Load frequency control method, system and equipment based on data driving and medium

Abstract

The invention relates to the technical field of automatic control, in particular to a load frequency control method, system and device based on data driving and a medium, and the method comprises the steps: firstly constructing a discrete time model for the operation of an ith regional power system, and calculating a local tracking error according to a control input parameter; constructing an action network, and calculating an unknown part estimation value of the discrete time model based on an error; constructing an output triggering condition through the local tracking error variation, judging whether the output triggering condition is met, and if so, processing the estimated value by using the sliding mode controller to obtain an updating instruction; and constructing an input trigger condition according to the control input parameter variation, and when the input trigger condition is met, operating the power system according to the updated control input parameter. The problems that communication resources are limited and a system model is uncertain are effectively solved, control performance glide caused by data delay is made up by means of a model uncertainty estimation technology, and an efficient and reliable solution is provided for load frequency control of a micro-grid system.

Description

Technical Field

[0001] The present invention relates to the field of automatic control technology, and in particular to a load frequency control method, system, device and medium based on data-driven. Background Art

[0002] In the operation of a microgrid system, load frequency control (LFC), as a key technology, aims to maintain the dynamic balance between the system load demand and frequency, and plays a crucial role in improving the overall operation performance and stability of the power system. However, due to the high complexity of its topological structure and the coupled influence of multiple factors in the operating environment, accurate system modeling faces huge challenges. Especially in large-scale microgrid scenarios, the problem of input delay is widespread, which further exacerbates the difficulty of system control. Therefore, researching a microgrid load frequency control strategy that can work under input delay conditions without relying on an accurate system model has become an important research topic in the current field of power system automation.

[0003] Meanwhile, considering the scarcity of communication resources in the microgrid communication architecture, the load frequency control strategy needs to balance communication efficiency and energy consumption optimization to reduce the number of communications, reduce communication energy consumption, and achieve the efficient and stable operation of the system. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a load frequency control method, system, device and medium based on data-driven, effectively solving the problems of limited communication resources and system model uncertainty. At the same time, using model uncertainty estimation technology to compensate for the degradation of control performance caused by data delay, providing an efficient and reliable solution for the load frequency control of the microgrid system. The method includes the following steps: Step S1: Construct a discrete-time model of the operation of the power system in the i-th area. According to the control input parameters received by the current power system, calculate the local tracking error through the discrete-time model. The local tracking error is the deviation between the actual area control error and the desired area control error of the system; Step S2: Construct an evaluation network and an action network, iteratively update the parameters of the action network according to the evaluation network. Based on the local tracking error, calculate the unknown part of the discrete-time model through the action network updated in each iteration to obtain an estimated value of the unknown part; Step S3: Construct an output trigger condition according to the change amount between the local tracking error and the local tracking error at the previous output trigger moment, and judge whether the current moment meets the output trigger condition: Otherwise, keep the estimated value of the unknown part and the local tracking error at the previous output trigger moment of the current discrete-time model unchanged, and return to continue executing step S1; If so, process the estimated value of the unknown part through a pre-constructed sliding mode controller to obtain updated control input parameters, and execute step S4; Step S4: Construct an input trigger condition based on the change in the control input parameter at different input moments, and determine whether the current moment satisfies the input trigger condition: If not, keep the control input parameter received by the power system in the current area unchanged, and return to continue executing step S1; If so, send the updated control input parameter to the power system in the current area, and the power system in the current area operates according to the updated control input parameter, and return to continue executing step S1.

[0005] In an embodiment of the present invention, in step S1, the method for constructing the discrete-time model of the operation of the power system in the i-th area is as follows: Represent the discrete-time model of the operation of the power system in the i-th area as: (1) where T represents the sampling period, h represents the input time delay, , , , are system matrices, represents the state vector, k represents the discrete time, represents the actual area control error at discrete time k, represents the input time constant, represents the control input parameter, represents the disturbance vector at discrete time k, , in addition, , , , are defined as , , , , Since the internal stability of the system equilibrium point is equivalent to the stability of the origin, that is, the disturbance vector , the discrete-time model representation in formula (1) is converted into an input delay system: (2) For the input delay system, if there is a control input at time k , there exists a bounded input delay function such that , substituting it into formula (2), we get: (3) where , is known, are the known parts of the system matrices and respectively, is the unknown but bounded part of the system matrices and . Formula (3) is further improved to a discrete-time model including the unknown part as follows: (4) where is the unknown part of the model, represents the generator inertia, represents the generator set damping coefficient, represents the synchronous torque coefficient, represents the steam turbine time constant, represents the control time constant, represents the frequency deviation coefficient, j represents the power system of the jth area, represents the droop characteristic.

