Method for cooperative suppression of short-circuit current and subsynchronous oscillation of offshore wind power flexible direct system

Abstract

The application provides a method for cooperatively suppressing short-circuit current and subsynchronous oscillation of an offshore wind power flexible direct system, comprising the following steps: step 1, determining common parameters for wideband coupling: for the problems of short-circuit current and subsynchronous oscillation, common parameters are determined through theoretical analysis and system topological structure of an offshore wind power flexible direct receiving end power grid; step 2, constructing a dynamic damping controller containing a deep double Q network (DDQN) according to the selected common parameters and a hierarchical design method: step 3, the upper optimization layer of the DDQN-based dynamic damping controller outputs the optimal adjustment instruction of the common parameters according to the system state, the lower execution layer receives the adjustment instruction of the upper layer, obtains a new common parameter value and outputs the new common parameter value to a current inner loop controller, and the current inner loop controller generates a modulation signal to control the output of the converter. The common parameters are optimized through the design of the dynamic damping controller, and the method realizes the cooperative suppression of the short-circuit current and the subsynchronous oscillation.

Description

Technical Field

[0001] This invention belongs to the field of stability control of new energy power systems, and specifically relates to a collaborative suppression method for short-circuit current-subsynchronous oscillation in offshore wind power flexible DC systems. Background Technology

[0002] The essence of subsynchronous oscillation is the phase mismatch of the positive / negative sequence impedance of the system in a specific frequency band, while the characteristics of short-circuit current (such as the short-circuit current decay rate and harmonic content) directly reflect the frequency domain distribution of the system impedance. In offshore wind power flexible DC systems, the physical characteristics of short-circuit current and subsynchronous oscillation (SSO) correspond to different frequency ranges. The rapid transient process of short-circuit current, especially the harmonic current generated by the switching of MMC submodules, is mainly concentrated in the 2-10kHz range. For example, the switching frequency of MMC is usually in the thousands of hertz, and high-frequency harmonic components will be introduced due to the switching of blocking or current limiting modes during a fault.

[0003] Subsynchronous oscillations (SSOs) typically occur in the 10-15 Hz frequency band and are closely related to the shaft torsional vibration frequency of doubly-fed wind turbines (e.g., 20-30 Hz) or the resonant characteristics of the power grid. A limitation of traditional research is that existing modeling methods often only address a single frequency band, neglecting the cross-frequency interaction between high-frequency harmonics and low-frequency oscillations, resulting in an incomplete mechanistic analysis.

[0004] In existing technologies, the short-circuit current mechanism analysis and subsynchronous oscillation suppression of offshore wind power connected to the MMC-HVDC grid are usually studied independently. The short-circuit current study and subsynchronous oscillation suppression are separated by mechanism, and the impedance coupling effect of the two in multiple frequency bands (kHz high frequency and Hz low frequency) is not considered, resulting in insufficient adaptability of control strategies. However, there is a complex dynamic coupling relationship between the short-circuit current and subsynchronous oscillation (SSO) of the flexible DC transmission system (MMC-HVDC). This relationship is mainly reflected in the impact of fault transient process on system stability, the interaction of control strategies, and the correlation of impedance characteristics.

[0005] In practice, the following problems exist: First, the mechanism is fragmented. The current dynamic characteristics of MMC-HVDC under short-circuit faults (such as high-frequency harmonic injection and negative sequence components) are coupled with the subsynchronous oscillation of wind turbines, but existing modeling methods do not fully reveal the interaction mechanism between the two. Second, parameter sensitivity is lacking. Traditional impedance models do not consider the nonlinear time-varying characteristics of system impedance under short-circuit conditions, resulting in insufficient adaptability of subsynchronous oscillation suppression strategies under fault scenarios. Short-circuit faults cause nonlinear distortion of system impedance, and existing linear models cannot accurately predict the risk of subsynchronous oscillations. Third, dynamic response is insufficient. The parameters of traditional damping controllers are fixed and cannot be dynamically adjusted under short-circuit faults, making it difficult to simultaneously suppress high-frequency current and damp the low-frequency oscillation. Summary of the Invention

[0006] The purpose of this invention is to provide a synergistic suppression method for short-circuit current-subsynchronous oscillation in offshore wind power flexible DC systems, achieving synergistic optimization of high-frequency harmonic suppression and low-frequency SSO damping. The technical solution adopted is as follows:

[0007] A method for synergistic suppression of short-circuit current-subsynchronous oscillation in an offshore wind power flexible DC system includes the following steps:

[0008] Step 1: Determine common parameters:

[0009] To address the issue of synergistic suppression of short-circuit current and subsynchronous oscillation (SSO), it is noted that the short-circuit current generates a large number of high-frequency components (high-frequency harmonics) during the initial transient process, with frequencies covering the 2-10kHz band.

