Multi-target speed-adjustable MPC self-optimization method

Abstract

The invention relates to a multi-target speed-adjustable MPC self-optimization method, and the method comprises the steps: constructing an economic optimization soft constraint, i.e., a second constraint, based on an appeal of economic optimization; based on the overall target tracking and economic optimization appeals, constructing a steady-state optimization target function and generating a problem of overall steady-state optimization; priority information of output constraints is introduced, an output constraint relaxation problem is constructed, the optimization target of the constraint relaxation problem enables the weighted sum of constraint relaxation amount to be minimum, a relaxation factor is solved, and the output constraints are updated; economic optimization priorities of the input variables are introduced, an economic optimization problem is constructed, an optimization target in the economic optimization problem enables the speed to meet specified requirements during economic optimization, and constraints of the input variables are updated; and based on the updated constraint, solving the problem of the whole steady-state optimization to obtain a solution of the input variable. The method can ensure that the input is economically optimized according to the specified stride under the condition that the system has the degree of freedom while the output constraint is met.

Description

Technical Field

[0001] The invention relates to the technical field of industrial control, and in particular to a multi-objective speed-adjustable MPC self-optimization method. Background Art

[0002] With the increasing requirements of lean production in process industries, APC (Advanced process control) not only has to meet the requirements of multivariable coordinated control stability, but also faces the demand for improved economic optimization benefits. Relying on stable control and pushing the control index edge can bring some economic benefits, which is the first level of benefit source; by solving self-optimization problems and performing optimal solutions in MPC (Model predictive control) steady-state optimization, the remaining benefits in the operating space can be further mined, which is the second level of economic benefit mining.

[0003] The predictive control technology of process industry generally adopts a two-layer structure. The upper layer is the steady-state optimization layer, which mainly uses the process steady-state model to solve the steady-state optimal solution that meets the process constraints and control objectives. The lower layer is the dynamic control layer, which takes the optimization result given by the steady-state optimization layer as the goal, calculates the optimal control output, and makes the process reach the final steady-state goal.

[0004] When the controller has insufficient degrees of freedom, steady-state optimization faces the problem of constraint trade-offs. When the controller has excess degrees of freedom, steady-state optimization faces the problem of recommending the local optimal point of the controller. Existing optimization schemes usually construct an optimization objective function and use LP (linear programming) and QP (quadratic programming) methods to solve the problem while satisfying steady-state constraints to obtain the optimal MV solution and achieve optimal economic performance during control. However, this method faces the following problems: First, the method of solving the optimization target cannot ensure the consistency of the direction of the two optimization results. In the face of model mismatch, noise interference, model gain changes, changes in raw material or product price coefficients, the results of the optimization solution have great uncertainty.

[0005] Second, when using the method of solving the optimization objective, when dealing with multi-objective optimization, it can only be integrated into one objective function, and the optimization result depends on the result of multi-objective parameter adjustment.

[0006] Third, the method of solving the problem using the optimization objective often optimizes to the optimal solution in one step. There is a large jump from a steady-state solution to the optimal solution, which can easily lead to unstable control.

[0007] Fourth, there is usually no obvious basis for judging from steady-state optimization to economic optimization. Regardless of whether there are degrees of freedom, economic optimization will be entered, which will lead to unnecessary calculations.

[0008] Fifth, there is no clear indication that the controller is performing economic optimization, making it impossible to give a reasonable explanation for the controller behavior. Summary of the invention

