A time synchronization-based wireless grid-connected control method and system

Abstract

This invention discloses a time-synchronized wireless grid-connected control method and system, relating to the field of power systems. The method includes: performing high-precision time synchronization to multiple grid-connected points to the same time base; collecting voltage, phase, and frequency parameters to generate local parameter sequences for each grid-connected point; collecting synchronization messages from the main grid and broadcasting them to multiple grid-connected points; performing time matching on the local parameter sequences and performing parameter difference analysis to obtain parameter differences between multiple main grids and grid-connected points; when preset grid-connected conditions are not met, calling a steady-state-transient parallel control module for analysis to generate voltage regulation and speed regulation signals; and outputting these signals to an automatic voltage regulator and speed control board to adjust the operating state until the preset grid-connected conditions are met. This application solves the technical problems of insufficient grid-connected synchronization control accuracy and poor deployment adaptability in existing distributed generation grid-connected control systems, achieving the technical effects of improving grid-connected synchronization control accuracy and enhancing the deployment adaptability of grid-connected control.

Description

Technical Field

[0001] This application relates to the field of power systems, and in particular to a wireless grid-connected control method and system based on time synchronization. Background Technology

[0002] Synchronization control between the grid connection point and the main grid in a distributed generation system is crucial for ensuring safe and stable grid connection, avoiding grid connection shocks and equipment failures, and directly determining the reliability and security of distributed generation grid connection. Currently, the industry mainly uses wired communication to transmit synchronization signals for grid connection control. Dedicated communication lines are laid to transmit main grid parameters to each grid connection point, and the operating status of the grid connection point is adjusted based on these parameters. However, existing technologies rely on wired cabling for signal transmission, which not only limits deployment flexibility due to cabling environment constraints but also lacks a high-precision clock calibration mechanism. The inherent deviation of the local clock at each grid connection point can easily cause errors in comparing parameters between the main grid and the grid connection point, thus reducing grid connection control accuracy and increasing system deployment and subsequent maintenance costs.

[0003] At present, distributed generation grid connection control suffers from technical problems such as insufficient grid connection synchronization control accuracy and poor deployment adaptability. Summary of the Invention

[0004] This application provides a time-synchronized wireless grid-connected control method and system. It employs high-precision time synchronization of multiple grid-connected points in a target area to the same microsecond-level time reference. Based on this reference, it collects voltage, phase, and frequency parameters of each grid-connected point and generates a local parameter sequence with time tags. The main grid controller collects the main grid parameters and generates a synchronization message, which is broadcast wirelessly to each grid-connected point. Each grid-connected point matches its local parameter sequence according to the timestamp in the synchronization message and calculates the parameter difference between the main grid and the grid-connected point. If the parameter difference does not meet the grid connection conditions, each grid-connected point uses a built-in steady-state-transient parallel control module to analyze the parameter difference and generate voltage and speed regulation signals. These signals are output to the corresponding automatic voltage regulator and speed control board of the grid-connected point to adjust the unit's operating state until the parameter difference meets the grid connection conditions, thus completing the grid connection. This technical approach solves the technical problems of insufficient grid connection synchronization control accuracy and poor deployment adaptability in existing distributed generation grid-connected control systems, achieving the technical effect of improving grid connection synchronization control accuracy and enhancing the deployment adaptability of distributed generation grid-connected control.

[0005] This application provides a time-synchronized wireless grid-connected control method, comprising: performing high-precision time synchronization on multiple grid-connected points in a target area, synchronizing the local clock modules of each grid-connected point to the same microsecond-level time reference, and collecting voltage, phase, and frequency parameters of each grid-connected point in real time based on the time reference to generate a sequence of local parameters of multiple grid-connected points with time tags; using a main network controller to collect synchronization messages of the main network in the target area, and broadcasting the synchronization messages to multiple grid-connected points through a wireless communication module; and multiple grid-connected points controlling each other based on the timestamps in the received synchronization messages. The local parameter sequence is time-matched, and parameter difference analysis is performed to obtain multiple main grid-connection point parameter differences. When the multiple main grid-connection point parameter differences do not meet the preset grid connection conditions, multiple steady-state and transient parallel control modules built into the multiple grid connection points are invoked to parse the multiple main grid-connection point parameter differences and generate multiple voltage regulation signals and multiple speed regulation signals. The multiple voltage regulation signals and multiple speed regulation signals are output to the automatic voltage regulators and speed control boards of the multiple grid connection points respectively to adjust the operating status until the multiple grid connection points meet the preset grid connection conditions and the grid connection operation is completed.

[0006] In a possible implementation, the following processing is performed: each of the multiple steady-state-transient parallel control modules includes a steady-state control branch and a transient control branch; the network parameters of the steady-state control branch are kept frozen; the initial network parameters of the transient control branch are consistent with those of the steady-state control branch, and can be optimized and updated locally and downloaded from the main network.

[0007] In a possible implementation, the following processing is performed: obtaining multiple historical grid-connected control log sets for multiple grid-connected points within a historical window; preprocessing the multiple historical grid-connected control log sets to construct a mapping training sample set; training the steady-state control branch using the mapping training sample set until training converges, obtaining the network parameters of the trained steady-state control branch; distributing the steady-state control branch network parameters to multiple steady-state control branches corresponding to the multiple steady-state-transient parallel control modules, and initializing the network parameters of the multiple transient control branches.

[0008] In a possible implementation, the multiple historical grid-connected control log sets are preprocessed to construct a mapping training sample set, and the following processing is performed: using the main grid-connection point parameter difference as an index, gradient aggregation is performed on the multiple historical grid-connected control log sets to determine multiple aggregated historical grid-connected control log sets; density iteration filtering is performed on the voltage regulation signals and speed regulation signals in each set of the multiple aggregated historical grid-connected control log sets to determine multiple filtered voltage regulation signals and multiple filtered speed regulation signals; the mean of the main grid-connection point parameter difference in each set of the multiple aggregated historical grid-connected control log sets is calculated to obtain multiple main grid-connection parameter difference mean values; using the multiple main grid-connection parameter difference mean values ​​as input and the multiple filtered voltage regulation signals and multiple filtered speed regulation signals as outputs, a mapping training sample set is constructed.

[0009] In a possible implementation, using the main grid-connection point parameter difference as an index, gradient aggregation is performed on the multiple historical grid-connection control log sets to determine multiple aggregated historical grid-connection control log sets. The following processing is then performed: a main grid-connection point parameter difference coordinate system is constructed, using voltage difference as the first coordinate axis, phase difference as the second coordinate axis, and frequency parameter difference as the third coordinate axis; data is extracted from the multiple historical grid-connection control log sets using the main grid-connection point parameter difference as an index, and the extracted data is input into the main grid-connection point parameter difference coordinate system to obtain multiple parameter difference coordinate points; gradient aggregation is performed on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets; mapping aggregation is performed on the multiple historical grid-connection control log sets based on the multiple aggregated parameter difference coordinate point sets to obtain multiple aggregated historical grid-connection control log sets.

