Charging pile collaborative control method and system

Abstract

The present application provides a charging pile collaborative control method and system. Specifically, the present application obtains the charging migration sequence data of electric vehicle users, parses and generates the migration preference matrix and charging interval distribution of charging piles across service areas; and simultaneously measures the acoustic wave delay data and the charging pile voltage fluctuation sequence. The acoustic wave delay is converted into a delay gradient matrix, and combined with the voltage fluctuation correlation coefficient matrix, an environmental correlation matrix is ​​generated. Based on the matrix, super-threshold charging piles are aggregated into charging clusters and the boundary coordinates are output. The migration preference, charging interval distribution and boundary coordinates are input into the prediction model, and the future vehicle density distribution map and the predicted value of the average stay time are output. The load priority vector is calculated in combination with the environmental correlation matrix. Using this vector as the weight, the total power of the service area is allocated to each charging pile, and a power adjustment instruction is issued. The present application realizes the dynamic matching of power distribution of the power grid with the real physical environment and user needs.

Description

Technical Field

[0001] The present application relates to the technical field of smart grid and new energy vehicle infrastructure, and in particular to a charging pile collaborative control method and system. Background Art

[0002] In centralized electric vehicle charging scenarios, such as highway service areas and urban fast-charging stations, the temporal and spatial imbalance of user charging behavior often leads to resource mismatches, where some charging piles are crowded while adjacent ones remain idle. A collaborative control method is urgently needed that can perceive the physical environmental correlations between charging piles, accurately predict dynamic charging demand, and automatically optimize the allocation strategy based on the available power of the grid to improve overall charging efficiency and grid stability in the service area.

[0003] Currently, a typical solution uses a time-series prediction model based on historical charging data. This model collects hourly charging power records from each charging station over the past week, trains a long-short-term memory neural network to predict the load demand of each charging station in the future, and then prioritizes and allocates more power to high-load charging stations based on the predicted values. This solution replaces traditional manual scheduling with machine learning and a power pre-allocation mechanism, achieving basic intelligent control.

[0004] However, this solution does not fully consider the physical environment constraints between charging piles, such as the differences in signal propagation characteristics between devices caused by the distribution of obstacles, and the coupling effect of power grid fluctuations in adjacent charging piles. At the same time, it only relies on historical charging data and lacks dynamic analysis of user migration path preferences and charging interval distribution, resulting in insufficient accuracy of the prediction model in sudden traffic migration scenarios. Summary of the Invention

[0005] The present application provides a charging pile collaborative control method and system to solve the problem of insufficient accuracy in sudden traffic migration scenarios in the prior art.

[0006] In a first aspect, the present application provides a charging pile collaborative control method, comprising:

[0007] Acquire charging migration sequence data of electric vehicle users, analyze the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measure the acoustic wave delay dataset at a fixed distance between charging piles, and simultaneously collect the voltage fluctuation value sequence of each charging pile;

[0008] Converting the acoustic wave delay data set into a delay gradient matrix, calculating the correlation coefficient matrix between the voltage fluctuation value sequences, and fusing the delay gradient matrix and the correlation coefficient matrix to generate an environmental correlation matrix between charging piles;

[0009] According to the environmental correlation matrix, the charging piles whose index values ​​exceed the set threshold are dynamically aggregated into charging clusters through a clustering algorithm, and the boundary coordinates of each charging cluster are output;

[0010] The migration preference matrix, charging interval distribution, and boundary coordinates are input into a prediction model, and a vehicle density distribution map and average dwell time prediction value for each charging cluster in a future time window are output; based on the vehicle density distribution map, average dwell time prediction value, and the environmental correlation matrix of the charging clusters, a load priority vector for each charging cluster is calculated;

[0011] The load priority vector is used as the allocation weight, and the total available power of the service area is proportionally allocated to each charging pile, and a charging pile power adjustment instruction is generated and issued for execution.

[0012] Optionally, the migration preference matrix, charging interval distribution, and boundary coordinates are input into a prediction model, and a vehicle density distribution map and average dwell time prediction value of each charging cluster in a future time window are output. Based on the vehicle density distribution map, average dwell time prediction value, and the environmental correlation matrix of the charging cluster set, a load priority vector of each charging cluster is calculated, including:

[0013] Extracting the boundary coordinates of each charging cluster, obtaining the change trend of the number of user groups entering and leaving the boundary coordinate areas within a continuous time window from the migration preference matrix, and extracting the interval characteristics of adjacent charging events in the charging interval distribution;

[0014] Input the boundary coordinates, quantity change trend and interval characteristics into a time series prediction model, and output a vehicle density distribution map per unit area and a predicted average stay time of the charging cluster in a future time window;

[0015] Extracting the highest density value from the vehicle density distribution map, multiplying the highest density value by the average dwell time prediction value to obtain a demand heat value, and averaging the environmental correlation scores of all charging pile combinations in the charging cluster to obtain an environmental correlation strength;

[0016] The demand heat value and the environment association strength are superimposed according to preset weights to obtain a comprehensive load index, and the comprehensive load indexes of all charging clusters are normalized to obtain a load priority vector.

[0017] Optionally, converting the acoustic wave delay data set into a delay gradient matrix, calculating the correlation coefficient matrix between the voltage fluctuation value sequences, fusing the delay gradient matrix and the correlation coefficient matrix to generate an environmental correlation matrix between charging piles, including:

[0018] Taking any two charging piles as a charging pile combination, obtain multiple acoustic wave delay data sets at a fixed distance between the charging pile combination;

[0019] Calculating the difference between the maximum delay value and the minimum delay value in the acoustic wave delay data as the acoustic wave delay variation amplitude, and generating a delay gradient matrix according to the acoustic wave delay variation amplitude;

[0020] Obtaining a voltage fluctuation value sequence of the charging pile combination within the same time window, and calculating a correlation coefficient matrix of the voltage fluctuation value sequence;

[0021] The values ​​in the delay gradient matrix and the correlation coefficient matrix are weighted and superimposed according to a preset ratio to generate the environmental relevance score of the group of charging piles, and the environmental relevance scores of all the charging pile combinations are arranged to form an environmental relevance matrix.

[0022] Optionally, the values ​​in the delay gradient matrix and the correlation coefficient matrix are weighted and superimposed according to a preset ratio to generate the environmental relevance score of the group of charging piles, and the environmental relevance scores of all the charging pile combinations are arranged to form an environmental relevance matrix, including:

[0023] Extracting the acoustic wave delay variation amplitude corresponding to the charging pile combination from the delay gradient matrix, and multiplying the acoustic wave delay variation amplitude by a first preset coefficient to obtain an acoustic wave impact component;

[0024] Extracting the fluctuation correlation degree values ​​of the two groups of voltage fluctuation value sequences of the charging pile combination from the correlation coefficient matrix, and multiplying the fluctuation correlation degree values ​​by a second preset coefficient to obtain a voltage influence component;

[0025] The acoustic wave influence component and the voltage influence component are added to generate the environment relevance score of the charging pile combination, and the environment relevance scores of all charging pile combinations are arranged into a symmetrical square matrix in the order of charging pile numbers to form the environment relevance matrix.

