Transient frequency support control method considering low voltage penetration characteristic of converter

Abstract

The invention discloses a transient frequency support control method considering the low voltage penetration characteristic of a converter, and relates to the field of electrical engineering, and the method comprises the steps: collecting the voltage data and frequency data of a power grid through a real-time monitoring sensor; based on the acquired power grid voltage data, carrying out converter voltage low-pass-through characteristic detection to obtain a voltage low-pass-through state; according to the voltage low penetration state, starting an adaptive response strategy, and adjusting a power instruction; based on the power grid frequency data, starting a virtual synchronous machine control algorithm according to the adjusted power instruction, and adjusting power output; a reinforcement learning algorithm is used to carry out cooperative adjustment on the voltage low-penetration adaptive response strategy and the virtual synchronous machine control algorithm, and the state of the power grid is recovered; according to the invention, through cooperative adjustment of the self-adaptive response strategy of the voltage low penetration state and the virtual synchronous machine control algorithm, the effect of the converter in power grid frequency adjustment is effectively improved, and the stability and reliability of the power grid frequency are enhanced.

Description

Technical Field

[0001] The invention relates to the field of electrical engineering, and in particular to a transient frequency support control method taking into account the voltage low-breakthrough characteristics of a converter. Background Art

[0002] In recent years, with the rapid development of new energy power generation technology, especially the access of renewable energy such as wind power and photovoltaic power, the dispatching and control difficulties faced by the power system are increasing. As a key device for the grid connection of renewable energy power generation, the converter plays a vital role. In the grid frequency regulation, the converter not only undertakes the task of regulating the grid frequency fluctuation, but also needs to ensure the stability of its voltage and power output. However, when the grid faces sudden load fluctuations, voltage sags, etc., the power output of the converter may be affected, especially in the state of voltage low penetration. The traditional converter control strategy is difficult to effectively respond to the instantaneous grid frequency fluctuations and power demand, which in turn affects the stability of the grid. Therefore, improving the converter voltage low penetration characteristics and its ability to support the grid frequency has become an important research topic to improve the stability of the power system.

[0003] The current grid frequency regulation methods mainly rely on the traditional synchronous generator inertial response and virtual synchronous machine control algorithms. However, these methods mostly rely on fixed response models and fail to fully consider the dynamic changes of the grid and the nonlinear response characteristics of the converter under voltage low penetration conditions. Especially during the period of converter voltage low penetration, the changes in grid frequency and voltage are highly dynamic, and traditional control algorithms are difficult to make accurate adjustments in a short time, resulting in unsatisfactory frequency support effects. At the same time, existing technologies usually lack flexible adaptive mechanisms and are difficult to cope with changes in complex grid environments. To address this problem, based on the combination of real-time grid monitoring data and adaptive control strategies, by introducing advanced algorithms such as virtual synchronous machines and machine learning, it is expected to achieve efficient frequency support for converters under voltage low penetration conditions. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a transient frequency support control method taking into account the voltage dropout characteristics of the converter to solve the problem of insufficient response speed and accuracy to frequency changes during voltage dropout in the prior art.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a transient frequency support control method taking into account the voltage low penetration characteristics of the converter, which includes collecting grid voltage data and grid frequency data through real-time monitoring sensors; based on the collected grid voltage data, detecting the voltage low penetration characteristics of the converter to obtain the voltage low penetration state; according to the voltage low penetration state, enabling an adaptive response strategy to adjust the power instruction; based on the grid frequency data, starting a virtual synchronous machine control algorithm according to the adjusted power instruction to adjust the power output; using a reinforcement learning algorithm to coordinately adjust the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm to restore the grid state.

[0008] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, wherein: the real-time monitoring sensor is used to collect the grid voltage data and grid frequency data, specifically,

[0009] The grid voltage data includes the instantaneous voltage value of each node in the grid, the fluctuation range and amplitude change of the grid voltage, and the instantaneous power of the grid;

[0010] The grid frequency data includes the instantaneous frequency of the grid and the deviation of the grid frequency from the standard frequency.

[0011] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, the converter voltage low penetration characteristics are detected based on the collected grid voltage data to obtain the voltage low penetration state. The specific steps are as follows:

[0012] Set the voltage low-breakthrough threshold and voltage recovery threshold;

[0013] Compare the voltage low-breakthrough threshold and the voltage recovery threshold with the grid voltage data to determine the voltage low-breakthrough state;

[0014] Transmitting the voltage low-break state to the dispatch control unit;

[0015] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, wherein: according to the voltage low penetration state, the adaptive response strategy is enabled to adjust the power instruction, and the specific steps are as follows:

[0016] Based on the voltage low penetration state and grid voltage and frequency data transmitted to the dispatch control unit, a voltage low penetration dynamic response model is established, and the expression is:

[0017]