[0010] In a possible implementation, gradient aggregation is performed on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets. The following processing is then performed: a first parameter difference coordinate point is randomly extracted from the multiple parameter difference coordinate points as a first gradient aggregation starting point; a neighborhood is constructed for the first gradient aggregation starting point according to a preset gradient to obtain a first starting neighborhood; the first starting neighborhood is updated by diffusion according to a preset gradient to determine a first diffusion neighborhood; when the neighborhood density of the first diffusion neighborhood and the neighborhood density of the first starting neighborhood satisfy the edge diffusion stopping condition, a first gradient aggregation neighborhood is obtained; a first aggregated parameter difference coordinate point set is constructed based on the parameter difference coordinate points in the first gradient aggregation neighborhood; after removing the first aggregated parameter difference coordinate point set from the main network-connection point parameter difference coordinate system, a second parameter difference coordinate point is randomly extracted again for gradient aggregation, until multiple parameter difference coordinate points are aggregated to obtain multiple aggregated parameter difference point sets.

[0011] In a possible implementation, the first starting neighborhood is updated by diffusion according to a preset gradient to determine the first diffusion neighborhood, and the following processing is performed: the coordinate point of the edge parameter difference that is farthest from the starting point of the first gradient aggregation in the first starting neighborhood is determined; the gradient from the starting point of the first gradient aggregation to the coordinate point of the edge parameter difference is used as the initial gradient, and the initial gradient is superimposed according to the preset gradient to determine the diffusion gradient; the neighborhood is constructed according to the diffusion gradient with the starting point of the first gradient aggregation as the center to determine the first diffusion neighborhood.

[0012] In a possible implementation, the following processing is performed: the main network collects network parameters of multiple transient control branches in the multiple steady-state-transient parallel control modules according to a preset period to obtain multiple periodic transient control branch network parameters; the multiple periodic transient control branch network parameters are weighted according to the processing effect to obtain weighted periodic transient control branch network parameters; the weighted periodic transient control branch network parameters are distributed to multiple transient control branches for network parameter updates.

[0013] In a possible implementation, the following processing is performed: the wireless communication module adopts one or more combinations of Wi-Fi, ZigBee, LoRa, 4G or 5G; the synchronization message includes a timestamp field, a main network voltage field, a main network frequency field and a main network phase field.

[0014] This application also provides a time-synchronized wireless grid-connected control system, comprising: a high-precision time synchronization module, used to perform high-precision time synchronization for multiple grid-connected points in a target area, synchronizing the local clock modules of each grid-connected point to the same microsecond-level time reference, and collecting voltage, phase, and frequency parameters of each grid-connected point in real time based on the time reference to generate a sequence of local parameters for multiple grid-connected points with time tags; a synchronization message broadcasting module, used to collect synchronization messages from the main network in the target area using the main network controller, and broadcast the synchronization messages to multiple grid-connected points through a wireless communication module; and a parameter difference analysis module, used by multiple grid-connected points to analyze the timestamps in the received synchronization messages. The system performs time matching on the local parameter sequences of the multiple grid-connected points, executes parameter difference analysis, and obtains multiple main grid-connected point parameter differences. A parameter difference parsing module is used to, when the parameter differences of the multiple main grid-connected points do not meet the preset grid connection conditions, call multiple steady-state and transient parallel control modules built into the multiple grid-connected points to parse the parameter differences, generating multiple voltage regulation signals and multiple speed regulation signals. A grid connection operation module is used to output the multiple voltage regulation signals and multiple speed regulation signals to the automatic voltage regulators and speed control boards of the multiple grid-connected points respectively for operating status adjustment until the multiple grid-connected points meet the preset grid connection conditions, thus completing the grid connection operation.

[0015] This application proposes a time-synchronized wireless grid-connected control method and system. First, high-precision time synchronization is performed on multiple grid-connected points in a target area, synchronizing the local clock modules of each point to the same microsecond-level time reference. Based on this time reference, voltage, phase, and frequency parameters of each grid-connected point are collected in real time, generating a sequence of local parameters for multiple grid-connected points with time tags. Next, the main network controller collects synchronization messages from the main network in the target area and broadcasts these messages to the multiple grid-connected points via a wireless communication module. Then, based on the timestamps in the received synchronization messages, the multiple grid-connected points control the system accordingly. The system performs time matching of local parameter sequences at each grid-connected point, executes parameter difference analysis, and obtains multiple main grid-connected point parameter differences. When multiple main grid-connected point parameter differences do not meet the preset grid connection conditions, it calls multiple steady-state and transient parallel control modules built into each of the multiple grid-connected points to parse the multiple main grid-connected point parameter differences, generating multiple voltage regulation signals and multiple speed regulation signals. Finally, the multiple voltage regulation signals and multiple speed regulation signals are output to the automatic voltage regulators and speed control boards of the multiple grid-connected points respectively for operating status adjustment until the multiple grid-connected points meet the preset grid connection conditions, completing the grid connection operation. Through the above process, the method and system proposed in this application achieve the technical effect of improving the accuracy of grid connection synchronization control and enhancing the deployment adaptability of distributed generation grid connection control. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating a time-synchronized wireless network control method provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a time-synchronized wireless grid-connected control system provided in an embodiment of this application.

[0019] Explanation of reference numerals in the attached diagram: High-precision time synchronization module 10, Synchronization message broadcasting module 20, Parameter difference analysis module 30, Parameter difference parsing module 40, Grid connection operation module 50. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides a time-synchronized wireless network control method, such as... Figure 1 As shown, the method includes: Step S100: Perform high-precision time synchronization on multiple grid-connected points in the target area, synchronize the local clock modules of each grid-connected point to the same microsecond-level time reference, and collect the voltage, phase and frequency parameters of each grid-connected point in real time based on the time reference to generate a sequence of local parameters of multiple grid-connected points with time tags.

[0022] Specifically, a BeiDou or GPS high-precision time synchronization module is configured for each grid connection point. Using the PTP (Precision Time Protocol) or NTP (Network Time Protocol) of the time synchronization module, the local crystal oscillator clock module of each grid connection point is calibrated in real time with microsecond-level accuracy. The calibration deviation is controlled within ±1 microsecond, ensuring that the clocks of all grid connection points are synchronized to the unified UTC time reference. Voltage transformers, phase detectors, and frequency detectors are deployed at each grid connection point, with a fixed acquisition period set. Based on the calibrated local clock, a unique time tag is added to each acquired voltage, phase, and frequency parameter, accurate to the microsecond. Continuously acquired data from the same grid connection point is then arranged in ascending order by time tag to generate a structured local parameter sequence. The sequence data format uses key-value pairs, where the key is the time tag and the value is the combination of voltage, phase, and frequency parameters at the corresponding time.

[0023] Step S200: The main network controller collects the synchronization messages of the main network in the target area and broadcasts the synchronization messages to multiple grid connection points via a wireless communication module. The wireless communication module uses one or more combinations of Wi-Fi, ZigBee, LoRa, 4G, or 5G. The synchronization messages include a timestamp field, a main network voltage field, a main network frequency field, and a main network phase field.