[0010] Specifically, short-circuit current is a fault phenomenon mainly composed of power frequency components and superimposed with high-frequency transient components. What this invention aims to suppress is the rate of change of the power frequency and high-frequency components in the short-circuit current.

[0011] Subsynchronous oscillations belong to the low-frequency band, usually in the range of 10-50Hz. Cross-frequency band cooperative suppression requires the determination of common parameters. Focusing on the system topology and theoretical analysis of offshore wind power directly connected to the receiving-end grid via flexible grid connection, we can qualitatively explain how common parameters affect the changing trends of both.

[0012] Step 2: Based on common parameters, construct the dynamic damping controller in layers using a layered design approach:

[0013] The dynamic damping controller includes an upper optimization layer and a lower execution layer connected in sequence;

[0014] The upper optimization layer takes the real-time state of the system as input and outputs the optimal adjustment instructions for common parameters.

[0015] The lower execution layer takes the modulation command from the upper layer as input and outputs the common parameter values ​​that are adjusted in real time.

[0016] Traditional controllers with fixed common parameters struggle to ensure transient short-circuit fault ride-through capability while maintaining small-signal stability in the subsynchronous oscillation frequency band. In contrast, dynamic damping controllers can dynamically optimize common parameters based on the real-time state of the system and divide them into two parts: an upper optimization layer and a lower execution layer.

[0017] Step 3: The dynamic damping controller transmits the common parameter values ​​that are adjusted in real time to the current inner loop controller.

[0018] Preferably, the common parameter is: the proportional parameter of the current inner loop control. ;

[0019] For short-circuit current (high-frequency transient): increase It can significantly improve the response speed and bandwidth of the current loop and effectively suppress the rise rate and peak value of the short-circuit current, which is a positive effect.

[0020] For subsynchronous oscillations (stable at low frequencies): increase It may introduce negative damping in the subsynchronous frequency band, leading to a deterioration of the phase margin, which may trigger or exacerbate subsynchronous oscillations, posing a potential risk.

[0021] Preferably, step 1 specifically includes the following steps:

[0022] Step 1A: Establish the system topology for offshore wind power to be directly connected to the receiving-end power grid via flexible connection;

[0023] Step 1B: Based on the system topology, determine the common parameters through theoretical analysis.

[0024] Preferably, in step 2, the upper optimization layer makes optimal adjustment instructions for common parameters based on the deep dual-Q network model DDQN and sends them to the lower execution layer; the lower execution layer receives the adjustment instructions, updates the parameters and outputs them to the current inner loop controller.

[0025] Preferably, the deep dual-Q network model includes a state space, an action space, and a reward function.

[0026] Preferably, the reward function for:

[0027] ;

[0028] in, , , - The weighting coefficient can be adjusted according to the degree of importance attached to different controlled objects;

[0029] - The amplitude of the subsynchronous oscillation current is extracted through real-time Fourier transform (FFT), which is a direct quantitative characteristic of the subsynchronous oscillation.

[0030] - Rate of change of short-circuit current;

[0031] - The fluctuation of DC voltage is an important indicator used to evaluate the overall stability of the system.

[0032] Preferably, in step 3, the optimal adjustment command is to adjust the proportional gain of the inner current loop. Adjustment amount .

[0033] Preferably, the real-time status variables of the system include: current. of d, q Axial component, DC side voltage DC voltage fluctuation Active power reactive power , amplitude of subsynchronous oscillation current and short-circuit current change rate .

[0034] Preferably, in step 2, the specific method of real-time adjustment is as follows: the adjustment command is converted into a new parameter value, and the new parameter value is written into the parameter register of the current inner loop controller.

[0035] Preferably, the dynamic damping controller is integrated into the control circuit of the offshore wind power flexible DC system in the form of a control algorithm, and the parameter execution module provides the ability to rewrite the parameter register value of the original current inner loop controller of the offshore wind power flexible DC system to execute instructions.