[0018] Among them, P dy (t) is the power command after dynamic adjustment at time t, P0 is the initial power setting value, Represents the fast response term to voltage deviation, V ref is the set voltage reference value, V(t) is the current collected voltage value, α is the voltage adjustment sensitivity coefficient, is the integral response term to the frequency change rate, β is the frequency adjustment gain coefficient, f g (t′) is the grid frequency at time t′, is the grid frequency f g (t′) is the partial derivative with respect to time t′, where t is the current time, t′ is the dynamic time of voltage breakdown, t0 is the starting point of the integration time, and d is the symbol of total differential;

[0019] According to the different stages of voltage breakdown, the parameters α and β are optimized to enhance the response effect;

[0020] According to the voltage low wear dynamic response model output P dy (t), combined with the actual power demand, the multivariable control algorithm is enabled for coordinated regulation, and the expression is:

[0021]

[0022] Among them, P out (t) is the output power command, γ is the power dynamic allocation weight, P dy (t) is the power command after dynamic adjustment at time t, κ is the frequency deviation correction coefficient, f ref is the reference frequency, f g (t′) is the real-time collected grid frequency, t is the current time, t′ is the dynamic time of voltage low penetration, and t0 is the starting point of the integration time;

[0023] Sending output power instructions to the converter controller to control power output;

[0024] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, wherein: according to the different stages of the voltage low penetration state, the parameters α and β are optimized to enhance the response effect, and the specific steps are as follows:

[0025] When voltage breakdown occurs in the early stage, adjust the parameters α and β to achieve rapid recovery;

[0026] When the voltage gradually recovers, adjust the parameters α and β to balance the voltage recovery and frequency support;

[0027] When the voltage returns to the normal range, adjust the parameters α and β to correct the power output;

[0028] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, wherein: based on the grid frequency data, according to the adjusted power command, the virtual synchronous machine control algorithm is started to adjust the power output, and the specific steps are:

[0029] Simulate the inertia and damping characteristics of the synchronous generator to obtain the inertia and damping compensation power, the expression is:

[0030]

[0031] Among them, ΔP sync (t) is the inertia and damping compensation power, is the angular velocity change rate of the virtual synchronous machine, J is the equivalent inertia coefficient, D is the damping coefficient, ω(t) is the instantaneous angular velocity of the virtual synchronous machine, ω ref is the reference angular velocity;

[0032] Based on the output power command and the inertia and damping compensation power, the final output power command is obtained, and the expression is:

[0033] P fin (t) = P out (t)+ΔP sync (t);

[0034] Calculate error data, design feedback optimization controller, and integrate feedback power compensation;

[0035] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter according to the present invention, the specific steps of calculating error data, designing a feedback optimization controller, and comprehensively compensating feedback power are as follows:

[0036] Calculate the power error and frequency error, the expression is:

[0037] ΔP(t)=P fin (t)-P act (t),Δf(t)=f ref -f g (t);

[0038] Among them, ΔP(t) is the power error, Δf(t) is the frequency error, P act (t) is the actual power grid power, f g (t) is the actual grid frequency;

[0039] Based on the error data, a feedback optimization controller is designed, and the expression is:

[0040]

[0041] Where N is the proportional gain coefficient, M is the integral gain coefficient, ΔP fb (t) is feedback power compensation;

[0042] According to the final output power command and feedback power compensation, calculate the feedback compensation P last (t), P last (t) is sent to the converter controller to drive the power output.

[0043] As a preferred solution of the transient frequency support control method taking into account the voltage low penetration characteristics of the converter described in the present invention, wherein: using a reinforcement learning algorithm, the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm are coordinated to restore the grid frequency. The specific steps are:

[0044] Define the grid state space S t and action space A t ;

[0045] Design the reward function, the expression is:

[0046]

[0047] Among them, Δf(t) is the frequency error, ΔP(t) is the power error, ΔV(t) is the voltage error, is the frequency change rate, λ1 is the power error weight coefficient, λ2 is the voltage error weight coefficient, and λ3 is the frequency change rate weight coefficient;

[0048] Select the PPO reinforcement learning algorithm and conduct reinforcement learning training;

[0049] After the training is completed, the model is deployed and the various parameters in the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm are adjusted dynamically in real time.

[0050] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the transient frequency support control method taking into account the converter voltage low-breakthrough characteristics as described in the first aspect of the present invention is implemented.

[0051] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the transient frequency support control method taking into account the converter voltage low-breakthrough characteristics as described in the first aspect of the present invention.

[0052] The beneficial effects of the present invention are as follows: the present invention utilizes high-precision sensors to collect grid voltage and frequency data in real time, and provides dynamic feedback for subsequent control. By detecting the voltage low-through characteristics of the converter, the abnormal state of the grid is identified in real time, and the adaptive response strategy is started to adjust the power command to quickly respond to grid changes. Combined with the virtual synchronous machine control algorithm, the characteristics of the synchronous generator are simulated, the inertia and damping support of the grid are enhanced, and the frequency stability is further improved. The response strategy and the control algorithm are collaboratively optimized through the reinforcement learning algorithm, and the control parameters can be dynamically adjusted to enhance the adaptive ability of the grid under complex working conditions. It effectively enhances the role of the converter in grid frequency regulation, enhances the stability and reliability of the grid frequency, and has good technical feasibility and potential application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0054] Figure 1 This is a flow chart of the transient frequency support control method taking into account the voltage low-breakthrough characteristics of the converter in Example 1.

