CN106295132A - A kind of air separation plant varying duty optimization method based on mould plate technique - Google Patents

Abstract

A variable load optimization method for air separation plant based on template technology, characterized in that the variable load optimization method is: by setting up a dynamic template containing variable marks and air separation plant mechanism models, before each real-time optimization calculation from real-time The real-time data and load information are obtained from the database, and then the template engine is used to analyze the template to generate the corresponding mechanism model under the current load, and the generated real-time data is embedded in it, and finally the optimization solver is called to optimize the mechanism model; During the load production process, the template parameters can be corrected in time according to the implementation of the operating variables on the premise that the optimized model structure remains unchanged, so as to adapt to the ever-changing variable load production environment; it has the advantages of simple implementation, small environmental dependence, and simple and clear principles , good stability and real-time performance, easy to realize on the computer, and good flexibility, can better meet the characteristics of large-scale air separation equipment variable load production requirements and so on.

Description

A variable load optimization method for air separation plant based on template technology

technical field

The invention relates to the field of real-time optimization research of air separation equipment in chemical and metallurgical industries, and in particular relates to a variable load optimization method for air separation equipment based on template technology.

Background technique

Air separation equipment is a large-scale equipment widely used in metallurgy, chemical industry, petrochemical, urban municipal engineering, medical treatment, aerospace and other fields. It is closely related to modern industry, especially various high-tech industries. Its development scale and technical status have become a measure of a country's An important symbol of China's industrial and technological development level. In recent years, with the rapid increase in demand for industrial gases such as oxygen, nitrogen, and argon, the demand for air separation units is also increasing. It is estimated that by 2015, the total demand for air separation units in my country will be 10.5 million to 11.4 million Nm3/h oxygen equivalent, and the average annual new demand will be 1.04 million to 1.14 million Nm3/h oxygen equivalent.

However, in industrial production, gas demand is not fixed, but cyclical, staged, and intermittent, which leads to large changes in the production load of air separation units to adapt to changes in demand. Taking iron and steel enterprises as an example, due to the particularity of the process (such as converter top blowing, intermittent oxygen use, continuous use of blast furnace oxygen enrichment, pulverized coal injection), the instantaneous oxygen consumption is large and the time is not continuous; The size of the converter is different, and the cycle of the peak and trough of the oxygen consumption is also different, resulting in an extremely unbalanced demand for the oxygen consumption. The user's unbalanced use of oxygen and the complex characteristics of the air separation process lead to the lack of automatic load-variable production optimization technology, and the production load cannot be adjusted in time, resulting in production fluctuations and accompanying failures, and at the same time causing a large amount of energy consumption and Economic losses. Therefore, how to comprehensively apply advanced process modeling and optimization control technology to realize the optimization control system with variable load technology as the core has become an urgent need in today's air separation industry.

Contents of the invention

The object of the present invention is to overcome the deficiencies in the prior art, and provide a simple implementation, small environmental dependence, simple and clear principle, good stability and real-time performance, easy to realize on the computer, and good flexibility, can be more A variable-load optimization method for air separation plants based on template technology that satisfies the variable-load production requirements of large-scale air separation plants well.

The object of the present invention is achieved through the following technical solutions: a method for optimizing the variable load of an air separation plant based on template technology, wherein the method for variable load optimization is: by setting up a dynamic template containing variable marks and the mechanism model of the air separation plant , obtain real-time data and load information from the real-time database before each real-time optimization calculation, then parse the template through the template engine to generate the corresponding mechanism model under the current load, embed the generated real-time data into it, and finally call the optimization solver to optimize the mechanism The model is optimized and solved; in the variable load production process of the air separation unit, the template parameters can be corrected in time according to the implementation of the operating variables on the premise that the structure of the optimized model remains unchanged, so as to adapt to the constantly changing variable load production environment.

As preferably: described variable load optimization method, specifically comprises the steps:

Step 1: Establish a group of optimization models corresponding to different loads of the air separation unit according to the material balance equation, and the expression form of each optimization model in this group is as follows:

optimize the target:

z=FobjFGOX,FLOX,FGAN,FLIN,σi,Δii=1...4

Restrictions:

Fcon1FGOX,FLOX,FGAN,FLIN,Y＝0

FuFGOX,FLOX,FGAN,FLIN,Y＜Up

FlFGOX,FLOX,FGAN,FLIN,Y＞Low

Among them, FGOX is the flow rate of oxygen product, FLOX is the flow rate of liquid oxygen, FGAN is the flow rate of nitrogen product, FLIN is the flow rate of liquid nitrogen, σi represents the correction coefficient, Δi represents the deviation between the actual value and the calculated output value of the previous stage, and Fcon1 represents the equality constraint, Fu represents the upper bound constraint equation, Fl represents the lower bound constraint equation, Y represents the operating variable, Up represents the upper bound constraint vector of the left variable, and Low represents the lower bound constraint vector of the left variable;

Step 2: Establish a dynamic working template Template=T for real-time optimization of the air separation plant, which includes: variable labels, differential variable labels, condition labels, constants, and optimization model parameters.

