Self-adaptive intelligent control method and equipment for direct current transmission fault ride-through of wind power plant

Abstract

The invention discloses a self-adaptive intelligent control method and device for wind power plant DC power transmission fault ride-through, and the method comprises the steps: constructing a wind power plant AC-DC power transmission system integrated model which comprises a wind power plant and a modular multilevel converter; constructing an artificial neural network model, taking the system real-time state vector as input, and taking the decision result vector as output; the action time interval of the control strategy is set as the direct current transmission fault ride-through time, the control strategy is applied to the wind power plant alternating current-direct current transmission system integrated model, and parameters of the artificial neural network model are updated; and deploying the parameter-updated artificial neural network model in a fault ride-through occasion. The method can solve the problems of system instability and equipment off-grid caused by power imbalance and overvoltage in the DC transmission fault period of the existing wind power plant, can be applied to the fault occasion of the flexible DC transmission system with wind power access, automatically detects the fault, guides the sending end converter station of the wind power plant to carry out fault ride-through, and reduces the equipment and economic loss.

Description

Technical Field

[0001] This invention belongs to the field of power systems and their automation, and relates to fault ride-through technology for DC transmission in wind farms, and particularly to an adaptive intelligent control method and device for fault ride-through of DC transmission in wind farms. Background Technology

[0002] With the large-scale development of offshore wind power and large-scale wind power bases, high-voltage direct current (HVDC) transmission technology based on voltage source converters has become the mainstream solution for wind farm grid connection. However, when a short-circuit fault occurs in the DC line, the system voltage drops sharply, and the huge fault current may lead to severe power imbalance, thereby damaging the converter equipment and even causing the entire wind farm to disconnect from the grid, seriously threatening the safe operation of the power system. Therefore, fault ride-through technology requires that when a specific fault occurs on the DC side, the wind farm not only can operate continuously without disconnecting from the grid, but also needs to provide reactive power to the grid to support voltage recovery. Current research directions mainly include improving converter topology, designing DC circuit breakers, and optimizing coordinated control strategies to ensure the stable operation of the wind power system. However, both improving the hardware topology and adding circuit breakers significantly increase system cost and complexity. In contrast, optimizing coordinated control strategies does not require changing the main circuit hardware structure. By fully exploring and improving the response and collaborative potential of the existing control resources of the system, it provides a solution for fault ride-through that is both economical and adaptable.

[0003] However, existing DC fault ride-through technologies based on optimized coordinated control strategies mainly rely on fixed threshold criteria and preset control sequences. These methods exhibit significant limitations when facing complex operating conditions such as different fault impedances, fault locations, and variable wind power output: (1) Fixed thresholds are difficult to accurately capture all fault characteristics, which may lead to misjudgment or missed judgment; (2) Preset rigid control sequences lack online adjustment capabilities, and control performance deteriorates significantly when dealing with unexpected fault scenarios, making it difficult to meet the higher requirements of adaptability and reliability of fault ride-through control for power systems connected to large-scale wind farms. Therefore, it is urgent to design an adaptive intelligent control method for DC transmission fault ride-through in large-scale wind farms that balances effectiveness and robustness, and has continuous optimization, autonomous decision-making, and multi-objective collaborative capabilities. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive intelligent control method and device for DC transmission fault ride-through in wind farms. This method can solve the problems of system instability and equipment disconnection caused by power imbalance and overvoltage during DC transmission faults in existing wind farms. It can be applied to fault situations in flexible DC transmission systems connected to wind power, automatically detect faults and guide the sending-end converter station of the wind farm to perform fault ride-through, thereby reducing equipment and economic losses.

[0005] To achieve the above objectives, the solution of the present invention is:

[0006] An adaptive intelligent control method for DC transmission fault ride-through in wind farms includes the following steps:

[0007] Step 1: Construct an integrated AC-DC transmission system model for the wind farm. This model includes the wind farm and a modular multilevel converter. The output end of the wind farm is connected to the AC bus through a full-power converter, the input end of the wind farm is connected to the AC side of the modular multilevel converter, and the DC side of the modular multilevel converter is connected to the DC transmission line.

[0008] Step two: Construct an artificial neural network model, which uses the system's real-time state vector As input, with the decision result vector The output is the artificial neural network model, which consists of computational neurons, and the weights of each computational neuron constitute the parameter matrix of the artificial neural network model. ;

[0009] Step 3: Set the effective time interval of the control strategy to the DC transmission fault ride-through time. Set the initial time for the control strategy to take effect. The control strategy is applied to the integrated model of the AC-DC transmission system of the wind farm, and the parameters of the artificial neural network model are updated based on the results.