[0006] In an embodiment of the present invention, the calculation method of the local tracking error is as follows: (5) where represents the desired area control error.

[0007] In an embodiment of the present invention, the method for constructing the evaluation network is as follows: Based on the control input parameters received by the power system in the current area and the local tracking error, calculate the instantaneous cost value through the constructed instantaneous cost function: (6) Construct the performance index function for the operation of the power system in the ith area according to the instantaneous cost value: (7) Simplify to , and convert the performance index function of formula (7) into the Bellman equation: (8) Convert the Bellman equation of formula (8) into a performance index function calculated iteratively using a neural network: (9) Among them, and are positive constants, and represent performance metric functions, represents the data composed of and ; represents the activation function, represents the weights of the evaluation network, represents the dimension, represents the number of neurons; represents the performance metric value obtained by iterative calculation of the evaluation network at discrete time k, is the discount factor.

[0008] In one embodiment of the present invention, the method for iteratively updating the parameters of the action network according to the evaluation network is as follows: Define the temporal difference error of the evaluation network as: (10) Based on the temporal difference error, construct the loss function of the evaluation network as follows: (11) With the goal of minimizing the loss function value of the evaluation network, adjust the weights of the evaluation network using the following formula: (12) Among them, represents the weights obtained by iterative training of the evaluation network at discrete time (k + 1), represents the weights obtained by iterative training of the evaluation network at discrete time k, represents the weights obtained by iterative training of the evaluation network at discrete time (k - 1), represents the loss function value calculated by the evaluation network at discrete time k; represents the performance metric value obtained by iterative calculation of the evaluation network at discrete time (k - 1), represents the learning rate of the evaluation network.

[0009] In one embodiment of the present invention, the training method of the action network is: Take the performance metric value function of the power system operation in the i-th region constructed by the evaluation network as the error function of the action network, and construct the loss function of the action network according to the error function (13) With the goal of minimizing the loss function value of the action network, adjust the weights of the action network using the following formula: (14) where, represents the weight iteratively calculated by the action network at the discrete time (k + 1), represents the weight iteratively calculated by the action network at the discrete time k, represents the estimated value of the unknown part iteratively calculated by the action network at the discrete time k; is the learning rate of the action network, represents the activation function, represents the dimension, represents the number of neurons; represents the data composed of the control input parameter and the local tracking error at the discrete time k, represents the activation function, represents the weight of the evaluation network, represents the number of neurons.

[0010] In an embodiment of the present invention, the method for obtaining the updated control input parameter is as follows: Calculate the estimated value of the unknown part according to the action network : (15) Construct a sliding mode controller, and the sliding mode controller processes the estimated value of the unknown part through the following formula to obtain the updated control input parameter: (16) where, , , , are sliding mode controller parameters and satisfy , is a sliding mode controller parameter greater than zero, , , and satisfy , is the set sliding mode surface, represents the local tracking error.

[0011] In an embodiment of the present invention, construct an output trigger condition according to the change amount between the local tracking error and the local tracking error at the previous output trigger moment, as follows: (17) where, represents the output trigger moment of the power system in the i-th area, represents the input trigger moment of the power system in the i-th area, , represents the local tracking error at discrete time k, represents the local tracking error at the previous output trigger moment of is a constant greater than zero, is the measurement standard trigger threshold.

[0012] In an embodiment of the present invention, an input trigger condition is constructed according to the change amount of the control input parameter at different input moments, as follows: (18) wherein, represents the input trigger moment of the power system in the i-th area, represents the previous input trigger moment of represents the moment when the control input parameter received by the system, is the measurement standard trigger threshold, is the trigger coefficient.

