System operation control method, device, equipment and readable storage medium

Abstract

The present application discloses a system operation control method, apparatus, device and readable storage medium, the method comprising: obtaining corresponding output data based on input data in an initial steady-state operating point; solving a control performance optimization problem based on the mean and variance of the input data, the mean and variance of the output data, and obtaining a new steady-state operating point and control performance parameters; controlling the system using an MPC method based on current adjustment parameters to obtain an actual steady-state operating point; updating the current adjustment parameters based on the actual steady-state operating point input data, output data and the control performance optimization problem, and returning to execute the step of solving the control performance optimization problem until the control performance parameters meet preset conditions. The above technical solution disclosed in the present application continuously obtains the direction of control performance improvement from the operating data by adopting an iterative learning control method, and controls the system according to this direction to improve the control performance, thereby improving the economic performance of the system.

Description

Technical Field

[0001] The present application relates to the field of process control technology, and more specifically, to a system operation control method, apparatus, device, and readable storage medium. Background Art

[0002] As a representative technology of advanced control strategies, MPC (Model Predictive Control) methods can easily handle process variable constraints and coordinate the optimization of multivariable control systems, and have been widely used. However, the promotion and implementation of MPC present several pressing challenges. First, a comprehensive and accurate return on investment analysis is required before implementation, which is a key concern for enterprises. Second, MPC controller performance can gradually degrade over time due to changes in factors such as raw material supply, production targets, operating conditions, and the production environment, preventing the expected maximum benefit from being achieved.

[0003] To address the above issues, the process industry needs a systematic and feasible economic performance evaluation solution to detect and diagnose the performance degradation or failure of the MPC controller in real time, and then improve the economic performance of the MPC system to ensure that the operation of the process control system meets the given performance indicators. The current method for evaluating the economic performance of process systems is mainly the LQG (Linear Quadratic Gaussian) performance evaluation benchmark method, which introduces the controller effect into the quadratic performance index and can easily set constraints on the manipulated variables, thereby obtaining a relatively more objective and feasible result. However, when the LQG method is applied to a multi-input multi-output system, it can only perform economic performance evaluation, but cannot provide how to adjust the controller to improve the control performance and thus improve the economic performance of the system.

[0004] In summary, how to improve control performance to enhance the economic performance of the system is a technical problem that currently needs to be solved urgently by those skilled in the art. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a system operation control method, device, equipment and readable storage medium for improving control performance to enhance the economic performance of the system.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] A system operation control method, comprising:

[0008] Obtain corresponding output data according to the input data in the initial steady-state operating point;

[0009] Solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain the new steady-state operating point and control performance parameters;

[0010] Controlling the system using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point;

[0011] The current adjustment parameters are updated according to the input data, output data and the control performance optimization problem of the actual steady-state operating point, and the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data is returned to be executed until the control performance parameters meet the preset conditions.

[0012] Preferably, controlling the system using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point includes:

[0013] Within the running steps, the distributed system is controlled by adopting a distributed MPC control algorithm based on Nash optimality according to the current adjustment parameters corresponding to each subsystem in the distributed system, so as to track the actual steady-state operating point corresponding to the new steady-state operating point.

[0014] Preferably, a distributed MPC control algorithm based on Nash optimality is used to control the distributed system, including:

[0015] Solving an objective function related to the optimization variables of each subsystem in the distributed system at each moment, and obtaining a solution for the optimization variables of each subsystem;

[0016] Determining whether each subsystem has reached an iterative convergence standard according to the solution of the optimization variable of each subsystem or the number of iterations of the solution;

[0017] If so, the solution of the optimization variables of each of the subsystems is passed to other subsystems as known information, and the step of solving an objective function related to the optimization variables of the subsystem itself for each subsystem in the distributed system at each moment is returned to be executed until the solution of each subsystem under the distributed MPC control algorithm at the current moment is obtained, and the solution of each subsystem under the distributed MPC control algorithm at the current moment is implemented in the distributed system.

[0018] Preferably, the control performance optimization problem is specifically an economic optimization problem:

[0019]

[0020] st

[0021]

[0022]

[0023]

[0024]

[0025] Among them, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data, is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ij Represents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the jth input data.

[0026] Preferably, updating the current adjustment parameters according to the input data and output data of the actual steady-state operating point and the control performance optimization problem includes:

[0027] According to the mean and variance of the input data of the actual steady-state operating point, the mean and variance of the output data and the economic optimization problem, the current regulation parameters corresponding to each subsystem in the distributed system are updated using an ILC-based weight coefficient adjustment strategy.

[0028] Preferably, according to the mean and variance of the input data of the actual steady-state operating point, the mean and variance of the output data and the economic optimization problem, the current adjustment parameters corresponding to each subsystem in the distributed system are updated using an ILC-based weight coefficient adjustment strategy, including:

[0029] Convert the economic optimization problem into a standard form of economic optimization problem:

[0030]

[0031] st

[0032] A[Δu s ,Δy s ] T ≤B

[0033] Among them, c is the matrix coefficient after the economic optimization problem is transformed into the standard form, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions in the standard form of the economic optimization problem;

[0034] Find the input constraints B corresponding to all inputs in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all outputs y , and record the minimum value MIN u and MIN y :

[0035]

[0036]

[0037] in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * The optimal solution obtained based on the economic optimization problem;

[0038] If MIN u Less than MIN y , then write φ in the sequence used to record the direction of change of economic performance parameters t =1, if MIN u Not less than MINy , then write φ in the sequence t =-1;

[0039] If the sequence always maintains the same value, then according to λ t+1 =λ t *(1+l(φ t )) Update the adjustment parameters corresponding to the mth subsystem; where λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, and l2 is the second learning rate;

[0040] If the sequence does not always maintain the same value, the current adjustment parameter corresponding to the m-th subsystem is updated using a bisection method.