[0009] In view of the above shortcomings and deficiencies of the prior art, the present invention provides a multi-objective speed-adjustable MPC self-optimization method, which solves the problem of defining low-priority soft constraints in the steady-state optimization process and can realize the importance of multiple economic optimization objectives. In order to achieve the above purpose, the main technical solutions adopted by the present invention include: In a first aspect, an embodiment of the present invention provides a multi-objective speed-adjustable MPC self-optimization method, comprising: S100, pre-establishing a first model and a first constraint set of a steady-state optimization problem for a specific optimization process of an industrial process; S200, based on the demand for economic optimization, construct an economic optimization soft constraint with a speed adjustment factor, namely the second constraint (economic constraint); S300, constructing a steady-state optimization objective function based on the overall target tracking and economic optimization requirements; generating the entire steady-state optimization problem (including the objective function and the constraints) according to the first model, the first constraint set and the second constraint; S400, introducing the priority information of the output constraint in the first constraint set, and constructing an output constraint relaxation problem, wherein the optimization objective of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts, calculate the relaxation factor, and update the output constraint; Introduce the economic optimization priority of the input variables in the second constraint, construct an economic optimization problem, and the optimization goal in the economic optimization problem is to make the speed during economic optimization meet the specified requirements, and update the constraints of the input variables; Based on the updated constraints, the entire steady-state optimization problem is solved to obtain the solution of the input variables.

[0010] Optionally, the S200 includes: The objective function J in the entire steady-state optimization problem to be solved is: ; Among them, the parameters represents the cost of the i-th operation variable, parameter Represents the penalty for the i-th operation variable deviating from its value at the previous moment, cost parameter represents the unit penalty cost of the ith manipulated variable deviating from the set point, and the variable represents the measured value of the i-th operating variable at the current time k, and the parameter represents the unit penalty for the jth controlled variable deviating from the set point, the variable represents the open-loop steady-state value of the jth controlled variable at the current time k, and the variable represents the set value of the jth controlled variable; represents the steady-state optimization increment of the ith input, represents the set value of the i-th input variable, represents the steady-state optimization increment of the j-th output; The complete steady-state optimization constraints in the entire steady-state optimization problem to be solved include: ; in, Represents the model gain of the i-th input to the j-th output; Respectively represent the lower and upper limits of the steady-state optimization increment of the j-th output; Represents the lower and upper limits of the position of the j-th output; represents the lower and upper bounds of the steady-state optimization increment for the ith input; Represents the lower and upper limits of the position of the i-th input; Represents input Speed ​​adjustment factor, according to the input maximization or minimization demand, enables the corresponding economic optimization soft constraint.

[0011] Optionally, the constraints in the entire steady-state optimization problem in S300 include three types of constraints: the first type is the input stride constraint and the input upper and lower limit constraints; the second type is the output stride constraint and the output upper and lower limit constraints; the third type is the economic optimization constraint / economic constraint; and / or, In S400, the second type of constraint priority is set according to the control demand and the demand for economic optimization; the third type of constraint priority is lower than the second type of constraint priority; Specifically, construct the output constraint relaxation problem and define the non-negative relaxation factor of each output constraint as , and the corresponding relaxation weights are ,The optimization objective J1 of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts;

[0012] ; Solving for J1, we get After the value is set, the output constraints are updated and the updated output constraints are used as hard constraints.

[0013] Optionally, the method further comprises: S500, judging the necessity of economic optimization for solving the optimization target in the constraint relaxation problem; Specifically, if the relaxation factor corresponding to any constraint is There are values ​​greater than 0, indicating that the output constraint is relaxed, the controller has no degrees of freedom, and does not need to enter economic optimization; If there is at least one output constraint and both relaxation factors are equal to 0, it means that the constraint has not been relaxed. At this time, there is room for economic optimization and the next step of economic optimization is entered.

[0014] Optionally, the method further comprises: S600, if it is necessary to enter into economic optimization, then enter into economic optimization and construct an economic optimization problem; According to the economic optimization requirements of different variables, different priorities are set to achieve the order of variable optimization; Specifically, let the economic optimization priority of all variables be , and define the non-negative relaxation factor of each economic optimization constraint as , the specific constraints are selected according to the direction of variable optimization. If the variable has no economic optimization requirements, then the constraint does not exist, that is, there is no corresponding The purpose of the optimization objective J2 is to make the speed meet the specified requirements during economic optimization; ; ; Solve for J2 and obtain the relaxation factor The value of If the input variable still has the degree of freedom for economic optimization, the corresponding It must be 0. If the input variable has no degrees of freedom, the output Must be greater than 0.

