Load frequency control method and system based on multi-region networked multi-source power generation system

Abstract

The invention provides a load frequency control method and system based on a multi-region networked multi-source power generation system, and relates to the technical field of power system automation, and the method comprises the steps: a trigger receives a k moment load frequency control error sent by a current region wind turbine power system, judges whether an event triggering condition is satisfied, and meanwhile, controls the load frequency control error; and the evaluation network receives the load frequency control error at the k moment and the control input and output value function at the # imgabs0 # moment, and updates the network weight. When the event triggering condition is met, the value function output by the evaluation network is sent to the action network, and the control input is updated. According to the method, the load frequency of the wind turbine power system can be controlled without a wind turbine power system model, and the problems that power network communication resources are limited and the wind turbine power system is easily influenced by the external environment are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and in particular to a load frequency control method and system based on a multi-region networked multi-source power generation system. Background Art

[0002] In the operation of the power system, frequency stability plays a key role in system safety. Once the power system becomes unstable, it will not only interfere with the normal power consumption of users and cause damage to system equipment, but in serious cases it may also cause the power grid to collapse, leading to large-scale power outages, causing huge economic losses and adverse effects on society. Load frequency control (LFC), as the core technology of the power system, aims to maintain the stability of the system frequency and ensure that power exchange between regions is carried out in an orderly manner according to the established plan. In a complex power system with multiple regions interconnected and multiple power sources connected, the main goal of load frequency control is to achieve a precise match between power generation and power consumption in each region, while stabilizing the frequency of the entire system within the permitted range.

[0003] With the country's vigorous promotion of clean energy power generation, wind power has become a dominant force in the field of clean energy power generation. However, the structure of wind turbine power systems is complex, and the construction of accurate mathematical models is difficult. In addition, the system frequency is easily affected by wind speed fluctuations, which makes its load frequency control face many challenges. In addition, wind turbine power systems are widely distributed, which significantly increases the cost of data transmission. In view of this, it is urgent to develop a load frequency robust control method suitable for multi-regional interconnected multi-source power generation systems. This method should have the characteristics of not relying on accurate power system models, effectively reducing wind speed interference, and fully considering the limited communication resources of the power network. It has become an important research direction in the current power field. Summary of the Invention

[0004] To this end, an embodiment of the present invention provides a load frequency control method and system based on a multi-region networked multi-source power generation system, which is used to solve the problem in the prior art that it is difficult to obtain an accurate model of the power system, and there are environmental disturbances and limited communication resources, which makes load frequency control in each region very difficult.

[0005] To solve the above problems, an embodiment of the present invention provides a load frequency control method based on a multi-region networked multi-source power generation system, the method comprising:

[0006] Step S1: A trigger receives the load frequency control error at time k sent by the wind turbine power system in the current region, and determines whether the event trigger condition is met based on the load frequency control error and the dynamic event trigger function. Simultaneously, an evaluation network receives the load frequency control error at time k and the control input, updates the weights of the evaluation network, and outputs a value function.

[0007] Step S2: If the event triggering condition is met, the output of the evaluation network is used as the estimated error of the behavior network, the weight of the behavior network is updated, and the control input for the next moment is generated based on the load frequency control error at time k and sent to the actuator; if the triggering condition is not met, the weight of the behavior network and the control input remain unchanged;

[0008] Step S3: the actuator executes the new control input, and the wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger;

[0009] By cyclically executing steps S1-S3, the load frequency control error of each region in the multi-region networked multi-source power generation system is stabilized within a preset range.

[0010] Preferably, the discretized model of the wind turbine power system is:

[0011]

[0012] Where k is the discrete moment, ACE i (k) is the real-time output trajectory, x i (k)= is the system state vector, Δf i is the frequency increment, ΔP tie,i is the pipeline power increment, ΔP mi is the motor output power increment, ΔP gi is the governor valve position increment, Δω ti for is the disturbance vector, ΔP di is the load disturbance increment, Δv i is the wind speed disturbance, N is the total number of regions;

[0013]

[0014] C i =[β i 1 0 0 0];

[0015]

[0016] Where D i is the unit damping coefficient of the generator, M i is the generator moment of inertia, R i is the generator speed reduction coefficient, H ωi is the inertia constant, T gi is the speed regulator time constant, Δω ti is the mechanical speed, kwpi ,k i ,k vwi is the torque coefficient, T ti is the turbine time constant, β i is the frequency offset factor, and T represents the sampling period of the discrete-time control system.

