Self-adaptive synchronous control method for grid-forming inverter with consideration of dynamic performance and stability

Abstract

The invention relates to the technical field of grid-connected inverter networking synchronization control, and relates to a networking inverter adaptive synchronization control method considering dynamic performance and stability. The method comprises the following steps: firstly, injecting a harmonic signal with a specific frequency into a power grid from an inverter side, acquiring a voltage and current signal corresponding to the harmonic signal with the specific frequency at a grid-connected point, and calculating to obtain power grid side impedance; then, converting a grid-connected inverter adaptive control optimization problem into a Markov decision process based on the power grid side impedance; and finally, substituting the Markov decision process into a reinforcement learning algorithm framework for reverse training to obtain an optimal control law of the grid-connected inverter based on hybrid synchronous control, and controlling the grid-forming inverter based on the optimal control law. The invention provides the self-adaptive synchronous control method for the grid-forming inverter, which gives consideration to the dynamic performance and the stability, so that an inverter grid-connected system can operate stably, has enough stability margin and also has good dynamic performance.

Description

Technical Field

[0001] The present application relates to the technical field of grid-connected inverter network synchronization control, and in particular to a grid-connected inverter adaptive synchronization control method that takes into account both dynamic performance and stability. Background Art

[0002] As the proportion of new energy sources connected to the grid increases, the proportion of power electronic equipment connected to the current new power system is also gradually increasing. Among them, the grid-connected inverter is the key equipment for the access of new energy sources to the grid, and the synchronous control technology that ensures the synchronous operation of the grid-connected inverter and the grid plays a vital role. At present, from the perspective of the types of synchronous control strategies adopted, grid-connected inverters mainly include grid-following inverters based on phase-locked synchronization and grid-building inverters based on power synchronization. However, due to the strong coupling characteristics between the synchronous loop of the grid-connected inverter and the grid impedance, different types of synchronous control loops have complex interactions with grids of different strengths, resulting in a series of synchronous instability / oscillation problems in the system.

[0003] At present, in response to the system instability problem caused by the interaction between the synchronization loop and the grid impedance, existing studies have pointed out that inverters based on phase-locked synchronization are prone to instability under weak grid conditions, while inverters based on power synchronization are prone to instability under strong grid conditions. In order to expand the applicability of the synchronization control strategy under different grid strengths, some studies have further proposed a hybrid synchronization control strategy combining phase-locked synchronization and power synchronization, and verified through simulation that the inverter grid-connected system based on this control strategy has better stability under a wide range of grid strengths. Subsequent studies have further analyzed the stability of the inverter grid-connected system based on hybrid synchronization control from a theoretical level. Other studies have established the second-order synchronization characteristic equation of the inverter grid-connected system based on hybrid synchronization control, pointing out that the phase-locked loop branch will increase the equivalent damping of the system, thereby improving the transient stability of the converter. Some studies have also analyzed the local sensitivity of different control parameters to the small disturbance oscillation mode of the system based on the impedance model. However, under the premise of ensuring the stable operation of the system under a wide range of grid strengths, the selection rules of the hybrid synchronization control parameters of the grid-connected inverter based on hybrid synchronization control are still relatively complex, and the selected parameters are difficult to take into account the dynamic performance constraints of the system.

[0004] Therefore, in the relevant technology, there is an urgent need for a method that can online adjust the synchronous control parameters of the grid-connected inverter based on hybrid synchronous control, so that the grid-connected inverter has sufficient stability margin under a wide range of grid strength, while taking into account the dynamic performance constraints of the system, to achieve adaptive hybrid synchronous control of the grid-connected inverter. Summary of the invention

[0005] Based on this, it is necessary to provide an adaptive synchronous control method for grid-connected inverters that takes into account both dynamic performance and stability and can online adjust the synchronous control parameters of the grid-connected inverters based on hybrid synchronous control to address the above technical problems.

[0006] In a first aspect, the present application provides a method for adaptive synchronization control of a grid-connected inverter that takes into account both dynamic performance and stability. The method comprises: A harmonic signal of a specific frequency is injected into the grid from the inverter side, and a voltage and current signal corresponding to the harmonic signal of the specific frequency is collected at the grid connection point to calculate the grid side impedance; Based on the grid-side impedance, the grid-connected inverter adaptive control optimization problem is converted into a Markov decision process; The Markov decision process is substituted into the reinforcement learning algorithm framework for reverse training to obtain the optimal control law of the grid-connected inverter based on hybrid synchronous control, and the grid-connected inverter is controlled based on the optimal control law.