[0010] Step four: Deploy the updated artificial neural network model in a fault-crossing scenario.

[0011] In step one above, the instantaneous active power injected into the AC bus by the wind farm In the synchronously rotating dq coordinate system, it has the following form.

[0012] ,

[0013] in, These are the d-axis and q-axis components of the AC bus voltage, respectively. These are the d-axis and q-axis components of the fan output current, respectively.

[0014] The DC voltage at the sending end of the modular multilevel converter DC active power They have the following relationship:

[0015] ,

[0016] in, This represents the DC line current.

[0017] In step one above, an equivalent aggregation model of the wind farm is constructed, aggregating all wind turbines in the farm into equivalent single-unit models. The wind turbines convert kinetic energy in the air into mechanical energy, and their output mechanical power... It is determined by the following wind energy capture formula,

[0018] ,

[0019] in, air density, The radius of the wind turbine blades. Wind speed; The wind energy utilization coefficient is the tip speed ratio. and pitch angle The function.

[0020] In step two above, electrical data of the system operation are collected in real time to construct the system's real-time state vector. as follows,

[0021] ,

[0022] in, These are the AC busbars at the wind farm outlet. , , Three-phase voltage per unit value This refers to the per-unit value of the DC output voltage of the modular multilevel converter. These are the per-unit values ​​of active and reactive power output from the wind farm, respectively. These are the per-unit values ​​of active and reactive power transmitted by a DC transmission system.

[0023] In step two above, the decision result vector Defined as,

[0024] ,

[0025] in, This is a reference value for the power transfer of the converter station. The number of AC power-consuming resistors used. The grid connection time of the AC power-consuming resistor.

[0026] In step three above, the DC transmission fault ride-through time is determined according to the following method:

[0027] Real-time monitoring of DC side voltage ,when Record the time when the following fault occurrence criteria are triggered. The fault start time.

[0028] ,

[0029] in, DC rated voltage, This is the voltage drop factor;

[0030] When DC side voltage If the following fault termination criteria are triggered, record the time. This is the fault end time.

[0031] ,

[0032] in, The voltage recovery coefficient, This is the minimum holding time for voltage stabilization.

[0033] The specific process of step three above is as follows:

[0034] Step 31, set the parameter matrix of the artificial neural network model as follows: ,in, The parameter matrix is The Middle OK Column parameters, These represent the depth and width of the artificial neural network model, respectively. , Integer and , ;

[0035] Step 32, construct the total An independent integrated AC-DC power transmission model for a wind farm, in which To determine the number of available AC-DC transmission models for wind farms, obtain the initial values ​​for each model. and ;

[0036] Step 33, the parameter matrix of the artificial neural network model copy Part, marked as,

[0037] ,

[0038] For each Construct the following adaptive normal distribution parameter generator:

[0039] ,

[0040] in, With a mean of 0 and a standard deviation of 0 Controllable normal distribution Generate parameters for adaptive normal distribution. , It serves as a generator for both forward and reverse normal distribution parameters;

[0041] Using the forward and backward model parameters generated by the aforementioned adaptive normal distribution parameter generator, calculate the forward and backward decision result vectors. and as follows,

[0042] ,

[0043] in, The state mean vector, Let be the state standard deviation vector. The output of the artificial neural network model is the parameter of the positive model. The output of the artificial neural network model for the inverse model parameters;

[0044] Total A vector of decision outcomes and The data are fed into the integrated AC-DC transmission system model of the wind farm, the time is updated, and the real-time state vector of the system is remeasured. and And calculate accordingly and Update the decision result vector and Repeated updates , , , , , until Calculate the fault isolation objective function of the artificial neural network model. and Update the parameter matrix of the artificial neural network model and real-time failure time as follows,

[0045] ,

[0046] Update the system's real-time state vector and mean vector and standard deviation vector as follows,

[0047] ,

[0048] Update the system's real-time state vector and The standard deviation is as follows:

[0049] ,

[0050] Step 34: Repeat step 33 until the requirements for DC transmission fault ride-through of the wind farm under the specified conditions are met.