[0017] Preferably, the load frequency control error includes:

[0018] The load frequency control error is defined as:

[0019] e i (k)=ACE i (k)-ACE d (k);

[0020] Where, e i (k) is the load frequency control error, ACE d (i) is the desired output trajectory and is always set to a constant 0, ACE i (k) is the real-time output trajectory;

[0021] The dynamic equation of load frequency control error is:

[0022] e i (k+1)=C i A i x i (i)+C i B i u i (i)+C i D i w i (i) = f i (x i (k),u i (i),w i (k)).

[0023] Preferably, the dynamic event triggering function includes:

[0024] Define the time set when the i-th region meets the event triggering conditions as s means The number of times the event trigger condition is met at a certain moment, s = 0, 1, 2, ..., R, where R represents the number of times the event trigger condition is met;

[0025] The calculation formula of the dynamic event trigger function at time k in the i-th region is:

[0026]

[0027] The dynamic variable θ of the i-th region at time k i (ks ) is calculated as:

[0028]

[0029] The i-th region k s The calculation formula of the event trigger error at time is:

[0030]

[0031] Where, i (k) is the dynamic event triggering function of the i-th region at time k, θ i (k) is the dynamic variable of the i-th region at time k, θ i (k-1) is the dynamic variable of the i-th region at time k-1, is the event triggering error at time k in the i-th region, is the event triggering error of the i-th region at time k-1, The control error setting value of the target area in the i-th area is The difference of load frequency error at each moment, e i (k) is the difference between the target area control error setting value of the i-th area and the load frequency error at time k, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s-th time, α i is the setting constant greater than 0 in the i-th region, μ i is the setting constant greater than 0 in the i-th region, λ i is a set constant greater than 0 for the i-th region, and where f i (x i (k),u i (k),w i (k)) represents the load frequency control error, x i (k),u i (k),w i (k) are the system state vector, control input, and disturbance vector respectively.

[0032] Preferably, the event triggering condition is:

[0033]

[0034] Where, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s+1th time, i (k) is the i-th region k s Dynamic events trigger functions at the moment.

[0035] Preferably, the value function is defined as:

[0036]

[0037] And the optimal value function Satisfies the discrete-time Hamilton-Jacobi-Bellman equation:

[0038]

[0039] Where V i is the value function, e i (k),u i (k),w i (k) are load frequency control error, control input, and disturbance vector, respectively. Q i ,M i ,P i is a set constant greater than 0, and η and γ are set constants greater than 0 and less than 1.

[0040] Preferably, the evaluation network and the behavior network are constructed based on the Weierstrass approximation theorem to estimate the optimal value function and the optimal control input respectively:

[0041]

[0042] Where, is the estimated value of the value function, Indicates the load frequency control error e i (k) and control input u i (k) composed of input data, and Represent the activation function and weight vector of the evaluation network respectively, and Represent the activation function and weight vector of the behavior network, n c and n a Represents the number of neurons.

[0043] Preferably, the error functions of the evaluation network and the behavior network are defined as:

[0044]

[0045] in,

[0046] In order to minimize the error function of the evaluation network and the behavior network, the update function of the weight vector is designed based on the gradient descent rule:

[0047]

[0048]

[0049] Where, ρ ci ,ρ ai is the learning rate.

[0050] An embodiment of the present invention further provides a load frequency control system based on a multi-region networked multi-source power generation system. The system is used to implement the above-mentioned load frequency control method based on a multi-region networked multi-source power generation system, specifically comprising:

[0051] a trigger module, configured to receive a load frequency control error at time k sent by a wind turbine power system in the current region, and determine whether an event trigger condition is met based on the load frequency control error and a dynamic event trigger function;

[0052] Evaluation network module, used to receive the load frequency control error and control input at time k, update the weight of the evaluation network, and output the value function;

[0053] The behavioral network module is used to use the output of the evaluation network as the estimated error of the behavioral network, update the weight of the behavioral network, and generate the control input for the next moment based on the load frequency control error at moment k and send it to the actuator if the event triggering condition is met. If the triggering condition is not met, the behavioral network weight and control input remain unchanged.