[0007] Optionally, in one embodiment of the present application, the specific formula for injecting a harmonic signal of a specific frequency from the inverter side into the power grid is:

[0008] in, It represents the modulation signal of the grid-connected inverter after adding the harmonic signal of a specific frequency. Indicates the modulation signal of the original grid-connected inverter. Represents a harmonic signal of a specific frequency, Indicates the ratio of the injected harmonic signal frequency to the rated frequency.

[0009] Optionally, in one embodiment of the present application, the calculation formula of the grid side impedance is:

[0010] in, is the grid impedance complex vector at the injected harmonic frequency, is the harmonic voltage at the grid connection point, is the grid-side harmonic voltage, is the harmonic current on the grid side, is the resistance on the grid side, is the inductance on the grid side, is an imaginary unit, for Nominal angular frequency of the subharmonic signal.

[0011] Optionally, in one embodiment of the present application, converting the grid-connected inverter adaptive control optimization problem into a Markov decision process based on the grid-side impedance includes: Define the state space of the Markov decision process based on the grid-side inductance value; Define the action space of Markov decision process based on the control parameters of hybrid synchronous loop; defining a reward space, and determining a reward value of a Markov decision process based on the state space and the action space; Define the state transition probability and reward discount rate, and form a replay experience pool based on the state space, action space and reward space; An evaluation value is obtained based on the grid-side inductance value in the state space and the hybrid synchronous loop control parameter in the action space, and the quality of the hybrid synchronous loop control parameter is determined based on the evaluation value.

[0012] Optionally, in one embodiment of the present application, the specific formula of the reward space is:

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019] in, The corresponding elements of a decision cycle, and They are The corresponding reward and penalty values, Substitute the characteristic roots after action in different states The damping ratio, are all characteristic roots of the inverter grid-connected system based on hybrid synchronous control, is the characteristic root located above the real axis among the first pair of dominant characteristic roots of the system, is the characteristic root located above the real axis in the second pair of dominant characteristic roots of the system, Characteristic root With characteristic root The expected value of the real distance, Characteristic root With characteristic root The opposite of the expected value of the imaginary axis distance.

[0020] Optionally, in one embodiment of the present application, substituting the Markov decision process into a reinforcement learning algorithm framework for reverse training includes: Randomly extract training samples from the playback experience pool; Determining a value network based on the training samples; Updating current network parameters by gradient descent based on the loss function of the value network; The network parameters of the neural network are updated by gradient descent based on the loss function of the policy network; The temperature coefficient representing the importance of entropy to the reward is updated by gradient ascent; The target value network is updated using a soft update method, and the control strategy is updated based on the updated target value network.

[0021] The above-mentioned adaptive synchronous control method for grid-connected inverters taking into account both dynamic performance and stability, first, injects a harmonic signal of a specific frequency into the grid from the inverter side, collects the voltage and current signals corresponding to the harmonic signal of the specific frequency at the grid-connected point, and calculates the grid-side impedance; then, based on the grid-side impedance, the adaptive control optimization problem of the grid-connected inverter is converted into a Markov decision process; finally, the Markov decision process is substituted into the reinforcement learning algorithm framework for reverse training to obtain the optimal control law of the grid-connected inverter based on hybrid synchronous control, and the grid-connected inverter is controlled based on the optimal control law. In other words, an adaptive synchronous control method for grid-connected inverters taking into account both dynamic performance and stability is proposed, which is used to realize the online adjustment of the synchronous control parameters in the hybrid synchronous loop of the grid-connected inverter. The adaptive control can avoid the complex process of selecting the hybrid synchronous control parameters under different grid strengths, and under a wide range of grid strengths, it can ensure that the inverter grid-connected system can not only operate stably, but also have sufficient stability margin and good dynamic performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 An application environment diagram of a grid-connected inverter adaptive synchronization control method that takes into account both dynamic performance and stability in one embodiment; Figure 2 It is a flow chart of an adaptive synchronization control method of a grid-connected inverter that takes into account both dynamic performance and stability in one embodiment; Figure 3 A schematic diagram of the structure of a hybrid synchronous inverter grid-connected system in one embodiment; Figure 4 is a schematic diagram of an adaptive control strategy in one embodiment; Figure 5 It is a schematic diagram of the test results of the optimal control rate corresponding to the damping ratio under different grid-side impedances in one embodiment; Figure 6A simulation waveform diagram of an adaptive control system in one embodiment, (a) output value of an adaptive control module, (b) power simulation waveform of a grid-connected system; Figure 7 The system simulation waveforms using different control strategies in one embodiment, (a) under a strong power grid, (b) under a weak power grid; Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0024] The grid-connected inverter adaptive synchronization control method that takes into account both dynamic performance and stability provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or it can be placed on the cloud or other network servers. Among them, the terminal can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0025] In one embodiment, Figure 2 As shown in the figure, a grid-connected inverter adaptive synchronization control method is provided that takes into account both dynamic performance and stability. Figure 1 The server in the example is used to illustrate the following steps: S201: injecting a harmonic signal of a specific frequency into the power grid from the inverter side, collecting voltage and current signals corresponding to the harmonic signal of the specific frequency at the grid connection point, and calculating the grid side impedance.