[0015] Optionally, the S100 includes: S101, establishing a controlled variable constraint set for steady-state optimization, wherein the controlled variable constraint set includes: a steady-state optimization increment lower limit and upper limit constraint of the controlled variable, and a position limit constraint; S102, establishing a set of operating variable constraints for steady-state optimization, wherein the set of operating variable constraints includes: a lower limit and an upper limit constraint of a steady-state optimization increment of the operating variable, and a position limit constraint; S103: Establish a first model of a steady-state optimization problem, wherein the controlled variable constraint set and the manipulated variable constraint set constitute the first constraint set.

[0016] Optionally, for the process control of the heavy oil fractionation tower, the controlled variable constraint set in the first constraint set includes: ; The operation variable constraint set in the first constraint set includes: ; The first model includes: ; in, Represents the model gain of the i-th input to the j-th output; represents the steady-state optimization increment of the j-th output, Respectively represent the lower and upper limits of the steady-state optimization increment of the j-th output, represents the predicted future steady-state value of the j-th output at time k, , Represents the lower and upper limits of the position of the j-th output; represents the steady-state optimization increment of the j-th input, Represent the lower and upper limits of the steady-state optimization increment of the j-th input, respectively. represents the current value of the jth input, Represents the lower and upper limits of the position of the j-th input.

[0017] Alternatively, if the optimization goal is to increase the extraction volume of the top product of the distillation tower to the maximum value , the side extraction volume of the distillation tower increases to the maximum value , and the reflux heat load at the bottom of the distillation tower Reduce to a minimum value; Then, the economic optimization constraints include: , For input Speed ​​adjustment factor, and i is 1, 2, 3; Alternatively, enter To track a specific value, the economic optimization constraints include: ; in, is the measured value of the i-th input at time k, represents the steady-state optimization increment of the ith input, is the set value, Represent the lower and upper limits of the steady-state optimization increment of the ith input, respectively. Speed ​​adjustment factors for upper and lower increment limits for steady state optimization.

[0018] In a second aspect, an embodiment of the present invention further provides a control device for a multi-objective speed-adjustable MPC self-optimization process for an industrial process, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program stored in the memory, specifically executing any of the multi-objective speed-adjustable MPC self-optimization methods described above.

[0019] In a third aspect, the present invention further provides a computer-readable storage medium, comprising: a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the multi-objective speed-adjustable MPC self-optimization method described in any one of the above are implemented. The beneficial effects of the present invention are as follows: In the embodiment of the present invention, a new economic optimization implementation strategy is adopted to implement the technical route of self-optimization. In the steady-state optimization problem, by adopting a soft constraint problem with a low priority to replace the method of solving the optimization objective function in the economic optimization stage, and introducing a speed adjustment factor, adjusting the weight and other coefficients, the following advantages can be achieved: (1) Control constraints are satisfied first, and economic optimization is entered only when there are degrees of freedom, thereby ensuring that economic optimization is performed under the premise of achieving the current control. (2) Economic optimization is achieved by means of constraints, and the direction of economic optimization can be accurately controlled. (3) Through multiple economic optimization constraints of different levels, the importance of multiple economic optimization objectives can be better weighed, and the optimization of a certain adjustment means can be completed before entering the optimization of the next adjustment means. (4) By introducing the speed factor, the pace of economic optimization can be adjusted to ensure smooth control in the economic optimization link. (5) Combined with the judgment of the constraint relaxation amount of the conventional steady-state optimization problem, it can be judged in advance whether to enter the self-optimization link, thereby reducing unnecessary calculations. (6) By satisfying the constraints in the self-optimization link, it can be judged whether the controller has entered the economic optimization stage, thereby providing a certain explanation for the controller behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic flow chart of a multi-objective speed-adjustable MPC self-optimization method provided by an embodiment of the present invention; Figure 2 A schematic flow chart of a multi-objective speed-adjustable MPC self-optimization method provided by another embodiment of the present invention; Figure 3 A schematic diagram of the structure of a control device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below through specific implementation modes in conjunction with the accompanying drawings.