[0036] Compared with the prior art, the advantages of the present invention are:

[0037] To address the coupling effect of short-circuit current and subsynchronous oscillation (SSO) in offshore wind power grid-connected systems via MMC-HVDC, and to develop a collaborative suppression strategy based on this mechanism, a hierarchical control architecture and dynamic strategy learning via deep dual-Q network (DDQN) are employed to achieve multi-band impedance decoupling, fault scenario adaptation, and rapid response, thereby improving the transient stability of the system.

[0038] This invention focuses on the synergistic study of short-circuit current and subsynchronous oscillation (SSO) suppression in offshore wind power flexible DC systems, in order to solve problems such as fragmented mechanisms, insufficient dynamic response, and lack of nonlinear modeling in traditional methods. Attached Figure Description

[0039] Figure 1 A technical roadmap for a coordinated suppression strategy of short-circuit current and subsynchronous oscillation in offshore wind power flexible DC systems;

[0040] Figure 2 Topology diagram of the flexible-vertical system;

[0041] Figure 3 This is a flowchart of dynamic damping control.

[0042] Figure 4 This is a schematic diagram of the DDQN algorithm. Detailed Implementation

[0043] The following description, with reference to schematic diagrams, provides a more detailed account of the synergistic suppression method for short-circuit current-subsynchronous oscillation in offshore wind power flexible DC systems of the present invention. Preferred embodiments of the invention are illustrated. It should be understood that those skilled in the art can modify the invention described herein while still achieving its advantageous effects. Therefore, the following description should be understood as being of general knowledge to those skilled in the art and is not intended to limit the invention.

[0044] 1. Coupled analysis of short-circuit current and subsynchronous oscillation dynamic characteristics

[0045] The dynamic coupling characteristics of short-circuit current and subsynchronous oscillation, i.e., their interaction, will be explained in this invention from the following two aspects:

[0046] One factor is the short-circuit current triggering SSO (Short-Circuit Current Response). When a short-circuit fault occurs in the system, the rapid rise in short-circuit current can lead to voltage drops and converter lockout, thereby altering the system's impedance characteristics. This nonlinear change in impedance may disrupt the original damping balance, inducing shaft torsional vibration or control loop instability, thus triggering or exacerbating SSO. Furthermore, when a fault occurs in a flexible DC system, the controller's dynamic response (such as current limiting strategies) may change the converter's control parameters, thereby affecting the damping effect of SSO.

[0047] Secondly, there is the impact of SSO on short-circuit current. Subsynchronous oscillation can cause generator speed fluctuations, which in turn affect the converter's current control loop through electromechanical coupling. For example, the frequency shift caused by SSO may delay the dynamic response of the phase-locked loop (PLL), resulting in distortion of the converter's output current, which is superimposed on the short-circuit current to form high-frequency harmonic components.

[0048] Figure 1 In this context, "short-circuit current mechanism" refers to "the triggering mechanism of short-circuit current on SSO".

[0049] The "low-frequency SSO mechanism" refers to "the effect of SSO on short-circuit current".

[0050] 2. Qualitative analysis of common parameters

[0051] High-frequency harmonics in short-circuit currents (such as harmonics generated by MMC submodule switching) are mainly concentrated in the 2-10kHz frequency band, while subsynchronous oscillations (SSOs) typically occur in the 10-50Hz frequency band. On the one hand, high-frequency harmonics in short-circuit currents may couple to the low-frequency band through the converter control circulating current, interfering with the stability of the phase-locked loop and current controller, and reducing the SSO damping ratio. On the other hand, low-frequency oscillations caused by SSOs can change the phase characteristics of the system's equivalent impedance, leading to current waveform distortion during short-circuit fault recovery and prolonging the fault clearing time.

[0052] Therefore, this invention proposes a cross-frequency band analysis method, which determines common parameters through the system topology and theoretical analysis of offshore wind power directly connected to the receiving-end power grid via flexible connection, thereby qualitatively analyzing the impact of common parameters on short-circuit current and SSO.

[0053] Regarding the impact of short-circuit current: AC short-circuit faults are large disturbances and nonlinear transient processes. Current inner-loop control belongs to the double closed-loop control of the control circuit in flexible transmission systems. Figure 2 The outer loop voltage control + inner loop current control mechanism primarily aims to quickly and accurately track current commands. However, in the event of a short circuit, a significant error signal arises between the current command value of the outer voltage loop and the actual current value. Meanwhile, the proportional gain of the inner current loop... K p It directly determines the strength of the controller's response to error signals, thereby affecting the system's effective bandwidth. To regulate the system's response to short-circuit current.