[0026] In the present application embodiment, first, Figure 3 As shown in the figure, it is a schematic diagram of the structure of the hybrid synchronous inverter grid-connected system. The grid can be composed of a three-phase Thevenin equivalent voltage source and impedance It indicates that, and In addition, the strength of the power grid can be described by the short circuit ratio (SCR), where 5 and 1 represent strong and weak power grids respectively. Assuming that the grid-side line is purely inductive, by changing Changing SCR, the range is [0.0098H, 0.046H]. The hardware circuit of the inverter is composed of switch tubes S1-S6 and LC filter. The inverter control strategy based on hybrid synchronization includes hybrid synchronization loop, AC voltage loop, current loop and pulse width modulation module. Among them, the hybrid synchronization loop is the main difference between this control strategy and other control strategies. The hybrid synchronization loop consists of a phase-locked loop synchronization link and a power synchronization link. q The sum of the outputs of the shaft voltage feedback signal and the power feedback signal is used as the inverter phase angle signal The superscript “ s "and" c "To distinguish the system dq Coordinate Systems and Controls dq Variables in a coordinate system.

[0027] In one embodiment of the present application, the specific formula for injecting a harmonic signal of a specific frequency from the inverter side into the power grid is:

[0028] in, It represents the modulation signal of the grid-connected inverter after adding the harmonic signal of a specific frequency. Indicates the modulation signal of the original grid-connected inverter. Represents a harmonic signal of a specific frequency, Indicates the ratio of the injected harmonic signal frequency to the rated frequency.

[0029] The voltage and current signals corresponding to the harmonic signals of the specific frequency are collected at the grid connection point. The specific formula is:

[0030]

[0031]

[0032]

[0033]

[0034]

[0035] in, is the grid voltage, is the grid-side current, is the harmonic voltage at the grid connection point, is the harmonic current on the grid side, and are the positive and negative sequence voltages at the grid connection point, and are the positive and negative sequence currents on the grid side, is the Laplace operator, is an imaginary unit, for The cutoff angular frequency of the sub-specific frequency harmonic signal, for The rated angular frequency of the subharmonic signal, is the cutoff frequency of the complex coefficient filter, is the rated angular frequency of the complex coefficient filter.

[0036] Afterwards, the complex vector of the grid-side impedance is calculated from the collected harmonic signals. In one embodiment of the present application, the calculation formula of the grid-side impedance is:

[0037] in, is the grid impedance complex vector at the injected harmonic frequency, is the harmonic voltage at the grid connection point, is the grid-side harmonic voltage, is the harmonic current on the grid side, is the resistance on the grid side, is the inductance on the grid side, is an imaginary unit, for Nominal angular frequency of subharmonic signal.

[0038] Considering that the grid side is an ideal voltage source, the specific grid side resistance and inductance are calculated from the 𝛼𝛽 components of each harmonic signal. The specific formula is as follows:

[0039] in, and The grid voltage is Quantity and Quantity, and The grid-side current is Quantity and Quantity.

[0040] S203: Converting the grid-connected inverter adaptive control optimization problem into a Markov decision process based on the grid-side impedance.

[0041] In an embodiment of the present application, the grid-connected inverter adaptive control optimization problem is transformed into a Markov decision process to collect reward values ​​for different actions taken by the inverter under different grid-connected system states.

[0042] Specifically, in one embodiment of the present application, converting the grid-connected inverter adaptive control optimization problem into a Markov decision process based on the grid-side impedance includes: S301: Define a state space of a Markov decision process based on the grid-side inductance value.