[0024] Specifically, the main grid controller deploys high-precision voltage transformers, phase acquisition modules, and frequency detection modules to collect main grid parameters in real time at the same acquisition cycle as the grid connection points. Simultaneously, the main grid controller's built-in high-precision time synchronization unit adds a microsecond-level timestamp, identical to that of the grid connection points, to each collected main grid parameter. The main grid controller's synchronization message generation module encapsulates the timestamp, main grid voltage, main grid phase, and main grid frequency into corresponding fields according to a preset message format. The message is structured using JSON format, with fields named TimeStamp, GridVoltage, GridPhase, and GridFrequency. The synchronization message is broadcast via the main grid controller's wireless communication module. This module has a built-in multi-communication protocol conversion chip, allowing selection of the communication method based on the actual scenario. For example, LoRa protocol is used for outdoor long-distance distributed grid connection points, Wi-Fi protocol for indoor short-distance grid connection points, and 4G / 5G protocol for remote areas without broadband. When multiple protocols are combined, channel allocation avoids signal interference.

[0025] In step S300, multiple grid-connected points perform time matching on their local parameter sequences based on the timestamps in the received synchronization messages, and perform parameter difference analysis to obtain multiple main network-grid-connected point parameter differences.

[0026] Specifically, after each grid connection point receives a synchronization message via its wireless communication module, the message parsing module extracts the microsecond-level timestamp from the message. A parameter matching algorithm searches the local parameter sequence for local parameter data that perfectly matches the timestamp. If a timestamp discrepancy exists, linear interpolation is used to calculate the corresponding local parameter compensation value. The difference calculation module performs parameter difference calculations: voltage difference is calculated as the absolute value of the local voltage minus the main grid voltage; frequency difference is calculated as the absolute value of the local frequency minus the main grid frequency; and phase difference is calculated as the absolute value of the phase difference between the local phase and the main grid phase. All difference results are rounded to two decimal places. This set of differences represents the main grid-grid connection point parameter difference for that grid connection point.

[0027] Step S400: When the parameter differences between multiple main grid-connection points do not meet the preset grid connection conditions, multiple steady-state and transient parallel control modules built into the multiple grid connection points are invoked to analyze the parameter differences between the multiple main grid-connection points, generating multiple voltage regulation signals and multiple speed regulation signals. Each steady-state and transient parallel control module includes a steady-state control branch and a transient control branch; the network parameters of the steady-state control branch are kept frozen; the initial network parameters of the transient control branch are consistent with those of the steady-state control branch, and can be locally optimized and updated, and downloaded from the main grid.

[0028] Specifically, preset grid connection conditions are set, such as voltage difference ≤ 5%, frequency difference ≤ 0.5Hz, and phase difference ≤ 5°. The difference judgment module of the grid connection point compares the parameter difference obtained in step S300 with the preset conditions. If any difference exceeds the threshold, it is determined that the grid connection conditions are not met. Then, the steady-state-transient parallel control module built into the grid connection point is triggered. This module is an FPGA-based hardware control module that integrates two independent but interconnected control branches. Both the steady-state control branch and the transient control branch adopt a three-layer feedforward neural network architecture. The input layer has 3 neurons, corresponding to voltage difference, phase difference, and frequency difference. The hidden layer has 16 neurons, and the output layer has 2 neurons, corresponding to voltage regulation signal and speed regulation signal. The neural network parameters, such as weights and biases, of the steady-state control branch are fixed through hardware programming after training and convergence, remaining frozen without modification. The initial network parameters of the transient control branch are completely identical to those of the steady-state control branch through data copying. This branch is also configured with a local parameter update interface and a main network parameter receiving interface, allowing for parameter updates via local gradient descent algorithms or overwriting updates using parameters from the main network. Both branches perform parallel analytical calculations on the parameter differences, and a weighted fusion strategy is used to fuse the outputs of the two branches, ultimately generating standardized voltage regulation and speed control signals. The signal types include analog and digital signals, adaptable to different types of automatic voltage regulators and speed control boards.

[0029] In step S500, the multiple voltage regulation signals and multiple speed regulation signals are respectively output to the automatic voltage regulators and speed control boards of multiple grid connection points to adjust their operating status until the multiple grid connection points meet the preset grid connection conditions, thus completing the grid connection operation.

[0030] Specifically, the signal type is adapted and converted through the built-in signal output module at the grid connection point. If the voltage regulation signal and speed regulation signal are digital or analog, they are output directly. If they are switching signals, they are converted into electrical signals compatible with the automatic voltage regulator and speed control board through a level conversion circuit. Analog signals use a 4-20mA standard current signal, and switching signals use a 0-5V standard level signal. Shielded cables are used for signal transmission to reduce interference. After receiving the voltage regulation signal, the automatic voltage regulator adjusts the generator set's output voltage by changing the excitation current of the excitation winding. The voltage regulation accuracy is controlled within ±0.5V, and the adjustment response time is ≤100 milliseconds. It tracks the value changes of the voltage regulation signal in real time for dynamic excitation adjustment. After receiving the speed regulation signal, the speed control board adjusts the generator set's rotor speed by adjusting the generator set's fuel supply or the prime mover's speed control mechanism, thereby changing the output frequency. The speed regulation accuracy is controlled within ±0.1Hz, and the adjustment response time is ≤150 milliseconds. During the adjustment process, the local parameter acquisition module at the grid connection point maintains a fixed acquisition cycle, continuously acquiring the adjusted voltage, phase, and frequency parameters, repeatedly performing parameter matching and difference calculation operations, and obtaining the latest parameter difference between the main grid and the grid connection point in real time. The difference judgment module continuously compares the latest parameter difference with the preset grid connection conditions. If any parameter difference does not meet the conditions, the state adjustment continues through voltage regulation and speed regulation signals. If all parameter differences meet the preset conditions, the signal output module stops sending adjustment signals, and the grid connection execution unit triggers the grid connection circuit breaker to close, completing the grid connection operation between the grid connection point and the main grid. After all grid connection points complete the above process in sequence, the grid connection operation of the entire distributed generation system is completed.

[0031] In one possible implementation, step S400 further includes step S410, acquiring multiple sets of historical grid-connected control logs for multiple grid-connected points within a historical window. Specifically, a historical window is set, such as the past 6 months. Through the local data storage modules of each grid-connected point and the grid-connected data center of the main network, historical grid-connected control logs for all grid-connected points within this time window are collected. Log collection uses a data synchronization protocol to upload locally stored logs to the main network data center for unified aggregation. The historical grid-connected control logs are structured data. Each log entry contains a unique grid-connected point number, a grid-connected timestamp, the difference in parameters between the main network and the grid-connected point, voltage regulation signals, speed regulation signals, grid-connected adjustment effects, and whether grid-connected conditions are met, among other fields. All historical logs for each grid-connected point form an independent log set.

[0032] Step S420: Preprocess the multiple historical grid-connected control log sets to construct a mapping training sample set. Specifically, the historical grid-connected control log sets are cleaned by using an outlier detection algorithm to remove logs with voltage difference, frequency difference, and phase difference exceeding reasonable ranges, logs missing key fields, and logs with ineffective grid-connected regulation effects. The cleaned valid logs are then standardized by mapping voltage difference, frequency difference, phase difference, voltage regulation signal, and speed regulation signal to a 0-1 value range using the min-max standardization formula: Standardized value = (Original value - Minimum value) / (Maximum value - Minimum value). Finally, the data is integrated according to the input-output mapping relationship, using the main grid-grid connection point parameter difference as the sample input and the corresponding voltage regulation signal and speed regulation signal as the sample output to construct a structured mapping training sample set. The sample set is divided into a training set, a validation set, and a test set in a 7:2:1 ratio and stored in CSV format, with each row representing one sample and columns for voltage difference, phase difference, frequency difference, voltage regulation signal, and speed regulation signal, respectively.