[0055] Figure 2 This is a flow chart of optimizing the parameters of the voltage low breakdown state in Example 1. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0059] Example 1, reference Figure 1 and Figure 2, which is the first embodiment of the present invention, provides a transient frequency support control method taking into account the voltage low-breakthrough characteristics of the converter, comprising:

[0060] S1: Collect grid voltage data and grid frequency data through real-time monitoring sensors.

[0061] Specifically, the following steps are included:

[0062] S1.1: Select real-time monitoring sensors with high precision, high stability and fast response capabilities.

[0063] It should be understood that high precision, high stability and fast response capabilities can ensure the accuracy of collected data, avoid control failures caused by errors, reduce the impact of environmental factors on sensor performance, and ensure that the control module can respond to sudden grid transient events in a timely manner.

[0064] Furthermore, the voltage sensor range needs to cover the rated voltage range of the power grid (such as 380V or 220V) and support the measurement of sudden fluctuations. The frequency sensor can monitor 50HZ frequency fluctuations in real time with a resolution of 0.01Hz.

[0065] S1.2: Deploy real-time monitoring sensors at the input and output of the converter to collect data.

[0066] Specifically, additional sensors are deployed at critical loads and main feeders to enhance monitoring capabilities.

[0067] S1.2.1: The collected data include grid voltage data and grid frequency data.

[0068] Specifically, the grid voltage data includes the instantaneous voltage value of each node in the grid, the fluctuation range and amplitude change of the grid voltage, and the instantaneous power of the grid.

[0069] The grid frequency data includes the instantaneous frequency of the grid and the deviation of the grid frequency from the standard frequency.

[0070] S1.3: Transmit the collected data to the central processing unit.

[0071] Among them, the collected data needs to be calibrated and time-synchronized in real time to ensure the accuracy of the collected data and its consistency in the time dimension to facilitate subsequent analysis.

[0072] The data transmission protocol selects RS485 communication protocol to ensure the integrity and real-time nature of the data.

[0073] S1.4: Preprocess the data in the central processing unit.

[0074] Specifically, data cleaning is required to remove invalid data and outliers. Then the data is filtered to remove high-frequency noise. Finally, the data is normalized and standardized to facilitate the processing and calculation of subsequent algorithms.

[0075] Preferably, through the design of the above steps, accurate monitoring of the power grid state can be ensured, and the power grid voltage and frequency data can be acquired and transmitted in real time, thereby making the frequency support control method more accurate and reliable. High-precision, stable and fast-response sensors avoid the influence of external environmental interference while ensuring data quality, and improve the ability to respond to sudden power grid events. In addition, through the preprocessing of real-time data, data quality can be improved, and the execution of subsequent control algorithms can be further optimized to ensure that the power grid maintains stable operation in dynamic changes.

[0076] S2: Based on the collected grid voltage data, the converter voltage low-breakthrough characteristic detection is performed to obtain the voltage low-breakthrough status.

[0077] Specifically, the following steps are included:

[0078] S2.1: Set the voltage low-breakthrough threshold and voltage recovery threshold.

[0079] It should be noted that, according to the grid operation standard, the voltage undershoot threshold is set at 80% of the rated voltage, and the voltage recovery threshold is set at 90% of the rated voltage.

[0080] S2.2: Compare the voltage undershoot threshold and the voltage recovery threshold with the grid voltage data to determine the voltage undershoot state.

[0081] Specifically, the method for determining the voltage low penetration state is that when the grid voltage data is less than the voltage low penetration threshold and lasts for more than 100ms, it is determined to be a voltage low penetration state (marked as "low penetration mode"). When the grid voltage data is greater than or equal to the voltage recovery threshold and lasts for more than 100ms, it is determined to be a recovery state (marked as "normal mode").

[0082] S2.3: Transmit the voltage low-breakdown status to the dispatch control unit.

[0083] Preferably, through standardized threshold setting and accurate state judgment, the voltage low penetration and recovery state of the power grid can be detected efficiently and accurately, and transmitted to the dispatch control unit in a timely manner. This not only helps to quickly identify and respond to power grid faults, but also reduces the negative impact of power grid fluctuations and misjudgments, and improves the stability and reliability of power grid operation. Accurate determination of the voltage low penetration state, confirmation of the recovery state, and timely transmission optimize the power grid control strategy and improve overall adaptability and response speed.

[0084] S3: According to the voltage low breakout state, enable the adaptive response strategy and adjust the power command.