[0026] Optionally, the boundary coordinates, quantity change trend, and interval characteristics are input into a time series prediction model to output a vehicle density distribution map per unit area and a predicted average stay time of the charging cluster in a future time window, including:

[0027] Calculate the coverage area of ​​the rectangular area enclosed by the boundary coordinates of each charging cluster, convert the number change trend into a sequence of changes in the number of users at the same time of day, and process the interval feature as a typical time difference between adjacent charging behaviors;

[0028] Combining the coverage area, user number change sequence, and typical time difference into input data blocks, arranging the input data blocks in chronological order and feeding them into a time series prediction model, thereby simultaneously obtaining the total number of vehicles and the average vehicle occupancy time in each future time period through the model;

[0029] The total number of vehicles in each future time period is divided by the coverage area of ​​the corresponding time period to form a time-series sequence of the number of vehicles per unit area as a vehicle density distribution map, and the average occupancy time of the vehicles is used as the average stay time prediction value.

[0030] Optionally, the total available power in the service area is proportionally distributed to each charging pile using the load priority vector as the distribution weight, and a charging pile power adjustment instruction is generated and issued for execution, including:

[0031] Obtain the total available power value of the service area power grid, extract the weight coefficient of each charging cluster from the load priority vector, multiply the total available power value by the weight coefficient of each charging cluster, and obtain the power allocation value of each charging cluster:

[0032] Obtain the total number of charging piles in each charging cluster, and divide the power allocation value by the total number of charging piles to obtain a power setting value for each charging pile;

[0033] An adjustment instruction including an identifier and the power setting value is created for each charging pile, and the adjustment instruction is sent to the corresponding charging pile controller to perform power adjustment.

[0034] Optionally, according to the environmental correlation matrix, charging piles whose index values ​​exceed a set threshold are dynamically aggregated into charging clusters through a clustering algorithm, and the boundary coordinates of each charging cluster are output, including:

[0035] Setting a minimum connection threshold of the environment correlation score in the environment correlation matrix, and initializing each charging pile as an independent charging cluster;

[0036] Identify charging pile association relationships greater than a minimum connection threshold in the environmental association matrix, extract the two charging piles corresponding to each association relationship that meets the conditions, and if the two charging piles belong to different charging clusters, merge them into a charging cluster set, and iterate the operation until no new charging cluster set is generated;

[0037] Obtain the composition information of all current charging clusters, extract the geographic coordinates of the charging piles in the charging cluster, integrate all the geographic coordinates into a geographic coordinate set, calculate the minimum and maximum values ​​of the horizontal and vertical coordinates of the geographic coordinate set to form boundary coordinates, and output the boundary coordinates of each charging cluster.

[0038] In a second aspect, the present application provides a charging pile collaborative control system, comprising:

[0039] An acquisition module acquires charging migration sequence data of electric vehicle users, analyzes the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measures the acoustic wave delay dataset at a fixed distance between charging piles, and simultaneously collects the voltage fluctuation value sequence of each charging pile;

[0040] a calculation module, which converts the acoustic wave delay data set into a delay gradient matrix, calculates a correlation coefficient matrix between the voltage fluctuation value sequences, and fuses the delay gradient matrix with the correlation coefficient matrix to generate an environmental correlation matrix between charging piles;

[0041] an aggregation module, which dynamically aggregates charging piles whose index values ​​exceed a set threshold into charging clusters based on the environmental correlation matrix through a clustering algorithm, and outputs the boundary coordinates of each charging cluster;

[0042] a prediction module that inputs the migration preference matrix, the charging interval distribution, and the boundary coordinates into a prediction model, outputs a vehicle density distribution map and an average dwell time prediction value for each charging cluster in a future time window, and calculates a load priority vector for each charging cluster based on the vehicle density distribution map, the average dwell time prediction value, and the environmental correlation matrix of the charging clusters;

[0043] The generation module uses the load priority vector as the distribution weight, distributes the total available power of the service area to each charging pile in proportion, generates a charging pile power adjustment instruction, and issues it for execution.

[0044] In a third aspect, the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a charging pile collaborative control method as described in the first aspect above.

[0045] In a fourth aspect, the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, a charging pile collaborative control method as described in the first aspect is implemented.

[0046] This application builds a foundation for environmental perception and behavior prediction by acquiring user charging migration sequences and charging pile physical status data; then, it integrates the acoustic delay gradient and voltage fluctuation correlation to generate an environmental correlation matrix, quantifies the physical space obstacle distribution and grid coupling strength between charging piles, and reveals the implicit correlation between devices; then, based on the matrix, it dynamically aggregates highly correlated charging piles into charging clusters and outputs boundary coordinates to form a collaborative control physical unit; then, it combines migration preferences, charging interval distribution, and spatial boundary input prediction models to output future vehicle density distribution maps and estimated dwell time values ​​to achieve refined spatiotemporal demand modeling; finally, it couples demand heat with environmental correlation strength to calculate the load priority vector, drives the total power of the service area to be accurately distributed to each charging pile according to the weight ratio, and simultaneously issues power adjustment instructions to completely eliminate the resource mismatch problem of local overload and idleness in the power grid. The entire process breaks through the limitations of traditional solutions that separate environmental variables from user behavior, corrects demand prediction deviations with acoustic / voltage environmental correlation data, and achieves three-dimensional collaborative optimization of physical environment constraints, dynamic load demand, and grid resources.

[0047] Furthermore, by extracting the boundary coordinates of each charging cluster and combining the user number change trend and adjacent event interval characteristics of the charging interval distribution from the migration preference matrix, the model is fed into a time series prediction model to output a future vehicle density distribution map and dwell time predictions. The highest density value in the density distribution map is then multiplied by the dwell time estimate to obtain the demand heat value. The average environmental correlation score of all charging piles within the cluster is taken to obtain the environmental correlation strength. The two are weighted and normalized to form a load priority vector. This approach couples user behavior prediction with the physical environment correlation strength driven by spatial boundaries, quantifying the comprehensive load indicators of charging clusters and providing a weighted basis for power allocation that integrates dynamic demand and environmental constraints. This significantly improves the spatiotemporal precision of resource allocation and addresses the grid zoning mismatch problem caused by the lack of environmental variables in traditional solutions.

[0048] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0050] Figure 1 A flowchart of a charging pile collaborative control method provided by the present application is shown;

[0051] Figure 2A scenario diagram showing a charging pile collaborative control method provided by the present application is shown;

[0052] Figure 3 The following is a schematic diagram showing the structure of a charging pile collaborative control system provided by the present application;

[0053] Figure 4 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0054] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0055] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0056] Researchers have found that traditional solutions ignore differences in signal propagation between charging stations due to the distribution of obstacles, such as acoustic wave delay characteristics, and the coupling effects of grid fluctuations on adjacent devices, leading to local overloads or resource waste. Furthermore, they rely on static historical charging data and lack dynamic analysis of user cross-region migration path preferences and charging interval patterns. This leads to severely insufficient prediction accuracy in bursty traffic scenarios, such as holiday migration peaks, resulting in power allocation imbalances and grid stability risks. Therefore, a collaborative control method that integrates physical environment constraints with user behavior dynamics is urgently needed.

[0057] In response to the above problems, the present invention proposes a collaborative control method for charging piles, the core of which is to build a dynamic response mechanism through multi-source data fusion. Specifically, the acoustic wave delay data set and voltage fluctuation sequence at a fixed distance between charging piles are first synchronously collected, and the environmental correlation matrix is ​​generated by fusion to quantify the physical environment and power grid coupling relationship between the devices; secondly, the user charging migration sequence data is analyzed to extract the migration preference matrix and charging interval distribution, and the prediction model is input with the charging cluster boundary coordinates to output the vehicle density distribution and dwell time prediction value in the future period; finally, based on the load priority vector, the vehicle density, dwell time and environmental correlation are comprehensively considered to dynamically allocate the total power of the service area. This method thoroughly solves the problem of physical environment constraints, accurately captures the signal propagation differences and power grid coupling effects, and at the same time realizes dynamic response to the user's sudden migration behavior. Actual measurements show that the risk of local overload of the power grid is reduced, breaking through the control bottleneck under the traditional single data dimension.