[0033] Step S430: Train the steady-state control branch using the mapped training sample set until training converges, obtaining the parameters of the trained steady-state control branch network. Specifically, the steady-state control branch uses a three-layer feedforward neural network. Using the training set of the mapped training sample set as input, the predicted output of the neural network is calculated through forward propagation. The activation function for forward propagation is the ReLU function. The hidden layer calculation formula is: Hidden layer output = ReLU(Input layer data × Input layer - Hidden layer weight + Hidden layer bias). The output layer calculation formula is: Output layer data = Hidden layer output × Hidden layer - Output layer weight + Output layer bias. Mean squared error is used as the loss function, and backpropagation is performed using stochastic gradient descent. The initial learning rate is set to 0.01, and a learning rate decay strategy is adopted, multiplying the learning rate by 0.9 every 100 training epochs. The batch size is set to 32. During training, the model accuracy is verified using a validation set. Training convergence is determined when the rate of change of the validation set loss value over 50 consecutive epochs is ≤0.001, and the prediction accuracy of the test set is ≥95%. After training converges, the input-hidden layer weights, hidden layer biases, hidden-output layer weights, and output layer biases of the neural network are extracted. This type of data is the parameter of the trained steady-state control branch network.

[0034] Step S440 involves distributing the steady-state control branch network parameters to the multiple steady-state control branches corresponding to the multiple steady-state-transient parallel control modules, and initializing the network parameters of the multiple transient control branches. Specifically, the main network distributes the steady-state control branch network parameters to all grid-connected points via a broadcast and acknowledgment method through a wireless communication link. The parameters are encapsulated in binary format. Each grid-connected point returns an acknowledgment message to the main network after receiving the parameters. The main network retransmits the parameters to grid-connected points that do not return acknowledgment messages to ensure the integrity of parameter distribution. After receiving the parameters, each grid-connected point permanently programs the network parameters into the storage unit of the steady-state control branch through the FPGA's hardware programming interface, completing the parameter solidification of the steady-state control branch. Simultaneously, through the data copy interface, the network parameters programmed into the steady-state control branch are completely copied to the parameter register of the transient control branch, overwriting the initial default parameters of the transient control branch, completing the network parameter initialization of the transient control branch. After initialization, the weights and biases of the transient control branch are completely consistent with those of the steady-state control branch.

[0035] In one possible implementation, the multiple historical grid-connected control log sets are preprocessed to construct a mapping training sample set. Step S420 further includes step S421, using the main grid-connection point parameter difference as an index, gradient aggregation is performed on the multiple historical grid-connected control log sets to determine multiple aggregated historical grid-connected control log sets. Specifically, the voltage difference, phase difference, and frequency difference between the main grid and the grid-connection point are used as core index fields, and a feature threshold for gradient aggregation is set for each index field. A density-based clustering algorithm is used, with the three parameter differences as feature dimensions and the neighborhood radius set as the comprehensive gradient threshold, to cluster the historical grid-connected control log sets. Log data in the same cluster after clustering are determined to be feature-similar data. The log data in the same cluster are integrated, retaining the core fields of the logs and calculating the feature mean of the data within the cluster. The integrated data of each cluster forms an aggregated historical grid-connected control log set. Isolated abnormal log data after clustering are directly removed.

[0036] Step S422: Density iterative filtering is performed on the voltage regulation signals and speed regulation signals in each of the multiple aggregated historical grid-connected control log sets to determine multiple selected voltage regulation signals and multiple selected speed regulation signals. Specifically, for each aggregated historical grid-connected control log set, the voltage regulation signal dataset and speed regulation signal dataset are extracted. The mean drift filtering algorithm is used to process the two datasets separately. The kernel function of the mean drift is a Gaussian kernel function, and the kernel bandwidth is determined by cross-validation. The drift vector of each data point is calculated based on the kernel function. Through multiple iterations, the data points drift towards the density center. The iteration terminates when the magnitude of the drift vector is ≤0.001. After the iteration is completed, the signal data within the 95% confidence interval around the density center is determined as valid data. The arithmetic mean of the valid data is calculated. This mean is the selected voltage regulation signal and selected speed regulation signal corresponding to the aggregate set. Abnormal signal data outside the confidence interval are directly discarded.

[0037] Step S423: Calculate the mean of the main grid-connection point parameter difference within each of the multiple aggregated historical grid-connection control log sets to obtain multiple main grid-connection parameter difference mean values. Specifically, for each aggregated historical grid-connection control log set, extract the voltage difference, phase difference, and frequency difference data from all logs. Calculate the mean of the three parameter differences using the arithmetic mean method. The calculation formula is: mean = sum of all data for a certain parameter difference within the set / number of data for that parameter difference within the set. The calculation result is rounded to two decimal places. Integrate the mean voltage difference, mean phase difference, and mean frequency difference of each set to obtain the corresponding main grid-connection parameter difference mean.

[0038] Step S424: Using the average difference between the multiple main grid and grid-connected parameters as input, and multiple selected voltage regulation signals and multiple selected speed regulation signals as output, a mapping training sample set is constructed. Specifically, for all aggregated historical grid-connected control log sets, a one-to-one input-output mapping relationship is established. Each aggregated set corresponds to one training sample. The sample input is the average difference between the main grid and grid-connected parameters of that set, and the sample output is the selected voltage regulation signal and selected speed regulation signal of that set. All samples are structurally integrated in the order of input first and output last. A unique sample number is added to each sample, and a mapping training sample set is constructed in tabular form. The number of samples in the sample set is consistent with the number of aggregated historical grid-connected control log sets. At the same time, the data format of the sample set is validated to ensure that there are no mapping errors or missing data samples.

[0039] In one possible implementation, using the main grid-connection point parameter difference as an index, gradient aggregation is performed on the multiple historical grid-connection control log sets to determine multiple aggregated historical grid-connection control log sets. Step S421 further includes step S4211, constructing a main grid-connection point parameter difference coordinate system by using voltage difference as the first coordinate axis, phase difference as the second coordinate axis, and frequency parameter difference as the third coordinate axis. Specifically, a Cartesian three-dimensional coordinate system is used to construct the main grid-connection point parameter difference coordinate system, setting the origin of the coordinate system to (0,0,0), representing that the parameter difference between the main grid and the grid-connection point is 0. The voltage difference is set as the X-axis (first coordinate axis), the phase difference as the Y-axis (second coordinate axis), and the frequency difference as the Z-axis (third coordinate axis). Each coordinate point (X,Y,Z) in the coordinate system uniquely corresponds to a set of main grid-connection point parameter differences. The coordinate system is constructed using MATLAB's three-dimensional coordinate system drawing tool, and the parameter configuration file of the coordinate system is saved.

[0040] Step S4212: Using the main grid-connection point parameter difference as an index, data is extracted from the multiple historical grid-connection control log sets, and the extracted data is input into the main grid-connection point parameter difference coordinate system to obtain multiple parameter difference coordinate points. Specifically, using the voltage difference, phase difference, and frequency difference fields in the historical grid-connection control logs as indexes, a data extraction script extracts three sets of parameter difference data from each log in batches from all historical logs. The extracted parameter difference data is format-converted to a three-dimensional coordinate system coordinate point format (X,Y,Z). A data import tool imports all coordinate point data into the constructed main grid-connection point parameter difference coordinate system. The coordinate system automatically generates visual points at corresponding positions based on the coordinate point data. Each point represents the parameter difference data of a historical log, and all generated points are multiple parameter difference coordinate points. Simultaneously, the coordinate points are associated with the corresponding log numbers for easy traceability.