Step 3: Obtain the key parameters reflecting the production load and stage target setting values from the on-site real-time database as needed, and then use this data to analyze the dynamic work template established in step 2, and replace the tags in the work template with reflective Real-time information on actual operating conditions and embedding it in the optimization model for the final solution;

Step 4: calling the solver to optimize and solve the optimization model of the current working condition of the reaction air separation plant generated in step 3;

Step 5: Judging whether the solution is successful according to whether the optimization proposition calculation of the operating point of the air separation plant described in step 4 converges; if the optimization proposition calculation converges, then skip to step 6; if the calculation process does not converge, the solution fails or the solution time exceeds the optimization time T Then skip to the following step eight;

Step 6: Input the calculation results of the optimal operating point of the air separation plant into the on-site real-time database;

Step 7: After waiting for the operator to confirm that the working conditions are stable, calculate the deviation between the output variable and the actual value in the previous stage and obtain the variable load target setting value in the next stage and return to step 3; until the load variable process is completed, the operator exits the system ;

Step 8: Wait for the operator to re-input the variable load target setting value of the next stage.

As a preference: in the step 2, the variable mark parameter indicates that the template needs to use real-time production data to replace the variable when parsing, and the differential variable parameter mark represents that the template needs to use the current value of the real-time production data when parsing. The variable that is replaced after the sampling value at the previous moment is differentiated; the condition tag parameter indicates whether the template uses the content set in the condition tag according to the conditions set by the tag during parsing; the constant parameter indicates that the mathematical model is related to the process The algebraic constant of ; the optimization model parameter represents the optimization model that best matches the current load in the optimization model set established in step 1.

The beneficial technical effect of the present invention is:

1) The present invention transmits the production real-time data into the data model to carry out optimization calculation by template, under the situation that traditional method can't carry out real-time optimization, this method has realized the real-time optimization of air separation production process;

2) The present invention is simple to realize and has little environmental dependence;

3) The principle of the present invention is simple and clear, easy to implement on a computer, and has good flexibility, which can better meet the complex requirements of variable load production of large-scale air separation equipment.

Description of drawings

Figure 1 is a schematic diagram of a real-time optimization system based on template technology for an air separation plant.

detailed description

The present invention will be described in more detail below with reference to the accompanying drawings of the present invention. However, the present invention can be embodied in many different forms, and therefore it should not be considered limited to the examples set forth in the specification. Describe the specific implementation process of the present invention.

As shown in Fig. 1, the present invention includes OPC (OLE for Process Control) Server, a set of special solver (GAMS) for nonlinear programming problems, and a working template for real-time optimization with and transmission of production data. First, the on-site generated data is obtained through the OPC Server, and then the process mechanism model file corresponding to the current actual working condition is obtained through the template analysis engine working template. Finally, the GAMS solver is called to optimize the solution to obtain a series of optimal operating variables. Finally, these optimal operating variable values are sent to the on-site APC advanced controller through the OPC Server as the optimal setting value.

A kind of variable load optimization method of air separation plant based on template technology described in the present invention, this variable load optimization method is: by setting up the dynamic template that contains variable mark and mechanism model of air separation plant, before each real-time optimization calculation, from real-time The real-time data and load information are obtained from the database, and then the template engine is used to analyze the template to generate the corresponding mechanism model under the current load, and the generated real-time data is embedded in it, and finally the optimization solver is called to optimize the mechanism model; During the load production process, the template parameters can be corrected in time according to the implementation of the operating variables on the premise that the optimized model structure remains unchanged, so as to adapt to the ever-changing variable load production environment.