[0054] The actuator module is used to execute the new control input. The wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger module.

[0055] An embodiment of the present invention further provides a computer storage medium storing a computer software product, wherein the computer software product includes a number of instructions for enabling a computer device to execute the above-mentioned load frequency control method based on a multi-region networked multi-source power generation system.

[0056] It can be seen from the above technical solutions that the present invention has the following beneficial effects:

[0057] (1) This invention constructs a dual neural network architecture consisting of an evaluation network and a behavior network, dynamically updates network weights using an online learning mechanism (e.g., gradient descent), and generates an optimal control strategy directly based on real-time load frequency error and control input. This overcomes the modeling difficulties associated with wind power generation systems due to their complex structure and environmental disturbances (e.g., sudden changes in wind speed), significantly improving control robustness.

[0058] (2) This invention designs a communication strategy based on a dynamic event-triggered function, triggering control input updates only when the error exceeds a preset threshold, thus reducing unnecessary data transmission. In widely distributed wind power generation systems, this strategy avoids bandwidth pressure caused by continuous high-frequency communication and adapts to scenarios with limited power network communication resources.

[0059] (3) The present invention achieves global frequency stability by coordinating the regional control errors and tie-line power deviations of each region. This solves the frequency instability problem caused by inter-regional power imbalance in multi-source power generation systems and improves the overall reliability of the system under dynamic disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the implementation cases of the present invention or the technical solutions in the prior art, the following is a brief description of the drawings required for use in the embodiments. By referring to the drawings, the features and advantages of the present invention will be more clearly understood. The drawings are schematic and should not be understood as limiting the present invention in any way. Those skilled in the art can derive other drawings based on these drawings without inventive effort. Among them:

[0061] Figure 1 This is a structural diagram of a load frequency control method based on a multi-region networked multi-source power generation system according to the present invention;

[0062] Figure 2 A flow chart of a load frequency control method based on a multi-region networked multi-source power generation system provided by the present invention;

[0063] Figure 3 Schematic diagram of a networked multi-source power generation system consisting of three regions in the present invention;

[0064] Figure 4 A schematic diagram of the power network delay in the networked multi-source power generation system consisting of three regions in the present invention;

[0065] Figure 5 This is a schematic diagram of the event triggering interval under static load in the present invention;

[0066] Figure 6 is a frequency deviation curve under static load in the present invention;

[0067] Figure 7 This is a schematic diagram of event triggering intervals under dynamic load in the present invention;

[0068] Figure 8 This is a frequency deviation curve under dynamic load in the present invention. DETAILED DESCRIPTION

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0070] Example 1

[0071] In order to solve the problem that it is difficult to know the accurate model of the power system in the existing technology, and there are environmental disturbances and limited communication resources, which makes it very difficult to control the load frequency in each region, such as Figure 1 、 2 As shown, the present invention proposes a load frequency control method based on a multi-region networked multi-source power generation system, the method comprising:

[0072] Step S1: A trigger receives the load frequency control error at time k sent by the wind turbine power system in the current region, and determines whether the event trigger condition is met based on the load frequency control error and the dynamic event trigger function. Simultaneously, an evaluation network receives the load frequency control error at time k and the control input, updates the weights of the evaluation network, and outputs a value function.

[0073] Step S2: If the event triggering condition is met, the output of the evaluation network is used as the estimated error of the behavior network, the weight of the behavior network is updated, and the control input for the next moment is generated based on the load frequency control error at time k and sent to the actuator; if the triggering condition is not met, the weight of the behavior network and the control input remain unchanged;

[0074] Step S3: the actuator executes the new control input, and the wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger;

[0075] By cyclically executing steps S1-S3, the load frequency control error of each region in the multi-region networked multi-source power generation system is stabilized within a preset range.

