A heat supply decoupling control method and system based on disturbance suppression

Abstract

The application discloses a heat supply decoupling control method and system based on disturbance suppression, comprising the following steps: establishing a high-dimensional linear model of a heat exchange station system based on Koopman operator theory; calculating a control rate of a Koopman MPC controller output based on the high-dimensional linear model of the heat exchange station system through a model predictive control method; equivalent coupling effect to a disturbance compensation value output by a Koopman state observer based on the high-dimensional linear model of the heat exchange station system; combining the disturbance compensation value with the control rate to obtain a final decoupling control rate, so as to obtain a decoupling controller of a temperature control loop and a flow control loop of the heat exchange station system. The application solves the problems of slow regulation speed, poor regulation effect and large decoupling difficulty caused by the mutual coupling of the temperature and flow regulation of the heat exchange station.

Description

Technical Field

[0001] The present invention relates to the technical field of heating decoupling control, and in particular to a heating decoupling control method and system based on disturbance suppression. Background Art

[0002] Energy efficiency optimization of centralized heating systems has become a key issue in urban infrastructure upgrades. As a key node in energy transmission and distribution, heat exchange stations exhibit typical complex dynamic characteristics such as multivariable coupling and nonlinear time lag. Their control quality directly affects 15%-25% of the overall energy consumption of the heating system. Current control research faces three technical bottlenecks: First, traditional PID and improved algorithms are difficult to effectively cope with the strong coupling characteristics of the heating network, and multivariable coordinated control suffers from the contradiction between response lag and overshoot; second, the decoupling method based on the mechanism model is limited by the modeling error of the complex heat transfer process of the heat exchange station, and the decoupling effect is significantly attenuated under actual working conditions; third, the online learning ability and stability of existing data-driven methods under time-varying conditions have not yet formed a reliable theoretical guarantee.

[0003] In recent years, nonlinear system modeling based on Koopman operator theory has provided a new approach to addressing these challenges. This theory, through data-driven dimensionality-increasing mapping, transforms nonlinear dynamic systems into observable systems in an infinite-dimensional linear space. Compared to traditional Taylor expansion linearization methods, it demonstrates greater generalization and predictive accuracy in modeling heat transfer dynamics in heat exchange stations. Integrating this with the model predictive control (MPC) framework enables the construction of a predictive decoupling controller with multi-step rolling optimization, effectively compensating for phase lag caused by system time delay.

[0004] Decoupling control is widely used in industries such as petroleum, chemical engineering, and steelmaking. As industrial processes advance, control systems become increasingly complex, and fluctuations in one variable can affect others. A heat exchange station is a typical nonlinear, strongly coupled system. Directly controlling a single channel without decoupling can lead to slow control, deviations between the control value and the actual setpoint, and even safety issues such as hydraulic imbalance. Therefore, research on decoupling heating systems is necessary. Current research in the industrial sector focuses on three key areas: 1. Compensation decoupling schemes based on traditional decoupling, including feedforward, feedback, and diagonal matrices; 2. Multivariable coupled adaptive control systems based on advanced control methods such as internal model control, adaptive control, and predictive control; and 3. Intelligent decoupling methods, such as neural networks and fuzzy algorithms. In summary, due to the complexity of real-world industrial environments, traditional decoupling methods require high model accuracy, making them less effective. Currently, both adaptive and intelligent decoupling control require further verification of system stability and convergence. In addition, although decoupling technology has been applied in chemical industry, papermaking, distillation tower, AC motor and other fields, there is still little research on decoupling technology for heating network in the heating industry. It is necessary to conduct research on decoupling design for centralized heating. Summary of the Invention

[0005] (1) Technical issues to be resolved

[0006] Based on the above problems, the present invention provides a heat supply decoupling control method and system based on disturbance suppression to solve the problems of slow control speed, poor control effect and great decoupling difficulty caused by the mutual coupling of temperature and flow control in the heat exchange station.