[0085] Specifically, the following steps are included:

[0086] S3.1: Based on the voltage low penetration state and grid voltage data and grid frequency data transmitted to the dispatch control unit, a voltage low penetration dynamic response model is established, and the expression is:

[0087]

[0088] Among them, P dy (t) is the power command after dynamic adjustment at time t, P0 is the initial power setting value, Represents the fast response term to voltage deviation, V ref is the set voltage reference value, V(t) is the current collected voltage value, α is the voltage adjustment sensitivity coefficient, is the integral response term to the frequency change rate, β is the frequency adjustment gain coefficient, f g (t′) is the grid frequency at time t′, is the grid frequency f g (t′) is the partial derivative with respect to time t′, where t is the current time, t′ is the dynamic time of voltage breakdown, t0 is the starting point of the integration time, and d is the symbol of total differential.

[0089] It should be noted that Used to quickly respond to voltage changes, when the voltage is lower than the reference value, the model can quickly adjust the power command. The historical impact of frequency changes is calculated by integration, and the power output is gradually adjusted to stabilize the grid frequency.

[0090] S3.2: According to the different stages of the voltage breakdown state, optimize the parameters α and β to enhance the response effect.

[0091] Specifically, the following steps are included:

[0092] S3.2.1: When voltage breakdown occurs in the early stage, adjust the parameters α and β to achieve rapid recovery;

[0093] Specifically, when the grid voltage is just below the threshold (enters the low-breakdown state), the parameter α needs to be increased to increase the voltage deviation response sensitivity, and the frequency response term β also needs to be increased to ensure a rapid recovery of the grid frequency.

[0094] S3.2.2: When the voltage gradually recovers, adjust the parameters α and β to balance the voltage recovery and frequency support;

[0095] Specifically, when the voltage gradually recovers to a normal range, the sensitivity coefficient α and the frequency gain coefficient β are gradually reduced to avoid over-regulation and excessive power fluctuation.

[0096] S3.2.3: When the voltage returns to the normal range, adjust the parameters α and β to correct the power output;

[0097] Specifically, when the voltage returns to a normal range, the frequency sensitivity coefficient α and the frequency gain coefficient β are reduced to avoid over-response.

[0098] S3.3: P output according to the voltage low wear dynamic response model dy (t), combined with the actual power demand, the multivariable control algorithm is enabled for coordinated regulation, and the expression is:

[0099]

[0100] Among them, P out (t) is the output power command, γ is the power dynamic allocation weight, P dy (t) is the power command after dynamic adjustment at time t, κ is the frequency deviation correction coefficient, f ref is the reference frequency, f g (t′) is the real-time collected grid frequency, t is the current time, t′ is the dynamic time of voltage low penetration, and t0 is the starting point of the integration time.

[0101] It should be noted that when the power command is adjusted, the power grid frequency and the reference frequency f ref The power command is weighted and adjusted by the frequency deviation correction coefficient κ. This helps to balance the impact of the voltage low penetration state on the power output. According to the severity and duration of the voltage low penetration, the γ value is dynamically adjusted to flexibly balance the proportion of voltage and frequency response.

[0102] S3.4: Send the output power command to the converter controller to control the power output.

[0103] Specifically, the calculated output power command P out (t) includes dynamic correction of voltage and frequency response. The power command is sent to the converter controller through the RS485 protocol control interface. The converter adjusts the output according to the received power command to ensure that the power matches the grid frequency and stability requirements. The converter control module monitors the output changes in real time to ensure that it is within the normal range.

[0104] Preferably, through a sophisticated dynamic response model, adaptive parameter adjustment, multivariable control algorithm, and real-time power command transmission, the model can quickly and stably adjust power output when voltage drops and frequency fluctuates in the power grid. While ensuring the stability of the power grid, it reduces the risk of over-response and power fluctuations, and improves the overall regulation and emergency response capabilities of the power grid.

[0105] S4: Based on the grid frequency data and the adjusted power command, the virtual synchronous machine control algorithm is started to adjust the power output.

[0106] Specifically, the following steps are included:

[0107] S4.1: Simulate the inertia and damping characteristics of the synchronous generator to obtain the inertia and damping compensation power, the expression is:

[0108]

[0109] Among them, ΔP sync (t) is the inertia and damping compensation power, is the angular velocity change rate of the virtual synchronous machine, J is the equivalent inertia coefficient, D is the damping coefficient, ω(t) is the instantaneous angular velocity of the virtual synchronous machine, ω ref is the reference angular velocity.

[0110] Specifically, when the grid frequency fluctuates, the virtual synchronous machine will simulate the inertial response and offset the instantaneous frequency fluctuation through the inertial compensation power to avoid the sharp fluctuation of the grid frequency. The damping part is used to balance the overshoot and oscillation of the frequency and quickly restore the frequency to normal levels. This compensation function makes the virtual synchronous machine similar to a traditional generator in the short term, which can provide fast power adjustment and mitigate grid fluctuations.

[0111] S4.2: Based on the output power command and the inertia and damping compensation power, the final output power command is obtained, and the expression is:

[0112] P fin (t) = P out (t)+ΔP sync (t);

[0113] Among them, P fin (t) is the final output power command, which is the power command after combining inertia and damping compensation. out (t) is the output power command, which indicates the power output after the voltage dropout model. ΔP sync (t) is the inertia and damping compensation power, which can be quickly adjusted when the grid frequency changes to ensure grid stability.