[0013] Based on the same inventive concept, the present invention also provides a data-driven load frequency control system for implementing the steps of the data-driven load frequency control method described above. The data-driven load frequency control system includes the following modules: A model construction module for constructing a discrete-time model of the operation of the power system in the i-th area, and calculating the local tracking error through the discrete-time model according to the control input parameter received by the current power system. The local tracking error is the deviation between the actual area control error and the desired area control error of the system; An unknown part calculation module for constructing an evaluation network and an action network, iteratively updating the parameters of the action network according to the evaluation network, and calculating the unknown part of the discrete-time model through the action network updated by each iteration based on the local tracking error to obtain an unknown part estimation value; A first trigger module for constructing an output trigger condition according to the change amount between the local tracking error and the local tracking error at the previous output trigger moment, and determining whether the current moment satisfies the output trigger condition: if not, keeping the unknown part estimation value and the local tracking error at the previous output trigger moment of the current discrete-time model unchanged; if so, processing the unknown part estimation value through a pre-constructed sliding mode controller to obtain an updated control input parameter; The second trigger module is used to construct an input trigger condition according to the change amount of the control input parameter at different input times, and determine whether the current time meets the input trigger condition: if not, keep the control input parameter received by the power system in the current area unchanged; if so, send the updated control input parameter to the power system in the current area, and the power system in the current area operates according to the updated control input parameter.

[0014] The present invention also provides an electronic device, which includes a processor, a memory, and a bus system. The processor and the memory are connected through the bus system. The memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the data-driven load frequency control method described above.

[0015] The present invention also provides a computer storage medium, which stores a computer software product. The computer software product includes several instructions for causing a computer device to execute the data-driven load frequency control method described above.

[0016] The above technical solutions of the present invention have the following advantages compared with the prior art: The present invention deeply couples the intelligent decision-making efficiency of reinforcement learning, the fast convergence characteristics and strong robustness presented by the fixed-time nonsingular terminal sliding mode control algorithm, and the efficient communication management ability realized by the dual-channel dynamic event triggering strategy to construct a collaborative and efficient integrated control system. This innovative integration not only solves the problems of limited communication resources and uncertain system models, but also significantly improves the system's ability to handle data delay, providing an efficient, reliable and comprehensive solution for the load frequency control of the microgrid system, with extremely high practical value and broad application prospects. Description of the Drawings

[0017] In order to make the content of the present invention easier to be clearly understood, the following further details the present invention according to the specific embodiments of the present invention and in conjunction with the drawings, where Figure 1 is a flowchart of a data-driven load frequency control method provided in an embodiment of the present invention; Figure 2 is a structural diagram of the dynamic model of the power system in the i-th area; Figure 3 is a specific implementation flowchart of the control method of the power system in the i-th area; Figure 4 is a three-area interconnected microgrid system; Figure 5is the triggering interval between the input and output channels under a step - changed load condition, where (a) represents the triggering interval of the input channel and (b) represents the triggering interval of the output channel; Figure 6 is the dynamic response of the load - frequency deviation under different control strategies of the present invention, Control Method A, and Control Method B under a step - changed load condition; Figure 7 is the triggering interval between the input and output channels under a dynamically - changed load condition, where (a) represents the triggering interval of the input channel and (b) represents the triggering interval of the output channel; Figure 8 is the dynamic response of the load - frequency deviation under different control strategies of the present invention, Control Method A, and Control Method B under a dynamically - changed load condition; Figure 9 is a structure diagram of a data - driven load - frequency control system provided in an embodiment of the present invention; Explanation of reference numerals in the accompanying drawings of the specification: 100, model construction module; 200, unknown part calculation module; 300, first triggering module; 400, second triggering module. Detailed implementation manners

[0018] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and be able to implement it, but the embodiments cited are not intended to limit the present invention.

[0019] Embodiment 1: Refer to Figure 1 As shown, the present invention provides a data - driven load - frequency control method, which includes the following steps: Step S1: Construct a discrete - time model for the operation of the power system in the i - th area. According to the control input parameters received by the current power system, calculate the local tracking error through the discrete - time model. The local tracking error is the deviation between the actual area control error and the desired area control error of the system; Step S2: Use the reinforcement learning algorithm to construct an evaluation network and an action network, iteratively update the parameters of the action network according to the evaluation network, and calculate the unknown part of the discrete - time model through the action network updated each time based on the local tracking error to obtain an estimated value of the unknown part; Step S3: Construct an output triggering condition according to the change amount between the local tracking error and the local tracking error at the previous output triggering moment, and determine whether the current moment satisfies the output triggering condition: Otherwise, keep the estimated value of the unknown part and the local tracking error at the last output trigger moment of the current discrete-time model unchanged, and return to continue executing step S1; If so, process the estimated value of the unknown part through a pre-constructed fixed-time nonsingular terminal sliding mode controller to obtain updated control input parameters, and execute step S4; Step S4: Construct an input trigger condition based on the change in the control input parameters at different input moments, and determine whether the current moment satisfies the input trigger condition: Otherwise, keep the control input parameters received by the power system in the current area unchanged, and return to continue executing step S1; If so, send the updated control input parameters to the power system in the current area, and the power system in the current area operates according to the updated control input parameters, and return to continue executing step S1.

