1 Introduction The distillation column plays an important role in the petrochemical industry. It is a widely used mass transfer heat transfer device in the production process of chemical companies and oil refineries. Its purpose is to separate the components in the mixture to meet the specified requirements. purity. The essence of the rectification process is to utilize the different volatility of each component in the mixture, that is, the different saturated vapor pressures of the components at the same temperature, so that the light components in the liquid phase are transferred to the vapor phase. The heavy components in the vapor phase are transferred to the liquid phase to achieve the purpose of separation. Therefore, the operation of the rectification device is directly related to the economic benefits of the petrochemical enterprise.
The fault diagnosis of the distillation tower has become an important part of chemical production. Due to the many factors affecting the production operation and product quality of the distillation tower, the relationship is particularly complicated, which affects the safe production of petrochemical enterprises. In recent years, with the development of computer and artificial intelligence, neural network technology and the theory of distillation process (such as tower process calculation, tray mechanical calculation and calculation of auxiliary equipment of tower), how to computer and distillation process The combination of the theory and the organic application to the fault diagnosis of the distillation column is one of the concerns of refining companies. Because of this, extensive research has been carried out on the fault diagnosis of the distillation process at home and abroad, and most of them have adopted the method of establishing mathematical models. This paper discusses in detail the design of the fault diagnosis of the distillation column.
2 Scheme design for fault diagnosis of distillation tower 2.1 Establishing expert system for fault diagnosis of distillation tower This method mainly includes four parts: fault input, interpretation-based learning model, quantitative deep knowledge base and result output to domain knowledge base. First enter a fault, then call the quantitative deep knowledge base, start the learning model based on the interpreted learning model, and obtain a learning. The description is directly used for fault diagnosis. After the Zui, the description is stored in the domain knowledge base, where the quantitative deep knowledge base needs Constantly updated to complement new troubleshooting methods.
2.2 Diagnostic steps 2.2.1 Establishing equipment files No matter what equipment has potential failures, the use of equipment is a process from potential failure to functional failure. The fault development curve is shown in Figure 3. In order to carry out economical and effective maintenance of the rectification tower, it is necessary to establish equipment files, track the operation status of the rectification tower, understand the historical records of the components of the rectification tower and their operating states, so as to adopt regular inspection and corresponding The fault diagnosis method detects potential faults before the functional failure of the distillation tower and avoids the occurrence of functional faults.
2.2.2 Determination of detection methods The main faults are low tray efficiency, low bottom temperature, high reflux temperature and flooding. In response to the above faults, the application of 7-ray field scanning detection technology can diagnose and eliminate faults, optimize operating conditions and extend operating cycles, and provide important data or scientific basis for refining and chemical companies to guide production.
2.2.3 Determining the detection cycle and maintenance time of the distillation column The determination of the monitoring cycle of each component of the rectification column depends on the life of the rectification column itself. Here, the average ultimate life of the rectification column is T, then the detection The period Tc=T/nd, where nd is the dispersion coefficient. The determination of nd depends on the environmental conditions in which the device is operating. The maintenance time of the rectification tower is determined according to the condition of the rectification tower after the detection. Here, the large operation state of the rectification tower is specified as the large value of the zui which can be achieved by the parameters of the rectification tower before the failure occurs, one of which is one of them. When several parameters exceed this value, it is considered that the distillation column has reached the limit life and needs to be repaired.
2.2.4 Establish performance prediction and file management After each inspection of the distillation column, record the detection, and determine the availability of the device (ie the remaining life of the device) and the time of the next detection. Therefore, performance prediction must be performed. The digital modeling of equipment performance prediction is as follows:
Let X{X1,X2,...,X. } is a time series in which the length N reflects the operating state of the rectification tower N times, X; for a certain characteristic parameter describing the operating state of the mechanical equipment, Xi∈Rn(i_1,...,N), it is now predicted after k steps The operating state of the mechanical equipment, that is, xN+k is obtained. The intrinsic function relationship F between the state xN+k at the (N+k)th time and the first N states is expressed as:
XN+k=F(Xl,X2,...,X.)
(1) The life prediction of the distillation column can be expressed as finding F by establishing a mathematical model, so that 任 ∈>0, with ||FF(X)-FF(X)||=||xN+k-XXN +k||
<∈>(2) where F(X)=xN+k actual state value, FF(x)=xN+k prediction model value.
The file management system includes the overall state of the rectification column, the performance of each component of the rectification column, and economic and technical evaluation. The system flow is shown in Figure 4.
3 Conclusion With the rapid development of modern science and technology, modern fault diagnosis systems are constantly improving. For example, artificial intelligence, neural network and dynamic simulation software can prevent and avoid accidents in time and accurately, so maintenance technology, The organic combination of the obstacle diagnosis technology and the computer remote monitoring technology is the development direction of fault diagnosis for refining and chemical enterprises in the future.