[0041] Preferably, the current adjustment parameter corresponding to the m-th subsystem is updated using a dichotomy method, including:

[0042] If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t, φ t = -1, then use λ t+1 =(λ t +λ t-1 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-1 is the adjustment parameter corresponding to the mth subsystem at the t-1th iteration;

[0043] If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t-2, φ t-2 =-1,φ t-1 =1,φ t =1, then use λ t+1 =(λ t +λ t-2 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-2 is the adjustment parameter corresponding to the mth subsystem at the t-2th iteration.

[0044] A system operation control device, comprising:

[0045] An obtaining module, used for obtaining corresponding output data according to input data in the initial steady-state operating point;

[0046] A solution module is used to solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters;

[0047] A control module is used to control the system using an MPC method according to current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point;

[0048] An updating module is used to update the current adjustment parameters according to the input data and output data of the actual steady-state operating point and the control performance optimization problem, and return to execute the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet the preset conditions.

[0049] A system operation control device, comprising:

[0050] memory for storing computer programs;

[0051] A processor is used to implement the steps of the system operation control method as described in any one of the above when executing the computer program.

[0052] A readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the above-mentioned system operation control methods.

[0053] The present application provides a system operation control method, apparatus, device and readable storage medium, wherein the method includes: obtaining corresponding output data based on input data in an initial steady-state operating point; solving a control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters; controlling the system using an MPC method based on current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point; updating the current adjustment parameters based on the input data, output data and control performance optimization problem of the actual steady-state operating point, and returning to execute the step of solving the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet preset conditions.

[0054] The above-mentioned technical solution disclosed in the present application utilizes the mean and variance of the input data in the steady-state operating point and the mean and variance of the output data to solve the control performance optimization problem to obtain a new steady-state operating point and new control performance parameters, and uses the MPC method to control the system according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point. Then, the current adjustment parameters are updated according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and the above process is repeated until the control performance meets the preset conditions. That is, the present application continuously obtains the direction of control performance improvement from the operating data by adopting an iterative learning method, and controls the system according to the direction of control performance improvement, thereby effectively improving the control performance and then effectively improving the economic performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0056] Figure 1 A flowchart of a system operation control method provided in an embodiment of the present application;

[0057] Figure 2 Schematic diagram of the carbonization decomposition process;

[0058] Figure 3 A flowchart of a fast dynamic matrix control method based on an ORC waste heat recovery system provided in an embodiment of the present application;

[0059] Figure 4 This is a rendering of the economic performance improvement provided by the embodiment of the present application;

[0060] Figure 5 A schematic diagram of the structure of a system operation control device provided in an embodiment of the present application;

[0061] Figure 6 A schematic diagram of the structure of a system operation control device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] The core of this application is to provide a system operation control method, device, equipment and readable storage medium for improving control performance to enhance the economic performance of the system.

[0063] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0064] See also Figure 1 , which shows a flow chart of a system operation control method provided by an embodiment of the present application. The system operation control method provided by an embodiment of the present application may include:

[0065] S11: Obtain corresponding output data according to the input data in the initial steady-state operating point.

[0066] In the present application, first the initial parameters of the system and the learning rate of iterative learning control can be selected, so that system operation control is carried out based on these parameters.Wherein, the initial parameters of the selected system can include the current adjustment parameters of the initial steady-state operating point and the system (when initially selecting, the current adjustment parameters of the system are also the initial adjustment parameters, for the convenience of unifying with the follow-up, referred to as the current adjustment parameters of the system herein). Steady-state operating point is the parameter that a system can stably operate, after specifying the input and output of the system, it is possible to obtain that the system reaches stability when the input and output are equal to what, now, the value of the input and output is exactly the steady-state operating point, and adjustment parameters can greatly affect the last data of input and output, and then affect the variance and mean of input and output, ultimately affecting control performance, therefore, obtaining adjustment parameters and regulating adjustment parameters is just unusually important. In addition, it should be noted that the system mentioned in the present application can specifically be a large-scale MPC system.

[0067] After the initial steady-state operating point is selected, the input data of the selected initial steady-state operating point may be put into the system for operation to obtain output data corresponding to the input data of the initial steady-state operating point.

[0068] In addition, the mean and variance of the input data can be calculated based on the specific values ​​of each input at multiple moments in the selected initial steady-state operating point, and the mean and variance of the output data can be calculated based on the specific values ​​of the output data corresponding to the input data in the initial steady-state operating point at multiple moments. The initial control performance parameters can be calculated using the mean of the input data and the mean of the output data, so that the initial control performance parameters and the control performance parameters calculated in each subsequent iteration can be used as the basis for whether to stop the iteration, and it is convenient to obtain relevant information on control performance improvement based on the initial control performance parameters and the control performance parameters calculated in each subsequent iteration.

[0069] S12: Solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters.

[0070] Based on step S11, the constructed control performance optimization problem can be solved based on the mean and variance of the input data in the steady-state operating point and the mean and variance of the corresponding output data to obtain a new steady-state operating point and new control performance parameters, thereby facilitating the use of the new steady-state operating point as a guide for system operation control and facilitating the determination of whether the improvement in system control performance meets preset conditions based on the obtained control performance parameters. Since the new steady-state operating point is obtained based on the control performance optimization problem, subsequent control of the system can be carried out in the direction of control performance optimization, thereby effectively improving the control performance of the system and thus improving the economic performance of the system.