[0114] S4.3: Calculate error data, design feedback optimization controller, and integrate feedback power compensation.

[0115] It should be understood that by calculating the power error and the frequency error, the model can optimize the power compensation to ensure that the final power output matches the actual demand of the grid.

[0116] Specifically, the following steps are included:

[0117] S4.3.1: Calculate the power error and frequency error using the following expressions:

[0118] ΔP(t)=P fin (t)-P act (t),Δf(t)=f ref -f g (t);

[0119] Among them, ΔP(t) is the power error, Δf(t) is the frequency error, P act (t) is the actual power grid power, f g (t) is the actual grid frequency.

[0120] Furthermore, calculating the power error ΔP(t) can help determine whether the module provides enough power to meet the grid demand. Calculating the frequency error Δf(t) can detect whether the grid frequency deviates from the reference frequency and further derive the required frequency compensation. Based on the power error and frequency error, the control strategy of the virtual synchronous machine can be further adjusted to ensure the consistency of the output power with the grid demand.

[0121] S4.3.2: Based on the error data, design a feedback optimization controller, expressed as:

[0122]

[0123] Where N is the proportional gain coefficient, M is the integral gain coefficient, ΔP fb (t) is the feedback power compensation.

[0124] Furthermore, the proportional term N·ΔP(t) directly responds to the current power error, allowing the power output to be adjusted quickly to reduce the error. The long-term error is accumulated and compensated, deviation is eliminated and accuracy is improved. The feedback controller can enhance stability and reduce grid frequency fluctuation and power fluctuation.

[0125] S4.3.3: Calculate the feedback compensation P according to the final output power command and feedback power compensation last (t), P last (t) is sent to the converter controller to drive the power output.

[0126] Specifically, calculate the feedback compensation P last (t) is expressed as,

[0127] P last (t) = P fin (t)+ΔP fb (t);

[0128] By calculating the feedback compensation P last(t), the module can further optimize the power output command according to the actual state of the power grid.

[0129] Preferably, by simulating inertia and damping response, drastic changes in grid frequency fluctuations are avoided, and the grid's anti-interference ability is enhanced. Real-time power error feedback control ensures that the power output always meets grid demand, avoids power over-regulation or under-regulation, and improves the accuracy of grid frequency and power. The multi-dimensional adjustment mechanism works together to ensure that the grid can quickly and smoothly restore a stable state in a dynamic environment, improving the grid's adaptive ability. By introducing the virtual synchronous machine control algorithm, the grid's stability and power regulation accuracy are effectively enhanced, providing an accurate and effective grid power regulation solution.

[0130] S5: Use the reinforcement learning algorithm to coordinate the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm to restore the power grid state.

[0131] Specifically, the following steps are included:

[0132] S5.1: Define the grid state space S t and action space A t .

[0133] Specifically, the state space S t Represents various state variables of the power grid, describing the current working state of the power grid, including the following elements: frequency error Δf(t), power error ΔP(t), voltage error ΔV(t), frequency change rate The current power output of the power grid P(t) and the voltage of the power grid V(t).

[0134] State space S t A multi-dimensional vector is formed, which integrates the real-time operation data of the power grid and reflects the current health status of the power grid.

[0135] Action Space A t Represents the different actions that the agent can choose in the current state. Each action in the action space will determine the adjustment direction and amplitude of each parameter in the power grid control strategy.

[0136] The main parameters include: sensitivity coefficient α for adjusting power command, frequency response gain coefficient β, power dynamic allocation weight γ, frequency deviation correction coefficient k, inertia coefficient J of virtual synchronous machine, and damping coefficient D of virtual synchronous machine.

[0137] Action Space A t It is expressed as:

[0138] A t =(α,β,γ,κ,J,D);

[0139] The action space determines how to adjust the voltage low penetration adaptive response strategy and virtual synchronous machine control strategy at each moment through discrete or continuous adjustment of these parameters to achieve smooth operation of the power grid.

[0140] S5.2: Design the reward function, expressed as:

[0141]

[0142] Among them, Δf(t) is the frequency error, ΔP(t) is the power error, ΔV(t) is the voltage error, is the frequency change rate, λ1 is the power error weight coefficient, λ2 is the voltage error weight coefficient, and λ3 is the frequency change rate weight coefficient.

[0143] Furthermore, the frequency error penalty term -|Δf(t)|: Through the frequency error penalty, the reward function encourages the agent to minimize the grid frequency deviation. Excessive frequency error will reduce the reward and prompt the converter to adjust the frequency.

[0144] Power error term -λ1·|ΔP(t)|: The penalty term for power error encourages the converter to adjust the power output to make it as close to the actual demand of the grid as possible. The weight coefficient λ1 controls the impact of the power error.

[0145] Voltage error penalty term -λ2·|ΔV(t)|: Voltage error affects stability. The reward function penalizes the voltage error to encourage the converter to optimize voltage regulation and ensure that the voltage is as stable as possible within the ideal range.