The fault diagnosis of the distillation tower has become an important part of chemical production. Due to the many factors affecting the production operation and product quality of the distillation tower, the relationship is particularly complicated, which affects the safe production of petrochemical enterprises. In recent years, with the development of computer and artificial intelligence, neural network technology and the theory of distillation process (such as tower process calculation, tray mechanical calculation and calculation of auxiliary equipment of tower), how to computer and distillation process The combination of the theory and the organic application to the fault diagnosis of the distillation column is one of the concerns of refining companies. Because of this, extensive research has been carried out on the fault diagnosis of the distillation process at home and abroad, and most of them have adopted the method of establishing mathematical models. This paper discusses in detail the design of the fault diagnosis of the distillation column.
2 Scheme design for fault diagnosis of distillation tower 2.1 Establishing expert system for fault diagnosis of distillation tower This method mainly includes four parts: fault input, interpretation-based learning model, quantitative deep knowledge base and result output to domain knowledge base. First enter a fault, then call the quantitative deep knowledge base, start the learning model based on the interpreted learning model, and obtain a learning. The description is directly used for fault diagnosis. After the Zui, the description is stored in the domain knowledge base, where the quantitative deep knowledge base needs Constantly updated to complement new troubleshooting methods.
2.2 Diagnostic steps 2.2.1 Establishing equipment files No matter what equipment has potential failures, the use of equipment is a process from potential failure to functional failure. The fault development curve is shown in Figure 3. In order to carry out economical and effective maintenance of the rectification tower, it is necessary to establish equipment files, track the operation status of the rectification tower, understand the historical records of the components of the rectification tower and their operating states, so as to adopt regular inspection and corresponding The fault diagnosis method detects potential faults before the functional failure of the distillation tower and avoids the occurrence of functional faults.
2.2.2 Determination of detection methods The main faults are low tray efficiency, low bottom temperature, high reflux temperature and flooding. In response to the above faults, the application of 7-ray field scanning detection technology can diagnose and eliminate faults, optimize operating conditions and extend operating cycles, and provide important data or scientific basis for refining and chemical companies to guide production.
2.2.3 Determining the detection cycle and maintenance time of the distillation column The determination of the monitoring cycle of each component of the rectification column depends on the life of the rectification column itself. Here, the average ultimate life of the rectification column is T, then the detection The period Tc=T/nd, where nd is the dispersion coefficient. The determination of nd depends on the environmental conditions in which the device is operating. The maintenance time of the rectification tower is determined according to the condition of the rectification tower after the detection. Here, the large operation state of the rectification tower is specified as the large value of the zui which can be achieved by the parameters of the rectification tower before the failure occurs, one of which is one of them. When several parameters exceed this value, it is considered that the distillation column has reached the limit life and needs to be repaired.
2.2.4 Establish performance prediction and file management After each inspection of the distillation column, record the detection, and determine the availability of the device (ie the remaining life of the device) and the time of the next detection. Therefore, performance prediction must be performed. The digital modeling of equipment performance prediction is as follows:
Let X{X1,X2,...,X. } is a time series in which the length N reflects the operating state of the rectification tower N times, X; for a certain characteristic parameter describing the operating state of the mechanical equipment, Xi∈Rn(i_1,...,N), it is now predicted after k steps The operating state of the mechanical equipment, that is, xN+k is obtained. The intrinsic function relationship F between the state xN+k at the (N+k)th time and the first N states is expressed as:
XN+k=F(Xl,X2,...,X.)
(1) The life prediction of the distillation column can be expressed as finding F by establishing a mathematical model, so that 任 ∈>0, with ||FF(X)-FF(X)||=||xN+k-XXN +k||
<∈>(2) where F(X)=xN+k actual state value, FF(x)=xN+k prediction model value.
The file management system includes the overall state of the rectification column, the performance of each component of the rectification column, and economic and technical evaluation. The system flow is shown in Figure 4.
3 Conclusion With the rapid development of modern science and technology, modern fault diagnosis systems are constantly improving. For example, artificial intelligence, neural network and dynamic simulation software can prevent and avoid accidents in time and accurately, so maintenance technology, The organic combination of the obstacle diagnosis technology and the computer remote monitoring technology is the development direction of fault diagnosis for refining and chemical enterprises in the future.
Products Of Livestock Reproduction
Livestock Reproduction,Custom 18" Sanitary Sleeve,21" Sanitary Sleeve,New Design Sanitary Sleeve
Jinan Mucho Commercial Inc. , https://www.muchovet.com