[0020] As can be seen from the above technical solutions, based on the fact that there are many factors in the power system that are difficult to accurately model, such as the random fluctuations of loads and the uncertainties of equipment parameters, etc., the present invention uses a reinforcement learning algorithm to construct an action network, calculates the unknown part of the system and obtains an estimated value, and can effectively estimate these unknown parts during the continuous interaction process, thereby making up for the deficiencies of the model and improving the reliability of control.

[0021] In steps S3 and S4, a dual-channel dynamic event-triggering strategy is used to construct the input trigger condition and the output trigger condition. The output trigger condition is constructed based on the change in the local tracking error, avoiding unnecessary updates of the control input parameters. Only when the error change reaches a certain degree, will the update of the estimated value of the unknown part and the adjustment of the control input parameters be triggered, reducing the waste of computing resources and improving the control efficiency. At the same time, when the trigger condition is not met, the previous estimated value and error are kept unchanged, avoiding system fluctuations caused by frequent adjustments and enhancing the stability of the system. The input trigger condition is constructed based on the change in the control input parameters at different input moments, further reducing the number of transmissions of the control input parameters. Only when the change in the control input parameters reaches a certain threshold, will the updated control input parameters be sent to the power system, reducing the communication burden and control cost. This dynamic trigger mechanism can make full use of limited communication resources on the premise of ensuring the control effect and improve the overall performance of the system.

[0022] As Figure 2 shown, the dynamic model of the power system in the i-th area is constructed as: (1) Wherein, represents the load frequency deviation, represents the tie-line power deviation, represents the generator output power deviation, represents the load change, represents the governor valve position deviation, represents the load disturbance and control input; represents the generator inertia, represents the damping coefficient of the generator set, represents the control time constant, represents the synchronous torque coefficient, represents the steam turbine time constant, represents the input time constant, represents the frequency deviation coefficient, represents the actual area control error, and j represents the power system of the jth area, represents the droop characteristic.

[0023] Furthermore, in the actual power system operation scenario, due to the difficulty in obtaining its accurate model, frequent network delays, and limited communication resources, the load frequency control faces great challenges. To address these problems, the present invention innovatively designs a fixed-time nonsingular terminal sliding mode control algorithm based on reinforcement learning. At the same time, a dual-channel dynamic event-triggering strategy is developed, which effectively compensates for the negative impact brought by data delay through accurate estimation of model uncertainty, and then solves the problems of limited communication resources and system model uncertainty. The specific implementation process is shown in Figure 3 .

[0024] Based on Euler's approximation law, the method of converting the dynamic model of the above power system into a discrete-time model is as follows: The discrete-time model of the power system operation in the ith area is expressed as: (2) where T represents the sampling period, h represents the input time delay, , , , are system matrices, represents the state vector, k represents the discrete time, represents the actual area control error at discrete time k, represents the control input parameter at discrete time k, represents the disturbance vector at discrete time k, , in addition, , , , are defined as: , , , , Since the internal stability of the system equilibrium point is equivalent to the stability of the origin, that is, the perturbation vector , the discrete-time model representation described by formula (2) is converted into an input-delay system: (3) For the input-delay system, if there exists a control input at time k, then there exists a bounded input-delay function such that . Substituting it into formula (3), we get: (4) where , are the known parts in the system matrices , respectively, and is the unknown but bounded part in the system matrices , . Formula (4) is further improved into a discrete-time model including the unknown part as follows: (5) where is the unknown part of the model.

[0025] In this embodiment, the calculation method of the defined local tracking error is as follows: (6) where represents the desired regional control error, usually set to 0.