[0054] ;

[0055] For the current loop bandwidth, L The inductance of the grid-side converter reactor. Inner current loop proportional gain. The larger the bandwidth The higher the value, the faster the system response and the earlier it can begin limiting short-circuit current. This value generates a powerful and immediate control command, a modulation signal, that causes the converter MMC to rapidly change its output voltage to cope with grid voltage changes caused by current surges. This effectively limits the rate of rise of the short-circuit current, thereby further limiting the peak value of the short-circuit current. Lower values... An excessively high value will cause the system to react slowly and fail to effectively limit the short-circuit current.

[0056] Regarding the impact of subsynchronous oscillations (SSO): Unlike short-circuit current, SSO is a small-signal stability problem that occurs in the 10-50Hz subsynchronous frequency band, specifically the frequency band where the ratio of the converter output impedance to the grid impedance meets specific conditions. Current inner loop proportional gain. This will significantly alter the phase characteristics of the converter's output impedance in the subsynchronous frequency band. When the output impedance exhibits negative damping characteristics at a certain subsynchronous frequency, and the grid impedance is inductive, the energy exchange between them will create subsynchronous oscillations. Higher... K p The value may lead to a deterioration of impedance characteristics, providing negative damping, thereby stimulating or exacerbating subsynchronous oscillations (SSO).

[0057] The above theoretical analysis shows that the proportional gain of the current inner loop... It is a key common parameter that significantly affects both short-circuit current and subsynchronous oscillation characteristics, but its effects are contradictory:

[0058] For short-circuit current (high-frequency transient): increase It can significantly improve the response speed and bandwidth of the current loop and effectively suppress the rise rate and peak value of the short-circuit current, which is a positive effect.

[0059] For subsynchronous oscillations (stable at low frequencies): increase It may introduce negative damping in the subsynchronous frequency band, leading to a deterioration of the phase margin, which could trigger or exacerbate subsynchronous oscillations, posing a potential risk.

[0060] This inherent contradiction indicates that adopting a fixed... Traditional PI controllers struggle to simultaneously guarantee short-circuit fault ride-through capability and small-signal stability of subsynchronous oscillations. Therefore, this invention proposes a method that intelligently optimizes common parameters based on the real-time system status. A dynamic damping controller, combined with a deep dual-Q network, is used to solve the problem of synergistic suppression of short-circuit current-subsynchronous oscillation.

[0061] 3. Design of Dynamic Damping Controller

[0062] The dynamic damping controller is the core execution unit of this invention. Traditional PI damping controllers have slow response speeds, cannot automatically adjust parameters or achieve multi-objective coordination, and can only suppress harmonics or subsynchronous oscillations individually. In contrast, the dynamic damping controller has a fast response speed and dynamically optimizes common parameters based on real-time conditions. The current inner loop controller generates a modulation signal. This allows for the control of the converter output, achieving coordinated suppression of short-circuit current and subsynchronous oscillation (SSO).

[0063] Table 1. Comparison of differences between dynamic damping controllers and traditional PI damping controllers

[0064] index Traditional PI damping controller This invention is a dynamic damping controller. Parameter applicability Fixed gain, relying on manual parameter tuning Online learning, adapting to environmental changes Multi-objective collaboration Single objective (harmonic suppression or SSO) Synergistic suppression of high-frequency harmonics and low-frequency SSO

[0065] The core execution unit design can be divided into two parts: an upper optimization layer and a lower execution layer, as detailed below. Figure 3 As shown:

[0066] 1) Working principle and function of the upper optimization layer

[0067] The upper optimization layer, as the core decision-making unit of the dynamic damping controller, is positioned to perceive the global operating state of the flexible DC system in real time and output optimization adjustment instructions for key common parameters based on a pre-trained deep reinforcement learning model, thereby achieving the goal of synergistic suppression of short-circuit current and subsynchronous oscillation. It employs a reinforcement learning algorithm based on a Dueling Double Deep Q-Network (DDQN) as its intelligent decision engine, with the upper optimization layer, also based on the DDQN, acting as the "intelligent brain," making globally optimal decisions based on the real-time system state using the trained DDQN network.