[0076] As can be seen from the above technical solution, the present invention proposes a load frequency control method for a multi-region interconnected multi-source power generation system. A trigger receives the load frequency control error transmitted by the wind turbines in the current region and determines whether the event triggering conditions are met based on established rules. Simultaneously, an evaluation network obtains the load frequency control error and control input and updates the network weights accordingly. If the triggering conditions are met, the evaluation network output serves as the estimated error of the behavior network, which is used to update the action network weights, thereby generating the next control input and transmitting it to the actuator. If the conditions are not met, the behavior network and control input remain unchanged. The actuator receives the new control input and implements the control action. The wind turbines in the current region feed back the load frequency control error at the next moment to the trigger. This method effectively controls the load frequency of a wind power system without requiring knowledge of the power system's internal model. By designing a dynamic event triggering function, it effectively addresses the challenge of limited communication resources in the power network. By leveraging a neural network controller, it effectively addresses external environmental disturbances. This method significantly reduces frequency fluctuations between regions, significantly improving the stability and reliability of the power system.

[0077] In step S1, the trigger receives the load frequency control error at time k sent by the wind turbine power system in the current area, and determines whether the event trigger condition is met based on the load frequency control error and the dynamic event trigger function; at the same time, the evaluation network receives the load frequency control error and control input at time k, updates the weights of the evaluation network, and outputs the value function.

[0078] The present invention models the wind turbine power system in region i as follows:

[0079]

[0080] Where, T ij is the synchronous torque coefficient between area i and area j, D i is the unit damping coefficient of the generator, M i is the generator moment of inertia, R i is the generator speed reduction coefficient, H ωi is the inertia constant, T gi is the speed regulator time constant, Δω ti is the mechanical speed, k wpi ,κ i ,k vwi is the torque coefficient, T ti is the turbine time constant, β i is the frequency offset factor, Δf i is the frequency increment, ΔP tie,i is the pipeline power increment, ΔP mi is the motor output power increment, ΔP giis the governor valve position increment, ΔP ci is the control input increment, ΔP di is the load disturbance increment, Δv i is the wind speed disturbance, and N is the total number of regions.

[0081] Furthermore, based on Euler's approximation law, the discrete system of the wind turbine power system is:

[0082]

[0083] Where k is the discrete moment, ACE i (k) is the real-time output trajectory, is the system state vector, u i (k) = ΔP ci (k) is the control input, P ci (k) is the control input increment, is the perturbation vector;

[0084] H i =[0 0 0 1 / T gi 0] T , C i =[β i 1 0 0 0];

[0085]

[0086] Where T represents the sampling period of the discrete-time control system.

[0087] Furthermore, the load frequency control error of the present invention includes:

[0088] The load frequency control error is defined as:

[0089] e i (k)=ACE i (k)-ACE d (k);

[0090] Where, e i (k) is the load frequency control error, ACE d (k) is the desired output trajectory and is always set to a constant 0, ACE i (k) is the real-time output trajectory;

[0091] The dynamic equation of load frequency control error is:

[0092] e i (k+1)=C i A ix i (k)+C i B i u i (k)+C i D i w i (k) = f i (x i (k),u i (k),w i (k)).

[0093] Furthermore, the dynamic event triggering function of the present invention includes:

[0094] Define the time set when the i-th region meets the event triggering conditions as s means The number of times the event trigger condition is met at a certain moment, s = 0, 1, 2, ..., R, where R represents the number of times the event trigger condition is met;

[0095] The calculation formula of the dynamic event trigger function at time k in the i-th region is:

[0096]

[0097] The dynamic variable θ of the i-th region at time k i (k s ) is calculated as:

[0098]

[0099] The i-th region k s The calculation formula of the event trigger error at time is:

[0100]

[0101] Where, i (k) is the dynamic event triggering function of the i-th region at time k, θ i (k) is the dynamic variable of the i-th region at time k, θ i (k-1) is the dynamic variable of the i-th region at time k-1, is the event triggering error at time k in the i-th region, is the event triggering error of the i-th region at time k-1, The control error setting value of the target area in the i-th area is The difference of load frequency error at each moment, e i (k) is the difference between the target area control error setting value of the i-th area and the load frequency error at time k, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s-th time, α iis the setting constant greater than 0 in the i-th region, μ i is the setting constant greater than 0 in the i-th region, λ i is a set constant greater than 0 for the i-th region, and where f i (x i (k),u i (k),w i (k)) represents the load frequency control error, x i (k),u i (k),w i (k) are the system state vector, control input, and disturbance vector respectively.