[0022] Example 1

[0023] Figure 1 and Figure 2 They are respectively flowchart diagrams of a multi-objective speed-adjustable MPC self-optimization method in Example 1 of the present invention. The method of this embodiment includes the following steps: S100. For a specific optimization process of an industrial process, a first model and a first constraint set of a steady-state optimization problem are pre-established.

[0024] For example, for the current working point of the industrial production process, the basic constraint objective function (i.e., the first model) and basic constraint conditions (i.e., the first constraint set) are constructed according to the upper and lower limits of multiple input variables, the incremental step, the control constraints of the output variables, and the incremental step conditions.

[0025] S200, based on the demand for economic optimization, construct an economic optimization soft constraint with a speed adjustment factor, namely the second constraint (i.e., economic constraint); S300, constructing a steady-state optimization objective function based on the overall target tracking and economic optimization requirements; generating the entire steady-state optimization problem (including the objective function and the constraints) according to the first model, the first constraint set and the second constraint; S400, introducing the priority information of the output constraint in the first constraint set, and constructing an output constraint relaxation problem, wherein the optimization objective of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts, calculate the relaxation factor, and update the output constraint; Introduce the economic optimization priority of the input variables in the second constraint, construct an economic optimization problem, and the optimization goal in the economic optimization problem is to make the speed during economic optimization meet the specified requirements, and update the constraints of the input variables; Based on the updated constraints, the entire steady-state optimization problem is solved to obtain the solution of the input variables.

[0026] It can be understood that in step S300, the constraints in the entire steady-state optimization problem include three types of constraints: the first type is the input step constraint and the input upper and lower limit constraints; the second type is the output step constraint and the output upper and lower limit constraints; the third type is the economic optimization constraint / economic constraint.

[0027] In S400, the second type of constraint priority is set according to the control demand and the demand for economic optimization; the third type of constraint priority is lower than the second type of constraint priority; Specifically, construct the output constraint relaxation problem and define the non-negative relaxation factor of each output constraint as , and the corresponding relaxation weights are The optimization objective J1 of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts.

[0028]

[0029] ; Solving for J1, we get After the value is set, the output constraints are updated and the updated output constraints are used as hard constraints.

[0030] Accordingly, the above method may further include the following steps S500 and S600: S500, judging the necessity of economic optimization for solving the optimization target in the constraint relaxation problem; Specifically, if the relaxation factor corresponding to any constraint is There are values ​​greater than 0, indicating that the output constraint is relaxed, the controller has no degrees of freedom, and does not need to enter economic optimization; If there is at least one output constraint and both relaxation factors are equal to 0, it means that the constraint has not been relaxed. At this time, there is room for economic optimization and the next step of economic optimization is entered.

[0031] S600, if it is necessary to enter into economic optimization, then enter into economic optimization and construct an economic optimization problem; According to the economic optimization requirements of different variables, different priorities are set to achieve the order of variable optimization; Specifically, let the economic optimization priority of all variables be , and define the non-negative relaxation factor of each economic optimization constraint as , the specific constraints are selected according to the direction of variable optimization. If the variable has no economic optimization requirements, then the constraint does not exist, that is, there is no corresponding The purpose of the optimization objective J2 is to make the speed meet the specified requirements during economic optimization; ; ; Solve for J2 and obtain the relaxation factor The value of If the input variable still has the degree of freedom for economic optimization, the corresponding It must be 0. If the input variable has no degrees of freedom, the output Must be greater than 0.

[0032] The above method belongs to a new technical route for self-optimization of economic optimization implementation strategy. In steady-state optimization problems, by adopting the definition of low-priority soft constraint problems to replace the method of solving the optimization objective function in the economic optimization stage, and introducing parameters such as constraint priority and speed adjustment factor, combined with the judgment of optimization solution results, the following advantages can be achieved: (1) Give priority to satisfying control constraints and enter economic optimization only when there are degrees of freedom, thereby ensuring that economic optimization is performed under the premise of achieving control constraints; (2) Economic optimization can be achieved through constraint methods, which can accurately control the economic optimization method; (3) By using economic optimization constraints at different levels, the importance of multiple economic optimization objectives can be better weighed, and the optimization of a certain adjustment measure can be completed before moving on to the optimization of the next adjustment measure.