[0068] like Figure 4 As shown, a deep double-Q network includes several basic elements such as a state space, an action space, and a reward function.

[0069] (1) State space: In order to comprehensively characterize the operation status and stability risks of the flexible DC transmission system, this patent selects some state characteristic parameters to form a state vector. S t It defines the dimensions by which the intelligent agent (upper optimization layer) perceives the environment (flexible and direct system):

[0070] ;

[0071] In the above formula, For flexible DC grid-connected current d, q The shaft component reflects the current power delivery status and the instantaneous state of the AC side current;

[0072] The active and reactive power transmitted reflect the system's operating condition point;

[0073] This is the DC-side voltage, reflecting the steady state of the DC system;

[0074] It represents the fluctuation of DC voltage and is an important indicator used to evaluate the overall stability of the system.

[0075] The amplitude of the subsynchronous oscillation current is extracted by real-time Fourier transform (FFT). This value is a direct quantitative feature of the subsynchronous oscillation and is one of the core elements in the state space.

[0076] It is the rate of change of short-circuit current, which is a key characteristic quantity for sensing high-frequency transient faults such as short-circuit current.

[0077] When a disturbance occurs in a flexible DC system, the current... It is called three-phase grid-connected fault current or three-phase grid-connected disturbance current;

[0078] Right now In Three-phase grid-connected fault current The amplitude.

[0079] The disturbance occurs because a short-circuit fault and a subsynchronous oscillation fault occur simultaneously. It includes both short-circuit current and subsynchronous oscillation components, but the processing methods in the topology diagram are different. The subsynchronous oscillation component uses frequency domain analysis (FFT) to extract the amplitude of the subsynchronous oscillation current. The short-circuit current component is analyzed using time-domain analysis for three-phase grid-connected fault currents. The amplitude is differentiated to obtain the rate of change of short-circuit current. .

[0080] "Subsynchronous oscillation current amplitude" "That is, the amplitude of the subsynchronous oscillation component."

[0081] Throughout the text, "short-circuit current" refers to the three-phase grid-connected fault current. The short-circuit current component in the circuit.

[0082] "Subsynchronous oscillation" refers to the fault current in a three-phase grid-connected system. The subsynchronous oscillation component in the middle.

[0083] This state space contains the characteristic information of both subsynchronous oscillations and short-circuit faults, enabling the agent to simultaneously perceive these two types of stability problems at different frequency bands and time scales, providing an information basis for subsequent cooperative control strategies.

[0084] (2) Action space: Action space It defines the operations that the upper optimization layer can perform, and the output of the upper optimization layer is an instruction for adjusting common parameters:

[0085] ;

[0086] In the above formula, This is the adjustment amount for the proportional gain of the inner loop current, a common parameter.

[0087] This action space The proportional gain of the inner loop current, which directly affects the original common parameter of the system, K p The upper optimization layer dynamically adjusts common parameters through different action combinations, thereby changing the impedance characteristics and response speed of the system, ultimately achieving the goal of both suppressing subsynchronous oscillations and limiting short-circuit current.

[0088] (3) Reward function: Reward function This invention serves as a guiding force for DDQN learning. It designs a multi-objective weighted reward function to simultaneously penalize subsynchronous oscillations and short-circuit currents.

[0089] ;

[0090] In the above formula, since the harm of subsynchronous oscillation is linearly related to its amplitude, therefore - This is a linear penalty term used to penalize subsynchronous oscillations, with the oscillation amplitude... The larger the value, the smaller the reward and the greater the punishment, thus incentivizing the agent to take action to inhibit the action.

[0091] The harm caused by short-circuit faults increases exponentially with the rate of change of current, therefore -( ) 2 The squared penalty term amplifies extreme dangerous situations and significantly incentivizes the agent to prioritize and quickly handle short-circuit faults. It severely punishes the huge short-circuit current rise rate, which increases sharply when a short-circuit fault occurs. This incentivizes the agent to act quickly to limit the short-circuit current.

[0092] -| | Used to penalize DC voltage fluctuations, ensuring that the agent does not jeopardize the stability of the DC side when optimizing AC side issues, thus guaranteeing the overall stability of the system;

[0093] , , These are weighting coefficients, which can be adjusted according to the degree of importance attached to different controlled objects, for example... It can increase during a short-circuit fault to quickly limit the magnitude of the short-circuit current.