[0051] In step three above, the fault isolation objective function of the artificial neural network model The objective function is defined as follows for each time point within the fault period: The sum,

[0052] ,

[0053] Wherein, the objective function at each time point Defined as,

[0054] ,

[0055] in, For the AC output power of the modular multilevel converter, For the reference power of the modular multilevel converter, This refers to the voltage to ground at the DC output of the modular multilevel converter. This is the reference voltage for the DC output of the modular multilevel converter.

[0056] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor; the processor executes the computer program to implement the steps of the adaptive intelligent control method for DC transmission fault ride-through in wind farms as described above.

[0057] A computer-readable storage medium storing a computer program; when executed by a processor, the computer program implements the steps of the adaptive intelligent control method for DC transmission fault ride-through in wind farms as described above.

[0058] After adopting the above scheme, this invention addresses the limitations of traditional DC transmission fault ride-through methods based on fixed thresholds and preset control sequences, which suffer from poor adaptability and difficulty in handling complex and variable fault conditions. These limitations prevent direct application to DC transmission fault ride-through scenarios involving a high proportion of large-scale wind power, the focus of this invention. Therefore, this invention proposes an adaptive intelligent control method. This method interacts with the environment and makes online decisions through an artificial neural network strategy model. The designed artificial neural network strategy model is fully trained, and its deep network structure can extract AC-DC transmission fault characteristics and map them to generate optimized coordinated control commands. This invention has the advantages of autonomously sensing system status, real-time adaptive adjustment of control strategies, and no need for precise mathematical models, thus improving the robustness and reliability of the system under various fault types and different operating conditions. Attached Figure Description

[0059] Figure 1 This is a general structural diagram of an adaptive intelligent control method for DC transmission fault ride-through in wind farms proposed in this invention. The diagram includes: 1-Integrated model of AC-DC transmission system of wind farm; 2-Equivalent aggregation model of wind farm; 3-Wind power; 4-Energy-consuming resistor; 5-Modular multilevel converter; 6-Metallic loop; 7-DC transmission line; 8-Subsequent DC transmission and converter section; 9-State vector; 10-Adaptive and intelligent process of control method; 11-Artificial neural network model; 12-Action vector.

[0060] Figure 2 yes Figure 1 10 - A detailed explanation of the adaptive and intelligent process of the control method, wherein 101 - the parameter matrix of the current artificial neural network model, 102 - a set of parameters randomly generated by adaptive normal distribution, 103 - a set of parameters randomly generated by adaptive normal distribution, 104 - a set of parameters randomly generated by adaptive normal distribution, 105 - the update process of the parameter matrix of the artificial neural network model, and 106 - the parameter matrix of the artificial neural network model after a single update;

[0061] Figure 3 This is a schematic diagram of power droop control. Detailed Implementation

[0062] The technical solution and beneficial effects of the present invention will be described in detail below with reference to the accompanying drawings.

[0063] like Figure 1 As shown, the adaptive intelligent control method for DC transmission fault ride-through in wind farms proposed in this invention should follow these steps in its implementation:

[0064] (1) Establish an integrated model of the AC-DC transmission system of the wind farm; specifically, the model is organically coupled from three parts: wind farm, modular multilevel converter and system network;

[0065] An equivalent aggregation model of the wind farm is established, aggregating all wind turbines in the field into equivalent single-unit models. The wind turbines convert kinetic energy in the air into mechanical energy, and their output mechanical power... Determined by the following wind energy capture formula

[0066]

[0067] in, air density, The radius of the wind turbine blades. Wind speed; The wind energy utilization coefficient is the tip speed ratio. and pitch angle The function;

[0068] The equivalent aggregation model of the wind farm connects to the AC bus via a full-power converter. The grid-side converter employs grid voltage-oriented vector control. A mathematical model of the full-power converter is established, and its output voltage equation in the synchronously rotating dq coordinate system has the following form.

[0069]

[0070] in, , , , These are the d-axis and q-axis components of the AC bus voltage and the wind turbine output current, respectively. , For the filter resistors and inductors of the grid-side converter, The angular frequency of the power grid. , These are the d-axis and q-axis components of the voltage at the point of common coupling.

[0071] Three-phase AC voltage output from the wind farm The instantaneous active power injected into the AC system is determined by the modulation wave of the modular multilevel converter and the voltage of the DC transmission line. In the synchronously rotating dq coordinate system, it has the following form.

[0072]

[0073] The DC-side dynamics of the modular multilevel converter are dominated by the DC line equations, and its sending-end DC voltage... DC active power It has the following relationship

[0074]

[0075] Simultaneously, the modular multilevel converter exchanges reactive power with the AC system during operation, and the reactive power... This refers to the reactive component defined in the state vector.