[0026] Furthermore, an evaluation network is constructed, including: S21: Based on the control input parameters received by the power system in the current region and the local tracking error, calculate the instantaneous cost value through the constructed instantaneous cost function: (7) S22: Construct the performance index function for the operation of the power system in the i-th region according to the instantaneous cost value: (8) S23: Simplify to , and convert the performance index function of formula (8) into the Bellman equation: (9) S24: According to the universal approximation property of the neural network, the Bellman equation in formula (9) is converted into a performance metric function for iterative calculation using the neural network: (10) where and are positive constants, and represent the performance metric function, represents the local tracking error, represents the data composed of and ; represents the activation function, represents the weights of the evaluation network, represents the dimension, represents the number of neurons; represents the value of the performance metric obtained by iterative calculation of the evaluation network at discrete time k, is the discount factor.

[0027] The method for iteratively updating the parameters of the action network according to the evaluation network is as follows: S25: Take the performance metric value function of the operation of the power system in the i-th region constructed by the evaluation network as the error function of the action network, and construct the loss function of the action network according to the error function : (11) S26: With the goal of minimizing the loss function value of the action network, adjust the weights of the action network using the following formula: (12) where represents the weights obtained by iterative calculation of the action network at discrete time (k + 1), represents the weights obtained by iterative calculation of the action network at discrete time k, represents the estimated value of the unknown part obtained by iterative calculation of the action network at discrete time k; is the learning rate of the action network, represents the activation function, represents the number of neurons; represents the data composed of the control input parameters and the local tracking error at discrete time k, represents the activation function, represents the weights of the evaluation network, represents the number of neurons.

[0028] Further, the training method of the evaluation network is as follows: Define the time difference error of the evaluation network as: (13) Construct the loss function of the evaluation network based on the time difference error as follows: (14) With the goal of minimizing the loss function value of the evaluation network, adjust the weights of the evaluation network using the following formula: (15) Among them, represents the weight obtained by iterative training of the evaluation network at the discrete time (k + 1), represents the weight obtained by iterative training of the evaluation network at the discrete time k, represents the weight obtained by iterative training of the evaluation network at the discrete time (k - 1), represents the loss function value calculated by the evaluation network at the discrete time k; represents the performance index value obtained by iterative calculation of the evaluation network at the discrete time (k - 1), represents the learning rate of the evaluation network.

[0029] Further, in step S3, construct the output trigger condition according to the change amount between the local tracking error and the local tracking error at the previous output trigger moment, as follows: (16) Among them, represents the output trigger moment of the power system in the i-th area, represents the input trigger moment of the power system in the i-th area, , represents the local tracking error at the discrete time k, represents the local tracking error at the previous output trigger moment of; is a constant greater than zero, is the measurement standard trigger threshold. Among them, when the condition of is satisfied, the system can achieve synchronous triggering of output and input.

[0030] Further, based on the constructed output trigger condition, judge whether the current moment satisfies the output trigger condition: If not, keep the estimated value of the unknown part and the local tracking error at the previous output trigger moment of the current discrete time model unchanged, and return to continue to execute step S1; If so, process the estimated value of the unknown part through a pre-constructed sliding mode controller to obtain updated control input parameters, and execute step S4; the method for obtaining the updated control input parameters is as follows: Calculate the estimated value of the unknown part according to the action network : (17) Construct a fixed-time nonsingular terminal sliding mode controller, and the sliding mode controller processes the estimated value of the unknown part through the following formula , to obtain updated control input parameters: (18) where 、 、 、 are sliding mode controller parameters and satisfy , is a sliding mode controller parameter greater than zero, , , and satisfy , is the set sliding mode surface.

[0031] In step S4, construct an input trigger condition according to the change of the control input parameter at different input times, as follows: (19) where represents the input trigger time of the power system in the i-th region, represents the previous input trigger time of represents the time when the system receives the control input parameter, is the measurement standard trigger threshold, is the trigger coefficient.

[0032] Furthermore, according to the constructed input trigger condition, determine whether the current time satisfies the input trigger condition: If not, keep the control input parameter received by the power system in the current region unchanged, that is , , and return to continue executing step S1; If so, send the updated control input parameter to the power system in the current region, and the power system in the current region operates according to the updated control input parameter, and return to continue executing step S1.

[0033] Next, use a specific experiment to verify the effectiveness of the method described in the present invention, with Figure 4Taking the power system structure shown as an example, the main parameters of the model are designed as follows: , , , , , , , , , , , , .