[0094] This reward function quantifies the goal of "cooperative inhibition" into an optimizable mathematical problem, allowing the agent to try different actions to maximize its long-term cumulative reward and thus learn the optimal cooperative inhibition strategy.

[0095] 2) Working principle and function of the lower execution layer

[0096] The upper optimization layer, acting as the command execution terminal for the dynamic damping controller, is positioned to reliably, safely, and seamlessly output the optimization commands generated by the upper optimization layer to the existing control system of the flexible-conductive system. This enables the dynamic updating of common parameters within the controller, thereby translating intelligent decisions into actual system behavior. Its input is the optimal parameter adjustment command issued by the upper optimization layer. The output is a write operation to the parameter register of the current inner loop controller in the standard control system of the flexible DC system, directly changing the common parameters. The values ​​during operation are used to achieve coordinated suppression of subsynchronous oscillation and short-circuit current.

[0097] The lower layer adopts an execution principle based on safety limiting and integration, and its workflow is as follows:

[0098] (1) Command reception: Real-time reception of optimal parameter adjustment commands from the upper optimization layer ;

[0099] (2) Parameter update: Perform simple algebraic operations to calculate the new parameter values: ;

[0100] (3) Safety limit: The calculated new parameter value K p_new The parameters are compared with and limited by the preset engineering safety boundary to ensure that the parameters always operate within an absolutely safe range;

[0101] (4) Parameter writing: The final processed safety parameter values ​​are directly written into the parameter memory of the current inner loop controller, overwriting the original common parameter values. .

[0102] like Figure 2 Subsynchronous oscillations generally occur on the wind farm side, but they can be transmitted to the receiving-end grid via high-voltage transmission, thus affecting the receiving-end grid. The most direct effect is on the grid-connected three-phase instantaneous current of the receiving-end grid (all of which are mentioned in existing literature). Short-circuit current faults generally occur at the PCC point on the grid side, i.e., the common coupling point, which also affects the grid-connected three-phase instantaneous current of the receiving-end grid.

[0103] Therefore, when a malfunction occurs, This transforms into a three-phase grid-connected fault current, which includes a subsynchronous oscillation component and a short-circuit current component. Frequency domain analysis is then performed on the subsynchronous oscillation component, i.e., FFT is used to analyze the three-phase grid-connected fault current. Extract the amplitude of the subsynchronous oscillation current The short-circuit current component was analyzed in the time domain, specifically the three-phase grid-connected fault current. The amplitude is differentiated to obtain the rate of change of short-circuit current. .

[0104] Based on the above analysis, Figure 2 An overview of the overall framework:

[0105] Figure 2 In this context, PI refers to a PI controller, which includes an integral element and a proportional element. The proportional element and the current inner loop control the proportional parameter. Related.

[0106] This typically appears in the decoupling calculation module of the current inner loop controller, and its value is derived from the grid angular frequency. and the inductance of the grid-side converter reactor The product of.

[0107] When the flexible DC system is in steady-state operation, the outer loop control of the DC voltage compares the DC-side voltage. Its reference value Output active current reference values ​​respectively and reactive current reference value ,

[0108] At the same time, current (Instantaneous current of three-phase grid connection) ） pass Park The transformation is from a three-phase stationary coordinate system to a system that rotates synchronously with the grid voltage. dq Two components in the rotating coordinate system and The angle output by the phase-locked loop θ It connects the three-phase stationary coordinate system with dq The key to rotating the coordinate system is that it is related to the grid voltage vector. θ The angles are completely synchronized.

[0109] Next, the inner current loop control receives commands from the outer loop. , and rotational components After modulation by the controller, a modulated signal is finally generated. And convert it into a three-phase modulated signal. The signal is fed into the MMC converter, thereby controlling the output of the converter's MMC.

[0110] When a disturbance occurs in the flexible DC system, a short-circuit fault and subsynchronous oscillation will occur. , These parameters can change drastically, and the dynamic damping controller captures these changes.

[0111] Flexible DC system real-time current monitoring (Three-phase grid-connected fault current), DC voltage The original quantities are equal, and the rotational components are analyzed using FFT. Extract the amplitude of the subsynchronous oscillation current ,and and It is calculated based on the original quantities, and all these quantities together form the state vector. This process is all completed in the upper optimization layer of the dynamic damping controller.