[0076] (2) Establish a DC transmission fault ride-through model; specifically, establish a dynamic model that includes active power balancing measures. The core of this model is to characterize how to quickly consume excess power and coordinate the system operating point during a fault. Energy-consuming devices are configured on the DC side of the sending-end converter station. The model needs to accurately simulate how, within milliseconds of detecting a fault, it puts in the corresponding number and power level of resistors according to adaptive control commands to quickly consume the excess active power that the wind farm cannot send to the grid;

[0077] Simultaneously, the model integrates voltage-power droop control to adjust the wind farm output, such as... Figure 3 As shown, under the condition of keeping the frequency constant, the correction amount of its active power reference value is as follows:

[0078]

[0079] in, For grid-side AC power, The reference frequency for the grid-side voltage. The reference frequency for the AC power grid. For drooping power, This is the droop coefficient.

[0080] (3) Real-time perception of system operation status; Specifically, this invention forms a distributed sensor network by deploying synchronous phasor measurement units at key nodes and a high-speed data acquisition system to achieve wide-area synchronous monitoring of the power grid status; The above-mentioned measurement units are installed at the wind farm outlet bus, the AC side of the sending-end converter station, and the DC outlet of the sending-end converter station, respectively. They receive synchronous clock signals through the built-in navigation and positioning system to ensure that the data of all measurement points have a unified time scale, and the sampling rate should meet the requirements of transient process analysis; The raw physical quantities collected in real time by the measurement units include: the three-phase instantaneous voltage of the AC bus at the wind farm outlet. With three-phase instantaneous current The voltage to ground at the DC output of the sending-end converter station DC line current All the above quantities are transmitted to the central processing unit in real time via the substation network after being collected.

[0081] (4) Construct the AC-DC transmission operation status of the wind farm; specifically, the data collected and transmitted in step (3) is subjected to the following per-unit processing:

[0082]

[0083]

[0084] in, This represents the per-unit value of the three-phase AC voltage. This represents the per-unit value of the three-phase alternating current. This is the reference value for three-phase AC voltage. This is the reference value for three-phase alternating current;

[0085] Subsequently, calculations were performed based on the standardized instantaneous values ​​and the model in step (1). The following system real-time state vector is constructed for subsequent fault detection and intelligent control.

[0086]

[0087] (5) Detect DC transmission faults and define the fault time interval; specifically, based on the system real-time state vector described in step (4). Real-time detection, type identification, and precise time interval definition of transient faults in DC transmission systems are performed, and the ground voltage at the DC outlet of the sending-end converter station is continuously monitored. Dynamic behavior, when Record the time when the following fault occurrence criteria are triggered. Fault start time

[0088]

[0089] in DC rated voltage, The voltage sag factor is usually set to 0.

[0090]

[0091] The system enters the fault crossing process and continuously monitors the real-time system state vector described in step (4). Until the system monitors the ground voltage at the DC outlet of the sending-end converter station. If the following fault termination criteria are triggered, the fault disturbance is determined to have ended.

[0092]

[0093] in The voltage recovery coefficient is typically set to...

[0094]

[0095] This is the minimum holding time for voltage stability, used to prevent misjudgments caused by voltage oscillations; this time is recorded. The fault ride-through time interval for DC transmission is defined as the fault end time. .

[0096] (6) Constructing a control strategy framework; specifically, the control strategy is based on an artificial neural network model. The system consists of three parts: a fault isolation objective function, a closed-loop time series model, and the artificial neural network model based on the system's real-time state vector. As input, with the decision result vector For output

[0097]

[0098] The decision result vector Defined as

[0099]

[0100] in This is a reference value for the power transfer of the converter station. The number of AC power-consuming resistors used. For the grid connection time of the AC power-consuming resistor, the artificial neural network model is composed of computational neurons, and the weights of each computational neuron constitute the parameter matrix of the artificial neural network model. The fault isolation objective function The objective function is defined as follows for each time point within the fault period: The sum

[0101]

[0102] Wherein, the objective function at each time point Defined as

[0103]

[0104] The closed-loop time series model is based on discrete time series, with artificial neural network model and wind farm AC-DC transmission model described in step (1) as the main components, and system real-time state vector. Decision outcome vector and objective function at each time step For input / output framework;

