factory digital transformation, production optimization has always been one of the important goals of the eternal pursuit of smart factories. In the service planning of the factory’s full life cycle, the production operation optimization plan is placed in the most important position.
2 Hierarchical concepts and characteristics of industrial process optimization
The optimization system of the industrial process industry includes four levels: planning, scheduling, operation, and control. Different optimization techniques are required at different levels. The number of optimization objectives designed for different levels and the complexity of optimization variables are also different. Different levels use different optimization techniques. The optimization techniques are also different. In the planning and scheduling layer, there are many management processes involved, and the variable data is huge. Therefore, the optimization uses the mechanism model combined with the linear programming method to optimize, such as the optimization of the production execution process, the optimization of the raw material warehousing process, etc. The optimization of the operation layer can take the process mechanism model combined with the big data model for optimization, such as the automation of operation procedures, advanced operation screen, alarm optimization, etc.; the optimization of the control layer, due to the requirement of fast speed, mostly adopts simple mechanism model or data model or establishes prediction. Model optimization, such as loop self-learning tuning optimization, variable prediction control, soft measurement, etc.
3. Industrial software is the basis for process optimization
Although there are different levels of optimization, and the technologies used at different levels are different, the implementation of optimization requires a common foundation, that is, industrial software that integrates industrial industry knowledge. Industrial software is the soul of digitalization. Industry 4.0 is commonly known as the “softening industrial revolution”. Compared with Industry 3.0, the biggest difference lies in the wide application of various professional and advanced industrial software. Industrial software is not the same as ordinary commercial software. Professional industrial software needs to integrate operational technology (OT), information technology (IT), automation technology (AT) and equipment technology (ET) in one, tightly integrated with industrial processes to maximize value for production optimization. Therefore, the process of industrial process optimization is, to a large extent, the process in which various industrial software are applied at various levels to exert their value.
4 The technical path of industrial process optimization realization
For a factory to realize the optimization of the overall production process, it is not simply a matter of stacking various industrial software. It needs to make overall planning according to the technical foundation of different factories. The author suggests that the overall design idea of the production optimization scheme should be “overall planning from top to bottom and realization layer by layer from bottom to top”. The following figure shows the typical technical path of production process optimization in the chemical industry, which is mainly divided into the following steps to realize step by step:
The first step is to establish the software technology foundation for factory process optimization: first, realize the informatization of production computers, let many on-site manual operations enter the automation platform, and realize the transformation from external operations to internal operations, also called “operation softening”; Instruments and valve equipment connected to the control system are intelligent, and the monitoring and management of the health information of these equipment is realized through industrial equipment management software (AMS), also known as “equipment maintenance softening”; then establish a factory digital virtual platform, such as a digital twin, Provide a software verification platform for subsequent control operation optimization, commonly known as “platform softening”.
The second step is to realize control optimization: In this step, we must first realize the management of abnormal conditions, determine the boundaries of interlocking, alarming and normal operation range, follow the alarm management standards, and realize the rational optimization of alarms and the optimization of interlocking management with the help of the alarm master database software platform. Minimize the proportion of abnormal state of factory production, reduce the practice of operators to deal with abnormal working conditions, thereby reducing the operating load; secondly, try to automate internal operations as much as possible, use digital loop optimization tool software, and use advanced self-learning to quickly improve the control loop of the entire production line. The automation rate reduces the proportion of manual operations in internal operations, so as to further operate the load. After the automation rate is increased, intelligent control strategies such as cascade, split-range, feedforward, override, and adaptive control are used for some complex loops to further optimize the loop. performance, and consider using APC for modeling optimization for loops with serious mutual coupling phenomenon to reduce production fluctuations, achieve production card edge optimization, and improve production efficiency.
The third step is to realize the automation of operating procedures at the system-wide level: its main solutions include state-based control (SBC) and state-based alarm (SBA), that is, dynamic rationalization of alarms, and high-performance operation screen solutions. These software solutions can simplify operations. process, reduce operational load, reduce the chance of operational errors, increase yield and quality and thus increase production efficiency. At present, the vision of “one-button driving” and “black screen operation” that are often mentioned in the digital field is realized based on the automation technology of operating procedures, but it should be remembered that if the automation of operating procedures must be based on “softening of operation”, “automation of control” is good can be achieved on the basis of.
The fourth step is to optimize the overall process of the whole plant: the main solutions at this level include automatic production scheduling, production execution business flow automation, asset management digitalization, etc. The optimization software platform has MES to open up business data islands and realize production scheduling. , online optimization of business flows such as planning and scheduling, raw material in and out of storage. In addition, there are big data platforms such as OT data warehouse (DW), data mart (DM) or data lake. Based on this, a data center is established for the entire factory; Collect, process, store, and calculate the data, and at the same time unify the standards and calibers. After the data is unified, it is stored in a standard form to form a big data asset layer to meet the needs of front-end data analysis and applications. Then the big data mining professional tool software mines and optimizes the data, and realizes various analysis and applications as needed. For example, multivariate online statistics, batch optimization analysis, root cause analysis, business flow management, etc., finally realize the data-driven production process, the production management mechanism of data decision-making, and finally realize the optimization of the whole plant business flow and obtain the best production efficiency.
Based on professional industrial digital software, overall planning and layer-by-layer optimization using different technologies are the current reliable technical paths to realize industrial process optimization.