[0058] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0059] Figure 1 A flowchart of a charging pile collaborative control method is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0060] 101. Obtain charging migration sequence data of electric vehicle users, analyze the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measure the acoustic wave delay dataset at a fixed distance between charging piles, and simultaneously collect the voltage fluctuation value sequence of each charging pile;

[0061] In the above steps, the charging migration sequence data of electric vehicle users refers to the historical data set that records the charging operations of electric vehicle users at different time points, including charging pile location information, charging start and end timestamps, and service area identification. The migration preference matrix refers to a mathematical matrix statistically generated based on the user's charging behavior within a continuous time window, which represents the probability or preference of the user to migrate from the current service area charging pile to the charging pile in other service areas. The charging interval distribution refers to the statistical probability distribution of the time intervals between adjacent charging events of the user, which is used to describe the random characteristics of the charging frequency. The sound wave delay data set refers to the sound propagation time delay data set measured under the condition of a fixed physical distance between multiple charging piles, reflecting the influence of the propagation speed of sound waves in the air medium. The voltage fluctuation value sequence refers to a continuous data sequence that records the voltage value of the charging pile changing with time when it is in working state, capturing the dynamic characteristics of the power supply stability.

[0062] In this embodiment, the system first collects detailed historical records of each electric vehicle user's charging. These records include the user's identity, the specific location of the charging station, the specific time of charging start and end, and the service area of ​​the charging station. For example, the system finds the record of user U123, showing that he used the charging station numbered P1 (with known location coordinates) to charge from 8:00 to 10:00 on October 1, 2023. The system organizes these charging records of all users in chronological order to form a structured data set.

[0063] Next, the system will analyze two key pieces of information from this data set: one is the user's preference pattern for "jumping" between charging piles in different service areas, and the other is the distribution pattern of the interval time between two charges of the user.

[0064] When analyzing "jump" preferences, the system treats each service area as a point and calculates the probability that a user, after charging at one service area, will go to another service area for the next charge. This probability is represented by a numerical matrix, the migration preference matrix. The specific calculation method is to count the number of times a user goes directly from service area i to service area j to charge, and then divide it by the total number of times the user has traveled from service area i to all other locations to obtain the probability of jumping from i to j. The specific formula is: ,in, is the probability of migrating from the starting service area i to the target service area j, is the number of migrations, and k represents the indexes of all possible target service areas.

[0065] When analyzing charging interval patterns, the system counts the time intervals between two consecutive charging events for all users (for example, from the end of the first to the start of the second), and then uses a specific mathematical curve (gamma distribution) to describe the overall distribution characteristics of these time intervals, finding the most likely interval time and fluctuation range. For example, when calculating the charging interval distribution, a gamma distribution fitting algorithm is used, specifically extracting a dataset of time intervals between adjacent charging events and estimating the shape parameters using the maximum likelihood estimation method. and scale parameters Generate the cumulative distribution function (CDF), where the gamma distribution probability density function formula is expressed as , where t is the time interval (unit: minutes), is the shape parameter controlling the shape of the distribution, is the scale parameter representing the distribution expansion factor, The gamma function value is used to normalize the distribution, and the calculation results are saved for subsequent use. Once the optimal and This "gamma mold" becomes the optimal mathematical model for describing the charging interval patterns of this group of users. The horizontal axis (time value) corresponding to the highest point of this model curve is the statistically most frequent charging interval, and the width of this model curve intuitively reflects the fluctuation range of the interval time.

[0066] At the same time, in order to understand the impact of the physical environment around the charging piles, the system will install sound sending and receiving devices between selected charging piles (for example, a fixed distance of 50 meters); the system will let one pile send a sound signal and the other pile receive it, accurately measuring how long it takes for the sound to be sent and received (delay), and repeat the measurement multiple times (for example, 30 times) to take the average value to obtain a reliable sound propagation time data set.

[0067] Finally, to monitor the power supply during charging, the system installs a voltmeter at the charging station connection point, ensuring that voltage and sound measurements are synchronized (timed using the same clock). This voltmeter records voltage changes at a rapid rate (e.g., 1,000 times per second) and then calculates the average amplitude (standard deviation) of voltage fluctuations over a short period of time (a window), generating a time-varying voltage stability data series. This allows the system to simultaneously capture user charging behavior (jump preferences and interval patterns), physical environment characteristics (sound propagation delay), and power conditions (voltage fluctuations), providing a foundation for subsequent comprehensive analysis.

[0068] For example, in a research project with 300 electric vehicle owners, the system first collected complete charging records for each of them, covering approximately 10 times over a period of time (e.g., 15 consecutive weeks). The system then analyzed these records: it counted users' charging movements between eight different service areas and calculated an 8-row, 8-column matrix (the migration preference matrix). Each number in this matrix represented the probability that a user would charge in one service area and then move to another (e.g., the probability of switching from service area A to service area B was 0.3). Furthermore, the system used mathematical methods (gamma distribution) to fit a typical distribution model for charging intervals based on a total of more than 18,000 charging intervals for all users (e.g., user U123's first charge ended two days before the start of the second, user U456's charge ended five hours, and so on). For example, the system found that intervals between 8 and 24 hours had the highest probability. To measure the physical environment, the researchers selected a pair of charging stations 50 meters apart. One station emitted a sound, while the other received it. This experiment was repeated 30 times, accurately recording the propagation time of each sound (e.g., an average of 0.147 seconds), generating a stable data set of acoustic delays. Finally, while simultaneously measuring the sound, a voltmeter recorded the voltage at the charging station connection point 1,000 times per second (e.g., 219.5 volts at one moment and 220.1 volts at another). The researchers then calculated the average magnitude of the voltage fluctuations (standard deviation, e.g., 0.3 volts within a 0.1-second window) within each short period (e.g., every 0.1 second). This generated a 500-point voltage fluctuation curve (a sequence of voltage fluctuation values). All of this data—the user jump probability matrix, the distribution model of charging intervals, the sound propagation time over a distance of 50 meters, and the simultaneously recorded voltage fluctuation curves—was collected and prepared for subsequent analysis, including optimization of the charging network.

[0069] In the overall solution of step 101 above, the ability to comprehensively analyze the charging behavior of electric vehicles is achieved by obtaining the user's charging migration sequence data and parsing this data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window. At the same time, the acoustic wave delay data set at a fixed distance between charging piles is measured and the voltage fluctuation value sequence of each charging pile is synchronously collected. These integrated data streams jointly support the efficient monitoring and optimization of the dynamic behavior and potential problems of the charging system, and improve the reliability and prediction accuracy of the entire system.

[0070] 102. Convert the acoustic wave delay data set into a delay gradient matrix, calculate the correlation coefficient matrix between the voltage fluctuation value sequences, and fuse the delay gradient matrix and the correlation coefficient matrix to generate an environmental correlation matrix between charging piles;

[0071] Optionally, step 102 may specifically include the following steps:

[0072] 1021. Taking any two charging piles as a charging pile combination, obtain a multiple acoustic wave delay data set at a fixed distance between the charging pile combination;

[0073] 1022. Calculate the difference between the maximum delay value and the minimum delay value in the acoustic wave delay data as the acoustic wave delay variation amplitude, and generate a delay gradient matrix according to the acoustic wave delay variation amplitude;

[0074] 1023. Obtain a voltage fluctuation value sequence of the charging pile combination within the same time window, and calculate a correlation coefficient matrix of the voltage fluctuation value sequence;

[0075] 1024. Weightedly superimpose the values ​​in the delay gradient matrix and the correlation coefficient matrix according to a preset ratio to generate an environmental relevance score for the group of charging piles, and arrange the environmental relevance scores of all the charging pile combinations to form an environmental relevance matrix.