[0105] (7) Initialize the control strategy; specifically, set the effective time interval of the control strategy to the DC transmission fault ride-through time. Set the initial time for the control strategy to take effect. The parameter matrix of the artificial neural network model in step (6) is set as follows: ,in The parameter matrix is The Middle OK Column parameters, , The depth and width of the artificial neural network model are given. , Integer and , , construct total Each independent AC-DC transmission model of the wind farm is used to measure and calculate the initial values ​​of each model based on the situational awareness and state construction of the wind farm DC transmission system described in steps (3) and (4). and ;

[0106] (8) Train the control strategy to make it adaptive and intelligent, such as Figure 2 As shown; specifically, the artificial neural network model parameter matrix described in step (7) copy copies, marked as

[0107]

[0108] Based on this, for each Construct the following adaptive normal distribution parameter generator.

[0109]

[0110] in With a mean of 0 and a standard deviation of 0, Controllable normal distribution Generate parameters for adaptive normal distribution. , For generating parameters of the forward and reverse normal distributions,

[0111] Using the forward and backward model parameters generated by the aforementioned adaptive normal distribution parameter generator, calculate the forward and backward decision result vectors. and as follows

[0112]

[0113] in, The state mean vector, The standard deviation vector of the state;

[0114] Total A vector of decision outcomes and The data are fed into the AC-DC transmission model of the wind farm constructed in step (1) to update the time and remeasure the real-time state vector of the system. and And calculate accordingly and Update the decision result vector and Repeated updates , , , , , until The fault isolation objective function constructed in step (6) is calculated. and Update the parameter matrix of the artificial neural network model and real-time failure time as follows

[0115]

[0116] Update the system real-time state vector and mean vector and standard deviation vector as follows

[0117]

[0118] Update the system real-time state vector and The standard deviation is as follows

[0119]

[0120] Repeat step (8) until the requirements for DC transmission fault ride-through in wind farms are met in various situations;

[0121] (9) Deployment strategy; Specifically, the artificial neural network model that has been trained and tested and has adaptive and intelligent decision-making capabilities is deployed as the core control terminal in the fault ride-through control loop of the DC transmission system of the wind farm. The control terminal refers to an industrial-grade distributed computer with high-performance parallel computing capabilities in hardware. Its core function is to reliably and with low latency execute the forward propagation calculation of the neural network model that has been embedded in its memory in each control cycle.

[0122] This invention also provides another computer device, including a processor and a memory configured to store a computer program capable of running on the processor; wherein, when the processor is configured to run the computer program, it performs the method steps of the foregoing embodiments.

[0123] In practical applications, the aforementioned processor includes a Field-Programmable Gate Array (FPGA), and the processor can be a Central Processing Unit (CPU) or a Digital Signal Processor (DSP). It is understood that for different devices, the electronic devices used to implement the above-mentioned processor functions can also be other types, and this embodiment of the invention does not impose specific limitations.

[0124] The aforementioned memory can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or a combination of the above types of memory, and provides instructions and data to the processor.

[0125] In an exemplary embodiment, the present invention also provides a computer-readable storage medium for storing a computer program.

[0126] Optionally, the computer-readable storage medium can be applied to any method in the embodiments of the present invention, and the computer program causes the computer to execute the corresponding processes implemented by the processor in the various methods of the embodiments of the present invention. For the sake of brevity, these will not be described in detail here.

[0127] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take 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. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0128] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0129] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0131] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0132] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An adaptive intelligent control method for DC transmission fault ride-through in wind farms, characterized in that... Includes the following steps: Step 1: Construct an integrated AC-DC transmission system model for the wind farm. This model includes the wind farm and a modular multilevel converter. The output end of the wind farm is connected to the AC bus through a full-power converter, the input end of the wind farm is connected to the AC side of the modular multilevel converter, and the DC side of the modular multilevel converter is connected to the DC transmission line. Step two: Construct an artificial neural network model, which uses the system's real-time state vector As input, with the decision result vector The output is the artificial neural network model, which consists of computational neurons, and the weights of each computational neuron constitute the parameter matrix of the artificial neural network model. ; Step 3: Set the effective time interval of the control strategy to the DC transmission fault ride-through time. Set the initial time for the control strategy to take effect. The control strategy is applied to the integrated model of the AC-DC transmission system of the wind farm, and the parameters of the artificial neural network model are updated based on the results. Step four: Deploy the updated artificial neural network model in a fault-crossing scenario.