[0102] Furthermore, the event triggering conditions of the present invention are:

[0103]

[0104] Where, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s+1th time, i (k) is the i-th region k s Dynamic events trigger functions at the moment.

[0105] Furthermore, the value function is defined as:

[0106]

[0107] in Q i ,M i ,P i is a set constant greater than 0, and η and γ are set constants greater than 0 and less than 1.

[0108] The value function can be rewritten as the following Bellman equation:

[0109]

[0110] Once the control input and disturbance have achieved their respective objectives, the optimal value function The following discrete-time Hamilton-Jacobi-Bellman equation is satisfied:

[0111]

[0112] The optimal control input for each region i is expressed as:

[0113]

[0114] Furthermore, according to the Weierstrass approximation theorem, the present invention constructs an evaluation (neural) network and a behavior (neural) network to estimate the optimal value function and the optimal control input respectively:

[0115]

[0116] Where, is the estimated value of the value function, Indicates the load frequency control error e i (k) and control input u i (k) composed of input data, and Represent the activation function and weight vector of the evaluation (neural) network, and Represent the activation function and weight vector of the behavioral (neural) network, n c and n a Represents the number of neurons.

[0117] Furthermore, the error functions of the evaluation network and the behavior network of the present invention are defined as:

[0118]

[0119] in,

[0120] In order to minimize the error function of the evaluation network and the behavior network, the update function of the weight vector is designed based on the gradient descent rule:

[0121]

[0122]

[0123] Where, ρ ci ,ρ ai is the learning rate.

[0124] In step S2, if the event triggering condition is met, the output of the evaluation network is used as the estimated error of the behavioral network, the weight of the behavioral network is updated, and the control input at the next moment is generated based on the load frequency control error at moment k and sent to the actuator; if the triggering condition is not met, the behavioral network weight and control input remain unchanged.

[0125] Specifically, the update function of the weight vector of the network under event triggering is:

[0126]

[0127] In step S3, the actuator implements the new control input, and the wind turbine power system in the current region feeds back the load frequency control error at time k+1 to the trigger. By repeatedly executing steps S1-S3, the present invention stabilizes the load frequency control error in each region of the multi-region interconnected multi-source power generation system within a preset range (centered around zero).

[0128] The advantages of the present invention are described below with reference to specific cases. Figure 3 As shown, a networked multi-source power generation system consisting of three regions is constructed, and its system model is consistent with that of Example 1. The load frequency control method based on a multi-region networked multi-source power generation system proposed in the present invention is used to perform load frequency control operations for static loads and dynamic loads respectively.

[0129] Set the target area control error to ACE d = 0, and impose a static load of 0.02 pu and a wind speed disturbance Δv in each area i (k) = 0.01. Figure 5 A diagram of the event triggering interval under static load is presented. This diagram clearly shows the event triggering instants and the corresponding triggering intervals. The height of each dot in the diagram represents the time difference between the current triggering instant and the previous triggering instant. In 1200 operations, the three regions were triggered 223, 214, and 157 times, respectively. Figure 6 This is the load increment curve under static load. According to this figure, the system load frequency increment curve reaches a convergence state at t=10s.

[0130] Set the target area control error to ACE again d = 0, and apply Figure 4 The dynamic load and wind speed disturbance shown in FIG. 4 are shown in FIG. 4 . The load frequency control method based on the multi-region networked multi-source power generation system proposed in the present invention is also effective for dynamic loads. Figure 7 This diagram shows the event triggering interval under dynamic load, showing the event triggering moment and the corresponding triggering interval. The meaning of the point height is consistent with the static load case. In 1000 operations, the three areas were triggered 413 times, 268 times, and 305 times, respectively. Figure 8 The load increment curve under dynamic load shows the load frequency increment fluctuation curve. Figure 8 It can be determined that the system load frequency increment curve converges at t=20s.