[0076] Among them, step 1024 may specifically include the following processes: extracting the sound wave delay variation amplitude corresponding to the charging pile combination from the delay gradient matrix, multiplying the sound wave delay variation amplitude by a first preset coefficient to obtain the sound wave influence component; extracting the fluctuation correlation degree value of the two groups of voltage fluctuation value sequences of the charging pile combination from the correlation coefficient matrix, multiplying the fluctuation correlation degree value by a second preset coefficient to obtain the voltage influence component; adding the sound wave influence component and the voltage influence component to generate the environmental correlation score of the charging pile combination, and arranging the environmental correlation scores of all charging pile combinations into a symmetrical square matrix in the order of the charging pile numbers to form the environmental correlation matrix.

[0077] In the above steps, the acoustic delay dataset refers to a collection of sound propagation time delay data measured at a fixed distance, reflecting the dynamic characteristics of sound propagation. The delay gradient matrix is ​​a data structure generated based on the acoustic delay variation amplitude, representing the difference in the degree of acoustic influence between charging piles. The voltage fluctuation value sequence refers to a continuous data sequence of voltage values ​​collected at the charging pile end over time. The correlation coefficient matrix is ​​a numerical matrix calculated from the voltage fluctuation value sequence, representing the degree of fluctuation correlation between different sequences. The environmental correlation matrix is ​​a symmetric score matrix describing the comprehensive environmental correlation between charging piles, used to quantify the mutual influence of environmental and power factors. A charging pile combination refers to a data processing unit formed by two charging piles selected in pairs. The acoustic delay variation amplitude refers to the difference between the maximum and minimum acoustic delay values ​​in multiple acoustic delay measurements for a specific charging pile combination, representing the stability of acoustic propagation. The fluctuation correlation value is a numerical element in the correlation coefficient matrix, reflecting the strength of the correlation between two voltage series. The environmental correlation score is a single numerical value calculated for each charging pile combination, which comprehensively measures the influence of acoustic and voltage factors. The first preset coefficient is a fixed weight coefficient for the weighted superposition of acoustic factors. The second preset coefficient is a fixed weighting factor for the weighted superposition of voltage factors. The acoustic impact component is the component obtained by multiplying the acoustic delay variation amplitude by the first preset coefficient, quantifying the acoustic contribution. The voltage impact component is the component obtained by multiplying the fluctuation correlation value by the second preset coefficient, quantifying the voltage fluctuation contribution. The preset ratio refers to the weighting rule for the first and second preset coefficients during the weighted superposition process, which is used to adjust the influence ratio of each component.

[0078] In the embodiment of this application, the purpose of describing the entire process is to generate a matrix representing the environmental correlation between charging piles. This helps to understand the location of charging piles in the physical environment and the degree of mutual influence. The solution starts with the stored measurement data:

[0079] The first step is to obtain data for a combination of charging piles through step 1021. Specifically, any two charging piles are paired into a group (for example, charging piles A and B), and the delay measurement values ​​of multiple sound propagation between them (the time required for sound to propagate from one charging pile to another) are extracted. Here, it is assumed that each combination has multiple measurement data to analyze stability. For example, the delay data of charging piles A and B may be three measurements: 150 milliseconds, 180 milliseconds, and 130 milliseconds.

[0080] The second step is to calculate the variation of these delay data through step 1022. This is called the acoustic delay variation amplitude, which is obtained by taking the maximum value minus the minimum value in the delay data (such as the delay values ​​150, 180, and 130 in the above example, the maximum value 180 minus the minimum value 130 equals 50 milliseconds). The variation amplitudes of all combinations will be filled in a symmetric matrix (called the delay gradient matrix), where the charging pile numbers correspond to rows and columns. For example, the variation amplitude of charging pile AB, 50, is filled in the matrix row A and column B.

[0081] The third step is to obtain the numerical sequence of voltage changes for the same charging pile combination through step 1023 (record the voltage fluctuation value of each charging pile in the same period of time), and calculate the degree of correlation between these sequences (called the degree of fluctuation correlation). Here, the Pearson correlation coefficient formula is used for calculation. The formula involves dividing the covariance by the product of the standard deviations of the two sequences. The covariance measures the degree of coordination of sequence changes, and the standard deviation indicates the magnitude of sequence fluctuations (for example, the voltage sequence of charging pile A in the same time period is [0.5V, 0.6V, 0.55V], and the sequence of charging pile B is [0.52V, 0.58V, 0.53V]. First, calculate the average value of sequence A as 0.55V and the average value of sequence B as 0.543V. The covariance is the product of the sequence value minus the average value divided by the number of points minus one: ((0. The square of the standard deviation of sequence A is [(0.5-0.55)^2+(0.6-0.55)^2+(0.55-0.55)^2] / 2≈0.0025, and the square root is 0.05V. The standard deviation of sequence B is calculated similarly, which is approximately 0.04V. The correlation coefficient is the covariance 0.00155 divided by (0.05×0.04), which is equal to 0.775. The results of each combination of this calculation will be arranged into a symmetric matrix (correlation coefficient matrix). For example, the correlation coefficient of charging pile AB is 0.775, which is filled in the corresponding position of the matrix.

[0082] The fourth step is to generate an environmental relevance matrix by fusing the two pieces of information in step 1024: first, extract the variation amplitude of the charging pile combination (e.g., 50 for AB) from the delay gradient matrix and multiply it by a first weight coefficient (e.g., the first coefficient α is set to 0.4) to obtain the acoustic wave influence component S (S = 50 × 0.4 = 20). Second, extract the fluctuation correlation degree value (e.g., 0.775 for AB) from the correlation coefficient matrix and multiply it by a second weight coefficient (e.g., the second coefficient β is set to 0.6) to obtain the voltage influence component V (V = 0.775 × 0.6 ≈ 0.465). Then, simply add S and V to generate the environmental relevance score of the combination (20 + 0.465 ≈ 20.465). After repeating this for all charging pile combinations (e.g., AC, BC, etc.), these scores are organized into a symmetrical square matrix in order of the charging pile numbers (rows and columns correspond to the charging pile numbers, and the values ​​are the corresponding scores). For example, the environmental relevance score of the charging pile combination is filled in row A and column B to become part of the environmental relevance matrix and stored for later use. The entire process transforms raw measurements into meaningful correlation outputs by cyclically processing data, ensuring that each charging pile combination is consistently evaluated to facilitate subsequent analysis of its environmental relevance.

[0083] In the overall solution of step 102 above, by collecting acoustic wave delay data at a fixed distance between charging piles and converting it into a delay gradient matrix, the correlation coefficient matrix of the voltage fluctuation value sequence is calculated. Then, the core parameters in the two matrices, the acoustic wave delay variation amplitude and the voltage fluctuation correlation value, are weighted and fused according to a preset ratio to generate the environmental correlation score of the charging pile combination. Finally, by systematically arranging the scores of all combinations, an environmental correlation matrix in the form of a symmetrical square matrix is ​​constructed, achieving three-dimensional quantitative diagnosis capabilities for the coupling strength of environmental interference in the charging pile cluster. First, the environmental interference transmission path is reflected by the change in the stability of acoustic wave propagation. Second, the synchronization characteristics of voltage fluctuations are used to capture the abnormal correlation of the power grid state. Finally, based on the fusion analysis of multiple physical quantities, an acoustic and electrical joint diagnosis model is established, which simultaneously reveals the fault conduction risk and electromagnetic compatibility characteristics, providing the charging network with fault tracing and environmental risk warning capabilities, significantly improving the collaborative operation reliability and risk active defense accuracy of the charging facility group under complex working conditions.