[0041] Step S4213: Perform gradient aggregation on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets. Specifically, a gradient-based density clustering method is used to aggregate the parameter difference coordinate points. First, the core parameters of gradient aggregation are set, including the initial neighborhood gradient, density judgment threshold, and diffusion step size. Starting from a randomly selected coordinate point, an initial neighborhood is constructed around the starting point according to the initial neighborhood gradient. The density of coordinate points within the neighborhood is counted. If the density is greater than or equal to the density judgment threshold, a new neighborhood is constructed by diffusion outward according to the diffusion step size. The diffusion operation is repeated until the density of the new neighborhood is less than the density of the previous neighborhood. At this point, the coordinate points within all the diffused neighborhoods are aggregated into a group. The aggregated coordinate points are removed from the coordinate system, and the next unaggregated coordinate point is randomly selected. The above operation is repeated until all parameter difference coordinate points in the coordinate system have been aggregated. Each group of aggregated coordinate points is an aggregated parameter difference coordinate point set.

[0042] Step S4214: Map and aggregate the multiple historical grid-connected control log sets according to the multiple aggregated parameter difference coordinate point sets to obtain multiple aggregated historical grid-connected control log sets. Specifically, since each parameter difference coordinate point is associated with a corresponding historical log number, firstly, trace the historical logs corresponding to all coordinate points in each aggregated parameter difference coordinate point set through the log number. Extract all historical logs corresponding to the same aggregated coordinate point set from the original historical grid-connected control log set, integrate the data, retain all core fields of the logs, and add an aggregation number to the integrated log set, which is consistent with the corresponding aggregated coordinate point set number. Treat the integrated log data corresponding to each aggregated coordinate point set as an independent set, which is the aggregated historical grid-connected control log set. The number of aggregated log sets is exactly the same as the number of aggregated parameter difference coordinate point sets. At the same time, perform data verification on the aggregated log sets to ensure that there are no missing or duplicate logs.

[0043] In one possible implementation, gradient aggregation is performed on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets. Step S4213 further includes step S42131, randomly extracting a first parameter difference coordinate point from the multiple parameter difference coordinate points as the starting point of the first gradient aggregation. Specifically, a random sampling algorithm without replacement is used to sample all unmarked parameter difference coordinate points in the coordinate system, setting the sampling seed to the current system timestamp to ensure the randomness of the sampling. A coordinate point is randomly extracted from all coordinate points, its three-dimensional coordinate value is recorded, and the coordinate point is marked as selected to avoid subsequent duplicate sampling. This marked coordinate point is the starting point of the first gradient aggregation. Simultaneously, the starting point is visually labeled in the coordinate system to facilitate subsequent neighborhood construction operations.

[0044] Step S42132: Construct a neighborhood for the first gradient aggregation starting point according to a preset gradient to obtain a first starting neighborhood. Specifically, a preset gradient is set, and a cubic neighborhood is constructed in a three-dimensional coordinate system with the first gradient aggregation starting point (X0, Y0, Z0) as the center. The boundary range of the neighborhood is: X-axis ∈ [X0 - preset gradient, X0 + preset gradient], Y-axis ∈ [Y0 - preset gradient, Y0 + preset gradient], Z-axis ∈ [Z0 - preset gradient, Z0 + preset gradient]. All parameter difference coordinate points contained in this cube are the components of the first starting neighborhood. At the same time, the number of coordinate points in this neighborhood is counted, and the neighborhood density is calculated. Neighborhood density = number of coordinate points in the neighborhood / neighborhood volume.

[0045] Step S42133: Update the first starting neighborhood according to the preset gradient to determine the first diffusion neighborhood. Specifically, with the first gradient aggregation starting point as the center, expand the boundary of the first starting neighborhood according to the diffusion gradient to construct a new cube neighborhood. The boundary range of the new neighborhood is: X-axis ∈ [X0 - diffusion gradient, X0 + diffusion gradient], Y-axis ∈ [Y0 - diffusion gradient, Y0 + diffusion gradient], Z-axis ∈ [Z0 - diffusion gradient, Z0 + diffusion gradient]. All parameter difference coordinate points contained in this new cube are the first diffusion neighborhood. At the same time, the number of coordinate points in the first diffusion neighborhood is counted, and its neighborhood density is calculated. During the diffusion process, all coordinate points of the first starting neighborhood are retained, and the neighborhood range is only expanded outward.

[0046] Step S42134: When the neighborhood density of the first diffusion neighborhood and the neighborhood density of the first starting neighborhood satisfy the edge diffusion stopping condition, a first gradient aggregation neighborhood is obtained. Specifically, the neighborhood densities of the first starting neighborhood and the first diffusion neighborhood are calculated using the density calculation formula. The density values ​​of the two neighborhoods are compared. If the density of the first diffusion neighborhood is less than the density of the first starting neighborhood, the edge diffusion stopping condition is satisfied, and the first starting neighborhood is determined as the first gradient aggregation neighborhood. If the density of the first diffusion neighborhood is greater than or equal to the density of the first starting neighborhood, diffusion continues outward according to the diffusion gradient to construct a new diffusion neighborhood. The density calculation and comparison operation is repeated until the stopping condition is satisfied.

[0047] Step S42135: Construct a first aggregated parameter difference coordinate point set based on the parameter difference coordinate points in the first gradient aggregation neighborhood. Specifically, using a coordinate point extraction tool, extract all included parameter difference coordinate points within the cube boundary of the first gradient aggregation neighborhood in batches, recording the 3D coordinate values ​​and corresponding log numbers of each coordinate point. Sort all extracted coordinate point data in ascending order of coordinate values ​​to construct a structured dataset. The dataset includes fields such as coordinate point number, X-axis voltage difference, Y-axis phase difference, Z-axis frequency difference, and log number. Save this dataset and label it as aggregated parameter difference coordinate point set 1, which is the first aggregated parameter difference coordinate point set. Simultaneously, use a uniform color to label all coordinate points in this set in the coordinate system for easy identification.

[0048] Step S42136: After removing the first aggregated parameter difference coordinate point set from the main network-connection point parameter difference coordinate system, second parameter difference coordinate points are randomly extracted again for gradient aggregation until multiple parameter difference coordinate points are aggregated, resulting in multiple aggregated parameter difference coordinate point sets. Specifically, using a data removal script, all coordinate points in the first aggregated parameter difference coordinate point set are deleted from the original dataset of the coordinate system, and these points are marked as aggregated to ensure that subsequent sampling will not select duplicates. The no-replacement random sampling algorithm is then used again to randomly extract a coordinate point from the remaining unmarked and unaggregated parameter difference coordinate points in the coordinate system as the starting point for the second gradient aggregation. Steps S42132 to S42135 are repeated to construct the second aggregated parameter difference coordinate point set. Continue the above elimination, sampling, and aggregation operations until all parameter difference coordinate points in the coordinate system have been marked as aggregated. At this point, all the aggregated parameter difference coordinate point sets that have been constructed are the final multiple aggregated parameter difference coordinate point sets. The number of sets depends on the density distribution of coordinate points. The higher the density, the more coordinate points are in the set.