[0131] Based on the above-mentioned load frequency control experimental results for static loads and dynamic loads, it can be clearly seen that the load frequency control method for the multi-region networked multi-source power generation system provided by the present invention is effective for both static loads and dynamic loads, and can significantly reduce energy consumption in practical applications.

[0132] Example 2

[0133] The present invention provides a load frequency control system based on a multi-region networked multi-source power generation system. The system is used to implement the load frequency control method based on a multi-region networked multi-source power generation system according to the first embodiment, and specifically includes:

[0134] a trigger module, configured to receive a load frequency control error at time k sent by a wind turbine power system in the current region, and determine whether an event trigger condition is met based on the load frequency control error and a dynamic event trigger function;

[0135] Evaluation network module, used to receive the load frequency control error and control input at time k, update the weight of the evaluation network, and output the value function;

[0136] The behavioral network module is used to use the output of the evaluation network as the estimated error of the behavioral network, update the weight of the behavioral network, and generate the control input for the next moment based on the load frequency control error at moment k and send it to the actuator if the event triggering condition is met. If the triggering condition is not met, the behavioral network weight and control input remain unchanged.

[0137] The actuator module is used to execute the new control input. The wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger module.

[0138] A load frequency control system based on a multi-regional networked multi-source power generation system in this embodiment is used to implement the aforementioned load frequency control method based on a multi-regional networked multi-source power generation system. Therefore, the specific implementation method of the load frequency control system based on a multi-regional networked multi-source power generation system can be found in the embodiment section of the load frequency control method based on a multi-regional networked multi-source power generation system above. In order to avoid redundancy, it will not be repeated here.

[0139] Example 3

[0140] An embodiment of the present invention provides a computer storage medium storing a computer software product. The computer software product includes several instructions for enabling a computer device to execute the above-mentioned load frequency control method based on a multi-region networked multi-source power generation system.

[0141] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. 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 magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0142] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0143] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0144] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications derived therefrom remain within the scope of protection of the present invention.

Claims

1. A load frequency control method based on a multi-region networked multi-source power generation system, characterized in that: include: Step S1: A trigger receives the load frequency control error at time k sent by the wind turbine power system in the current region, and determines whether the event trigger condition is met based on the load frequency control error and the dynamic event trigger function. Simultaneously, an evaluation network receives the load frequency control error at time k and the control input, updates the weights of the evaluation network, and outputs a value function. Step S2: If the event triggering condition is met, the output of the evaluation network is used as the estimated error of the behavior network, the weight of the behavior network is updated, and the control input at the next moment is generated according to the load frequency control error at time k and sent to the actuator; If the trigger condition is not met, the behavior network weights and control inputs remain unchanged; Step S3: the actuator executes the new control input, and the wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger; By cyclically executing steps S1-S3, the load frequency control error of each region in the multi-region networked multi-source power generation system is stabilized within a preset range.

2. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 1, characterized in that: The discretized model of the wind turbine power system is: Where k is the discrete moment, ACE i (k) is the real-time output trajectory, x i (k)=[Δf i ΔP tie,i ΔP mi ΔP gi Δω ti ] T is the system state vector, Δf i is the frequency increment, ΔP tie,i is the pipeline power increment, ΔP mi is the motor output power increment, ΔP gi is the governor valve position increment, Δω ti is the mechanical speed, u i (k)=ΔP ci (k) is the control input, P ci (i) is the control input increment, is the disturbance vector, ΔP di is the load disturbance increment, Δv i is the wind speed disturbance, N is the total number of regions; C i ＝[β i 1 0 0 0]； Where D i is the unit damping coefficient of the generator, M i is the generator moment of inertia, R i is the generator speed reduction coefficient, H ωi is the inertia constant, T gi is the speed regulator time constant, Δω ti is the mechanical speed, k wpi ,κ i ,k vwi is the torque coefficient, T ti is the turbine time constant, β i is the frequency offset factor, and T represents the sampling period of the discrete-time control system.

3. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 2, characterized in that: The load frequency control error includes: The load frequency control error is defined as: e i (i)=ACE i (i)-ACE d (k); Where, e i (k) is the load frequency control error, ACE d (k) is the desired output trajectory and is always set to a constant 0, ACE i (k) is the real-time output trajectory; The dynamic equation of load frequency control error is: e i (k+1)=C i A i x i (i)+C i B i u i (i)+C i D i w i (i)=f i (x i (k),u i (i),w i (k))。 4. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 1, characterized in that: The dynamic event triggering function includes: Define the time set when the i-th region meets the event triggering conditions as s means The number of times the event trigger condition is met at a certain moment, s = 0, 1, 2, ..., R, where R represents the number of times the event trigger condition is met; The calculation formula of the dynamic event trigger function at time k in the i-th region is: The dynamic variable θ of the i-th region at time k i (k s ) is calculated as: The i-th region k s The calculation formula of the event trigger error at time is: Where, i (k) is the dynamic event triggering function of the i-th region at time k, θ i (k) is the dynamic variable of the i-th region at time k, θ i (k-1) is the dynamic variable of the i-th region at time k-1, is the event triggering error at time k in the i-th region, is the event triggering error of the i-th region at time k-1, The control error setting value of the target area in the i-th area is The difference of load frequency error at each moment, e i (k) is the difference between the target area control error setting value of the i-th area and the load frequency error at time k, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s-th time, α i is the setting constant greater than 0 in the i-th region, μ i is the setting constant greater than 0 in the i-th region, λ i is a set constant greater than 0 for the i-th region, and where f i (x i (k),u i (k),w i (k)) represents the load frequency control error, x i (k),u i (k),w i (k) are the system state vector, control input, and disturbance vector respectively.

5. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 4, characterized in that: The event triggering conditions are: Where, Indicates the corresponding moment when the i-th region meets the event triggering condition for the s+1th time, i (k) is the i-th region k s Dynamic events trigger functions at the moment.

6. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 1, characterized in that: The value function is defined as: And the optimal value function Satisfies the discrete-time Hamilton-Jacobi-Bellman equation: Where V i is the value function, e i (k),u i (k),w i (k) are load frequency control error, control input, and disturbance vector, respectively. Q i ,M i ,P i is a set constant greater than 0, and η, γ are set constants greater than 0 and less than 1.

7. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 6, characterized in that: The evaluation network and behavior network are constructed based on the Weierstrass approximation theorem to estimate the optimal value function and the optimal control input respectively: Where, is the estimated value of the value function, Indicates the load frequency control error e i (k) and control input u i (k) composed of input data, and Represent the activation function and weight vector of the evaluation network respectively, and Represent the activation function and weight vector of the behavior network, n c and n a Represents the number of neurons.

8. The load frequency control method based on a multi-region networked multi-source power generation system according to claim 7, characterized in that: The error functions of the evaluation network and the behavior network are defined as: in, In order to minimize the error function of the evaluation network and the behavior network, the update function of the weight vector is designed based on the gradient descent rule: Where, ρ ci ,ρ ai is the learning rate.

9. A load frequency control system based on a multi-region networked multi-source power generation system, characterized in that: The system is used to implement the load frequency control method based on a multi-region networked multi-source power generation system according to any one of claims 1 to 8, specifically comprising: a trigger module, configured to receive a load frequency control error at time k sent by a wind turbine power system in the current region, and determine whether an event trigger condition is met based on the load frequency control error and a dynamic event trigger function; Evaluation network module, used to receive the load frequency control error and control input at time k, update the weight of the evaluation network, and output the value function; The behavioral network module is used to use the output of the evaluation network as the estimated error of the behavioral network, update the weight of the behavioral network, and generate the control input for the next moment based on the load frequency control error at moment k and send it to the actuator if the event triggering condition is met. If the triggering condition is not met, the behavioral network weight and control input remain unchanged. The actuator module is used to execute the new control input. The wind turbine power system in the current area feeds back the load frequency control error at time k+1 to the trigger module.

10. A computer storage medium, characterized in that The computer storage medium stores a computer software product, which includes several instructions for enabling a computer device to execute the load frequency control method based on a multi-region networked multi-source power generation system according to any one of claims 1 to 8.

Images