[0033] (4) By introducing the speed factor, the pace of economic optimization can be adjusted to ensure smooth control in the economic optimization link.

[0034] (5) Combined with the judgment of the relaxation amount of the constraints of the steady-state optimization problem, it is possible to determine in advance whether to enter the self-optimization stage, thereby reducing unnecessary calculations.

[0035] (6) By satisfying the constraints of the self-optimization link, it is possible to determine whether the controller has entered the economic optimization stage, thereby providing a certain explanation for the controller behavior.

[0036] Example 2 Taking the process model of Shell heavy oil fractionation tower as an example, the process has three operating variables, Nu=3, which are the extraction amount of the top product of the fractionation tower , Sideline extraction volume of fractionation tower and the reflux heat load at the bottom of the distillation tower The process has three controlled variables, Ny=3, which are the components of the top product of the distillation tower. , fractionation tower side line extraction components and the reflux temperature at the bottom of the distillation tower ; Set the system input and output constraints to , the input-output steady-state optimization increment constraint is .

[0037] The first step: problem construction, that is, establishing the mathematical expression of the steady-state optimization problem.

[0038] First, a set of controlled variable constraints for steady-state optimization is established. These constraints are generally soft constraints, which are the lower and upper limit constraints of the steady-state optimization increment of the controlled variables and the position limit constraints.

[0039] (A1) In formula (A1), represents the steady-state optimization increment of the j-th output, Respectively represent the lower and upper limits of the steady-state optimization increment of the j-th output, represents the predicted future steady-state value of the j-th output at time k, Represents the lower and upper limits of the position of the j-th output. According to the problem assumption, , .

[0040] Secondly, a set of operating variable constraints for steady-state optimization is established. These constraints are generally hard constraints, which are the lower and upper limits of the steady-state optimization increment of the operating variables and the position limit constraints.

[0041] (A2) In formula (A2), represents the steady-state optimization increment of the j-th input, Represent the lower and upper limits of the steady-state optimization increment of the j-th input, respectively. represents the current value of the jth input, Represents the lower and upper limits of the position of the jth input. According to the problem assumption, , .

[0042] Next, establish the relationship between output and input, which is determined by the model relationship.

[0043] , (A3) In formula (A3), represents the model gain of the i-th input to the j-th output, represents the steady-state optimization increment of the j-th output, Respectively represent the lower and upper limits of the steady-state optimization increment of the j-th output, represents the predicted future steady-state value of the j-th output at time k, Represents the lower and upper limits of the position of the j-th output; represents the steady-state optimization increment of the j-th input, Represent the lower and upper limits of the steady-state optimization increment of the j-th input, respectively. represents the current value of the jth input, Represents the lower and upper limits of the position of the j-th input.

[0044] Step 2: Define the self-optimization goal.

[0045] Usually, the desired extraction rate of the top product of the distillation tower is The bigger the better, the side line extraction volume of the distillation tower The larger the better, and the reflux heat load at the bottom of the distillation tower The smaller the better. At this time, we use the soft constraint method to achieve our economic optimization needs and introduce the speed adjustment factor. Compared with the previous method of achieving economic optimization by establishing an economic optimization objective function, the constraint method has many advantages.

[0046] The corresponding economic optimization constraints are: (A4) In formula (A4), It is called the economic optimization speed adjustment factor, which takes a value between 0 and 1 and is used to achieve self-optimization speed regulation. The speed adjustment factor can set different values ​​for each variable.

[0047] Step 3: Define the steady-state optimization objective function.