[0146] Frequency change rate penalty The frequency change rate reflects the speed of change of the grid frequency. Too fast a frequency change may lead to instability. By penalizing the frequency change rate, the converter is encouraged to smooth the grid frequency change and avoid rapid frequency fluctuations.

[0147] λ1, λ2, and λ3 are weight coefficients for adjusting the influence of different error terms. Reasonable setting of these coefficients can balance the importance of various control objectives. For example, in the case of frequent voltage low penetration or frequency fluctuation, the weights of voltage error and frequency error can be increased to focus on the stability of voltage and frequency.

[0148] S5.3: Select the PPO reinforcement learning algorithm and perform reinforcement learning training.

[0149] Specifically, the reinforcement learning training process is:

[0150] Initialization: Initialize the neural network model and set a random initial strategy.

[0151] Simulation training: In a simulation environment, actions are selected based on the current policy and adjusted based on feedback from the grid (reward function).

[0152] Strategy update: Use the PPO algorithm to optimize the current strategy, update the strategy to obtain higher rewards, and gradually learn the optimal control parameter adjustment method.

[0153] Continuous iteration: Through continuous experimentation and learning, the model gradually improves its ability to adjust strategies under different grid conditions, and ultimately is able to optimize grid stability under various complex situations.

[0154] S5.4: After the training is completed, the model is deployed and the various parameters in the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm are adjusted in real time and dynamically.

[0155] Specifically, the reinforcement learning model will adjust the parameters α, β, γ, and k to make the power output more responsive when the voltage drops. By adjusting the inertia coefficient J and the damping coefficient D, the inertia and damping characteristics of the synchronous generator are simulated to quickly respond to frequency fluctuations, thus realizing the coordinated regulation of the voltage drop adaptive response strategy and the virtual synchronous machine control algorithm.

[0156] Preferably, the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm are coordinated and adjusted through the reinforcement learning algorithm, so as to realize the real-time optimization and dynamic control of the power grid state. The control strategy is automatically adjusted according to the real-time state of the power grid to ensure the stable operation of the power grid in a complex dynamic environment. By reducing the power, frequency and voltage errors and controlling the frequency change rate, the stability of the power grid is enhanced and the risk of power grid instability is reduced. Through the reinforcement learning model, the model can continuously optimize and adapt itself to improve the stability of the power grid in long-term operation. Automatic adjustment reduces manual intervention and improves control accuracy and response speed. It can cope with emergencies and complex environments in the power grid and ensure that the power grid maintains stable operation under changing conditions. An efficient and intelligent power grid control strategy is provided.

[0157] This embodiment also provides a computer device, which is suitable for the case of a transient frequency support control method taking into account the voltage low-thru characteristics of the converter, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the transient frequency support control method taking into account the voltage low-thru characteristics of the converter as proposed in the above embodiment.

[0158] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0159] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the transient frequency support control method taking into account the voltage low-breakthrough characteristics of the converter as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0160] In summary, the present invention achieves the following goals: The present invention utilizes high-precision sensors to collect grid voltage and frequency data in real time, and provides dynamic feedback for subsequent control. By detecting the voltage low-through characteristics of the converter, the abnormal state of the grid is identified in real time, and the adaptive response strategy is started to adjust the power command to quickly respond to grid changes. Combined with the virtual synchronous machine control algorithm, the characteristics of the synchronous generator are simulated, the inertia and damping support of the grid are enhanced, and the frequency stability is further improved. The response strategy and the control algorithm are collaboratively optimized through the reinforcement learning algorithm, so that the algorithm can dynamically adjust the control parameters and improve the adaptive ability of the grid under complex working conditions. It effectively improves the role of the converter in grid frequency regulation, enhances the stability and reliability of the grid frequency, and has good technical feasibility and potential application value.

[0161] Example 2, referring to Table 1, is the second example of the present invention. To further verify the technical solution of the present invention, experimental simulation data of a transient frequency support control method taking into account the voltage low penetration characteristics of the converter are provided.

[0162] This experiment designed a dynamic power adjustment strategy based on voltage dropout and frequency fluctuation. Through refined models and adaptive control, the power output of the power grid can be adjusted in real time. First, a dynamic response model is established based on the voltage and frequency data of the power grid, and the power command is quickly adjusted to cope with the impact of voltage dropout. At the same time, a multivariable control algorithm is used to dynamically optimize power distribution according to the changes in voltage and frequency to ensure the stability of the power grid frequency and voltage. By adjusting the control parameters, the algorithm avoids over-response and power fluctuations, and improves the regulation and emergency response capabilities of the power grid. Finally, the power command is accurately transmitted through the converter controller to ensure that the power grid can quickly and smoothly adjust the power output when voltage dropout and frequency fluctuation occur.

[0163] The sensor collected data is shown in Table 1 below:

[0164] Table 1 Sensor data collection table

[0165]

[0166] In addition, set the voltage reference value V ref =380V, initial power P0=1000W.