[0007] (2) Technical solution

[0008] In view of the above technical problems, the present invention provides a heat supply decoupling control method based on disturbance suppression, comprising:

[0009] S1. Establish a high-dimensional linear model of the heat exchange station system based on Koopman operator theory;

[0010] S2. Calculating the control rate output by the Koopman MPC controller using a model predictive control method based on a high-dimensional linear model of the heat exchange station system;

[0011] S3. Based on the high-dimensional linear model of the heat exchange station system, the coupling effect is equivalent to the disturbance compensation value output by the Koopman state observer;

[0012] S4. Combining the disturbance compensation value with the control rate to obtain a final decoupling control rate, thereby obtaining a decoupling controller for the temperature control loop and the flow control loop of the heat exchange station system.

[0013] Furthermore, the S1 includes:

[0014] S11. Determine the model input and output and collect data:

[0015] The model is a temperature control loop model or a flow control loop model. For the temperature control loop model, the state of the system is the temperature and the rate of change of the temperature, that is, ,in Indicates temperature, superscript Represents the transpose of the matrix. The input of the collected system is the secondary network water supply temperature corresponding to the change in the opening of the primary side water inlet valve;

[0016] For the flow control loop model, the state of the system is the flow rate and the rate of change of the flow rate, that is, ,The input of the collected system is the data of the secondary network regulating valve flow and circulating water flow;

[0017] Set the sampling frequency to collect the state and input of the system, and divide the resulting data set into a time series matrix , and , where n represents the number of rows in the matrix, is the number of samples, n=2;

[0018]

[0019] S12. State Dimensionality: Define Dimensionality Function , consisting of a set of linearly independent essential functions Composition, of which Represents the dimension after dimensionality increase, , Upgrade to:

[0020]

[0021] Before the dimension-raising function The dimension remains consistent with the original state, that is:

[0022]

[0023] S13, extended input: and Control input As an extension, we get:

[0024]

[0025] S14. Obtain the Koopman model of the system based on the Koopman operator: Obtain the Koopman operator by optimizing the problem Finite-dimensional approximation of , and construct a high-dimensional linear model of the temperature control loop and the flow control loop, namely the Koopman model;

[0026] Furthermore, By minimizing the following formula:

[0027]

[0028] The optimization problem is transformed into:

[0029] matrix , By minimizing the following formula,

[0030]

[0031] matrix By minimizing the following formula:

[0032]

[0033] Thus, the high-dimensional linear model of the temperature control loop and the flow control loop is obtained:

[0034]

[0035] in, , are the system matrix and input matrix of the high-dimensional linear model, , is the system output matrix in the original dimension, is the matrix in the high-dimensional linear model, represents the discrete time step of sampling;

[0036] Define the loss function:

[0037]

[0038] in, To control the time domain, For the prediction time domain, is the error weight, To control the weight, e is the error between the feedback value and the input.

[0039] Furthermore, the S2 includes:

[0040] According to the high-dimensional linear model, the state prediction equation in the prediction domain is expressed as:

[0041]

[0042] The output prediction equation of the system is expressed as:

[0043]

[0044] in, is the predicted output vector, is the input vector;

[0045] The loss function is expressed as:

[0046]

[0047] in is the reference input sequence, and The diagonal elements are and The diagonal matrix of , we get the final optimal control problem:

[0048]

[0049] in, , and finally we get a standard quadratic programming problem, which can find the global optimal solution, that is, in the prediction time domain Optimal position control sequence within , and The first element is directly used as the control rate of the controller output .

[0050] Furthermore, the S3 includes:

[0051] Let the state space equation form:

[0052]

[0053] The system state and disturbance are estimated using the following high-dimensional state observer based on the Koopman operator:

[0054]

[0055] in, represents the observation value of the high-dimensional state, represents the observer gain, shows the control rate without disturbance compensation, Represents the output of the observer, and the state error of the observer is:

[0056]

[0057] We can get:

[0058]

[0059] in , let the disturbance estimate be:

[0060]

[0061] but:

[0062]

[0063] Then the minimum variance solution of the equivalent input estimate is:

[0064]

[0065] in Then the disturbance compensation value is obtained .

[0066] Furthermore, in S4, the decoupling control rates of the temperature control loop and the flow control loop of the heat exchange station system are both:

[0067] .