[0048] The goal of the entire steady-state optimization is to achieve the best overall economic benefits while meeting the CV target tracking. At the same time, for variables that do not require economic optimization, the smaller the action, the better. Under the above constraints, the steady-state optimization objective function can be simplified to: (A5) In the above formula, the parameters represents the cost of the i-th operation variable, parameter Represents the penalty for the i-th operation variable deviating from its value at the previous moment, cost parameter represents the unit penalty cost of the ith manipulated variable deviating from the set point, and the variable represents the measured value of the i-th operating variable K at the current moment, and the parameter represents the unit penalty for the jth controlled variable deviating from the set point, the variable represents the open-loop steady-state value of the jth controlled variable at the previous moment, and the variable represents the set value of the jth controlled variable, represents the steady-state optimization increment of the ith input, represents the set value of the i-th input variable, Represents the steady-state optimization increment of the j-th output.

[0049] Step 4: Solve the overall optimization problem.

[0050] Based on the above steady-state optimization objective function and all constraints, the overall optimization problem can be written as follows: ; ; Step 5: Output constraint relaxation solution.

[0051] The above mentioned three types of constraints are involved. The first type is the input stride constraint and the input upper and lower limit constraints. This is a hard constraint and cannot be violated under any circumstances. The second type is the output stride constraint and the output upper and lower limit constraints. This is a soft constraint. The constraint priority can be set independently according to the user's control requirements. The third type is the economic optimization constraint. Economic optimization is performed while meeting the control requirements. The constraint priority is generally lower than the control demand priority.

[0052] Taking the constructed problem as an example, assume that the output constraint has a priority of 1, which is the highest priority among all soft constraints. Then the first step is to construct the following output constraint relaxation problem, defining the non-negative relaxation factor of each output constraint as , and the corresponding relaxation weights are The purpose of the optimization objective of the constraint relaxation is to minimize the weighted sum of the constraint relaxation amounts.

[0053] The different subscripts here represent the number of slack variables of the constraint corresponding to this output. For example, j=1, the subscripts represent .

[0054] ; ; Solve this problem and get the specific slack variable value, that is, After the value is set, the corresponding constraint is updated to a hard constraint.

[0055] Step 6: Necessity test of economic optimization.

[0056] Based on the above solution results, we first determine the necessity of economic optimization. If the relaxation factor corresponding to any constraint is If both of them have values ​​greater than 0, it means that the output constraint is relaxed, the controller has no degrees of freedom, and there is no need to enter economic optimization. If there is at least one CV constraint, and its two relaxation factors are both 0, it means that the constraint is not relaxed, there may be room for economic optimization, and the next step of economic optimization can be entered. This step can reduce unnecessary constraint relaxation solutions, and can greatly improve the calculation efficiency when the controller scale becomes larger.

[0057] Step 7: Solve the self-optimization problem.

[0058] Next, we construct an economic optimization problem to solve. Different priorities can be set for the economic optimization of different variables to achieve the order of variable optimization, which is more flexible than traditional objective function optimization. To simplify the explanation, this case assumes that the economic optimization priority of all variables is the same.

[0059] Define the non-negative relaxation factor of each economic optimization constraint as The optimization goal J2 is to minimize the weighted sum of the soft constraint violations of the economic optimization, so as to achieve the speed meeting the specified requirements during economic optimization.

[0060] ; Solve the above problem and get the relaxation factor Due to the setting of the optimization problem, if the input variable still has the degree of freedom for economic optimization, the corresponding =0, which means that the variable's economic optimization step must be approaching the optimal direction according to the specified speed adjustment factor. If the input variable has no degrees of freedom, that is, the variable is still in the stage of meeting the control requirements, then the output It must be greater than 0. So it can be The value of a variable is used to determine whether the variable enters the economic optimization stage, so that a clear explanation of the controller behavior can be given.

[0061] Step 8: Solve the steady-state optimization objective function.

[0062] ; ; In this constraint, the relaxation factor symbol has been updated, and * represents the actual solution, which is a specific value.

[0063] Solving the above steady-state optimization objective function, we get , which is the solution to the entire steady-state optimization problem. This method can ensure that the input is economically optimized at a specified step size while satisfying the output constraints and the system still has degrees of freedom.