[0071] S13: Control the system using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point.

[0072] Once the new steady-state operating point and the system's current adjustment parameters are obtained, the MPC method can be used to control the system's operation based on the current adjustment parameters. Considering the influence of factors such as noise, when the MPC method is used to control the system's operation based on the current adjustment parameters, the calculated new steady-state operating point is not reached. Instead, the actual steady-state operating point corresponding to the new steady-state operating point is obtained, allowing for iterative calculations based on the input data and output of the actual steady-state operating point.

[0073] It should be noted that before using the MPC method to control the system, it is necessary to first identify the MPC model. Different identification methods require different test data, and the methods for obtaining test data for identification include but are not limited to open-loop step tests, open-loop pseudo-random binary sequence tests, closed-loop excitation tests, etc.

[0074] S14: Update the current adjustment parameters according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and return to the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet the preset conditions.

[0075] After obtaining the actual steady-state operating point, the current adjustment parameters of the system can be updated according to the input data, output data and control performance optimization problem of the actual steady-state operating point, so that the updated current adjustment parameters are used as the current adjustment parameters. Among them, when updating the current adjustment parameters, the current adjustment parameters can be updated specifically according to the learning rate of the iterative learning control initially selected, so that the current adjustment parameters can be updated according to a certain learning rate at each iteration, and it is convenient to subsequently control the system according to the updated current adjustment parameters. In addition, the input data and output data of the actual steady-state operating point mentioned above can specifically be a relatively stable input data and output data after the actual steady-state operating point is obtained, so as to improve the accuracy of the update of the current adjustment parameters.

[0076] After the current adjustment parameters are updated and the updated current adjustment parameters are used as the current adjustment parameters, step S12 can be returned to execute, and steps S12 to S14 can be repeated until the obtained control performance parameters meet the preset conditions, wherein the preset conditions can specifically be that the control performance parameters reach a first threshold or that the control performance parameters remain stable for a period of time (specifically, the control performance parameters do not change, or the range of change of the control performance parameters during the period is less than a second threshold). It should be noted that when returning to step S12 and executing the step of solving the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data, the input data and output data mentioned here specifically refer to the input data and output data corresponding to the actual steady-state operating point obtained in the previous iteration, that is, in each iteration, the control performance optimization problem is solved according to the input data and output data of the actual steady-state operating point obtained in the previous iteration to obtain a new steady-state operating point and performance control parameters, so as to realize iterative control based on past data.

[0077] From the above process, it can be seen that the present application adopts an iterative learning control method to continuously obtain the direction of improving control performance in the running data until the control performance parameters meet the preset conditions, thereby effectively improving the control performance of the system, so that the economic performance of the system is effectively improved, avoiding the possible performance aging problem of the system over time and reducing the online burden of large-scale MPC systems, providing a performance improvement solution for the industrial control of the currently widely used large-scale MPC systems.

[0078] The above-mentioned technical solution disclosed in the present application utilizes the mean and variance of the input data in the steady-state operating point and the mean and variance of the output data to solve the control performance optimization problem to obtain a new steady-state operating point and new control performance parameters, and uses the MPC method to control the system according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point. Then, the current adjustment parameters are updated according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and the above process is repeated until the control performance meets the preset conditions. That is, the present application continuously obtains the direction of control performance improvement from the operating data by adopting an iterative learning method, and controls the system according to the direction of control performance improvement, thereby effectively improving the control performance and then effectively improving the economic performance of the system.

[0079] An embodiment of the present application provides a system operation control method, which uses an MPC method to control the system according to current adjustment parameters to obtain an actual steady-state operating point corresponding to a new steady-state operating point, and may include:

[0080] Within the running steps, the distributed system is controlled by a distributed MPC control algorithm based on Nash optimality according to the current adjustment parameters corresponding to each subsystem in the distributed system, so as to track the actual steady-state operating point corresponding to the new steady-state operating point.

[0081] In the present application, the system targeted may be a distributed system comprising multiple subsystems. Accordingly, the current adjustment parameter is specifically the current adjustment parameter corresponding to each subsystem in the distributed system. In this case, the initially selected current adjustment parameter is specifically the adjustment parameter in the objective function of each subsystem, wherein the objective function is as follows:

[0082]

[0083] Among them, m is the number of the subsystem, y represents the system output, u represents the system input, λ m is the adjustment parameter of the current system m, where the number of adjustment parameters is the same as the number of subsystems. If the number of subsystems is M (where the number of subsystems is also included in the initial parameters of the selected system), then the number of adjustment parameters is also M. is the steady-state mean value of the system output, is the steady-state mean of the system input, I is the weight coefficient, which here represents the unit matrix of all 1s, and J m Represents the objective function value corresponding to the mth subsystem.

[0084] In addition, the initial parameters of the selected system can also include the number of running steps, where the selection of running steps should be frequent enough to ensure that the distributed system can track the steady-state target value. In order to ensure the sequential execution of the distributed MPC algorithm, the initial parameters of the selected system, in addition to the parameters mentioned above, also include input and output constraints, control time domain, prediction time domain, the maximum number of iterations of the distributed algorithm, the running cycle, etc. Among them, the maximum number of iterations of the distributed algorithm should be determined according to the actual situation. Too large a maximum number of iterations may result in too long calculation time, and too small a maximum number of iterations may cause the algorithm to fail to converge. Therefore, the setting of the maximum number of iterations must first be subject to experimental links.

