Adaptive operation and maintenance method and system for charging stations based on XGBoost algorithm

Abstract

The present invention relates to the technical field of charging station management, and specifically to an adaptive operation and maintenance method and system for charging stations based on an XGBoost algorithm. According to the characteristic that abnormal charging pile equipment usually manifests as abnormal current fluctuation, the abnormal current fluctuation in the charging pile is first analyzed to determine the current abnormality parameter to reflect the abnormal equipment state of the charging pile; then, according to the characteristic that abnormal charging operation or frequent charging resulting in current transient pulses will aggravate equipment loss, the equipment loss degree of the charging pile is determined based on the current mutation situation and relevant parameters of the charging process; further, based on the correlation between equipment loss and equipment abnormality and the overall size of the current abnormality parameter and the equipment loss degree, an operation and maintenance characteristic value characterizing the operation and maintenance demand of the charging pile in the current operation and maintenance cycle of the charging station is comprehensively determined; thereby, the operation and maintenance characteristic value is introduced into the XGBoost algorithm, so that the division of the operation and maintenance priority of the charging piles in the charging station is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of charging station management, and in particular to a charging station adaptive operation and maintenance method and system based on an XGBoost algorithm. Background Art

[0002] With the increasing popularity of electric vehicles, demand for charging infrastructure is rapidly increasing. Efficient operation and service quality of charging stations require tiered O&M management. Traditional management methods rely on manual experience or simple rules, resulting in inefficient management. Existing technologies use the XGBoost algorithm to analyze multi-dimensional data (such as device status, environmental parameters, and user behavior) of charging piles within a station, dynamically assigning O&M priorities and enabling refined and preventative maintenance of each charging pile.

[0003] However, due to the fragmentation problem in the multi-dimensional data collection process, coupled with the nonlinear relationship between the charging pile's own status and operation and maintenance needs, data islands occur. Ultimately, the system makes an incorrect assessment of the actual maintenance needs of the charging pile, which ultimately leads to a mismatch of regional charging station operation and maintenance resources. That is, the existing technology has low accuracy in dividing the operation and maintenance priorities of charging piles in charging stations through the XGBoost algorithm, resulting in poor adaptive operation and maintenance effects of charging stations. Summary of the Invention

[0004] In order to solve the technical problem of low accuracy in prior art in prioritizing the operation and maintenance of charging piles in charging stations using the XGBoost algorithm, the present application aims to provide a method and system for adaptive operation and maintenance of charging stations based on the XGBoost algorithm. The technical solutions adopted are as follows:

[0005] The first aspect of the present application provides a charging station adaptive operation and maintenance method based on the XGBoost algorithm, comprising:

[0006] During the current operation and maintenance cycle of the charging station, the current data of each charging pile at all sampling times in each charging time period is obtained;

[0007] Determine the corresponding current anomaly parameters based on the abnormal fluctuations in the local current data in each charging time period; determine the corresponding transient impact intensity based on the distribution of the sudden changes in the local current data in each charging time period; and determine the corresponding equipment loss level based on the transient impact intensity, duration, and time interval of each charging time period.

[0008] Based on the correlation between the abnormal current parameters and the degree of equipment loss for each charging pile in each charging time period during the current charging station operation and maintenance cycle, as well as the overall size, the corresponding operation and maintenance characteristic values ​​are determined; and the charging station operation and maintenance classification is performed based on the operation and maintenance characteristic values ​​combined with the XGBoost algorithm.

[0009] Furthermore, the process of obtaining the abnormal current parameters includes:

[0010] Divide each charging time period into at least two time windows; wherein all time windows have the same length;

[0011] Determine the current abnormality of each time window based on the deviation between the overall magnitude of the current data in each time window and the preset rated current value;

[0012] The average value of the current abnormality of all time windows is used as the overall abnormality; the product of the negative correlation mapping value of the overall abnormality and the time length of the time window is used as the correction time length;

[0013] Each charging time period is divided into at least two correction windows, where the time length of the correction window is the correction time length; based on the principle of obtaining the current anomaly degree of each time window, the current anomaly degree of each correction window is calculated; and the accumulated value of the current anomaly degrees of all correction windows in each charging time period is used as the corresponding current anomaly parameter.

[0014] Furthermore, the process of obtaining the current abnormality degree includes:

[0015] The mean value of the current data at all sampling moments in each time window is taken as the corresponding current effective value; the difference between the current effective value and the preset rated current value is positively correlated to determine the current abnormality of each time window.