[0064] It should be noted that in the above formulas, the superscript s and superscript S have the same meaning and are displayed as S in some formulas.

[0065] Example 3 The embodiment of the present invention also provides a control device for a multi-objective speed-adjustable MPC self-optimization process in an industrial process, such as Figure 3 As shown, it includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program stored in the memory, specifically executing the multi-objective speed-adjustable MPC self-optimization method described in any of the above embodiments.

[0066] In addition, an embodiment of the present invention also provides a computer-readable storage medium, which includes: a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the multi-objective speed-adjustable MPC self-optimization method as described in any one of the embodiments are implemented.

[0067] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0068] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0069] In the present invention, unless otherwise clearly specified and limited, when a first feature is “on” or “below” a second feature, it may be that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediate medium. Moreover, when a first feature is “above”, “above” or “above” a second feature, it may be that the first feature is directly above or obliquely above the second feature, or it may simply mean that the first feature is higher in level than the second feature. When a first feature is “below”, “below” or “below” a second feature, it may be that the first feature is directly below or obliquely below the second feature, or it may simply mean that the first feature is lower in level than the second feature.

[0070] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0071] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may alter, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A multi-objective speed-adjustable MPC self-optimization method, characterized in that: include: S100, pre-establishing a first model and a first constraint set of a steady-state optimization problem for a specific optimization process of an industrial process; S200, based on the demand for economic optimization, construct an economic optimization soft constraint with a speed adjustment factor, i.e., the second constraint; S300, constructing a steady-state optimization objective function based on the overall target tracking and economic optimization requirements; generating the entire steady-state optimization problem according to the first model, the first constraint set and the second constraint; S400, introducing the priority information of the output constraint in the first constraint set, and constructing an output constraint relaxation problem, wherein the optimization objective of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts, calculate the relaxation factor, and update the output constraint; Introduce the economic optimization priority of the input variables in the second constraint, construct an economic optimization problem, and the optimization goal in the economic optimization problem is to make the speed during economic optimization meet the specified requirements, and update the constraints of the input variables; Based on the updated constraints, the entire steady-state optimization problem is solved to obtain the solution of the input variables.

2. The method according to claim 1, characterized in that The S200 includes: The objective function J in the entire steady-state optimization problem to be solved is: ； Among them, the parameters represents the cost of the i-th operation variable, parameter Represents the penalty for the i-th operation variable deviating from its value at the previous moment, cost parameter represents the unit penalty cost of the ith manipulated variable deviating from the set point, and the variable represents the measured value of the i-th operating variable at the current time k, and the parameter represents the unit penalty for the jth controlled variable deviating from the set point, the variable represents the open-loop steady-state value of the jth controlled variable at the current time k, and the variable represents the set value of the jth controlled variable; represents the steady-state optimization increment of the ith input, represents the set value of the i-th input variable, represents the steady-state optimization increment of the j-th output; The complete steady-state optimization constraints in the entire steady-state optimization problem to be solved include: ； in, represents the model gain of the i-th input to the j-th output; Respectively represent the lower and upper limits of the steady-state optimization increment of the j-th output; Represents the lower and upper limits of the position of the j-th output; represents the lower and upper bounds of the steady-state optimization increment for the ith input; Represents the lower and upper limits of the position of the i-th input; They represent the economic optimization speed adjustment factors of the i-th input respectively, and the corresponding economic optimization soft constraints are enabled according to the input maximization or minimization requirements.

3. The method according to claim 1 or 2, characterized in that: The constraints in the entire steady-state optimization problem in S300 include three types of constraints: the first type is the input stride constraint and the input upper and lower limit constraints; the second type is the output stride constraint and the output upper and lower limit constraints; the third type is the economic optimization constraint / economic constraint; and / or, In S400, the second type of constraint priority is set according to the control demand and the demand for economic optimization; the third type of constraint priority is lower than the second type of constraint priority; Specifically, construct the output constraint relaxation problem and define the non-negative relaxation factor of each output constraint as , and the corresponding relaxation weights are ,The optimization objective J1 of the constraint relaxation problem is to minimize the weighted sum of constraint relaxation amounts; ； ； Solve for J1 and get the relaxation factor After the value is set, the output constraints are updated and the updated output constraints are used as hard constraints.