[0068] Furthermore, the decoupling controllers of the temperature control loop and the flow control loop of the heat exchange station system are respectively:

[0069] Decoupling controller for the temperature control loop: The temperature setpoint x1(t) is input to the MPC controller, and the output of the Koopman MPC controller is Subtract the disturbance compensation value output by the disturbance estimator Then it is input into the variable frequency control board of the booster pump to control the output secondary side supply and return water temperature y1(t), and the feedback supply and return water temperature y1(t) is output to the Koopman state observer and Koopman MPC controller. The Koopman state observer outputs the feedback supply and return water temperature y1(t) to the disturbance estimator and Koopman MPC controller;

[0070] Decoupling controller of the flow control loop: The flow setpoint x2(t) is input to the MPC controller, and the output of the Koopman MPC controller is Subtract the disturbance compensation value output by the disturbance estimator The feedback of the circulating water flow rate y2(t) is then output to the Koopman state observer and the Koopman MPC controller. The Koopman state observer outputs the feedback of the circulating water flow rate y2(t) to the disturbance estimator and the Koopman MPC controller.

[0071] The present invention also discloses a heat supply decoupling control system based on disturbance suppression, comprising:

[0072] at least one processor; and at least one memory communicatively coupled to the processor, wherein:

[0073] The memory stores program instructions that can be executed by the processor, and the processor can execute the method by calling the program instructions.

[0074] The present invention also discloses a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions enable the computer to execute the method.

[0075] (3) Beneficial effects

[0076] The above technical solution of the present invention has the following advantages:

[0077] (1) The present invention establishes a high-dimensional linear model of the heat exchange station system through the Koopman operator theory, and then combines the model predictive control method to control the temperature and flow. At the same time, the high-dimensional linear model of the system is used to design a state observer, and the idea of ​​equivalent input interference is adopted to perform decoupling control to reduce the degree of coupling between the temperature and flow regulation channels, thereby solving the problem of slow control speed and poor control effect caused by the mutual coupling of temperature and flow control in the heat exchange station;

[0078] (2) The decoupling method of the present invention does not require the establishment of a mechanism model, does not require prior information about interference, and does not require high model accuracy. It realizes the decoupling control and interference suppression of the heat exchange station system without modeling the coupling between the temperature and flow channels, thereby reducing the difficulty of decoupling. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] The features and advantages of the present invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the present invention in any way. In the accompanying drawings:

[0080] Figure 1 Schematic diagram of a coupling model according to an embodiment of the present invention;

[0081] Figure 2 A schematic diagram of a decoupling strategy according to an embodiment of the present invention;

[0082] Figure 3 Schematic diagram of a Koopman decoupling controller based on disturbance rejection according to an embodiment of the present invention;

[0083] Figure 4 Schematic diagram of the structure of the decoupling controller of the heat exchange station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0084] The specific embodiments of the present application are described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0085] The present application provides a heat supply decoupling control method based on disturbance suppression. In the operation process of a central heat supply system, a heat exchange station transmits the energy generated at a heat source to various building users through the rotation of a heat exchange system and a circulating pump. The size and number of heat exchange stations are mainly determined by the required heat supply area and the building envelope characteristics. The process control parameters of the heat exchange station include the supply and return water temperature, flow rate, pressure, rotating speed of the pressurizing pump and circulating pump, etc. The working process of the heat exchange station is as follows: the water supply of the primary pipe network exchanges heat with the return water of the secondary pipe network through the plate heat exchanger, and the exchanged heat is transmitted to each building heat user through the secondary pipe network supply side under the driving of the circulating pump. Then the medium in the user room returns to the return circuit of the secondary pipe network again, preparing for the next heat exchange with the primary pipe network water supply side, and so on, to achieve the heat exchange purpose. Generally speaking, there are two main operating indicators in the heat exchange station system: the first indicator is the regulation of the secondary side outlet water temperature, also known as the quality regulation, which directly affects the temperature of the heat supply terminal; generally, the opening degree of the primary side inlet water flow valve is adjusted to control the flow of municipal hot water or hot steam on the primary side, so as to regulate the secondary side outlet water temperature, and the set value of the secondary side outlet water temperature is determined by the outdoor temperature. The second indicator is the regulation of the circulating water quantity of the secondary side user pipe network, also known as the quantity regulation, and the circulating water quantity is adjusted according to the actual use of heat energy, i.e. according to the heat energy load, without considering the influence of outdoor temperature; when the number of building heat terminal openings is small, the circulating water quantity can be reduced by reducing the circulating water pump frequency, so that the secondary side supply and return water temperature difference of the heat exchange station remains a constant value, and the heat transport efficiency of the water pump is improved through the operation mode of the large temperature difference and small flow air conditioning water system, so as to achieve the purpose of energy saving.