[0034] The main parameters of the controller are designed as follows: , , , , , , , , , , , , , , , , , , , , .

[0035] In addition, the initial weights of the action network are randomly selected from (0, 1), and the initial weights of the evaluation network are 0. The activation functions of the evaluation network and the action network are respectively selected as and .

[0036] Experiment 1: Set the desired area control error of the power systems in Region 1, Region 2, and Region 3 to , and apply a step-changing load. In Figure 5 (a) and (b), the event-triggering instants and the corresponding triggering intervals are shown, where the height of each point represents the time difference between the current triggering instant and the previous triggering instant. Among a total of 7200 sampling instants, a total of 2819 times are triggered, saving 39.15% of the energy. In Figure 6 , the load frequency deviation Under the action of the control method of the present invention, it converges rapidly to 0, and is compared with the existing data-driven control method A and model-based control method B, demonstrating the advantages of this method in terms of convergence speed and elimination of steady-state error.

[0037] Experiment 2: Set the desired area control error of the power systems in Area 1, Area 2, and Area 3 to , and apply a continuously changing load. The method proposed by the present invention is still effective for the changing load. In Figure 7 (a) and (b), the event-triggering instants and corresponding triggering intervals are shown, where the height of each point represents the time difference between the current triggering instant and the previous triggering instant. Among a total of 7200 sampling instants, a total of 3197 triggers occurred, saving 44.40% of the energy. In Figure 8 , the load frequency deviation Under the action of the control method of the present invention, it converges rapidly to 0, and is compared with the existing data-driven control method A and model-based control method B, demonstrating the advantages of this method in terms of convergence speed and elimination of steady-state error.

[0038] Example 2: Based on the same inventive concept as in Example 1, the present invention also provides a data-driven load frequency control system for implementing the steps of the data-driven load frequency control method described in Example 1. As Figure 9 shown, the data-driven load frequency control system includes the following modules: Model construction module 100, used to construct a discrete-time model of the operation of the power system in the i-th area, and calculate the local tracking error through the discrete-time model according to the control input parameters received by the current power system. The local tracking error is the deviation between the actual output and the desired output of the system; Unknown part calculation module 200, used to construct an evaluation network and an action network, iteratively update the parameters of the action network according to the evaluation network, and calculate the unknown part of the discrete-time model based on the local tracking error through the action network updated each time to obtain an estimated value of the unknown part; First trigger module 300, used to construct an output trigger condition according to the change amount between the local tracking error and the local tracking error at the previous output trigger moment, and judge whether the current moment meets the output trigger condition: if not, keep the estimated value of the unknown part and the local tracking error at the previous output trigger moment of the current discrete-time model unchanged; if so, process the estimated value of the unknown part through a pre-constructed sliding mode controller to obtain updated control input parameters; The second trigger module 400 is configured to construct an input trigger condition according to the change amount of the control input parameter at different input times, and determine whether the current time meets the input trigger condition: if not, keep the control input parameter received by the power system in the current area unchanged; if so, send the updated control input parameter to the power system in the current area, and the power system in the current area operates according to the updated control input parameter.

[0039] A load frequency control system based on data driving proposed in this embodiment is used to implement the foregoing load frequency control method based on data driving. Therefore, the specific implementation manners in the load frequency control system based on data driving can be seen in the embodiment part of the foregoing load frequency control method based on data driving. For example, the model construction module 100, the unknown part calculation module 200, the first trigger module 300, and the second trigger module 400 are respectively used to correspondingly implement steps S1, S2, and S3 in the load frequency control method based on data driving in Embodiment 1. Therefore, its specific implementation manners can be referred to the descriptions of the corresponding individual part embodiments. To avoid redundancy, it will not be elaborated here.

[0040] Embodiment 3:

[0041] The present invention also provides an electronic device, which includes a processor, a memory, and a bus system. The processor and the memory are connected through the bus system. The memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the load frequency control method based on data driving described in Embodiment 1.

[0042] Embodiment 4:

[0043] The present invention also provides a computer storage medium, which stores a computer software product. The computer software product includes several instructions for causing a computer device to execute the load frequency control method based on data driving described in Embodiment 1.