[0085] On the basis of the above, when the system is controlled by the MPC method according to the current adjustment parameters to obtain the actual steady-state operating point corresponding to the new steady-state operating point, the distributed system can be controlled by the distributed MPC control algorithm based on Nash optimality according to the current adjustment parameters corresponding to each subsystem in the distributed system within the number of running steps, so that the distributed system can track the actual steady-state operating point corresponding to the new steady-state operating point.

[0086] Through the above process, the operation control of the distributed system can be achieved using the distributed MPC algorithm.

[0087] An embodiment of the present application provides a system operation control method that uses a distributed MPC control algorithm based on Nash optimality to control a distributed system, which may include:

[0088] At each moment, an objective function related to the subsystem's own optimization variables is solved for each subsystem in the distributed system, and the solution of the optimization variables of each subsystem is obtained;

[0089] According to the solution of the optimization variable of each subsystem or the number of iterations of the solution, it is judged whether each subsystem has reached the iterative convergence standard;

[0090] If so, the solution of the optimization variables of each subsystem is passed to other subsystems as known information, and the step of solving an objective function related to the subsystem's own optimization variables for each subsystem in the distributed system at each moment is returned to execute until the solution of each subsystem under the distributed MPC control algorithm at the current moment is obtained, and the solution of each subsystem under the distributed MPC control algorithm at the current moment is implemented in the distributed system.

[0091] In this application, the specific process of controlling the distributed system using the distributed MPC control algorithm based on Nash optimality is as follows:

[0092] (1) At each moment, each subsystem solves an objective function related only to its own optimization variables and obtains the solution of each subsystem’s optimization variables. Wherein, l is the number of iterations of the distributed algorithm, i represents the subsystem number, and the objective function mentioned here can be specifically referred to the DMPC (distributed model predictive control) algorithm based on Nash optimality in the distributed MPC control algorithm, and the self-optimization variables mentioned here can generally be considered as the input or output of the current subsystem;

[0093] (2) Based on the solution of the optimization variable of each subsystem or the number of iterations of the solution, determine whether each subsystem has reached the iterative convergence standard, where the iterative convergence standard is: 1) Among them, ∈ i is the convergence value of the i-th subsystem; 2) the maximum number of iterations of the distributed algorithm is l = l max .

[0094] (3) If it is determined that each subsystem has reached the iterative convergence standard, the solution of the optimization variable of each subsystem is transferred to other subsystems as known information to facilitate the solution of other subsystems (the distributed MPC control algorithm based on Nash optimality is that in each iteration process, each subsystem regards the optimization variables of other subsystems as known quantities and solves them);

[0095] (4) Repeat steps (1)-(3) until the solution of each subsystem under the distributed MPC control algorithm at the current moment is obtained. because It contains future information of the control time domain length, and its first element (i.e., the solution of each subsystem under the distributed MPC control algorithm at the current moment) needs to be implemented into the distributed system to achieve control of the distributed system.

[0096] By adopting a distributed MPC algorithm to control it, the online calculation speed can be accelerated, thereby avoiding damage to the system caused by the inability to obtain a solution within one sampling period.

[0097] The embodiment of the present application provides a system operation control method, wherein the control performance optimization problem is specifically an economic optimization problem:

[0098]

[0099] st

[0100]

[0101]

[0102]

[0103]

[0104] Among them, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data, is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ij Represents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the jth input data.

[0105] In the present application, the control performance optimization problem may specifically be an economic optimization problem. Accordingly, the control performance parameters mentioned in the present application may specifically be economic performance parameters, so as to achieve the improvement of control performance and economic performance by using economic performance parameters as control performance parameters.

[0106] Specifically, the economic optimization problem mentioned above can be:

[0107]

[0108] st

[0109]

[0110]

[0111]

[0112]

[0113] In the above economic optimization problem, the superscript i represents the i-th input data or the i-th output data, the superscript j represents the j-th input data or the j-th output data, y represents the output data, u represents the input data, mean represents the mean, s represents the steady state, min represents the minimum value, and max represents the maximum value. Specifically, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data (its unit can be specifically yuan / unit), is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ij Represents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the jth input data. st represents the constraints. The first constraint is the constraint on the output, the second constraint is the constraint on the input, the third constraint is the constraint on the difference Δ of the input data, and the fourth constraint is the constraint on the difference Δ of the output data.

[0114] Among them, when solving the economic optimization problem, it is to find and Under what circumstances can J be maximized, so as to obtain the economic performance parameters, and can be used according to and The corresponding input data and output data under the new steady-state operating point are obtained.

[0115] It should be noted that in the above case, you can use The initial control performance parameters are calculated.

[0116] An embodiment of the present application provides a system operation control method, which updates current adjustment parameters based on input data, output data, and control performance optimization issues at an actual steady-state operating point, and may include:

[0117] According to the mean and variance of the input data, the mean and variance of the output data and the economic optimization problem at the actual steady-state operating point, the current regulation parameters corresponding to each subsystem in the distributed system are updated using the ILC-based weight coefficient adjustment strategy.

[0118] In the present application, when updating the current adjustment parameters based on the input data, output data and control performance optimization problem of the actual steady-state operating point, the mean and variance of each input data can be calculated based on the specific values ​​of each input at multiple moments in the actual steady-state operating point, and the mean and variance of each output data can be calculated based on the specific values ​​of each output at multiple moments in the actual steady-state operating point. When calculating the mean and variance of each input data and the mean and variance of each output data in the actual steady-state operating point, a relatively stable data segment after the distributed MPC control algorithm tracks the actual steady-state operating point can be selected to improve the accuracy of the mean and variance calculation. Then, based on the calculated mean and variance of the input data, the mean and variance of the output data and the economic optimization problem of the actual steady-state operating point, the current adjustment parameters corresponding to each subsystem in the distributed system can be updated using a weight coefficient adjustment strategy based on ILC (Iterative Learning Control), so that the updated adjustment parameters can be used to control the corresponding subsystem, thereby improving the economic performance of the distributed system.