[0043] In one embodiment of the present application, the state space of the Markov decision process is defined as S , the specific formula is:

[0044] in, For the The state of the inverter in a decision cycle, For the The grid-side inductance value calculated by the inverter impedance measurement module within a decision cycle.

[0045] S303: Defining an action space of a Markov decision process based on the hybrid synchronization loop control parameters.

[0046] In one embodiment of the present application, the action space of the Markov decision process is defined as A , the specific formula is:

[0047] in, For the The inverter operates within a decision cycle. and Respectively The proportional coefficient setting value of the power synchronization loop and the phase-locked loop in the inverter hybrid synchronization loop within a decision cycle.

[0048] S305: Define a reward space, and determine a reward value of a Markov decision process based on the state space and the action space.

[0049] In one embodiment of the present application, the return space R The specific formula is:

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056] in, The corresponding elements of a decision cycle, and They are The corresponding reward and penalty values, Substitute the characteristic roots after action in different states The damping ratio, are all characteristic roots of the inverter grid-connected system based on hybrid synchronous control, is the characteristic root located above the real axis among the first pair of dominant characteristic roots of the system, is the characteristic root located above the real axis in the second pair of dominant characteristic roots of the system, Characteristic root With characteristic root The expected value of the real distance is 20 here. Characteristic root With characteristic root The opposite of the expected value of the imaginary axis distance, here -12.

[0057] S307: Define the state transition probability and reward discount rate, and form a replay experience pool based on the state space, action space and reward space.

[0058] S309: Obtaining an evaluation value based on the grid-side inductance value in the state space and the hybrid synchronous loop control parameter in the action space, and determining the quality of the hybrid synchronous loop control parameter based on the evaluation value.

[0059] In one embodiment of the present application, the state transition probability of the Markov decision process is defined as P , that is, the grid-connected system is in state and actions In the case of Enter The probability of Due to the grid-side inductance It will not change due to the change of the inverter control parameters. The Markov decision process in the present invention is a single-step process. Define the reward discount rate γ of the Markov decision process. The reward discount rate of the Markov decision process in the present invention is 0. The specific formula is:

[0060] in, To end the The cumulative discounted return of the decision cycle, The corresponding elements of a decision cycle.

[0061] The tuple Store in replay experience pool D.

[0062] Assume that the inverter grid-connected system is in a certain state When the grid-side inductance is , inverter input hybrid synchronous loop control parameters , then the grid-connected system returns the evaluation value according to the input parameters , and will not return the next state value. Determine the quality of the input synchronization loop control parameters under the grid strength. The inverter repeats the above process and finally determines an optimal control strategy that can obtain the maximum reward in a given state action space. , the specific formula is:

[0063] in, is the optimal control strategy that can obtain the maximum reward in a given state-action space.

[0064] S205: Substituting the Markov decision process into a reinforcement learning algorithm framework for reverse training to obtain an optimal control law for a grid-connected inverter based on hybrid synchronous control, and controlling the grid-connected inverter based on the optimal control law.

[0065] In the embodiment of the present application, the constructed Markov decision process model is substituted into the corresponding reinforcement learning SoftActor Critic algorithm framework for offline training to finally obtain the optimal control law of the inverter.

[0066] Specifically, in one embodiment of the present application, substituting the Markov decision process into a reinforcement learning algorithm framework for reverse training includes: S401: randomly extracting training samples from the playback experience pool.

[0067] S403: Determine a value network based on the training samples.

[0068] S405: Update current network parameters by gradient descent based on the loss function of the value network.

[0069] S407: Update the network parameters of the neural network by gradient descent based on the loss function of the policy network.

[0070] S409: Update the temperature coefficient representing the importance of entropy to the reward by gradient ascent.

[0071] S411: Use a soft update method to update the target value network, and update the control strategy based on the updated target value network.

[0072] In one embodiment of the present application, first, a small batch of samples is randomly selected from the playback experience pool D, and then the two current value networks Q in the Critic network are calculated by the Bellman general recursive equation. The specific formula is:

[0073] in, It is the current value network in the Critic network of the Soft Actor Critic algorithm.

[0074] After that, the current network parameters are updated by gradient descent through the expression of the loss function of the current value network , the specific formula is:

[0075] in, is the current network parameter, For Take action in state The reward value, is the loss function of the value network.

[0076] Afterwards, the neural network is updated by gradient descent through the loss function of the policy network Network parameters , and select two in the current network The smaller value of the network output value, the specific formula is:

[0077]

[0078] in, For neural networks The network parameters, is random noise that follows a standard normal distribution, is the loss function of the policy network, for Network 1 value assessment value, for Network 2 value assessment value, For two The smaller value of the network output value, is the temperature coefficient, For the strategy network.