[0084] 103. Based on the environmental correlation matrix, dynamically aggregate charging piles whose index values ​​exceed a set threshold into charging clusters using a clustering algorithm, and output the boundary coordinates of each charging cluster;

[0085] Optionally, step 103 may specifically include the following steps:

[0086] 1031. Set a minimum connection threshold of the environment correlation score in the environment correlation matrix, and initialize each charging pile as an independent charging cluster;

[0087] 1032. Identify charging pile association relationships greater than a minimum connection threshold in the environmental association matrix, extract the two charging piles corresponding to each association relationship that meets the conditions, and if the two charging piles belong to different charging clusters, merge them into a charging cluster set. Iterate the operation until no new charging cluster set is generated.

[0088] 1033. Obtain the composition information of all current charging clusters, extract the geographic coordinates of the charging piles in the charging cluster, integrate all the geographic coordinates into a geographic coordinate set, calculate the minimum and maximum values ​​of the horizontal and vertical coordinates of the geographic coordinate set to form boundary coordinates, and output the boundary coordinates of each charging cluster.

[0089] In the above steps, the environmental correlation matrix refers to the symmetric score matrix output from step 102 that describes the comprehensive environmental correlation between charging piles. A clustering algorithm refers to a computational method used to group data, forming groups based on the relationships between data points. An index value refers to the environmental correlation score value in the environmental correlation matrix. Setting a threshold refers to presetting a fixed numerical limit to determine whether the correlation strength meets the criteria. Charging piles refer to electric vehicle charging facilities and equipment. Dynamic aggregation refers to the process of grouping data points according to rules. A charging cluster refers to a group of charging piles generated by the clustering algorithm. Boundary coordinates refer to location coordinate data describing the spatial extent of a charging cluster, typically including extreme longitude and latitude values. The minimum connection threshold refers to the specific lower limit of the environmental correlation score set in step 1031. An independent charging cluster refers to a group state in which each charging pile is initially treated as an independent individual. A charging pile association relationship refers to a pair of charging pile objects whose environmental correlation scores exceed a threshold. Geographic coordinates refer to location information data of a charging pile, such as longitude and latitude values. A geographic coordinate set refers to the aggregated geographic coordinates of all charging piles within a charging cluster. The minimum and maximum values ​​of the horizontal and vertical coordinates refer to the position boundary values ​​of all points in the direction of the coordinate axis and are used to define the spatial range.

[0090] In the embodiment of the present application, first, through step 1031, a minimum connection score (minimum connection threshold) is set based on the correlation score table (environmental correlation matrix) that describes the degree of environmental similarity between charging piles, and each charging pile is treated as an independent group (independent charging cluster).

[0091] Next, in step 1032, charging pile associations in the environmental association matrix that exceed a minimum connection threshold are identified and dynamically aggregated into charging clusters. Based on the thresholded matrix data, the initial charging cluster list data, and the threshold, a matrix scanning algorithm is used to traverse all charging pile pairs in the environmental association matrix. The environmental association score is checked to see if it exceeds the minimum connection threshold. Charging pile pairs that meet the criteria are then extracted using charging cluster relationship detection techniques. If the charging pile pairs belong to different charging clusters, their clusters are merged using a cluster merging algorithm. This process is iterated until no new mergers occur, and updated charging cluster data is output. In other words, all charging pile pairings with scores exceeding the set minimum score are found in the association score table in step 1031. For each matching pair of charging piles, a check is performed to see if they belong to different groups. If so, the two groups are merged into a single large group (charging cluster).

[0092] Finally, in step 1033, the composition information of all current charging clusters is obtained, and the geographic coordinates are extracted and integrated into a set, and the boundary coordinates are calculated. Based on the charging cluster data, a data query technique is used to read the list of charging piles in each charging cluster. A coordinate extraction algorithm is applied to the stored charging pile geographic information to obtain the geographic coordinates of all members. Coordinate integration techniques are used to aggregate all coordinates into a geographic coordinate set. A boundary calculation algorithm is used to calculate the minimum and maximum values ​​of the horizontal coordinates and the minimum and maximum values ​​of the vertical coordinates of this set to form boundary coordinate data. Finally, the boundary coordinates of each charging cluster are output to the data storage system. In other words, the process of searching for matching pairs and merging groups is repeated in step 1032 until no new matching pairs can be merged. Then, for each large group that is finally formed, obtain the specific location information (geographic coordinates) of all charging piles in the group and collect all these location points; find the easternmost and westernmost longitude values ​​(minimum and maximum values ​​of the horizontal coordinate) and the northernmost and southernmost latitude values ​​(minimum and maximum values ​​of the vertical coordinate) of this group of location points; use these four values ​​(east-west longitude boundary, north-south latitude boundary) to determine the rectangular area range (boundary coordinates) occupied by this large group of charging piles on the map, and finally output this range data for each large group.

[0093] In the overall solution of the above step 103, the spatial mapping capability of the environmental correlation of the charging facility network is realized. By setting the minimum connection threshold of the environmental correlation matrix, each charging pile is initialized as an independent charging cluster, and the correlation relationship above the threshold in the matrix is ​​dynamically scanned. When a pair of charging piles with strong correlation is identified, an iterative merging operation is performed on the nodes belonging to different charging clusters until a stable charging cluster architecture is formed; on this basis, the geographic coordinate set of all charging piles in the cluster is extracted, and the horizontal and vertical coordinate extreme value extraction algorithm is used to generate the coordinates of the rectangular bounding box, and finally the charging cluster boundary range with spatial clarity is output.

[0094] 104. Input the migration preference matrix, charging interval distribution, and boundary coordinates into a prediction model, output a vehicle density distribution map and average dwell time prediction value for each charging cluster in a future time window, and calculate a load priority vector for each charging cluster based on the vehicle density distribution map, average dwell time prediction value, and the environmental correlation matrix of the charging clusters.

[0095] Optionally, step 104 may specifically include the following steps:

[0096] 1041. Extract the boundary coordinates of each charging cluster, obtain the change trend of the number of user groups entering and exiting the boundary coordinate area within a continuous time window from the migration preference matrix, and extract the interval characteristics of adjacent charging events in the charging interval distribution;

[0097] 1042. Input the boundary coordinates, quantity change trend, and interval characteristics into a time series prediction model, and output a vehicle density distribution map per unit area and a predicted average stay time of the charging cluster in a future time window;

[0098] Among them, step 1042 may specifically include the following processes: calculating the coverage area of ​​the rectangular area enclosed by the boundary coordinates of each charging cluster, converting the quantity change trend into a user quantity change sequence at the same time period every day, and processing the interval feature into a typical time difference between adjacent charging behaviors; combining the coverage area, user quantity change sequence and typical time difference into an input data block, arranging the input data block in chronological order and sending it to a time series prediction model, through which the total number of vehicles and the average vehicle occupancy time values ​​in each future time period are simultaneously obtained; dividing the total number of vehicles in each future time period by the coverage area of ​​the corresponding time period to form a sequence of the number of vehicles per unit area arranged in chronological order as a vehicle density distribution map, and using the average vehicle occupancy time value as the average stay time prediction value.

[0099] 1043. Extract the highest density value in the vehicle density distribution map, multiply the highest density value by the predicted average stay time to obtain a demand heat value, and average the environmental correlation scores of all charging pile combinations in the charging cluster to obtain an environmental correlation strength.