[0086] The traditional control strategy includes two loops of secondary network temperature control and circulating water flow control, and models are established for the secondary network water supply temperature control system and the circulating water flow control system. Meanwhile, considering the coupling relationship between the two channels, a coupling model is established as follows: Figure 1As shown in the figure, the secondary outlet water temperature control loop and the secondary supply / return water temperature difference control loop of the heat exchange station are two coexisting control loops. However, in actual systems, the two control loops do not exist independently and influence and interfere with each other during their respective adjustments. In the secondary outlet water temperature control, valve adjustment not only affects the secondary outlet water temperature but also the secondary supply / return water temperature difference. In the secondary supply / return water temperature difference control, adjusting the circulating water pump's frequency control board to change the secondary temperature difference also affects the secondary outlet water temperature. Mechanism modeling is generally used when modeling, and field data is collected for parameter identification. However, the heating system itself is complex, and the control loops have significant inertia and time-varying lag. The traditional modeling method, which generates a first-order and second-order link plus lag model, ignores many factors and is difficult to describe the heating process.

[0087] Taking into account the nonlinearity existing in the heat exchange process, especially the characteristics such as inertia and time lag that change with the process, the present invention establishes a high-dimensional linear model of the heat exchange station system based on the Koopman operator theory, and combines it with the MPC algorithm for temperature / flow control. At the same time, the high-dimensional linear model of the system is used to design a disturbance observer, and the idea of ​​equivalent input disturbance is adopted for decoupling control to reduce the degree of coupling between the temperature and flow regulation channels.

[0088] The following are the detailed steps for implementing the heat supply decoupling control method based on Koopman operator theory in the present invention:

[0089] S1. Establish a high-dimensional linear model of the heat exchange station system based on Koopman operator theory;

[0090] The heat exchange station's secondary outlet water temperature control loop and the secondary supply / return water temperature difference control loop are two coexisting control loops. However, in actual systems, the two control loops do not exist independently; they influence and interfere with each other during their respective adjustments. In the secondary outlet water temperature control, when the valve is adjusted, not only the secondary outlet water temperature is affected, but also the secondary supply / return water temperature difference. In the secondary supply / return water temperature difference control, when the frequency control board of the circulating water pump is adjusted to change the secondary temperature difference, the secondary outlet water temperature is also affected. For the heat exchange station control system, the coupling effect is treated as a disturbance, and a separate disturbance compensation controller is designed. For both control loops, due to their strong hysteresis characteristics, it is difficult to obtain an accurate model using mechanism modeling. Modeling based on data-driven Koopman operator theory can obtain a high-dimensional linear model, which is convenient for designing controllers in combination with linear control algorithms.

[0091] Koopman operator theory seeks the optimal representation for transforming nonlinear systems into linear systems by projecting the state space into a high-dimensional Koopman space using appropriate lifting functions. Initially, this theory was used to model autonomous systems. For controlled systems like heat exchange stations, the EDMD algorithm is a simple and effective solution.

[0092] The temperature control and flow control processes of the heat exchange station system are both relatively complex nonlinear systems. Considering that periodic data is used in both modeling and control, the following discrete nonlinear system with control input is adopted:

[0093]

[0094] in, , They represent the system at discrete time steps The Koopman model of the temperature loop and flow loop of the heat exchange station system can be obtained by the following four steps, taking the mass control channel / temperature loop as an example:

[0095] S11. Determine the model input and output and collect data:

[0096] The model is a temperature control loop model or a flow control loop model. For the temperature loop model, taking into account the inertia and hysteresis of temperature change, the state of the system is the temperature and the rate of change of temperature, that is, ,in Indicates temperature, superscript Representing the transpose of the matrix, the first input of the system that needs to be collected is the secondary network water supply temperature corresponding to the change in the opening of the primary side water inlet valve, so as to obtain all the dynamic characteristics of the system as much as possible.

[0097] For the flow control loop model, the state of the system is the flow rate and the rate of change of the flow rate, that is, ,The input of the collected system is the data of secondary network regulating valve flow and circulating water flow.