[0167] The voltage low break-through threshold is set to 80% of the voltage reference value, that is, 304V.

[0168] The voltage recovery threshold is 90% of the voltage reference value, that is, 342V.

[0169] The adjustment rules for setting the voltage sensitivity coefficient α and the frequency gain coefficient β are:

[0170] Low voltage ride-through: When V(t) < 304V, α = 0.10 and β = 0.05.

[0171] Voltage recovery: When V(t) > 342V, α = 0.05 and β = 0.02.

[0172] Voltage within the recovery range 304V < V(t) < 342V: For every 10V increase, α and β each decrease by 0.01.

[0173] The adjusted data is specifically shown in Table 2 below:

[0174] Table 2 Coefficient Adjustment Table

[0175]

[0176] According to the low voltage ride-through dynamic response model, calculate the adjusted power command P dy (t), and the expression is:

[0177]

[0178] The specific process is as follows:

[0179] Calculate the voltage deviation response term

[0180] Calculate the frequency change rate response term

[0181] Add the two terms together to obtain P dy (t).

[0182] Taking time t = 0 and t = 1 as examples, calculate P dy (t) at each time point.

[0183] Time t = 0 seconds:

[0184] Grid voltage V(t) = 379V; Grid frequency f g (t) = 50Hz; Frequency change rate

[0185] Voltage deviation response term:

[0186]

[0187] Frequency change rate response term:

[0188] (because t0 = 0 and the frequency change rate is 0);

[0189] Dynamically adjusted power command:

[0190] P dy(0)=1000·0.9512+0=951.2W;

[0191] Time t = 1 second:

[0192] Grid voltage V(t) = 305V; grid frequency f g (t) = 49.92 Hz; frequency change rate

[0193] Voltage deviation response term:

[0194]

[0195] Frequency change rate response term:

[0196]

[0197] Dynamically adjusted power command:

[0198] P dy (1)=1000·0.00187+(-0.00432)=1.87-0.00432≈1.8657W;

[0199] Other times are calculated according to the above example.

[0200] The calculated dynamically adjusted power instructions are shown in Table 3:

[0201] Table 3 Dynamically adjust power instruction table

[0202] Time (seconds) Grid voltage (V) Dynamically adjusted power command (W) 0 379 951.2 1 305 1.8657 2 300 0.3315 3 338 73.29928 4 348 201.9024 5 357 316.6018

[0203] After that, the multivariable control algorithm is enabled to coordinate the adjustment and calculate the output power command P out (t), the expression is:

[0204]

[0205] Set the reference frequency f ref =50Hz; power dynamic allocation weight γ=0.5; frequency deviation correction coefficient k=100.

[0206] Taking time t = 0 and t = 1 as an example, calculate P at each time point out (t).

[0207] Time t = 0 seconds:

[0208] P dy (0) = 951.2;

[0209] From t0 = 0 to t = 0, there is no change, so the integral is 0;

[0210] P out (0)=0.5·951.2+(1-0.5)·0=0.5Pout951.2+0=475.6W;

[0211] Time t = 1 second:

[0212] P dy (1) = 1.8657;

[0213] From t0=0 to t=1,

[0214]

[0215] P out (1) = 0.5 1.8657 + (1 - 0.5) 8 = 0.5 1.8657 + 0.5 8 = 0.93285 +

[0216] 4 = 4.93285W;

[0217] Other times are calculated according to the above example.

[0218] The calculated output power instructions are shown in Table 4:

[0219] Table 4 Output power instruction table

[0220]

[0221] The calculated P out (t) is sent to the converter controller based on the collected frequency data and the calculated P out (t), start the virtual synchronous machine control algorithm, continue to adjust the power output, and finally use the reinforcement learning algorithm to coordinate the parameters in the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm to restore the power grid state.

[0222] Through the above steps and data tables, the experiment can be used to control power output. Through a sophisticated dynamic response model, adaptive parameter adjustment, multivariable control algorithm, and real-time power command transmission, it ensures that the converter can quickly and stably adjust power output when the grid voltage drops and the frequency fluctuates. While ensuring the stability of the grid, it reduces the risk of over-response and power fluctuations, and improves the overall regulation and emergency response capabilities of the grid.

[0223] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A transient frequency support control method taking into account the voltage dropout characteristics of a converter, characterized in that: include: Collect grid voltage data and grid frequency data through real-time monitoring sensors; Based on the collected grid voltage data, the converter voltage low-breakthrough characteristic detection is performed to obtain the voltage low-breakthrough status; According to the voltage low penetration state, the adaptive response strategy is enabled to adjust the power command; Based on the grid frequency data and the adjusted power command, the virtual synchronous machine control algorithm is started to adjust the power output; The reinforcement learning algorithm is used to coordinate the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm to restore the power grid state.

2. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: The real-time monitoring sensor is used to collect grid voltage data and grid frequency data, specifically, The grid voltage data includes the instantaneous voltage value of each node in the grid, the fluctuation range and amplitude change of the grid voltage, and the instantaneous power of the grid; The grid frequency data includes the instantaneous frequency of the grid and the deviation of the grid frequency from the standard frequency.

3. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: The method of detecting the voltage low penetration characteristics of the converter based on the collected grid voltage data and obtaining the voltage low penetration state comprises the following specific steps: Set the voltage low-breakthrough threshold and voltage recovery threshold; Compare the voltage low-breakthrough threshold and the voltage recovery threshold with the grid voltage data to determine the voltage low-breakthrough state; The voltage low-break status is transmitted to the dispatch control unit.

4. The transient frequency support control method taking into account the voltage dropout characteristics of the converter according to claim 1, characterized in that: According to the voltage low penetration state, the adaptive response strategy is enabled to adjust the power instruction. The specific steps are: Based on the voltage low penetration state and grid voltage and frequency data transmitted to the dispatch control unit, a voltage low penetration dynamic response model is established, and the expression is: Among them, P dy (t) is the power command after dynamic adjustment at time t, P0 is the initial power setting value, Represents the fast response term to voltage deviation, V ref is the set voltage reference value, V(t) is the current collected voltage value, α is the voltage adjustment sensitivity coefficient, is the integral response term to the frequency change rate, β is the frequency adjustment gain coefficient, f g (t′) is the grid frequency at time t′, is the grid frequency f g (t′) is the partial derivative with respect to time t′, where t is the current time, t′ is the dynamic time of voltage breakdown, t0 is the starting point of the integration time, and d is the symbol of total differential; According to the different stages of voltage breakdown, the parameters α and β are optimized to enhance the response effect; According to the voltage low wear dynamic response model output P dy (t), combined with the actual power demand, the multivariable control algorithm is enabled for coordinated regulation, and the expression is: Among them, P out (t) is the output power command, γ is the power dynamic allocation weight, P dy (t) is the power command after dynamic adjustment at time t, κ is the frequency deviation correction coefficient, f ref is the reference frequency, f g (t′) is the real-time collected grid frequency, t is the current time, t′ is the dynamic time of voltage low penetration, and t0 is the starting point of the integration time; The output power command is sent to the converter controller to control the power output.

5. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: According to the different stages of the voltage breakdown state, the parameters α and β are optimized to enhance the response effect. The specific steps are: When voltage breakdown occurs in the early stage, adjust the parameters α and β to achieve rapid recovery; When the voltage gradually recovers, adjust the parameters α and β to balance the voltage recovery and frequency support; When the voltage returns to the normal range, adjust the parameters α and β to correct the power output.

6. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: Based on the grid frequency data, according to the adjusted power command, the virtual synchronous machine control algorithm is started to adjust the power output. The specific steps are: Simulate the inertia and damping characteristics of the synchronous generator to obtain the inertia and damping compensation power, the expression is: Among them, ΔP sync (t) is the inertia and damping compensation power, is the angular velocity change rate of the virtual synchronous machine, J is the equivalent inertia coefficient, D is the damping coefficient, ω(t) is the instantaneous angular velocity of the virtual synchronous machine, ω ref is the reference angular velocity; Based on the output power command and the inertia and damping compensation power, the final output power command is obtained, and the expression is: P fin (t)=P out (t)+ΔP sync (t); Calculate error data, design feedback optimization controller, and integrate feedback power compensation.

7. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: The specific steps of calculating error data, designing feedback optimization controller, and comprehensive feedback power compensation are as follows: Calculate the power error and frequency error, the expression is: ΔP(t)=P fin (t)-P act (t),Δf(t)=f ref -f g (t); Among them, ΔP(t) is the power error, Δf(t) is the frequency error, P act (t) is the actual power grid power, f g (t) is the actual grid frequency; Based on the error data, a feedback optimization controller is designed, and the expression is: Where N is the proportional gain coefficient, M is the integral gain coefficient, ΔP fb (t) is feedback power compensation; According to the final output power command and feedback power compensation, calculate the feedback compensation P last (t), P last (t) is sent to the converter controller to drive the power output.

8. The transient frequency support control method taking into account the voltage breakdown characteristics of the converter according to claim 1, characterized in that: The reinforcement learning algorithm is used to coordinate the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm to restore the grid frequency. The specific steps are as follows: Define the grid state space S t and action space A t ; Design the reward function, the expression is: Among them, Δf(t) is the frequency error, ΔP(t) is the power error, ΔV(t) is the voltage error, is the frequency change rate, λ1 is the power error weight coefficient, λ2 is the voltage error weight coefficient, and λ3 is the frequency change rate weight coefficient; Select the PPO reinforcement learning algorithm and conduct reinforcement learning training; After the training is completed, the model is deployed and the various parameters in the voltage low penetration adaptive response strategy and the virtual synchronous machine control algorithm are adjusted dynamically in real time.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the transient frequency support control method taking into account the voltage low-breakthrough characteristics of the converter as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the transient frequency support control method taking into account the voltage low-breakthrough characteristics of the converter as described in any one of claims 1 to 8 are implemented.

Images