[0100] 1044. Superimpose the demand heat value and the environment association strength according to a preset weight to obtain a comprehensive load index, normalize the comprehensive load indexes of all charging clusters, and obtain a load priority vector.

[0101] In the above steps, the migration preference matrix is ​​a mathematical matrix generated based on user charging behavior within a continuous time window, representing the probability distribution of users migrating from a charging station in the current service area to another charging station in another service area. The charging interval distribution is the statistical probability distribution of the time intervals between adjacent charging events. The boundary coordinates refer to the spatial location range of the charging cluster, represented by the minimum and maximum values ​​of the horizontal and vertical coordinates. The prediction model refers to a time series prediction algorithm used for future trend analysis. The vehicle density distribution map is the sequence of vehicles per unit area in each future time period output by the prediction model, reflecting the vehicle distribution density within the charging cluster area. The average dwell time prediction value is the average time that vehicles occupy the charging cluster area in the future, output by the prediction model. The charging cluster refers to the set of charging station groups generated after clustering. The environmental correlation matrix is ​​a symmetric score matrix that describes the comprehensive environmental correlation between charging stations. The load priority vector is the normalized vector of the final calculated priority scores for each charging cluster, used to represent the order in which charging requests are processed. The area quantity change trend is the time series of the number of users entering and exiting the boundary coordinate area. The interval feature refers to the typical time difference between adjacent charging events in the charging interval distribution. The demand heat value is the product of the highest density value in the vehicle density distribution map and the predicted average dwell time, reflecting peak demand intensity. The environmental correlation strength is the average of the environmental correlation scores for all charging pile combinations within a charging cluster, representing an indicator of environmental stability. The comprehensive load index is a weighted composite score derived from the combination of the demand heat value and the environmental correlation strength. Normalization is the mathematical process of converting the comprehensive load index into a relative priority.

[0102] In the embodiment of the present application, the core of this step is to combine three types of data: user behavior, charging habits, and spatial range, to predict the vehicle density and usage time of each charging area in the future, and finally calculate the priority ranking of different areas.

[0103] First, for each assigned charging cluster (i.e., the charging pile group formed in step 103), extract its boundary coordinates (the rectangular area occupied by the group on the map, such as the easternmost / westernmost longitude and the southernmost / northernmost latitude). Simultaneously, extract the changing trends in the number of users entering and leaving the area over multiple consecutive time periods (e.g., 24 hours a day) from the migration preference matrix (a table that records the flow patterns of user groups between regions) (e.g., 10 people entering at 8:00 AM and 5 leaving at 9:00 AM). Also, extract typical interval characteristics from the charging interval distribution (statistical data on the time difference between two user charges) (e.g., most users charge every 45 minutes).

[0104] Next, these three types of data are input into the time series prediction model: first calculate the area of ​​the rectangle enclosed by the cluster boundary coordinates (for example, 100 meters east-west span × 60 meters north-south span = 6000 square meters), convert the change in the number of users into a daily number sequence at the same time period (for example, the number of users in the area at 9 am for three consecutive days [20, 22, 19]), and combine it with the typical charging interval (45 minutes) as the input data block. The model (such as LSTM) will output the expected total number of vehicles and average stay time in the area in the future period (for example, it is predicted that 25 vehicles will stay for 30 minutes at 9:00 on the fourth day); divide the predicted total number of vehicles by the cluster area to obtain the vehicle density distribution map at different time periods (for example, 25 vehicles / 6000m 2 ≈4.17 vehicles / 1,000 square meters), with the average dwell time directly output as the predicted value (30 minutes). The demand heat value is then calculated by extracting the highest density value in the density distribution map (e.g., 12 vehicles / 1,000 square meters during the morning rush hour) and multiplying it by the average dwell time (30 minutes). This gives 360 vehicle minutes / 1,000 square meters, reflecting the cumulative impact of the high-load period in the area. The average environmental correlation score (environmental correlation strength, for example, an average of 18.5) is also calculated for all pairwise combinations of charging piles within the cluster.

[0105] Finally, the demand heat value and environmental association strength are superimposed according to the preset weights (e.g., 360×0.7 weight + 18.5×0.3 weight = 252 + 5.55 = 257.55) to obtain the comprehensive load index of the cluster. The indicators of all charging clusters are normalized (compressed to the range of 0-1) to form a load priority vector (e.g., the two cluster indicators 257.55 and 180 are normalized to 0.82 and 0.57). The higher the value, the higher the scheduling priority.

[0106] In the overall solution of the above step 104, the pre-diagnosis optimization capability of the intelligent control of the charging load is realized. By collaboratively inputting the charging cluster boundary coordinates, the user migration preference matrix and the charging interval distribution characteristics into the time series prediction model, the vehicle density distribution and the average residence time prediction values ​​for each future time period are first output; then, the demand heat index is obtained based on the product of the peak density and the residence time in the vehicle density distribution diagram, and the mean value of the environmental correlation matrix within the charging cluster is simultaneously calculated as the facility group coupling strength; finally, the two core parameters are integrated into a comprehensive load index according to the preset weights, and the load priority vector is generated after normalization.

[0107] 105. Using the load priority vector as the allocation weight, the total available power of the service area is proportionally allocated to each charging pile, and a charging pile power adjustment instruction is generated and issued for execution.

[0108] Optionally, step 105 may specifically include the following steps:

[0109] 1051. Obtain the total available power value of the service area power grid, extract the weight coefficient of each charging cluster from the load priority vector, multiply the total available power value by the weight coefficient of each charging cluster, and obtain the power allocation value of each charging cluster:

[0110] 1052. Obtain the total number of charging piles in each charging cluster, and divide the power allocation value by the total number of charging piles to obtain a power setting value for each charging pile;

[0111] 1053. Create an adjustment instruction including an identifier and the power setting value for each charging pile, and send the adjustment instruction to the corresponding charging pile controller to perform power adjustment.

[0112] In the above steps, the load priority vector refers to the normalized priority vector output from step 104, which represents the weight coefficient of the urgency of the demand for each charging cluster. The total available power of the service area refers to the maximum total power supply value that the power grid system can allocate to the charging piles. The allocation weight refers to the normalized coefficient value corresponding to each charging cluster in the load priority vector. The charging cluster refers to the dynamic charging pile group generated in step 103. The power allocation value refers to the power resource value allocated to each charging cluster. The total number of charging piles refers to the number of charging facilities included in a single charging cluster. The power setting value refers to the specific power value ultimately allocated to a single charging pile. The adjustment instruction refers to the control command data packet containing the charging pile identifier and the power setting value. The charging pile controller refers to the hardware control module within the charging pile that performs power regulation. The identifier refers to the coded data that uniquely identifies the charging pile. Power adjustment refers to the operation process of the charging pile changing the power input according to the set value.

[0113] In the embodiment of the present application, first, the total available power value of the service area power grid is obtained through step 1051, and the power allocation weight coefficient value of the charging cluster is extracted based on the load priority vector: the normalized weight coefficient of each charging cluster is read from the load priority vector using data extraction technology. For example, the weight coefficient of charging cluster number C1 is 0.9, and the weight coefficient of charging cluster number C2 is 0.5; the total available power value of the service area is multiplied by the weight coefficient of each charging cluster using a multiplication algorithm, and the formula is expressed as charging cluster power allocation value = ,in is the total available power value of the power grid and the power unit is kW, In order to make the normalized weight coefficient of the corresponding charging cluster in the load priority vector a dimensionless value, a specific example of generating a power allocation value set is that when the total power is 100 kW, the allocation value of charging cluster C1 is 100 multiplied by 0.9, which is 90 kW, and the allocation value of charging cluster C2 is 100 multiplied by 0.5, which is 50 kW.