[0098] Set a suitable sampling frequency to collect the state and input of the system, and divide the obtained data set into a time series matrix , and , where n represents the number of rows in the matrix, is the number of samples, n =2.

[0099] S12. State Dimensionality: Define Dimensionality Function , consisting of a set of linearly independent essential functions Composition, of which Represents the dimension after dimensionality increase, , Upgrade to:

[0100] In order to fully include the original dimension information in the high-dimensional space, the dimension-raising function is preceded by The dimension remains consistent with the original state, that is:

[0101]

[0102] S13, extended input: The control input is not subjected to dimensionality increase processing, but is treated as an extension of the lifting state. and Control input As an extension, we get:

[0103]

[0104] Since the control input is not subjected to dimensionality increase, the resulting Koopman model still undergoes high-dimensional evolution under the stimulation of the original input.

[0105] S14. Obtain the Koopman model of the system based on the Koopman operator: Obtain the Koopman operator by optimizing the problem Finite-dimensional approximation of , and construct the Koopman model. It can be obtained by minimizing the following formula:

[0106]

[0107] Since there is no need to predict the control input, the last components, matrix Before Row Elements Can be decomposed into , so the optimization problem can be transformed into:

[0108]

[0109] matrix , By minimizing the above formula, the matrix It can be obtained by minimizing the following formula:

[0110]

[0111] At this point, we can obtain high-dimensional linear models of the temperature control loop and the flow control loop, which can be used to predict future states and design controllers:

[0112] Taking temperature regulation as an example, its Koopman high-dimensional prediction model is:

[0113]

[0114] in, , are the system matrix and input matrix of the high-dimensional linear model obtained above, ,in is the system output matrix in the original dimension, which is [1 0 0], is the matrix in the high-dimensional model. The control quantity under the dimension is obtained by performing rolling optimization at each time step using model predictive control;

[0115] Define the loss function of the high-dimensional linear model of the heat exchange station system:

[0116]

[0117] in, To control the time domain, For the prediction time domain, is the error weight, To control the weight, e is the error between the feedback value and the input.

[0118] S2. Based on the high-dimensional linear model of the heat exchange station system, the control rate output by the Koopman MPC controller is calculated by the model predictive control method. ;

[0119] The model predictive control method is MPC, Model Predictive Control. According to the high-dimensional linear model, the state prediction equation in the prediction domain can be expressed as:

[0120]

[0121] Since the predicted system output , then the output prediction equation of the system can be expressed as:

[0122]

[0123] in, is the predicted output vector, is the input vector.

[0124] The specific expression of the matrix in the output prediction equation is:

[0125]

[0126] The loss function can be expressed as:

[0127]

[0128] in is the reference input sequence, and The diagonal elements are and Since the first term on the right side of the equation is Regardless, the final optimal control problem is obtained:

[0129]

[0130] in, , and finally we get a standard quadratic programming problem, which can quickly find the global optimal solution, that is, in the prediction time domain Optimal position control sequence within Since the modeling process maintains the original dimension of the input, it can be The first element of is directly used as the control rate of the controller output.

[0131] The calculation method of the control quantity of the flow control loop and the temperature control loop is the same;

[0132] S3. Based on the high-dimensional linear model of the heat exchange station system, the coupling effect is equivalent to the disturbance compensation value output by the Koopman state observer. ;

[0133] Since there is a coupling relationship between the temperature control loop and the flow control loop of the heat exchange station, if the decoupling operation is not performed, directly controlling a single channel will lead to a series of safety issues such as slow control speed, deviation between the control value and the actual set value, and even hydraulic imbalance. The present invention adopts the idea of ​​disturbance suppression to regard the coupling effect in the system as a disturbance, such as Figure 2 shown. Figure 2 The left figure is a schematic diagram of the original system without decoupling; Figure 2 The right figure is a schematic diagram of decoupling in this embodiment. 、 Both represent input; 、 、 、 、 and Both represent transfer functions; and Both represent disturbances. The disturbance estimator generates an estimated disturbance value, which is then compensated at the input to suppress the influence of the other channel's coupling on the system. This approach enables decoupled control and disturbance suppression for the heat exchange station system without modeling the coupling between the temperature and flow channels.