[0119] An embodiment of the present application provides a system operation control method that updates the current adjustment parameters corresponding to each subsystem in a distributed system using an ILC-based weight coefficient adjustment strategy based on the mean and variance of input data and output data at an actual steady-state operating point and an economic optimization problem. The method may include:

[0120] Convert the economic optimization problem into a standard form economic optimization problem:

[0121]

[0122] st

[0123] A[Δu s ,Δy s ] T ≤B

[0124] Among them, c is the matrix coefficient after being converted into the standard form of economic optimization problem, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions in the standard form of the economic optimization problem;

[0125] Find the input constraints B corresponding to all inputs in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all outputs y , and record the minimum value MIN u and MIN y :

[0126]

[0127]

[0128] in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * The optimal solution obtained based on the economic optimization problem;

[0129] If MIN u Less than MIN y , then write φ in the sequence used to record the direction of change of economic performance parameters t =1, if MIN u Not less than MIN y , then write φ in the sequence t =-1;

[0130] If the sequence always maintains the same value, then according to λ t+1 =λ t *(1+l(φ t )) Update the adjustment parameters corresponding to the mth subsystem; where λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, and l2 is the second learning rate;

[0131] If the sequence does not always maintain the same value, the current adjustment parameter corresponding to the mth subsystem is updated using the bisection method.

[0132] In this application, based on the mean and variance of the input data at the actual steady-state operating point, the mean and variance of the output data, and the economic optimization problem, the specific process of updating the current adjustment parameters corresponding to each subsystem in the distributed system using the ILC-based weight coefficient adjustment strategy is as follows:

[0133] (1) Convert the economic optimization problem into a standard form of economic optimization problem:

[0134]

[0135] st

[0136] A[Δu s ,Δy s ] T ≤B

[0137] Among them, c is the matrix coefficient after the economic optimization problem is transformed into the standard form, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions of the standard form of the economic optimization problem (A can be called the first constraint coefficient matrix, and B can be called the second constraint coefficient matrix). A 1 、A 2 、A 3 is the first constraint coefficient submatrix contained in the first constraint coefficient matrix, and A 1 、A 2 、A 3 It is obtained by expressing the first constraint coefficient matrix as a matrix with 9 rows and 1 column, B 1 To B 9 is the second constraint coefficient submatrix contained in the second constraint coefficient matrix, and B 1 To B 9 It is obtained by expressing the second constraint coefficient matrix as a matrix with 9 rows and 1 column B 1 =[0 Ny×(Nu+Ny) ],

[0138]

[0139]

[0140] (2) Find the input constraint B corresponding to all input data in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all output data y , and record the minimum value MIN u (minimum value corresponding to the input constraint) and MIN y (Output the minimum value corresponding to the constraint):

[0141]

[0142]

[0143] in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * According to the optimal solution obtained from the economic optimization problem, m=1,2,…,M;

[0144] (3) Compare MIN u and MIN y ; If MIN u Less than MIN y , it means that the input constraint is restricted, thus affecting the economic performance. At this time, a φ is written into the sequence Φ used to record the direction of change of economic performance parameters. t =1, where t represents the number of iterations of iterative learning control; if MIN u Not less than MIN y , then write φ in the sequence t =-1;

[0145] (4) Observe the sequence Φ used to record the direction of change of economic performance parameters:

[0146] (4-1) If the sequence Φ maintains the same value from the beginning, it means that the economic performance is continuously improving. Then the adjustment parameter λ corresponding to the m-th subsystem is updated according to a certain learning rate. Specifically, we have:

[0147] λ t+1 =λ t *(1+l(φ t ))

[0148] in:

[0149]

[0150] λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, l2 is the second learning rate, l1 and l2 are used to adjust λ;

[0151] (4-2) On the contrary, it means that the direction of performance update has changed. It is necessary to use the bisection method to update the current adjustment parameters corresponding to the m-th subsystem so that the adjustment parameters λ can converge and find the better adjustment parameters λ, thereby improving the economic performance of the system.

[0152] An embodiment of the present application provides a system operation control method that uses a dichotomy method to update the current adjustment parameters corresponding to the mth subsystem, which may include:

[0153] If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t, φ t = -1, then use λ t+1 =(λ t +λ t-1 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-1 is the adjustment parameter corresponding to the mth subsystem at the t-1th iteration;

[0154] If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t-2, φ t-2 =-1,φ t-1 =1,φ t =1, then use λ t+1 =(λ t +λ t-2 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-2 is the adjustment parameter corresponding to the mth subsystem at the t-2th iteration.

[0155] In this application, the specific process of updating the current adjustment parameters corresponding to the mth subsystem using the dichotomy method is as follows:

[0156] (4-2-1) If in the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φt ], assuming φ p =1, p<t, φ t =-1, then use λ t+1 =(λ t +λ t-1 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; t-1 is the adjustment parameter corresponding to the mth subsystem at the t-1th iteration;

[0157] (4-2-2) If in the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ], assuming φ p =1, p<t-2, φ t-2 =-1,φ t-1 =1,φ t =1, then use λ t+1 =(λ t +λ t-2 ) / 2 updates the adjustment parameters corresponding to the mth subsystem, where λ t-2 is the adjustment parameter corresponding to the mth subsystem at the t-2th iteration.