4. The method according to claim 3, characterized in that: The method further comprises: S500, judging the necessity of economic optimization for solving the optimization target in the constraint relaxation problem; Specifically, if the relaxation factor corresponding to any constraint is There are values ​​greater than 0, indicating that the output constraint is relaxed, the controller has no degrees of freedom, and does not need to enter economic optimization; If there is at least one output constraint and both relaxation factors are equal to 0, it means that the constraint has not been relaxed. At this time, there is room for economic optimization and the next step of economic optimization is entered.

5. The method according to claim 4, characterized in that The method further comprises: S600, if it is necessary to enter into economic optimization, then enter into economic optimization and construct an economic optimization problem; According to the economic optimization requirements of different variables, different priorities are set to achieve the order of variable optimization; Specifically, let the economic optimization priority of all variables be , and define the non-negative relaxation factor of each economic optimization constraint as , the specific constraint is selected according to the direction of variable optimization. If the variable has no economic optimization requirement, then the constraint does not exist, that is, there is no corresponding ; The purpose of the optimization objective J2 is to make the speed meet the specified requirements during economic optimization; ； ； Solve for J2 and obtain the relaxation factor The value of If the input variable still has the degree of freedom for economic optimization, the corresponding It must be 0. If the input variable has no degrees of freedom, the output Must be greater than 0; The corresponding economic optimization constraints are updated, and the updated economic optimization constraints are used as hard constraints.

6. The method according to claim 1 or 2, characterized in that: The S100 includes: S101, establishing a controlled variable constraint set for steady-state optimization, wherein the controlled variable constraint set includes: a steady-state optimization increment lower limit and upper limit constraint of the controlled variable, and a position limit constraint; S102, establishing a set of operating variable constraints for steady-state optimization, wherein the set of operating variable constraints includes: a lower limit and an upper limit constraint of a steady-state optimization increment of the operating variable, and a position limit constraint; S103: Establish a first model of a steady-state optimization problem, wherein the controlled variable constraint set and the manipulated variable constraint set constitute the first constraint set.

7. The method according to claim 6, characterized in that: For the process control of the heavy oil fractionator, the controlled variable constraint set in the first constraint set includes: ; The operation variable constraint set in the first constraint set includes: ; The first model includes: ; in, represents the model gain of the i-th input to the j-th output; represents the steady-state optimization increment of the j-th output, Represent the lower and upper limits of the steady-state optimization increment of the j-th output, respectively. represents the predicted future steady-state value of the j-th output at time k, Represents the lower and upper limits of the position of the j-th output; represents the steady-state optimization increment of the j-th input, Represent the lower and upper limits of the steady-state optimization increment of the j-th input, respectively. represents the current value of the jth input, Represents the lower and upper limits of the position of the j-th input.

8. The method according to claim 7, characterized in that If the optimization target is the extraction rate of the top product of the distillation tower Increase to the maximum value, the distillation tower side line extraction Increased to the maximum value, and the reflux heat load at the bottom of the distillation tower Reduce to a minimum value; Then, the economic optimization constraints include: , For input Speed ​​adjustment factor; Alternatively, enter To track a specific value, the economic optimization constraints include: ； in, is the measured value of the i-th input at time k, represents the steady-state optimization increment of the ith input, is the set value, Represent the lower and upper limits of the steady-state optimization increment of the ith input, respectively. Speed ​​adjustment factors for upper and lower increment limits for steady state optimization.

9. A control device for a multi-objective speed-adjustable MPC self-optimization process in an industrial process, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program stored in the memory, specifically executing the multi-objective speed-adjustable MPC self-optimization method described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: It comprises: a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the multi-objective speed-adjustable MPC self-optimization method as described in any one of claims 1 to 8 are implemented.

Images