The variable load optimization method of the present invention specifically comprises the following steps:

Step 1: Establish a group of optimization models corresponding to different loads of the air separation unit according to the material balance equation, and the expression form of each optimization model in this group is as follows:

optimize the target:

z=FobjFGOX,FLOX,FGAN,FLIN,σi,Δii=1...4

Restrictions:

Fcon1FGOX,FLOX,FGAN,FLIN,Y＝0

FuFGOX,FLOX,FGAN,FLIN,Y＜Up

FlFGOX,FLOX,FGAN,FLIN,Y＞Low

Among them, FGOX is the flow rate of oxygen product, FLOX is the flow rate of liquid oxygen, FGAN is the flow rate of nitrogen product, FLIN is the flow rate of liquid nitrogen, σi represents the correction coefficient, Δi represents the deviation between the actual value and the calculated output value of the previous stage, and Fcon1 represents the equality constraint, Fu represents the upper bound constraint equation, Fl represents the lower bound constraint equation, Y represents the operating variable, Up represents the upper bound constraint vector of the left variable, and Low represents the lower bound constraint vector of the left variable;

Step 2: Establish a dynamic working template Template=T for real-time optimization of the air separation plant, which includes: variable labels, differential variable labels, condition labels, constants, and optimization model parameters.

Step 3: Obtain the key parameters reflecting the production load and stage target setting values from the on-site real-time database as needed, and then use this data to analyze the dynamic work template established in step 2, and replace the tags in the work template with reflective Real-time information on actual operating conditions and embedding it in the optimization model for the final solution;

Step 4: calling the solver to optimize and solve the optimization model of the current working condition of the reaction air separation plant generated in step 3;

Step 5: Judging whether the solution is successful according to whether the optimization proposition calculation of the operating point of the air separation plant described in step 4 converges; if the optimization proposition calculation converges, then skip to step 6; if the calculation process does not converge, the solution fails or the solution time exceeds the optimization time T Then skip to the following step eight;

Step 6: Input the calculation results of the optimal operating point of the air separation plant into the on-site real-time database;

Step 7: After waiting for the operator to confirm that the working conditions are stable, calculate the deviation between the output variable and the actual value in the previous stage and obtain the variable load target setting value in the next stage and return to step 3; until the load variable process is completed, the operator exits the system ;

Step 8: Wait for the operator to re-input the variable load target setting value of the next stage.

In step 2 of the present invention, the variable mark parameter indicates that the template needs to use real-time production data to replace the variable when parsing, and the difference variable parameter mark represents that the template needs to use the current value and previous value of real-time production data during parsing. The variable that is replaced after the sampling value at a moment is differentiated; the condition tag parameter indicates whether the template uses the content set in the condition tag according to the conditions set by the tag during parsing; the constant parameter indicates the process-related parameters in the mathematical model Algebraic constant; the optimization model parameter represents the optimization model that best matches the current load in the optimization model set established in step 1.

Example:

Taking the 20,000 air separation unit of a gas company in a steel plant in Jiangsu as an example, the template technology is used to optimize the production process of the air separation unit in real time, including the following steps in the computer system:

Step 1: Select output variables as product output FGOX, FLOX, FGAN and FLIN. The operating variables are FAIR, FTURBINE, FHPAIR, FWN2, FLAR, and FGAR, and the working condition optimization model is established according to the physical balance as follows:

Min:

z=sqrFLOX+δ1ΔLOX-FLOXSP1600+sqrFGOX+δ2ΔGOX-FGOXSP20000+sqrFLIN+δ3ΔLIN-FL

INSP1600+sqrFGAN+δ4ΔGAN-FGANSP40000

s.t

FkXi, Yj=0 (i=1...4, j=1...6, k=high, middle, low)

12000,0,24000,0T＜X＜[24000,1680,48000,1680]T

[61980,7250,17250,20000,50,400]T<Y<[123960,34500,34500,64000,600,400]T

Among them: FLOX represents the product liquid oxygen flow rate, FGOX represents the product oxygen flow rate, FLIN represents the product liquid nitrogen flow rate, FGAN represents the product nitrogen gas flow rate, and δ1, δ2, δ3, and δ4 represent the correction coefficients of the four output variables respectively (in this example, it is set to 0.92 ,0.95,1.02,1.17), ΔLOX, ΔGOX, ΔLIN, ΔGAN represent the deviation between the actual value of the product in the previous stage and the calculated output value, FAIR represents the total air flow, FTURBINE represents the expansion air flow, FHPAIR represents the high-pressure air flow, FWN2 represents Contaminated nitrogen flow, FLAR represents liquid argon flow, FGAR represents crude argon flow, X represents the output variable column vector, Y represents the operating variable column vector, Fk represents the optimization model corresponding to the typical load. The subscript k indicates the corresponding load segment. FLOXSP, FGOXSP, FLINSP, and FGANSP indicate that the current target output setting value is selected according to the actual load. The value table is as follows:

Table 1: Value table for different load segments

Note: The output target value of the stage corresponding to the load not in the load value table can be calculated by linear interpolation according to the load table.