[0158] The above process can facilitate finding a better adjustment parameter λ, thereby facilitating the improvement of the economic performance of the system.

[0159] Through the above technical solutions of the present application, it can be seen that: (1) the present application avoids various problems existing in the current LQG performance evaluation benchmark when applied to large-scale MPC systems; (2) the present application uses an iterative learning control method and a distributed model predictive control method based on Nash optimality to provide a solution for improving the economic performance of large-scale MPC systems; (3) the method of the present application is universal and can be widely applied to performance improvement scenarios of various large-scale MPC systems.

[0160] The following describes the technical solution of this application in detail using an actual industrial carbonization decomposition process. Figure 2 , which shows a schematic diagram of the carbonization decomposition process, and the chemical reaction formula is as follows:

[0161] 2NaOH+CO2=Na2CO3+H2O

[0162] 2NaAlO2+4H2O=2Al(OH)3↓+2NaOH

[0163] like Figure 3 , a flowchart of a fast dynamic matrix control method based on an ORC waste heat recovery system provided in an embodiment of the present application:

[0164] Step 1, select some parameters as follows:

[0165] The learning rate l1 = 0.75, l2 = 3. The initial adjustment parameters of each subsystem are 0.01. The number of subsystems is 5, the number of operation steps is 600, the operation cycle is selected as 10, and the maximum number of iteration steps is 20. The adjustment parameters in the objective function of each subsystem are as follows:

[0166]

[0167] Where i is the subsystem number, y represents the system output, and u is the system input. i is the adjustment parameter of the current subsystem i. The number of adjustment parameters is the same as the number of subsystems. If the number of subsystems is M, the number of adjustment parameters is also M.

[0168] Step 2: The economic performance function is selected as follows:

[0169] J=2y5-(u1+u2+u3+u4+u5)

[0170] Step 3: The specific economic optimization problem for updating the steady-state operating point is as follows:

[0171]

[0172] st

[0173]

[0174]

[0175]

[0176]

[0177] in, is the mean of the i-th output data, is the variance of the i-th output data. For the input, the subscripts min and represent the minimum and maximum values ​​respectively, and the subscript s represents the steady state. G ij Represents the output data y i To input data Δu j The steady-state gain, represents the probability of obtaining a normal distribution with a confidence level of 1-α,

[0178] Step 4: Use the distributed MPC control algorithm based on Nash optimality to control the subsystem so that it tracks the steady-state operating point. The specific process is described above and will not be repeated here.

[0179] Step 5: The specific implementation process of the ILC-based weight coefficient adjustment strategy is described in detail in the corresponding section above and will not be repeated here.

[0180] Finally, the proposed ILC-based distributed MPC system performance improvement method has the following effect on economic performance improvement: Figure 4 As shown, it shows the effect diagram of the economic performance improvement provided by the embodiment of the present application, wherein the horizontal axis is the operating cycle, the vertical axis is the economic performance, and its unit is yuan. It can be seen that this method can gradually adjust the economy towards the maximum economic performance until it converges to the maximum economic performance.

[0181] The present application also provides a system operation control device, see Figure 5 , which shows a schematic structural diagram of a system operation control device provided in an embodiment of the present application, which may include:

[0182] An obtaining module 51 is used to obtain corresponding output data according to the input data in the initial steady-state operating point;

[0183] A solving module 52 is used to solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters;

[0184] A control module 53 is configured to control the system using an MPC method according to current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point;

[0185] The updating module 54 is used to update the current adjustment parameters according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and return to the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet the preset conditions.

[0186] The embodiment of the present application provides a system operation control device, wherein the control module 53 may include:

[0187] The control unit is used to control the distributed system by adopting a distributed MPC control algorithm based on Nash optimality according to the current adjustment parameters corresponding to each subsystem in the distributed system within the running steps, so as to track the actual steady-state operating point corresponding to the new steady-state operating point.

[0188] An embodiment of the present application provides a system operation control device, wherein the control unit may include:

[0189] The solving subunit is used to solve an objective function related to the optimization variables of the subsystem itself for each subsystem in the distributed system at each moment, and obtain the solution of the optimization variables of each subsystem;

[0190] A judgment subunit, used to judge whether each subsystem has reached the iterative convergence standard according to the solution of the optimization variable of each subsystem or the number of iterations of the solution;

[0191] The return subunit is used to determine whether each subsystem has reached the iterative convergence standard based on the solution of the optimization variable of each subsystem or the number of iterations of the solution, then pass the solution of the optimization variable of each subsystem to other subsystems as known information, and return to execute the step of solving an objective function related to the subsystem's own optimization variable for each subsystem in the distributed system at each moment until the solution of each subsystem under the distributed MPC control algorithm at the current moment is obtained, and implement the solution of each subsystem under the distributed MPC control algorithm at the current moment into the distributed system.

[0192] The embodiment of the present application provides a system operation control device, wherein the control performance optimization problem is specifically an economic optimization problem:

[0193]

[0194] st

[0195]

[0196]

[0197]

[0198]

[0199] Among them, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data, is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ijRepresents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the jth input data.

[0200] In an embodiment of the present application, a system operation control device is provided, wherein the update module 54 may include:

[0201] The updating unit is used to update the current adjustment parameters corresponding to each subsystem in the distributed system using the ILC-based weight coefficient adjustment strategy according to the mean and variance of the input data of the actual steady-state operating point, the mean and variance of the output data and the economic optimization problem.