[0049] In one possible implementation, the first starting neighborhood is updated by diffusion according to a preset gradient to determine the first diffusion neighborhood. Step S42133 further includes step S421331, determining the edge parameter difference coordinate point in the first starting neighborhood that is farthest from the first gradient aggregation starting point. Specifically, the distance from each coordinate point in the first starting neighborhood to the first gradient aggregation starting point is calculated using the three-dimensional Euclidean distance calculation formula. The calculation results of all coordinate points in the neighborhood are traversed by a traversal algorithm to find the coordinate point with the largest distance value. This coordinate point with the largest distance is marked as the edge parameter difference coordinate point of the first starting neighborhood, and its three-dimensional coordinate value and distance to the starting point are recorded.

[0050] Step S421332: Using the distance from the starting point of the first gradient aggregation to the coordinate point of the edge parameter difference as the initial gradient, and superimposing the initial gradient according to the preset gradient, the diffusion gradient is determined. Specifically, the Euclidean distance from the starting point of the first gradient aggregation to the coordinate point of the edge parameter difference is determined as the initial gradient, and the value of the initial gradient is recorded. The diffusion gradient is equal to the preset gradient plus the initial gradient.

[0051] Step S421333: Using the first gradient aggregation starting point as the center, construct a neighborhood according to the diffusion gradient to determine the first diffusion neighborhood. Specifically, using the first gradient aggregation starting point as the center, construct a cubic diffusion neighborhood according to the diffusion gradient value in the three-dimensional coordinate system of the main network-connection point parameter difference. All parameter difference coordinate points contained within the boundary of this cube are components of the first diffusion neighborhood. Count the total number of coordinate points in the first diffusion neighborhood and calculate the neighborhood density.

[0052] In one possible implementation, step S400 further includes step S450, whereby the main network collects network parameters of multiple transient control branches in the multiple steady-state-transient parallel control modules according to a preset period, obtaining multiple periodic transient control branch network parameters. Specifically, the main network sets a parameter collection period, which can be adjusted according to the grid connection scenario. The collection period is triggered by the main network's timed task module. The main network sends a parameter collection command to all grid connection points via a wireless communication link. The command includes a collection timestamp and the grid connection point number. After receiving the command, each grid connection point extracts the current network parameters through the parameter reading interface of the transient control branch. Each grid connection point encapsulates the extracted network parameters with its own number and collection timestamp, and uploads them to the main network via a wireless communication link. The main network performs integrity verification on the uploaded parameters and performs a second collection on grid connection points that fail the verification. The main network summarizes the transient control branch network parameters uploaded by all grid connection points within the period, classifies and stores them according to the grid connection point number, and the transient control branch network parameters of each grid connection point form a group. The set of parameters of all grid connection points is the multiple periodic transient control branch network parameters.

[0053] Step S460: The parameters of the multiple periodic transient control branch networks are weighted according to their processing effects to obtain weighted periodic transient control branch network parameters. Specifically, evaluation indicators for grid-connected regulation processing effects are set, including parameter difference regulation rate, regulation accuracy, and grid connection success rate. The regulation effect of the transient control branch of each grid-connected point in each acquisition cycle is quantitatively scored, with a score range of 0-10 points, where a higher score indicates a better regulation effect. Based on the quantitative score, the weighting coefficient of the transient control branch network parameters of each grid-connected point is calculated. The weighting coefficient = score of the grid-connected point / sum of scores of all grid-connected points. The weighted average method is used to calculate the network parameters of the same type for all grid-connected points in the cycle. The calculation formula is: weighted parameter value = Σ (parameter value of a grid-connected point × weighting coefficient of the grid-connected point). The calculated network parameters of all types are the weighted periodic transient control branch network parameters for that cycle.

[0054] Step S470: The weighted periodic transient control branch network parameters are distributed to multiple transient control branches for network parameter updates. Specifically, the main network encapsulates the weighted transient control branch network parameters for this period in binary format. The encapsulation packet includes a parameter version number, parameter type, parameter data, and a checksum to ensure the accuracy of parameter transmission. The parameter packet is distributed to all grid connection points via the main network's wireless communication module using broadcast and targeted retransmission, prioritizing 5G / LoRa protocol for transmission. During transmission, CRC cyclic redundancy check is used for data verification. After receiving the parameter packet, each grid connection point first verifies the checksum. If the verification is successful, the parameter data is extracted, and the weighted network parameters are overwritten with the original parameters through the parameter update interface of the transient control branch. During the update process, the transient control branch pauses its operation and automatically resumes parallel operation after the update is completed. After the parameters are updated, each grid connection point returns an update confirmation message to the main network. The main network performs targeted retransmission for grid connection points that have not returned confirmation, ensuring that the transient control branches of all grid connection points complete the periodic parameter updates. The updated transient control branch will perform analytical calculations on the parameter difference between the main network and the grid connection point based on the new weighted network parameters, thereby improving the subsequent grid connection regulation effect.

[0055] This application embodiment employs high-precision time synchronization of multiple grid-connected points in the target area to the same microsecond-level time reference. Based on this reference, voltage, phase, and frequency parameters of each grid-connected point are collected and a local parameter sequence with time tags is generated. The main grid controller collects the main grid parameters and generates a synchronization message, which is broadcast to each grid-connected point via wireless communication. Each grid-connected point matches its local parameter sequence according to the timestamp in the synchronization message and calculates the parameter difference between the main grid and the grid-connected point. If the parameter difference does not meet the grid connection conditions, each grid-connected point analyzes the parameter difference through its built-in steady-state-transient parallel control module, generates voltage regulation and speed regulation signals, and outputs the signals to the automatic voltage regulator and speed control board of the corresponding grid-connected point to adjust the unit's operating status until the parameter difference meets the grid connection conditions, thus completing the grid connection. This technical approach solves the technical problems of insufficient grid connection synchronization control accuracy and poor deployment adaptability in existing distributed generation grid connection control, achieving the technical effect of improving grid connection synchronization control accuracy and enhancing the deployment adaptability of distributed generation grid connection control.

[0056] In the above text, refer to Figure 1 A wireless network connection control method based on time synchronization according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 A time-synchronization-based wireless grid-connected control system according to an embodiment of the present invention is described.

[0057] According to an embodiment of the present invention, a time-synchronization-based wireless grid-connected control system is used to solve the technical problems of insufficient grid synchronization control accuracy and poor deployment adaptability in existing distributed generation grid-connected control systems. It achieves the technical effect of improving grid synchronization control accuracy and enhancing the deployment adaptability of distributed generation grid-connected control. The time-synchronization-based wireless grid-connected control system includes: a high-precision time synchronization module 10, a synchronization message broadcasting module 20, a parameter difference analysis module 30, a parameter difference parsing module 40, and a grid-connected operation module 50.