[0114] Next, the power setting value of each charging pile is calculated in step 1052: based on the power allocation value set, the cluster information query technology is used to obtain the total number of charging piles in each charging cluster. For example, charging cluster C1 contains 3 charging piles and charging cluster C2 contains 5 charging piles. The power allocation value is divided by the total number of charging piles using the division algorithm. The formula is expressed as power setting value = ,in is the charging cluster power allocation value output in step 1051, The calculated value is used as the benchmark for single-pile power control. Specifically, the setpoint for a single pile in charging cluster C1 is 90kW divided by 3, which equals 30kW, and the setpoint for a single pile in charging cluster C2 is 50kW divided by 5, which equals 10kW. This process implements a hierarchical allocation of total grid power to single-pile power. For example, the execution process for charging cluster C1 illustrates this process: A weight of 0.9 is extracted from the load priority vector and multiplied by the total power of 100kW to obtain a cluster-level allocation of 90kW. This is then evenly distributed among the three charging piles in the cluster, resulting in a setpoint of 30kW per pile. This completes the refined scheduling of power resources based on load priority.

[0115] Finally, step 1053 creates an adjustment instruction containing an identifier and the power setting value for each charging pile, and sends the adjustment instruction to the corresponding charging pile controller to perform power adjustment. Based on the power setting value, an instruction construction technology is used to generate a data structure containing the charging pile unique identifier and the power setting value. A communication protocol encapsulation algorithm is applied to form a transmittable adjustment instruction data packet, which is sent to the target charging pile controller through the network transmission module. Finally, the charging pile controller performs the power adjustment operation to change the power input level.

[0116] In the overall solution of the above step 105, the total available power value of the service area power grid is obtained, the weight coefficient of each charging cluster is extracted from the load priority vector, and the power allocation value of each charging cluster is generated by multiplying the total available power by the weight coefficient; secondly, the total number of charging piles in each charging cluster is identified, and the power allocation value is evenly distributed to each charging pile to obtain a power setting value; finally, a digital adjustment instruction containing a unique device identifier and a set value is created for each charging pile, and sent to the charging pile controller through the control network to perform the power adjustment operation, forming a closed-loop control link from decision-making to execution.

[0117] The following is a complete embodiment of steps 101 to 105:

[0118] like Figure 2As shown in the figure, in the implementation of coordinated control in an electric vehicle charging service area, we first collected charging migration sequence data for 300 users. After analysis, we generated an 8×8 migration preference matrix and constructed a charging interval distribution based on 18,000 time interval samples. We also measured 30 acoustic delays between charging piles at a fixed distance of 50 meters and simultaneously collected a 500-point voltage fluctuation sequence for each pile. We then converted the acoustic delay data into a delay gradient matrix reflecting the amplitude of the variation, e.g., the delay difference between a pair of piles was 5 milliseconds. We calculated the correlation coefficient matrix of the voltage fluctuation series, e.g., the correlation coefficient for a pair of piles was 0.85. We then weighted the matrix with 0.4 times the delay variation amplitude and 0.6 times the correlation coefficient to generate an environmental correlation matrix, e.g., a score of 1.02 for a pair of piles. We then set a threshold of 0.8 and used dynamic clustering to group the eight charging piles into two charging clusters: C1, containing five piles, and C2, containing three piles. The boundary regions were then calculated based on the geographical coordinates of the pile groups.

[0119] The migration preference matrix, charging interval distribution, and boundary coordinates are then input into the prediction model. Using the characteristics of a 3,000-square-meter area, the daily vehicle count sequence, and a typical 180-minute charging interval, the model outputs vehicle density distributions of [0.016, 0.017, 0.0167] vehicles per square meter and an average dwell time of 45 minutes for the next three periods. The peak density value of 0.017 and the dwell time are multiplied to obtain a demand heat value of 0.765. Combined with the average environmental correlation score of 0.93 within the charging cluster, a weighting of 0.7:0.3 is used to obtain a charging cluster load index of 0.8175. This is then normalized to form a load priority vector of [0.56, 0.44]. Finally, based on a 5,000-kW grid total power allocation, C1 receives 2,800 kW (5,000 × 0.56), with each of its five charging piles configured at 560 kW. C2 receives 2,200 kW (5,000 × 0.44), with each of its three charging piles configured at 733.3 kW. The system generates power setting instructions and issues them for execution through the communication protocol, thus achieving dynamic coordinated control of charging resources.

[0120] Figure 3 A schematic diagram of the structure of a charging pile collaborative control system is provided for the embodiment of the present application. Figure 3 As shown, the system includes:

[0121] Acquisition module 31 acquires charging migration sequence data of electric vehicle users, analyzes the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measures the acoustic wave delay dataset at a fixed distance between charging piles, and simultaneously collects the voltage fluctuation value sequence of each charging pile;

[0122] A calculation module 32 converts the acoustic wave delay data set into a delay gradient matrix, calculates a correlation coefficient matrix between the voltage fluctuation value sequences, and fuses the delay gradient matrix with the correlation coefficient matrix to generate an environmental correlation matrix between charging piles;

[0123] Aggregation module 33, based on the environmental correlation matrix, dynamically aggregates charging piles whose index values ​​exceed a set threshold into charging clusters through a clustering algorithm, and outputs the boundary coordinates of each charging cluster;

[0124] A prediction module 34 inputs the migration preference matrix, the charging interval distribution, and the boundary coordinates into a prediction model, outputs a vehicle density distribution map and a predicted average dwell time for each charging cluster in a future time window, and calculates a load priority vector for each charging cluster based on the vehicle density distribution map, the predicted average dwell time, and the environmental association matrix of the charging clusters;

[0125] The generation module 35 uses the load priority vector as the distribution weight, distributes the total available power of the service area to each charging pile in proportion, generates a charging pile power adjustment instruction, and issues it for execution.

[0126] Figure 3 The charging pile collaborative control system can execute Figure 1 The implementation principle and technical effects of the charging pile collaborative control method described in the illustrated embodiment will not be elaborated here. The specific manner in which each module and unit performs operations in the charging pile collaborative control system in the above embodiment has been described in detail in the embodiment of the method and will not be elaborated here.

[0127] In one possible design, Figure 3 A charging pile cooperative control system of the embodiment shown can be implemented as a computing device, such as Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42;

[0128] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 42 .

[0129] The processing component 42 is used for the above Figure 1 The charging pile collaborative control method of the embodiment.

[0130] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0131] The storage component 41 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0132] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0133] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0134] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0135] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0136] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A charging pile collaborative control method according to the embodiment shown.

[0137] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0138] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0139] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer or server) to execute the methods described in each embodiment or certain portions of the embodiments.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A charging pile collaborative control method, characterized in that: include: Acquire charging migration sequence data of electric vehicle users, analyze the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measure the acoustic wave delay dataset at a fixed distance between charging piles, and simultaneously collect the voltage fluctuation value sequence of each charging pile; Converting the acoustic wave delay data set into a delay gradient matrix, calculating the correlation coefficient matrix between the voltage fluctuation value sequences, and fusing the delay gradient matrix and the correlation coefficient matrix to generate an environmental correlation matrix between charging piles; According to the environmental correlation matrix, the charging piles whose index values ​​exceed the set threshold are dynamically aggregated into charging clusters through a clustering algorithm, and the boundary coordinates of each charging cluster are output; The migration preference matrix, charging interval distribution, and boundary coordinates are input into a prediction model, and a vehicle density distribution map and average dwell time prediction value for each charging cluster in a future time window are output; based on the vehicle density distribution map, average dwell time prediction value, and the environmental correlation matrix of the charging clusters, a load priority vector for each charging cluster is calculated; The load priority vector is used as the allocation weight, and the total available power of the service area is proportionally allocated to each charging pile, and a charging pile power adjustment instruction is generated and issued for execution.