[0134] Taking the temperature control loop as an example, considering the disturbance and coupling of the temperature control channel, an equivalent input disturbance value can be defined in the input channel by adopting the idea of ​​equivalent input disturbance. ,like Figure 3 As shown, represents the disturbance estimator, Express the transfer function so that the effect of the disturbance on the output is the same as the effect of the original total disturbance on the output, and its state space equation form can be obtained:

[0135]

[0136] The system state and disturbance are estimated using the following high-dimensional state observer based on the Koopman operator:

[0137]

[0138] in, represents the observation value of the high-dimensional state, represents the observer gain, shows the control rate without disturbance compensation, Represents the output of the observer, and the state error of the observer is:

[0139]

[0140] We can get:

[0141]

[0142] in , let the disturbance estimate be:

[0143]

[0144] but:

[0145]

[0146] We can get:

[0147]

[0148] Then the minimum variance solution of the equivalent input estimate is:

[0149]

[0150] in .

[0151] The disturbance compensation value can be obtained .

[0152] S4, the disturbance compensation value The control rate of the controller output Combined, the final decoupling control rate is obtained, thereby obtaining the decoupling controller of the temperature control loop and the flow control loop of the heat exchange station system;

[0153] The decoupling control rates of the quality control channel / temperature control loop and the quantity control channel / flow control loop of the heat exchange station system are:

[0154]

[0155] The decoupling controller structure of the quality control channel / temperature control loop and the quantity control channel / flow control loop is designed as follows: Figure 3 and 4 As shown in the figure, the decoupling controller structure of the quality control channel / temperature control loop includes: inputting the temperature setting value x1(t) into the MPC controller, and the output of the Koopman MPC controller Subtract the disturbance compensation value output by the disturbance estimator The frequency conversion control board of the booster pump is then input to control the output secondary side supply and return water temperature y1(t), and the feedback supply and return water temperature y1(t) is output to the Koopman state observer and the Koopman MPC controller. The Koopman state observer outputs the feedback supply and return water temperature y1(t) to the disturbance estimator and the Koopman MPC controller; the decoupling controller structure of the quantity regulation channel / flow control loop includes: inputting the flow set value x2(t) into the MPC controller, and the output of the Koopman MPC controller Subtract the disturbance compensation value output by the disturbance estimator The feedback of the circulating water flow rate y2(t) is then output to the Koopman state observer and the Koopman MPC controller. The Koopman state observer outputs the feedback of the circulating water flow rate y2(t) to the disturbance estimator and the Koopman MPC controller.

[0156] Finally, it should be noted that the above-mentioned control method can be converted into software program instructions, which can be implemented by using a control system including a processor and a memory, or by computer instructions stored in a non-transitory computer-readable storage medium. The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform some steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0157] In summary, the above-mentioned heating decoupling control method and system based on disturbance suppression have the following beneficial effects:

[0158] (1) The present invention establishes a high-dimensional linear model of the heat exchange station system through the Koopman operator theory, and then combines the model predictive control method to control the temperature and flow. At the same time, the high-dimensional linear model of the system is used to design a state observer, and the idea of ​​equivalent input interference is adopted to perform decoupling control to reduce the degree of coupling between the temperature and flow regulation channels, thereby solving the problem of slow control speed and poor control effect caused by the mutual coupling of temperature and flow control in the heat exchange station;

[0159] (2) The decoupling method of the present invention does not require the establishment of a mechanism model, does not require prior information about interference, and does not require high model accuracy. It realizes the decoupling control and interference suppression of the heat exchange station system without modeling the coupling between the temperature and flow channels, thereby reducing the difficulty of decoupling.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations shall fall within the scope defined by the appended claims.