[0079] Afterwards, the temperature coefficient representing the importance of entropy to the reward is updated by gradient ascent , the specific formula is:

[0080] in, is the temperature coefficient representing the importance of entropy to the reward, is the average entropy, A target strategy.

[0081] Afterwards, the target value network Soft is updated through the soft update method Q , the specific formula is:

[0082] in, is the soft update rate.

[0083] According to the new Soft Q Network, by old strategy New Strategy Update, the specific formula is:

[0084] in, For the old strategy, For new strategies, For measurement and KL divergence of distribution differences, for The strategy adopted in the state Used to normalize the distribution.

[0085] The above process is repeated continuously, and finally the optimal control law of the grid-connected inverter based on hybrid synchronous control within the range of grid strength SCR∈[1,5] is obtained through offline training. The optimal control rate obtained by offline training is substituted into the adaptive control module. When the measured grid-side inductance value changes, after a delay period and the identification signal converges, it is substituted into the optimal control law, and the corresponding control parameters are read to update the controller, finally realizing the adaptive stable control of the inverter.

[0086] In the above-mentioned adaptive synchronous control method for grid-connected inverters that takes into account both dynamic performance and stability, first, a harmonic signal of a specific frequency is injected into the grid from the inverter side, and the voltage and current signals corresponding to the harmonic signal of the specific frequency are collected at the grid-connected point to calculate the grid-side impedance; then, the adaptive control optimization problem of the grid-connected inverter is converted into a Markov decision process based on the grid-side impedance; finally, the Markov decision process is substituted into the reinforcement learning algorithm framework for reverse training to obtain the optimal control law of the grid-connected inverter based on hybrid synchronous control, and the grid-connected inverter is controlled based on the optimal control law. In other words, an adaptive synchronous control method for grid-connected inverters that takes into account both dynamic performance and stability is proposed, which is used to realize the online adjustment of the synchronous control parameters in the hybrid synchronous loop of the grid-connected inverter. The adaptive control can avoid the complex process of selecting the hybrid synchronous control parameters under different grid strengths, and under a wide range of grid strengths, it can ensure that the inverter grid-connected system can not only operate stably, but also have sufficient stability margin and good dynamic performance.

[0087] The following is a specific example to illustrate the specific implementation steps of the grid-connected inverter adaptive synchronization control method of the present application that takes into account both dynamic performance and stability. As shown in Tables 1, 2 and 3 below, they are specific parameters of the example.

[0088] Table 1 Inverter grid-connected system parameters

[0089] Table 2 Basic parameters of the grid-side impedance measurement module

[0090] Table 3 Algorithm framework simulation parameters

[0091] like Figure 4 As shown in the figure, it is a schematic diagram of the adaptive control strategy. The adaptive control strategy of the grid-connected inverter based on grid impedance measurement is divided into two parts: offline training and online application. The Markov decision process is as follows: Assuming that the inverter grid-connected system is in a certain state When the grid-side inductance is , inverter input hybrid synchronous loop control parameters , then the grid-connected system returns the evaluation value according to the input parameters , and will not return the next state value. Determine the quality of the input synchronization control parameters under the strength of the power grid. The inverter repeats the above process continuously, and finally determines an optimal control strategy to obtain the maximum reward in the given state action space. Then, the constructed Markov decision process model is substituted into the corresponding Soft Actor Critic algorithm framework for offline training to finally obtain the optimal control law of the inverter. Then, the optimal control rate obtained by offline training is substituted into the adaptive control module. After the inverter side injects a harmonic signal of a specific frequency into the power grid, the voltage and current signals of this frequency are collected at the grid connection point, and the grid-side inductance is identified based on this. When the measured grid-side inductance value changes, after a delay period and the identification signal converges, it is substituted into the optimal control law, and the corresponding control parameters are read to update the controller, and finally the adaptive stable control of the inverter is realized.

[0092] like Figure 5 As shown in the figure, the test results of the optimal control rate corresponding to the damping ratio under different grid-side impedances are shown. It can be seen that under different grid strengths, the dominant characteristic root λ 1 , 2 The damping ratios are between 0.65 and 0.81, indicating that the inverters using this control ratio have good dynamic performance.