2. The method according to claim 1, characterized in that The migration preference matrix, the charging interval distribution, and the boundary coordinates are input into a prediction model, and a vehicle density distribution map and an average dwell time prediction value for each charging cluster in a future time window are output. Based on the vehicle density distribution map, the average dwell time prediction value, and the environmental correlation matrix of the charging clusters, a load priority vector for each charging cluster is calculated, including: Extracting the boundary coordinates of each charging cluster, obtaining the change trend of the number of user groups entering and leaving the boundary coordinate areas within a continuous time window from the migration preference matrix, and extracting the interval characteristics of adjacent charging events in the charging interval distribution; Input the boundary coordinates, quantity change trend and interval characteristics into a time series prediction model, and output a vehicle density distribution map per unit area and a predicted average stay time of the charging cluster in a future time window; Extracting the highest density value from the vehicle density distribution map, multiplying the highest density value by the average dwell time prediction value to obtain a demand heat value, and averaging the environmental correlation scores of all charging pile combinations in the charging cluster to obtain an environmental correlation strength; The demand heat value and the environment association strength are superimposed according to preset weights to obtain a comprehensive load index, and the comprehensive load indexes of all charging clusters are normalized to obtain a load priority vector.

3. The method according to claim 1, characterized in that Converting the acoustic wave delay data set into a delay gradient matrix, calculating the correlation coefficient matrix between the voltage fluctuation value sequences, fusing the delay gradient matrix and the correlation coefficient matrix to generate an environmental correlation matrix between charging piles, including: Taking any two charging piles as a charging pile combination, obtain multiple acoustic wave delay data sets at a fixed distance between the charging pile combination; Calculating the difference between the maximum delay value and the minimum delay value in the acoustic wave delay data set as the acoustic wave delay variation amplitude, and generating a delay gradient matrix according to the acoustic wave delay variation amplitude; Obtaining a voltage fluctuation value sequence of the charging pile combination within the same time window, and calculating a correlation coefficient matrix of the voltage fluctuation value sequence; The values ​​in the delay gradient matrix and the correlation coefficient matrix are weighted and superimposed according to a preset ratio to generate the environmental relevance score of the group of charging piles, and the environmental relevance scores of all the charging pile combinations are arranged to form an environmental relevance matrix.

4. The method according to claim 3, characterized in that The values ​​in the delay gradient matrix and the correlation coefficient matrix are weighted and superimposed according to a preset ratio to generate the environmental relevance score of the group of charging piles, and the environmental relevance scores of all the charging pile combinations are arranged to form an environmental relevance matrix, including: Extracting the acoustic wave delay variation amplitude corresponding to the charging pile combination from the delay gradient matrix, and multiplying the acoustic wave delay variation amplitude by a first preset coefficient to obtain an acoustic wave impact component; Extracting the fluctuation correlation degree values ​​of the two groups of voltage fluctuation value sequences of the charging pile combination from the correlation coefficient matrix, and multiplying the fluctuation correlation degree values ​​by a second preset coefficient to obtain a voltage influence component; The acoustic wave influence component and the voltage influence component are added to generate the environment relevance score of the charging pile combination, and the environment relevance scores of all charging pile combinations are arranged into a symmetrical square matrix in the order of charging pile numbers to form the environment relevance matrix.

5. The method according to claim 2, characterized in that Input the boundary coordinates, quantity change trend, and interval characteristics into a time series prediction model, and output a vehicle density distribution map per unit area and a predicted average stay time of the charging cluster in a future time window, including: Calculate the coverage area of ​​the rectangular area enclosed by the boundary coordinates of each charging cluster, convert the number change trend into a sequence of changes in the number of users at the same time of day, and process the interval feature as a typical time difference between adjacent charging behaviors; Combining the coverage area, user number change sequence, and typical time difference into input data blocks, arranging the input data blocks in chronological order and feeding them into a time series prediction model, thereby simultaneously obtaining the total number of vehicles and the average vehicle occupancy time in each future time period through the model; The total number of vehicles in each future time period is divided by the coverage area of ​​the corresponding time period to form a time-series sequence of the number of vehicles per unit area as a vehicle density distribution map, and the average occupancy time of the vehicles is used as the average stay time prediction value.

6. The method according to claim 1, characterized in that Using the load priority vector as the allocation weight, the total available power of the service area is proportionally allocated to each charging pile, and a charging pile power adjustment instruction is generated and issued for execution, including: Obtain the total available power value of the service area power grid, extract the weight coefficient of each charging cluster from the load priority vector, multiply the total available power value by the weight coefficient of each charging cluster, and obtain the power allocation value of each charging cluster: Obtain the total number of charging piles in each charging cluster, and divide the power allocation value by the total number of charging piles to obtain a power setting value for each charging pile; An adjustment instruction including an identifier and the power setting value is created for each charging pile, and the adjustment instruction is sent to the corresponding charging pile controller to perform power adjustment.

7. The method according to claim 1, characterized in that According to the environmental correlation matrix, the charging piles whose index values ​​exceed the set threshold are dynamically aggregated into charging clusters through a clustering algorithm, and the boundary coordinates of each charging cluster are output, including: Setting a minimum connection threshold of the environment correlation score in the environment correlation matrix, and initializing each charging pile as an independent charging cluster; Identify charging pile association relationships greater than a minimum connection threshold in the environmental association matrix, extract the two charging piles corresponding to each association relationship that meets the conditions, and if the two charging piles belong to different charging clusters, merge them into a charging cluster set, and iterate the operation until no new charging cluster set is generated; Obtain the composition information of all current charging clusters, extract the geographic coordinates of the charging piles in the charging cluster, integrate all the geographic coordinates into a geographic coordinate set, calculate the minimum and maximum values ​​of the horizontal and vertical coordinates of the geographic coordinate set to form boundary coordinates, and output the boundary coordinates of each charging cluster.

8. A charging pile collaborative control system, characterized in that: include: An acquisition module is used to obtain charging migration sequence data of electric vehicle users, analyze the charging migration sequence data to obtain the user's migration preference matrix and charging interval distribution across charging piles in the service area within a continuous time window, measure the acoustic wave delay dataset at a fixed distance between charging piles, and synchronously collect the voltage fluctuation value sequence of each charging pile; a calculation module, configured to convert the acoustic wave delay data set into a delay gradient matrix, calculate a correlation coefficient matrix between the voltage fluctuation value sequences, and fuse the delay gradient matrix with the correlation coefficient matrix to generate an environmental correlation matrix between charging piles; an aggregation module, configured to dynamically aggregate charging piles whose index values ​​exceed a set threshold into charging clusters using a clustering algorithm based on the environmental correlation matrix, and output the boundary coordinates of each charging cluster; a prediction module, configured to input the migration preference matrix, the charging interval distribution, and the boundary coordinates into a prediction model, output a vehicle density distribution map and an average dwell time prediction value for each charging cluster in a future time window, and calculate a load priority vector for each charging cluster based on the vehicle density distribution map, the average dwell time prediction value, and the environmental association matrix of the charging clusters; The generation module is used to allocate the total available power of the service area to each charging pile in proportion based on the load priority vector, generate a charging pile power adjustment instruction and issue it for execution.

9. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a charging pile collaborative control method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, a charging pile collaborative control method as described in any one of claims 1 to 7 is implemented.

Images