2. The method as described in claim 1, characterized in that: In step one, the instantaneous active power injected into the AC bus by the wind farm In the synchronously rotating dq coordinate system, it has the following form. ， in, These are the d-axis and q-axis components of the AC bus voltage, respectively. These are the d-axis and q-axis components of the fan output current, respectively. The DC voltage at the sending end of the modular multilevel converter DC active power They have the following relationship: ， in, This represents the DC line current.

3. The method as described in claim 1, characterized in that: In step one, an equivalent aggregation model of the wind farm is constructed, aggregating all wind turbines in the farm into equivalent single-unit models. The wind turbines convert kinetic energy in the air into mechanical energy, and their output mechanical power... It is determined by the following wind energy capture formula, ， in, air density, The radius of the wind turbine blades. Wind speed; The wind energy utilization coefficient is the tip speed ratio. and pitch angle The function.

4. The method as described in claim 1, characterized in that: In step two, electrical data of the system operation are collected in real time to construct a real-time system state vector. as follows, ， in, These are the AC busbars at the wind farm outlet. , , Three-phase voltage per unit value This refers to the per-unit value of the DC output voltage of the modular multilevel converter. These are the per-unit values ​​of active and reactive power output from the wind farm, respectively. These are the per-unit values ​​of active and reactive power transmitted by a DC transmission system.

5. The method as described in claim 1, characterized in that: In step two, the decision result vector Defined as, ， in, This is a reference value for the power transfer of the converter station. The number of AC power-consuming resistors put in. The grid connection time of the AC power-consuming resistor.

6. The method as described in claim 1, characterized in that: In step three, the DC transmission fault ride-through time is determined according to the following method: Real-time monitoring of DC side voltage ,when Record the time when the following fault occurrence criteria are triggered. The fault start time. ， in, DC rated voltage, This is the voltage drop factor; When DC side voltage If the following fault termination criteria are triggered, record the time. This is the fault end time. ， in, The voltage recovery coefficient, This is the minimum holding time for voltage stabilization.

7. The method as described in claim 1, characterized in that: The specific process of step three is as follows: Step 31, set the parameter matrix of the artificial neural network model as follows: ,in, The parameter matrix is The Middle OK Column parameters, These represent the depth and width of the artificial neural network model, respectively. , Integer and , ; Step 32, construct the total An independent integrated AC-DC power transmission model for a wind farm, in which To determine the number of available AC-DC transmission models for wind farms, obtain the initial values ​​for each model. and ; Step 33, the parameter matrix of the artificial neural network model copy Part, marked as, ， For each Construct the following adaptive normal distribution parameter generator: ， in, With a mean of 0 and a standard deviation of 0 Controllable normal distribution Generate parameters for adaptive normal distribution. , It serves as a generator for both forward and reverse normal distribution parameters; Using the forward and backward model parameters generated by the aforementioned adaptive normal distribution parameter generator, calculate the forward and backward decision result vectors. and as follows, ， in, The state mean vector, Let be the state standard deviation vector. The output of the artificial neural network model is the parameter of the positive model. The output of the artificial neural network model for the inverse model parameters; Total A vector of decision outcomes and The data are fed into the integrated AC-DC transmission system model of the wind farm, the time is updated, and the real-time state vector of the system is remeasured. and And calculate accordingly and Update the decision result vector and Repeated updates , , , , , until Calculate the fault isolation objective function of the artificial neural network model. and Update the parameter matrix of the artificial neural network model and real-time failure time as follows, ， Update the system's real-time state vector and mean vector and standard deviation vector as follows, ， Update the system's real-time state vector and The standard deviation is as follows: ， Step 34: Repeat step 33 until the requirements for DC transmission fault ride-through of the wind farm under the specified conditions are met.

8. The method as described in claim 7, characterized in that: In step three, the fault isolation objective function of the artificial neural network model The objective function is defined as follows for each time point within the fault period: The sum, ， Wherein, the objective function at each time point Defined as, ， in, For the AC output power of the modular multilevel converter, For the reference power of the modular multilevel converter, This refers to the voltage to ground at the DC output of the modular multilevel converter. This is the reference voltage for the DC output of the modular multilevel converter.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; characterized in that: When the processor executes the computer program, it implements the steps of the adaptive intelligent control method for DC transmission fault ride-through in wind farms as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program; characterized in that: When the computer program is executed by the processor, it implements the steps of the adaptive intelligent control method for DC transmission fault ride-through in wind farms as described in any one of claims 1 to 8.

Images