[0093] like Figure 6 As shown, this is the simulation waveform of the adaptive control system. Figure 6 (a) Figure 6 (b) are the output value of the adaptive control module and the grid-connected system power simulation waveform, respectively. Figure 6 (a) It can be seen that according to the adaptive control process and the optimal control rate obtained by offline training, the inverter can achieve control parameters under different grid strengths. and Adaptive selection of Figure 6 (b) It can be seen that the adaptive control strategy proposed in the present invention can enable the grid-connected inverter based on hybrid synchronous control to have stable grid-connected capability under a wide range of grid strengths.

[0094] like Figure 7 As shown, Figure 7 (a) and Figure 7 (b) The system simulation waveforms of fixed parameter control and adaptive control under strong and weak power grids, respectively. The parameters selected in the fixed parameter control are , and , As can be seen from the figure, under strong and weak power grids, when the power changes suddenly from 7.5kW to 9.3kW, compared with fixed parameter control, the dynamic characteristics of the grid-connected inverter based on hybrid synchronous control using adaptive control have smaller adjustment time and overshoot, and its output dynamic characteristics are optimal.

[0095] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0096] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, 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, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for adaptive synchronization control of a grid-connected inverter that takes into account dynamic performance and stability is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0097] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0098] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0099] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0100] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability, characterized in that: The method comprises: A harmonic signal of a specific frequency is injected into the grid from the inverter side, and a voltage and current signal corresponding to the harmonic signal of the specific frequency is collected at the grid connection point to calculate the grid side impedance; Based on the grid-side impedance, the grid-connected inverter adaptive control optimization problem is converted into a Markov decision process; The Markov decision process is substituted into the reinforcement learning algorithm framework for reverse training to obtain the optimal control law of the grid-connected inverter based on hybrid synchronous control, and the grid-connected inverter is controlled based on the optimal control law.

2. The grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability according to claim 1 is characterized in that: The specific formula for injecting a harmonic signal of a specific frequency from the inverter side into the power grid is: in, It represents the modulation signal of the grid-connected inverter after adding the harmonic signal of a specific frequency. Indicates the modulation signal of the original grid-connected inverter. Represents a harmonic signal of a specific frequency, Indicates the ratio of the injected harmonic signal frequency to the rated frequency.

3. The grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability according to claim 1, characterized in that: The calculation formula of the grid side impedance is: in, is the grid impedance complex vector at the injected harmonic frequency, is the harmonic voltage at the grid connection point, is the grid-side harmonic voltage, is the harmonic current on the grid side, is the resistance on the grid side, is the inductance on the grid side, is an imaginary unit, for Nominal angular frequency of the subharmonic signal.

4. The grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability according to claim 1 is characterized in that: The step of converting the grid-connected inverter adaptive control optimization problem based on the grid-side impedance into a Markov decision process includes: Define the state space of the Markov decision process based on the grid-side inductance value; Define the action space of Markov decision process based on the control parameters of hybrid synchronous loop; defining a reward space, and determining a reward value of a Markov decision process based on the state space and the action space; Define the state transition probability and reward discount rate, and form a replay experience pool based on the state space, action space and reward space; An evaluation value is obtained based on the grid-side inductance value in the state space and the hybrid synchronous loop control parameter in the action space, and the quality of the hybrid synchronous loop control parameter is determined based on the evaluation value.

5. The grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability according to claim 4 is characterized in that: The specific formula of the return space is: in, The corresponding elements of a decision cycle, and They are The corresponding reward and penalty values, Substitute the characteristic roots after action in different states The damping ratio, are all characteristic roots of the inverter grid-connected system based on hybrid synchronous control, is the characteristic root located above the real axis among the first pair of dominant characteristic roots of the system, is the characteristic root located above the real axis in the second pair of dominant characteristic roots of the system, Characteristic root With characteristic root The expected value of the real distance, Characteristic root With characteristic root The opposite of the expected value of the imaginary axis distance.

6. The grid-connected inverter adaptive synchronization control method taking into account both dynamic performance and stability according to claim 4 is characterized in that: Substituting the Markov decision process into the reinforcement learning algorithm framework for reverse training includes: Randomly extract training samples from the playback experience pool; Determining a value network based on the training samples; Updating current network parameters by gradient descent based on the loss function of the value network; The network parameters of the neural network are updated by gradient descent based on the loss function of the policy network; The temperature coefficient representing the importance of entropy to the reward is updated by gradient ascent; The target value network is updated using a soft update method, and the control strategy is updated based on the updated target value network.

Images