[0112] Among them, the three-phase grid-connected fault current is obtained through FFT. Extracted subsynchronous oscillation current amplitude .

[0113] The DDQN network in the upper optimization layer processes the state vector. Real-time analysis is performed to intelligently determine the optimal modulation command. The new parameter values ​​are then calculated by the lower execution layer. It is then written into the parameter memory of the current inner loop controller, overwriting the original common parameter values. .

[0114] Finally, the current inner loop control uses new parameter values. And combined with the phase-locked loop (PLL) provided θ The calculations are performed to generate a modulation signal, which in turn controls the output of the converter.

[0115] Dynamic damping controllers indirectly optimize flexible-conductive systems by dynamically optimizing common parameters. dq Overall dynamic performance in synchronous rotating coordinate system K p Ultimately, this achieves the goal of coordinated inhibition.

[0116] The lower-level execution layer designed in this patent has two major advantages. First, it offers high reliability. The built-in safety limiting logic ensures that even in the event of an abnormal DDQN output instruction, the common parameters will not exceed the safe range, greatly improving the reliability of the entire dynamic damping controller. Second, it boasts high integration. It interacts with the existing flexible DC system control circuit through parameter writing overwriting, rather than signal superposition, without requiring any changes to the original flexible DC system control circuit.

Claims

1. A method for synergistic suppression of short circuit current-sub synchronous oscillation of offshore wind power flexible direct system, characterized in that, The method comprises the following steps: Step 1, determining common parameters; Step 2, according to the common parameters, adopting a hierarchical design method to hierarchically construct a dynamic damping controller: The dynamic damping controller comprises a top optimization layer and a bottom execution layer connected in sequence; The input of the top optimization layer is real-time state quantity of the system, and the output is optimal adjustment instruction for the common parameters; The input of the bottom execution layer is modulation instruction of the top layer, and the output is real-time adjusted common parameter value; Step 3, the dynamic damping controller delivers the real-time adjusted common parameter value to the current inner loop controller; The common parameter is: current inner loop control proportional parameter ; The real-time status parameters of the system include: current. of d, q Axial component, DC side voltage DC voltage fluctuation Active power reactive power , amplitude of subsynchronous oscillation current and short-circuit current change rate ; Step 1 specifically comprises the following steps: Step 1A, establishing system topology structure of offshore wind power through flexible direct access to receiving end power grid; Step 1B, based on the system topology structure, determining common parameters through theoretical analysis.

2. The method for synergistic mitigation of short circuit current- subsynchronous oscillation of offshore wind power flexible direct system according to claim 1, characterized in that, In step 2, the top optimization layer makes optimal adjustment instruction for the common parameters based on a deep double Q network model DDQN and issues the instruction to the bottom execution layer; the bottom execution layer receives the adjustment instruction, updates the parameters and outputs to the current inner loop controller.

3. The method for synergistic mitigation of short circuit current- subsynchronous oscillation of offshore wind power flexible direct system according to claim 2, characterized in that, The deep double Q network model comprises state space, action space and reward function.

4. The method for synergistic mitigation of short circuit current- subsynchronous oscillation of offshore wind power flexible direct system according to claim 3, characterized in that, The reward function is: ； wherein, , , - weight coefficient, which can be adjusted according to the importance of different control objects; - a sub-synchronous oscillation current amplitude extracted by a real-time Fourier transform, FFT; - short circuit current rate of change; - the amount of fluctuation of the direct voltage.

5. The method for synergistic mitigation of short circuit current- subsynchronous oscillation of offshore wind power flexible direct system according to claim 1, characterized in that, In step 2, the optimal adjustment instruction is an adjustment amount of the proportional gain of the current inner loop . .

6. The method of synergistic mitigation of short circuit current - subsynchronous oscillation of offshore wind power flexible direct system according to claim 1, characterized in that, In step 2, the specific way of real-time adjustment is to convert the adjustment instruction into new parameter value and write the new parameter value into the parameter register of the current inner loop controller.

7. The method for synergistic mitigation of short circuit current- subsynchronous oscillation of offshore wind power flexible direct system according to claim 1, characterized in that, The dynamic damping controller is integrated in the control circuit of the offshore wind power flexible system in the form of control algorithm, and the parameter execution module provides rewriting of the parameter register value of the original current inner loop controller of the offshore wind power flexible system to execute the instruction.

Images