Claims

1. A heating decoupling control method based on disturbance suppression, characterized in that: include: S1. Establish a high-dimensional linear model of the heat exchange station system based on Koopman operator theory; S2. Calculating the control rate output by the Koopman MPC controller using a model predictive control method based on a high-dimensional linear model of the heat exchange station system; S3. Based on the high-dimensional linear model of the heat exchange station system, the coupling effect is equivalent to the disturbance compensation value output by the Koopman state observer; The S3 includes: Let the state space equation form: The system state and disturbance are estimated using the following high-dimensional state observer based on the Koopman operator: in, represents the observation value of the high-dimensional state, represents the observer gain, represents the control rate without disturbance compensation, Represents the output of the observer, and the state error of the observer is: We can get: in , let the disturbance estimate be: but: Then the minimum variance solution of the equivalent input estimate is: in , then the disturbance compensation value is obtained ; S4. Combining the disturbance compensation value with the control rate to obtain a final decoupling control rate, thereby obtaining a decoupling controller for the temperature control loop and the flow control loop of the heat exchange station system; The decoupling control rates of the temperature control loop and the flow control loop of the heat exchange station system are: ; The decoupling controllers of the temperature control loop and flow control loop of the heat exchange station system are: Decoupling controller for the temperature control loop: The temperature setpoint x1(t) is input to the MPC controller, and the output of the Koopman MPC controller is Subtract the disturbance compensation value output by the disturbance estimator Then it is input into the variable frequency control board of the booster pump to control the output secondary side supply and return water temperature y1(t), and the feedback supply and return water temperature y1(t) is output to the Koopman state observer and Koopman MPC controller. The Koopman state observer outputs the feedback supply and return water temperature y1(t) to the disturbance estimator and Koopman MPC controller; Decoupling controller of the flow control loop: The flow setpoint x2(t) is input to the MPC controller, and the output of the Koopman MPC controller is Subtract the disturbance compensation value output by the disturbance estimator The feedback of the circulating water flow rate y2(t) is then output to the Koopman state observer and the Koopman MPC controller. The Koopman state observer outputs the feedback of the circulating water flow rate y2(t) to the disturbance estimator and the Koopman MPC controller.

2. The heating decoupling control method based on disturbance suppression according to claim 1 is characterized in that: Said S1 comprises: S11. Determine the model input and output and collect data: The model is a temperature control loop model or a flow control loop model. For the temperature control loop model, the state of the system is the temperature and the rate of change of the temperature, that is, ,in Indicates temperature, superscript Represents the transpose of the matrix. The input of the collected system is the secondary network water supply temperature corresponding to the change in the opening of the primary side water inlet valve; For the flow control loop model, the state of the system is the flow rate and the rate of change of the flow rate, that is, ,The input of the collected system is the data of the secondary network regulating valve flow and circulating water flow; Set the sampling frequency to collect the state and input of the system, and divide the obtained data set into a time series matrix , and , where n represents the number of rows in the matrix, is the number of samples, n=2; S12. State Dimensionality Upgrading: Defining Dimensionality Upgrading Function , consisting of a set of linearly independent essential functions Composition, of which Represents the dimension after dimensionality increase, , Upgrade to: Before the dimension-raising function The dimension remains consistent with the original state, that is: S13, extended input: and Control input As an extension, we get: S14. Obtain the Koopman model of the system based on the Koopman operator: Obtain the Koopman operator by optimizing the problem Finite-dimensional approximation of , and construct a high-dimensional linear model of the temperature control loop and the flow control loop, namely the Koopman model.

3. The heating decoupling control method based on disturbance suppression according to claim 2 is characterized in that: The S14 specifically includes: By minimizing the following formula: The optimization problem is transformed into: matrix , By minimizing the following formula, matrix By minimizing the following formula: Thus, the high-dimensional linear model of the temperature control loop and the flow control loop is obtained: in, , are the system matrix and input matrix of the high-dimensional linear model, , is the system output matrix in the original dimension, is the matrix in the high-dimensional linear model, represents the discrete time step of sampling; Define the loss function: in, To control the time domain, For the prediction time domain, is the error weight, To control the weight, e is the error between the feedback value and the input.

4. The heat supply decoupling control method based on disturbance suppression according to claim 3 is characterized in that: The S2 includes: According to the high-dimensional linear model, the state prediction equation in the prediction domain is expressed as: The output prediction equation of the system is expressed as: in, is the predicted output vector, is the input vector; The loss function is expressed as: in is the reference input sequence, and The diagonal elements are and The diagonal matrix of , we get the final optimal control problem: in, , and finally we get a standard quadratic programming problem, which can find the global optimal solution, that is, in the prediction time domain Optimal position control sequence within , and The first element is directly used as the control rate of the controller output .

5. A heating decoupling control system based on disturbance suppression, characterized in that: include: at least one processor; and at least one memory communicatively connected to the processor, wherein: The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 4.

6. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the method according to any one of claims 1 to 4.

Images