[0044] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0045] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce means for implementing the functions specified in one or more flows and / or blocks in the flowchart. Figure 1 in one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks.

[0046] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means for implementing the functions specified in one or more flows and / or blocks in the flowchart. Figure 1 in one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks.

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows and / or blocks in the flowchart. Figure 1 in one or more flows and / or blocks Figure 1 or means for implementing the functions specified in one or more blocks.

[0048] Obviously, the above embodiments are only examples for clear illustration and are not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or variations can be made based on the above description. It is not necessary and impossible to enumerate all implementation manners here. And the obvious changes or variations derived therefrom are still within the protection scope of the present invention.

Claims

1. A data-driven load frequency control method, characterized in that: The following steps are involved: Step S1: construct a discrete time model of the power system operation in the i-th region, and calculate the local tracking error through the discrete time model according to the control input parameters received by the current power system, wherein the local tracking error is the deviation between the actual regional control error of the system and the expected regional control error; Step S2: constructing an evaluation network and an action network, iteratively updating the parameters of the action network according to the evaluation network, and calculating the unknown part of the discrete time model through the action network after each iterative update based on the local tracking error to obtain an estimated value of the unknown part; Step S3: constructing an output trigger condition according to the change between the local tracking error and the local tracking error at the last output triggering moment, and judging whether the output trigger condition is satisfied at the current moment: If not, keep the unknown part estimate value and local tracking error of the last output triggering moment of the current discrete-time model unchanged, and return to continue executing step S1; If yes, the unknown part estimation value is processed by the sliding mode controller constructed in advance to obtain the updated control input parameter, and step S4 is executed; Step S4: construct an input trigger condition according to the change of the control input parameter at different input times, and determine whether the input trigger condition is met at the current time: If not, keep the control input parameters received by the power system in the current area unchanged, and return to continue executing step S1; If so, the updated control input parameters are sent to the power system in the current area, the power system in the current area operates according to the updated control input parameters, and the process returns to continue executing step S1.

2. The data-driven load frequency control method according to claim 1, characterized in that: In step S1, the method for constructing a discrete time model of the power system operation in the i-th region is as follows: The discrete time model of the power system operation in the i-th region is expressed as: (1) Where T is the sampling period, h is the input time delay, , , , is the system matrix, represents the state vector at discrete time k, represents the actual area control error at discrete time k, represents the input time constant, Indicates the control input parameters, represents the disturbance vector at discrete time k, ,also, , , , is defined as: ， ， ， ， Since the internal stability of the system equilibrium point is equivalent to the origin stability, that is, the disturbance vector , the discrete-time model representation described in formula (1) is converted into an input delay system: (2) For the input delay system, if there is a control input at time k , then there exists a bounded input delay function Make , substituting it into formula (2), we get: (3) in, , The system matrix is , The known part of is the system matrix , The unknown but bounded part of , formula (3) is further improved to a discrete time model including the unknown part, as follows: (4) in, is the unknown part of the model, represents the generator inertia, represents the damping coefficient of the generator set, represents the synchronous torque coefficient, represents the turbine time constant, represents the control time constant, represents the frequency deviation coefficient, j represents the power system of the jth region, Indicates droop characteristics.

3. The data-driven load frequency control method according to claim 2, characterized in that: The local tracking error The calculation method is as follows: (5) in, represents the expected regional control error.

4. The data-driven load frequency control method according to claim 1, characterized in that: The method of constructing the evaluation network is as follows: Based on the control input parameters received by the power system in the current area and the local tracking error, the instantaneous cost value is calculated by constructing an instantaneous cost function: (6) The performance index function of the power system operation in the i-th region is constructed according to the instantaneous cost value: (7) Will Simplified to , convert the performance index function of formula (7) into the Bellman equation: (8) The Bellman equation of formula (8) is converted into a performance indicator function that is iteratively calculated using a neural network: (9) in, , is a normal number, and represents the performance indicator function, represents the local tracking error, Indicated by , The data composed of represents the activation function, represents the weight of the evaluation network, Represents the dimension, represents the number of neurons; represents the performance index value obtained by iterative calculation of the evaluation network at discrete time k, is the discount factor.