[0202] In an embodiment of the present application, a system operation control device is provided, wherein an update unit may include:

[0203] The conversion unit is used to convert the economic optimization problem into a standard form of economic optimization problem:

[0204]

[0205] st

[0206] A[Δu s ,Δy s ] T ≤B

[0207] Among them, c is the matrix coefficient after being converted into the standard form of economic optimization problem, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions in the standard form of the economic optimization problem;

[0208] The recording unit is used to find the input constraints B corresponding to all inputs in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all outputs y , and record the minimum value MIN u and MIN y :

[0209]

[0210]

[0211] in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * The optimal solution obtained based on the economic optimization problem;

[0212] Write unit, used if MIN u Less than MIN y , then write φ in the sequence used to record the direction of change of economic performance parameters t =1, if MIN u Not less than MIN y , then write φ in the sequence t =-1;

[0213] The first updating unit is used to update the sequence according to λ if the sequence always maintains the same value. t+1 =λ t *(1+l(φ t )) Update the adjustment parameters corresponding to the mth subsystem; where λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, and l2 is the second learning rate;

[0214] The second updating unit is configured to update the current adjustment parameter corresponding to the mth subsystem by using a bisection method if the sequence does not always maintain the same value.

[0215] In an embodiment of the present application, a system operation control device is provided, wherein the second updating unit may include:

[0216] The first updating subunit is used for the case where the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1p<t,φ t = -1, then use λ t+1 =(λ t +λ t-1 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-1 is the adjustment parameter corresponding to the mth subsystem at the t-1th iteration;

[0217] The second updating subunit is used for the case where the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1p<t-2,φ t-2 =-1,φ t-1 =1,φ t =1, then use λ t+1 =(λ t +λ t-2 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-2 is the adjustment parameter corresponding to the mth subsystem at the t-2th iteration.

[0218] The present application also provides a system operation control device, see Figure 6 , which shows a schematic structural diagram of a system operation control device provided in an embodiment of the present application, which may include:

[0219] Memory 61, for storing computer programs;

[0220] The processor 62, when used to execute the computer program stored in the memory 61, can implement the following steps:

[0221] The corresponding output data is obtained according to the input data in the initial steady-state operating point; the control performance optimization problem is solved according to the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters; the system is controlled using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point; the current adjustment parameters are updated according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data is returned to execute until the control performance parameters meet the preset conditions.

[0222] The present application also provides a readable storage medium, which stores a computer program. When the computer program is executed by a processor, the following steps can be implemented:

[0223] The corresponding output data is obtained according to the input data in the initial steady-state operating point; the control performance optimization problem is solved according to the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters; the system is controlled using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point; the current adjustment parameters are updated according to the input data, output data and control performance optimization problem of the actual steady-state operating point, and the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data is returned to execute until the control performance parameters meet the preset conditions.

[0224] The readable storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0225] For the description of the relevant parts of a system operation control device, equipment and readable storage medium provided in this application, please refer to the detailed description of the corresponding parts in a system operation control method provided in an embodiment of this application, and will not be repeated here.

[0226] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device comprising a series of elements are inherent to the elements. In the absence of further restrictions, the elements limited by the statement "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device comprising the elements. In addition, the above-mentioned technical solutions provided in the embodiments of the present application are not described in detail in accordance with the corresponding technical solutions in the prior art to achieve the same principle, so as to avoid excessive elaboration.

[0227] The above description of the disclosed embodiments will enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A system operation control method, characterized in that: include: Obtain corresponding output data according to the input data in the initial steady-state operating point; Solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain the new steady-state operating point and control performance parameters; Controlling the system using the MPC method according to the current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point; updating the current adjustment parameters according to the input data and output data of the actual steady-state operating point and the control performance optimization problem, and returning to the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet the preset conditions; Controlling the system using an MPC method according to current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point, including: controlling the distributed system using a distributed MPC control algorithm based on Nash optimality according to current adjustment parameters corresponding to each subsystem in the distributed system within the number of running steps to track the actual steady-state operating point corresponding to the new steady-state operating point; The control performance optimization problem is specifically an economic optimization problem: st Among them, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data, is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ij Represents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the j-th input data; Updating the current adjustment parameters according to the input data and output data of the actual steady-state operating point and the control performance optimization problem, including: updating the current adjustment parameters corresponding to each subsystem in the distributed system using an ILC-based weight coefficient adjustment strategy according to the mean and variance of the input data, the mean and variance of the output data of the actual steady-state operating point and the economic optimization problem; According to the mean and variance of the input data of the actual steady-state operating point, the mean and variance of the output data, and the economic optimization problem, a weight coefficient adjustment strategy based on ILC is used to update the current adjustment parameters corresponding to each subsystem in the distributed system, including: Convert the economic optimization problem into a standard form of economic optimization problem: st A[Δu s ,Δy s ] T ≤B Among them, c is the matrix coefficient after being converted into the standard form of economic optimization problem, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions in the standard form of the economic optimization problem; Find the input constraints B corresponding to all inputs in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all outputs y , and record the minimum value MIN u and MIN y : in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * The optimal solution obtained based on the economic optimization problem; If MIN u Less than MIN y , then write φ in the sequence used to record the direction of change of economic performance parameters t =1, if MIN u Not less than MIN y , then write φ in the sequence t =-1; If the sequence always maintains the same value, then according to λ t+1 =λ t *(1+l(φ t )) Update the adjustment parameters corresponding to the mth subsystem; where λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, and l2 is the second learning rate; If the sequence does not always maintain the same value, the current adjustment parameter corresponding to the m-th subsystem is updated using a bisection method.