[0058] The high-precision time synchronization module 10 is used to provide high-precision time synchronization for multiple grid-connected points in the target area, synchronizing the local clock modules of each grid-connected point to the same microsecond-level time reference, and collecting the voltage, phase, and frequency parameters of each grid-connected point in real time based on the time reference to generate a sequence of local parameters for multiple grid-connected points with time tags; the synchronization message broadcasting module 20 is used to collect the synchronization messages of the main network in the target area using the main network controller, and broadcast the synchronization messages to multiple grid-connected points through a wireless communication module; the parameter difference analysis module 30 is used for multiple grid-connected points to analyze the local parameter sequence of the multiple grid-connected points based on the timestamps in the received synchronization messages. The system performs time matching and parameter difference analysis to obtain parameter differences between multiple main grid and grid connection points. The parameter difference parsing module 40 is used to call multiple steady-state and transient parallel control modules built into the multiple grid connection points to parse the parameter differences when the parameter differences between multiple main grid and grid connection points do not meet the preset grid connection conditions, generating multiple voltage regulation signals and multiple speed regulation signals. The grid connection operation module 50 is used to output the multiple voltage regulation signals and multiple speed regulation signals to the automatic voltage regulators and speed control boards of the multiple grid connection points respectively to adjust their operating status until the multiple grid connection points meet the preset grid connection conditions, thus completing the grid connection operation.

[0059] The detailed description of the specific configuration of the parameter difference resolution module 40 is explained as follows: As mentioned above, the parameter difference resolution module 40 may further include: each of the multiple steady-state-transient parallel control modules includes a steady-state control branch and a transient control branch; the network parameters of the steady-state control branch are kept in a frozen state; the initial network parameters of the transient control branch are consistent with those of the steady-state control branch, and can be optimized and updated locally and downloaded from the main network.

[0060] The parameter difference resolution module 40 may further include: a historical grid-connected control log set acquisition unit for acquiring multiple historical grid-connected control log sets for multiple grid-connected points within a historical window; a preprocessing unit for preprocessing the multiple historical grid-connected control log sets to construct a mapping training sample set; a steady-state control branch training unit for training the steady-state control branch using the mapping training sample set until the training converges, obtaining the network parameters of the trained steady-state control branch; and a network parameter initialization unit for distributing the steady-state control branch network parameters to multiple steady-state control branches corresponding to the multiple steady-state-transient parallel control modules, and initializing the network parameters of the multiple transient control branches.

[0061] The preprocessing unit for the multiple historical grid-connected control log sets includes: a gradient aggregation subunit for performing gradient aggregation on the multiple historical grid-connected control log sets using the main grid-connection point parameter difference as an index to determine multiple aggregated historical grid-connected control log sets; a density iteration filtering subunit for performing density iteration filtering on the voltage regulation signals and speed regulation signals in each set of the multiple aggregated historical grid-connected control log sets to determine multiple filtered voltage regulation signals and multiple filtered speed regulation signals; a mean calculation subunit for calculating the mean of the main grid-connection point parameter difference in each set of the multiple aggregated historical grid-connected control log sets to obtain multiple main grid-connection parameter difference mean values; and a mapping training sample set construction subunit for using the multiple main grid-connection parameter difference mean values ​​as input and multiple filtered voltage regulation signals and multiple filtered speed regulation signals as outputs to construct a mapping training sample set.

[0062] Specifically, using the main grid-connection point parameter difference as an index, gradient aggregation is performed on the multiple historical grid-connection control log sets to determine multiple aggregated historical grid-connection control log sets. The gradient aggregation subunit may further include: a main grid-connection point parameter difference coordinate system construction component for constructing a main grid-connection point parameter difference coordinate system using voltage difference as the first coordinate axis, phase difference as the second coordinate axis, and frequency parameter difference as the third coordinate axis; a data extraction component for extracting data from the multiple historical grid-connection control log sets using the main grid-connection point parameter difference as an index, and inputting the extracted data into the main grid-connection point parameter difference coordinate system to obtain multiple parameter difference coordinate points; a gradient aggregation component for performing gradient aggregation on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets; and a mapping aggregation component for performing mapping aggregation on the multiple historical grid-connection control log sets based on the multiple aggregated parameter difference coordinate point sets to obtain multiple aggregated historical grid-connection control log sets.

[0063] The gradient aggregation component further includes: a first gradient aggregation starting point determination subcomponent, which randomly extracts a first parameter difference coordinate point from the plurality of parameter difference coordinate points as a first gradient aggregation starting point; a neighborhood construction subcomponent, which constructs a neighborhood for the first gradient aggregation starting point according to a preset gradient to obtain a first starting neighborhood; a diffusion update subcomponent, which performs diffusion update on the first starting neighborhood according to a preset gradient to determine a first diffusion neighborhood; a first gradient aggregation neighborhood determination subcomponent, which obtains a first gradient aggregation neighborhood when the neighborhood density of the first diffusion neighborhood and the neighborhood density of the first starting neighborhood satisfy the edge diffusion stopping condition; a first aggregation parameter difference coordinate point set construction subcomponent, which constructs a first aggregation parameter difference coordinate point set based on the parameter difference coordinate points in the first gradient aggregation neighborhood; and an iterative aggregation subcomponent, which removes the first aggregation parameter difference coordinate point set from the main network-connection point parameter difference coordinate system and then randomly extracts a second parameter difference coordinate point for gradient aggregation again, until all parameter difference coordinate points are aggregated to obtain multiple aggregation parameter difference point sets.

[0064] Specifically, the first starting neighborhood is updated by diffusion according to a preset gradient to determine the first diffusion neighborhood. The diffusion update sub-component may further include: an edge parameter difference coordinate point determination element for determining the edge parameter difference coordinate point in the first starting neighborhood that is farthest from the first gradient aggregation starting point; a diffusion gradient determination element for determining the diffusion gradient by superimposing the initial gradient from the first gradient aggregation starting point to the edge parameter difference coordinate point according to a preset gradient; and a first diffusion neighborhood determination element for constructing a neighborhood centered on the first gradient aggregation starting point according to the diffusion gradient to determine the first diffusion neighborhood.

[0065] The parameter difference resolution module 40 may further include: a periodic acquisition unit for the main network to acquire network parameters of multiple transient control branches in the multiple steady-state-transient parallel control modules according to a preset period, thereby obtaining multiple periodic transient control branch network parameters; a weighting unit for weighting the multiple periodic transient control branch network parameters according to the processing effect, thereby obtaining weighted periodic transient control branch network parameters; and a network parameter update unit for distributing the weighted periodic transient control branch network parameters to multiple transient control branches for network parameter updates.

[0066] The specific configuration of the synchronization message broadcast module 20 is described in detail below: As mentioned above, the synchronization message broadcast module 20 may further include: the wireless communication module adopts one or more combinations of Wi-Fi, ZigBee, LoRa, 4G or 5G; the synchronization message includes a timestamp field, a main network voltage field, a main network frequency field and a main network phase field.