Step 2: Replace the input variables and operating variables established in step 1 with the variable tags of the template to connect the mechanism model and actual production data. The corresponding replacement table is as follows:

Table 2: Variable label comparison table

Step 3: When the staff needs to optimize the production operation in real time, start the system.

Step 4: The template engine of the system collects production data through the OPC interface and uses the optimization model set established in step 1 to generate an optimization model corresponding to the current load.

Step 5: Judging whether the solution is successful or not according to whether the optimization proposition calculation of the operating point of the air separation plant corresponding to the current load generated in step 4 converges. If the optimization proposition calculation converges, skip to step six. If the calculation process does not converge, the solution fails or the solution takes more than 60 seconds, skip to step 7.

Step 6: Output the optimal operating variables obtained from the optimization calculation to the OPC server of the factory

Step 7: After waiting for the operator to confirm that the working conditions are stable, calculate the deviation between the output variable of the previous stage and the actual value and obtain the variable load target setting value of the next stage, and return to step 4. The operator exits the system until the load changing process is completed.

Step 8: Wait for the operator to input another set of stage target values.

As mentioned above, the invention can also be embodied in many different forms and therefore should not be considered limited to the embodiments set forth in the specification. The principle of the method adopted in the present invention is simple and clear, and is convenient to implement on a computer, has good flexibility, and can well meet the requirements of real-time optimization, rapidity, safety and the like of the air separation plant.

Claims (3)

1. A variable load optimization method for air separation plant based on template technology, characterized in that the variable load optimization method is: by setting up a dynamic template containing variable marks and air separation plant mechanism models, before each real-time optimization calculation Obtain real-time data and load information from the real-time database, then parse the template through the template engine to generate the corresponding mechanism model under the current load, embed the generated real-time data into it, and finally call the optimization solver to optimize the mechanism model; in the air separation During the variable load production process of the device, under the premise that the optimized model structure remains unchanged, the template parameters can be corrected in time according to the implementation of the operating variables to adapt to the constantly changing variable load production environment. 2. the variable load optimization method of air separation plant based on template technology according to claim 1, is characterized in that described variable load optimization method, specifically comprises the steps: Step 1: Establish a group of optimization models corresponding to different loads of the air separation unit according to the material balance equation, and the expression form of each optimization model in this group is as follows: Among them, FGOX is the flow rate of oxygen product, FLOX is the flow rate of liquid oxygen, FGAN is the flow rate of nitrogen product, FLIN is the flow rate of liquid nitrogen, σi represents the correction coefficient, Δi represents the deviation between the actual value and the calculated output value of the previous stage, and Fcon1 represents the equality constraint, Fu represents the upper bound constraint equation, Fl represents the lower bound constraint equation, Y represents the operating variable, Up represents the upper bound constraint vector of the left variable, and Low represents the lower bound constraint vector of the left variable; Step 2: Establish a dynamic working template Template=T for real-time optimization of the air separation plant, which includes: variable labels, differential variable labels, condition labels, constants, and optimization model parameters. Step 3: Obtain the key parameters reflecting the production load and stage target setting values from the on-site real-time database as needed, and then use this data to analyze the dynamic work template established in step 2, and replace the tags in the work template with reflective Real-time information on actual operating conditions and embedding it in the optimization model for the final solution; Step 4: calling the solver to optimize and solve the optimization model of the current working condition of the reaction air separation plant generated in step 3; Step 5: Judging whether the solution is successful according to whether the optimization proposition calculation of the operating point of the air separation plant described in step 4 converges; if the optimization proposition calculation converges, then skip to step 6; if the calculation process does not converge, the solution fails or the solution time exceeds the optimization time T Then skip to the following step eight; Step 6: Input the calculation results of the optimal operating point of the air separation plant into the on-site real-time database; Step 7: After waiting for the operator to confirm that the working conditions are stable, calculate the deviation between the output variable and the actual value in the previous stage and obtain the variable load target setting value in the next stage and return to step 3; until the load variable process is completed, the operator exits the system ; Step 8: Wait for the operator to re-input the variable load target setting value of the next stage. 3. the variable load optimization method of air separation plant based on template technology according to claim 2, is characterized in that in described step 2, described variable mark parameter expression template needs to use real-time production data to replace when analyzing Variable, difference variable parameter mark indicates that the template needs to use the current value of real-time production data and the sampling value of the previous moment to make a difference and replace the variable when parsing; the condition mark parameter represents the template that is set according to the mark when parsing The condition selects whether to use the content set in the condition mark; the constant parameter indicates the algebraic constant related to the process in the mathematical model; the optimization model parameter indicates the optimization model that best matches the current load in the optimization model set established in step 1.