5. The data-driven load frequency control method according to claim 4, characterized in that: The method for iteratively updating the parameters of the action network according to the evaluation network is: The temporal difference error of the evaluation network is defined as: (10) The loss function of the evaluation network is constructed based on the time difference error, as follows: (11) With the goal of minimizing the loss function value of the evaluation network, the weight of the evaluation network is adjusted using the following formula: (12) in, represents the weight obtained by iterative training of the evaluation network at discrete time (k+1), represents the weight obtained by iterative training of the evaluation network at discrete time k, represents the weight obtained by iterative training of the evaluation network at discrete time (k-1), represents the loss function value calculated by the evaluation network at discrete time k; It represents the performance index value obtained by iterative calculation of the evaluation network at discrete time (k-1). Represents the learning rate of the evaluation network.

6. The data-driven load frequency control method according to claim 1, characterized in that: The training method of the action network is: The performance index value function of the power system operation in the i-th region constructed by the evaluation network As the error function of the action network , according to the error function Construct the loss function of the action network: (13) With the goal of minimizing the loss function value of the action network, the weight of the action network is adjusted using the following formula: (14) in, represents the weights iteratively calculated by the action network at discrete time (k+1), represents the weight calculated iteratively by the action network at discrete time k, Represents the estimated value of the unknown part obtained by iterative calculation of the action network at discrete time k; is the learning rate of the action network, represents the activation function, Represents the dimension, represents the number of neurons; represents the data consisting of the control input parameters and local tracking error at discrete time k, represents the activation function, Represents the weight of the evaluation network.

7. The data-driven load frequency control method according to claim 6, characterized in that: The method to obtain the updated control input parameters is as follows: Calculate the unknown part estimate based on the action network : (15) Construct a sliding mode controller, which processes the unknown part estimate through the following formula: , and get the updated control input parameters: (16) in, , , , is the sliding mode controller parameter and satisfies , is a sliding mode controller parameter greater than zero, , , and meet , is the set sliding surface, represents the local tracking error.

8. The data-driven load frequency control method according to claim 1, characterized in that: The output trigger condition is constructed according to the change between the local tracking error and the local tracking error at the last output triggering moment, as follows: (17) in, represents the output triggering time of the power system in the ith region, represents the input triggering time of the power system in the ith region, , represents the local tracking error at discrete time k, express The local tracking error at the last output trigger time; is a constant greater than zero, is the measurement standard trigger threshold.

9. The data-driven load frequency control method according to claim 1 or 8, characterized in that: The input trigger condition is constructed according to the change of the control input parameter at different input times, as follows: (18) in, represents the input triggering time of the power system in the ith region, express The last input trigger time of Indicates time The control input parameters received by the system when is the measurement standard trigger threshold, is the trigger coefficient.

10. A data-driven load frequency control system, characterized in that: The method for implementing the data-driven load frequency control method according to any one of claims 1 to 9, wherein the data-driven load frequency control system comprises the following modules: A model building module is used to build a discrete time model of the power system operation in the i-th region, and calculate the local tracking error through the discrete time model according to the control input parameters received by the current power system, wherein the local tracking error is the deviation between the actual regional control error of the system and the expected regional control error; An unknown part calculation module is used to construct an evaluation network and an action network, iteratively update the parameters of the action network according to the evaluation network, and calculate the unknown part of the discrete time model through the action network after each iterative update based on the local tracking error to obtain an estimated value of the unknown part; A first trigger module is used to construct an output trigger condition according to the change between the local tracking error and the local tracking error at the last output trigger time, and judge whether the output trigger condition is met at the current moment: if not, keep the unknown part estimation value and the local tracking error at the last output trigger time of the current discrete time model unchanged; If yes, the unknown part estimate is processed by a sliding mode controller constructed in advance to obtain updated control input parameters; The second trigger module is used to construct an input trigger condition according to the change amount of the control input parameter at different input moments, and judge whether the input trigger condition is met at the current moment: if not, keep the control input parameter received by the power system in the current area unchanged; If so, the updated control input parameters are sent to the power system in the current area, and the power system in the current area operates according to the updated control input parameters.

11. An electronic device, characterized in that: The electronic device includes a processor, a memory and a bus system, the processor and the memory are connected through the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the data-driven load frequency control method described in any one of claims 1 to 9.

12. A computer storage medium, characterized in that: The computer storage medium stores a computer software product, and the computer software product includes a number of instructions for enabling a computer device to execute the data-driven load frequency control method according to any one of claims 1 to 9.

Images