[0067] The time-synchronization-based wireless grid-connected control system provided in this embodiment of the invention can execute the time-synchronization-based wireless grid-connected control method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0068] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A wireless network-connected control method based on time synchronization, characterized in that, The method includes: High-precision time synchronization is performed on multiple grid-connected points in the target area, synchronizing the local clock modules of each grid-connected point to the same microsecond-level time base, and collecting the voltage, phase and frequency parameters of each grid-connected point in real time based on the time base to generate a sequence of local parameters of multiple grid-connected points with time tags; The main network controller collects the synchronization messages of the main network in the target area and broadcasts the synchronization messages to multiple grid connection points through the wireless communication module; Based on the timestamps in the received synchronization messages, multiple grid connection points perform time matching on their local parameter sequences, execute parameter difference analysis, and obtain multiple main network-grid connection point parameter differences. When the parameter differences between multiple main grid and grid connection points do not meet the preset grid connection conditions, the multiple steady-state and transient parallel control modules built into the multiple grid connection points are invoked to analyze the parameter differences between the multiple main grid and grid connection points and generate multiple voltage regulation signals and multiple speed regulation signals. The multiple voltage regulation signals and multiple speed regulation signals are respectively output to the automatic voltage regulators and speed control boards of multiple grid connection points to adjust their operating status until the multiple grid connection points meet the preset grid connection conditions and the grid connection operation is completed.

2. The wireless network control method based on time synchronization as described in claim 1, characterized in that, Each of the multiple steady-state-transient parallel control modules includes a steady-state control branch and a transient control branch; The network parameters of the steady-state control branch remain frozen. The initial network parameters of the transient control branch are the same as those of the steady-state control branch, and local optimization updates and main network download updates are possible.

3. The wireless network control method based on time synchronization as described in claim 2, characterized in that, include: Retrieve a set of historical grid connection control logs for multiple grid connection points within a historical window; The multiple historical grid-connected control log sets are preprocessed to construct a mapping training sample set; The steady-state control branch is trained using the mapped training sample set until the training converges, and the parameters of the trained steady-state control branch network are obtained. The steady-state control branch network parameters are distributed to the multiple steady-state control branches corresponding to the multiple steady-state-transient parallel control modules, and the network parameters of the multiple transient control branches are initialized.

4. The wireless network control method based on time synchronization as described in claim 3, characterized in that, The multiple historical grid-connected control log sets are preprocessed to construct a mapping training sample set, including: Using the difference in parameters between the main network and the grid connection point as an index, gradient aggregation is performed on the multiple historical grid connection control log sets to determine multiple aggregated historical grid connection control log sets. Density iteration filtering is performed on the voltage regulation signals and speed regulation signals in each of the multiple aggregated historical grid-connected control log sets to determine multiple filtered voltage regulation signals and multiple filtered speed regulation signals. Calculate the mean of the main network-connection point parameter difference in each set of the multiple aggregated historical grid-connection control log sets to obtain the mean of multiple main network-connection parameter differences; Using the average difference between the multiple main grid and grid-connected parameters as input, and multiple selected voltage regulation signals and multiple selected speed regulation signals as output, a mapping training sample set is constructed.

5. The wireless network control method based on time synchronization as described in claim 4, characterized in that, Using the difference in parameters between the main network and the grid connection point as an index, gradient aggregation is performed on the multiple historical grid connection control log sets to determine multiple aggregated historical grid connection control log sets, including: A coordinate system for the main grid-connection point parameter difference is constructed by using the voltage difference as the first coordinate axis, the phase difference as the second coordinate axis, and the frequency parameter difference as the third coordinate axis. Using the main grid-connection point parameter difference as an index, data is extracted from the multiple historical grid-connection control log sets, and the extracted data is input into the main grid-connection point parameter difference coordinate system to obtain multiple parameter difference coordinate points; Gradient aggregation is performed on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets; Based on the multiple sets of coordinate points with different aggregation parameters, the multiple sets of historical grid-connected control logs are mapped and aggregated to obtain multiple aggregated historical grid-connected control log sets.

6. The wireless network control method based on time synchronization as described in claim 5, characterized in that, Gradient aggregation is performed on the multiple parameter difference coordinate points to determine multiple aggregated parameter difference coordinate point sets, including: Randomly extract the first parameter difference coordinate point from the plurality of parameter difference coordinate points as the starting point of the first gradient aggregation; According to the preset gradient, a neighborhood is constructed for the first gradient aggregation starting point to obtain the first starting neighborhood; The first starting neighborhood is updated by diffusion according to a preset gradient to determine the first diffusion neighborhood; When the neighborhood density of the first diffusion neighborhood and the neighborhood density of the first starting neighborhood satisfy the edge diffusion stopping condition, a first gradient aggregation neighborhood is obtained. Based on the parameter difference coordinate points in the first gradient aggregation neighborhood, construct a first aggregated parameter difference coordinate point set; After removing the first set of aggregated parameter difference coordinate points from the main network-connection point parameter difference coordinate system, the second set of parameter difference coordinate points is randomly extracted again for gradient aggregation until multiple parameter difference coordinate points are aggregated to obtain multiple sets of aggregated parameter difference coordinate points.

7. The wireless network control method based on time synchronization as described in claim 6, characterized in that, The first initial neighborhood is updated by diffusion according to a preset gradient to determine the first diffusion neighborhood, including: Determine the coordinate point of the edge parameter difference that is furthest from the first gradient aggregation starting point in the first initial neighborhood; The initial gradient is determined by superimposing the initial gradient from the first gradient aggregation starting point to the edge parameter difference coordinate point, and then superimposing the initial gradient according to the preset gradient. Centered on the first gradient aggregation starting point, a neighborhood is constructed according to the diffusion gradient to determine the first diffusion neighborhood.

8. The wireless network control method based on time synchronization as described in claim 1, characterized in that, include: The main network collects network parameters of multiple transient control branches in the multiple steady-state-transient parallel control modules according to a preset period, and obtains network parameters of multiple periodic transient control branches. The parameters of the multiple periodic transient control branch networks are weighted according to the processing effect to obtain weighted periodic transient control branch network parameters; The weighted periodic transient control branch network parameters are distributed to multiple transient control branches for network parameter updates.

9. The wireless network control method based on time synchronization as described in claim 1, characterized in that, The wireless communication module adopts one or more combinations of Wi-Fi, ZigBee, LoRa, 4G or 5G; The synchronization message includes a timestamp field, a main network voltage field, a main network frequency field, and a main network phase field.

10. A time-synchronized wireless grid-connected control system, characterized in that, The system is used to implement a time-synchronization-based wireless network control method according to any one of claims 1-9, the system comprising: The high-precision time synchronization module is used to provide high-precision time synchronization for multiple grid-connected points in the target area, synchronize the local clock modules of each grid-connected point to the same microsecond-level time reference, and collect the voltage, phase and frequency parameters of each grid-connected point in real time based on the time reference to generate a sequence of local parameters of multiple grid-connected points with time tags. The synchronization message broadcasting module is used to collect the synchronization messages of the main network in the target area using the main network controller, and broadcast the synchronization messages to multiple grid connection points through the wireless communication module; The parameter difference analysis module is used to perform time matching on the local parameter sequences of multiple grid-connected points based on the timestamps in the received synchronization messages, perform parameter difference analysis, and obtain the parameter differences between multiple main network and grid-connected points. The parameter difference analysis module is used to analyze the parameter differences of multiple main grid-connection points when the parameter differences of multiple main grid-connection points do not meet the preset grid connection conditions, and to generate multiple voltage regulation signals and multiple speed regulation signals. The grid connection operation module is used to output the multiple voltage regulation signals and multiple speed regulation signals to the automatic voltage regulators and speed control boards of multiple grid connection points to adjust their operating status until the multiple grid connection points meet the preset grid connection conditions and complete the grid connection operation.

Images