2. The system operation control method according to claim 1, characterized in that: The distributed system is controlled using a distributed MPC control algorithm based on Nash optimality, including: Solving an objective function related to the optimization variables of each subsystem in the distributed system at each moment, and obtaining a solution for the optimization variables of each subsystem; Determining whether each subsystem has reached an iterative convergence standard according to the solution of the optimization variable of each subsystem or the number of iterations of the solution; If so, the solution of the optimization variables of each of the subsystems is passed to other subsystems as known information, and the step of solving an objective function related to the optimization variables of the subsystem itself for each subsystem in the distributed system at each moment is returned to be executed until the solution of each subsystem under the distributed MPC control algorithm at the current moment is obtained, and the solution of each subsystem under the distributed MPC control algorithm at the current moment is implemented in the distributed system.

3. The system operation control method according to claim 1, characterized in that: The current adjustment parameters corresponding to the mth subsystem are updated using the bisection method, including: If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t, φ t = -1, then use λ t+1 =(λ t +λ t-1 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-1 is the adjustment parameter corresponding to the mth subsystem at the t-1th iteration; If the sequence Φ=[φ1,φ2,φ3,...φ t-2 ,φ t-1 ,φ t ] in φ p =1, p<t-2, φ t-2 =-1,φ t-1 =1,φ t =1, then use λ t+1 =(λ t +λ t-2 ) / 2 updates the adjustment parameters corresponding to the mth subsystem; where λ t-2 is the adjustment parameter corresponding to the mth subsystem at the t-2th iteration.

4. A system operation control device, characterized in that: include: An obtaining module, used for obtaining corresponding output data according to input data in the initial steady-state operating point; A solution module is used to solve the control performance optimization problem based on the mean and variance of the input data and the mean and variance of the output data to obtain a new steady-state operating point and control performance parameters; A control module is used to control the system using an MPC method according to current adjustment parameters to obtain an actual steady-state operating point corresponding to the new steady-state operating point; an updating module, configured to update the current adjustment parameters according to the input data and output data of the actual steady-state operating point and the control performance optimization problem, and return to executing the step of solving the control performance optimization problem according to the mean and variance of the input data and the mean and variance of the output data until the control performance parameters meet the preset conditions; The control module is specifically configured to control the distributed system using a distributed MPC control algorithm based on Nash optimality according to current adjustment parameters corresponding to each subsystem in the distributed system within the operating steps, so as to track an actual steady-state operating point corresponding to the new steady-state operating point; The control performance optimization problem is specifically an economic optimization problem: st Among them, J is the economic performance parameter, is the economic coefficient corresponding to the i-th output data, is the mean of the i-th output data, N y is the number of output data, is the economic coefficient corresponding to the j-th input data, is the mean of the jth input data, N u is the number of input data, is the i-th output data corresponding to the new steady-state operating point, is the difference between the value of the i-th output data corresponding to the new steady-state operating point and its mean value, is the jth input data corresponding to the new steady-state operating point, is the difference between the value of the jth input data corresponding to the new steady-state operating point and its mean value, represents the probability of obtaining a normal distribution with a confidence level of 1-α, is the variance of the i-th output data, G ij Represents the output data y i To input data Δu j The steady-state gain, is the minimum value of the i-th output data, is the maximum value of the i-th output data, is the minimum value of the j-th input data, is the maximum value of the j-th input data; The updating module is specifically configured to update the current adjustment parameters corresponding to each subsystem in the distributed system using an ILC-based weight coefficient adjustment strategy based on the mean and variance of the input data of the actual steady-state operating point, the mean and variance of the output data, and the economic optimization problem; The updating module is further used to convert the economic optimization problem into a standard form of economic optimization problem: st A[Δu s ,Δy s ] T ≤B Among them, c is the matrix coefficient after being converted into the standard form of economic optimization problem, Δu s is the vector consisting of the difference between the value of the input data corresponding to the new steady-state operating point and its mean value, Δy s is a vector consisting of the difference between the output data value corresponding to the new steady-state operating point and its mean value, A and B are the constraint coefficient matrices in the constraint conditions in the standard form of the economic optimization problem; Find the input constraints B corresponding to all inputs in the mth subsystem from the constraints in the standard form of the economic optimization problem. u Output constraints B corresponding to all outputs y , and record the minimum value MIN u and MIN y : in, represents the row corresponding to the output constraint in the mth subsystem, Indicates the row corresponding to the input constraint in the mth subsystem, B k 、A k represents the constraint coefficient corresponding to the kth row, x * The optimal solution obtained based on the economic optimization problem; If MIN u Less than MIN y , then write φ in the sequence used to record the direction of change of economic performance parameters t =1, if MIN u Not less than MIN y , then write φ in the sequence t =-1; If the sequence always maintains the same value, then according to λ t+1 =λ t *(1+l(φ t )) Update the adjustment parameters corresponding to the mth subsystem; where λ t is the current adjustment parameter of the mth subsystem at the tth iteration, λ t+1 is the updated current adjustment parameter of the mth subsystem corresponding to the t+1th iteration, l(φ t ) is the updated learning rate, l1 is the first learning rate, and l2 is the second learning rate; If the sequence does not always maintain the same value, the current adjustment parameter corresponding to the m-th subsystem is updated using a bisection method.

5. A system operation control device, characterized in that: include: memory for storing computer programs; A processor, configured to implement the steps of the system operation control method according to any one of claims 1 to 3 when executing the computer program.

6. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the system operation control method according to any one of claims 1 to 3 are implemented.

Images