[0016] Furthermore, the process of obtaining the transient impact strength includes:

[0017] Based on the local fluctuation similarity between adjacent sampling moments in time sequence, all similar fluctuation moments in each charging time period are screened out; all similar fluctuation moments are traversed to determine all similar fluctuation time periods; the similar fluctuation time periods are all continuous similar fluctuation moments and the first sampling moment before and the first sampling moment after the similar fluctuation time period are not similar fluctuation moments;

[0018] Determine the corresponding local fluctuation characteristic value based on the mean of the fluctuation change reference values ​​of all similar fluctuation moments in each similar fluctuation time period; determine the degree of fluctuation anomaly of each similar fluctuation time period based on the fluctuation deviation between the local fluctuation characteristic value of each similar fluctuation time period and the local fluctuation characteristic value of the adjacent similar fluctuation time period;

[0019] The corresponding transient impact intensity is determined according to the accumulated value of the fluctuation abnormality degree of all similar fluctuation time periods in each charging time period.

[0020] Furthermore, the process of obtaining the similar fluctuation moments includes:

[0021] In each charging time period, the standard deviation of the current data at all sampling moments within the preset neighborhood window of each sampling moment is taken as the corresponding local fluctuation degree; the difference between the local fluctuation degree at each sampling moment and the local fluctuation degree at the next sampling moment is taken as the fluctuation change reference value at each sampling moment; the sampling moment at which the fluctuation change reference value is less than the preset change threshold is taken as the similar fluctuation moment.

[0022] Furthermore, the process of obtaining the degree of fluctuation anomaly includes:

[0023] The local fluctuation characteristic value of the previous similar fluctuation time period of each similar fluctuation time period is used as the first reference fluctuation characteristic value; the local fluctuation characteristic value of the next similar fluctuation time period of each similar fluctuation time period is used as the second reference fluctuation characteristic value;

[0024] determining a degree of difference in fluctuations on both sides of each similar fluctuation time period based on a difference between the second reference fluctuation characteristic value and the first reference fluctuation characteristic value;

[0025] determining a reference fluctuation difference degree for each similar fluctuation time period according to a difference between the local fluctuation characteristic value for each similar fluctuation time period and the second reference fluctuation characteristic value;

[0026] The degree of fluctuation anomaly in each similar fluctuation time period is determined according to the ratio between the reference fluctuation difference degree and the fluctuation difference degree on both sides.

[0027] Furthermore, the process of obtaining the degree of equipment loss includes:

[0028] The time interval between each charging time period and the previous charging time period is used as a reference charging interval;

[0029] The device loss degree of each charging time period is determined according to the product of the negative correlation mapping value of the reference charging interval, the time length of each charging time period, and the transient impact intensity.

[0030] Furthermore, the process of obtaining the operation and maintenance characteristic value includes:

[0031] Arrange the equipment loss degrees of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an equipment loss degree sequence; arrange the current abnormality parameters of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an abnormal current parameter sequence; normalize the Pearson correlation coefficient between the equipment loss degree sequence and the current abnormality parameter sequence to determine the loss abnormality correlation;

[0032] During the current operation and maintenance cycle of the charging station, the accumulated values ​​of the equipment loss degree in all charging time periods are normalized to determine the cumulative loss degree; the accumulated values ​​of the overload capacity parameters in all charging time periods are normalized to determine the cumulative overload parameter;

[0033] Normalization is performed based on the product of the cumulative loss degree, the cumulative overload parameter, and the loss anomaly correlation to determine the operation and maintenance characteristic value of each charging pile in the current operation and maintenance cycle of the charging station.

[0034] Furthermore, the process of performing charging station operation and maintenance classification based on the operation and maintenance characteristic values ​​in combination with the XGBoost algorithm includes:

[0035] The operation and maintenance feature value is input as a feature into the XGBoost model, and based on the output of the XGBoost model, the operation and maintenance risk level of each charging pile in the current operation and maintenance cycle of the charging station is determined.

[0036] In a second aspect, the present application provides a charging station adaptive operation and maintenance system based on the XGBoost algorithm, the system comprising:

[0037] The data acquisition module is used to obtain the current data of each charging pile at all sampling times in each charging time period during the current charging station operation and maintenance cycle;

[0038] The parameter determination module is used to determine the corresponding current anomaly parameters based on the abnormal fluctuation of the local current data in each charging time period; determine the corresponding transient impact intensity based on the sudden change distribution of the local current data in each charging time period; and determine the corresponding equipment loss degree based on the transient impact intensity, charging time period duration, and charging time interval of each charging time period;

[0039] The charging station operation and maintenance grading module is used to determine the corresponding operation and maintenance characteristic values ​​based on the correlation between the current abnormality parameters and the degree of equipment loss corresponding to each charging time period in the current charging station operation and maintenance cycle, as well as the overall size; and to perform charging station operation and maintenance grading based on the operation and maintenance characteristic values ​​combined with the XGBoost algorithm.

[0040] In a third aspect, the present application provides a computer device comprising a memory and a processor. The memory is configured to store computer program code, and the processor is configured to call and execute the computer program code from the memory to perform the method of the first aspect or any embodiment of the first aspect of the present application.

[0041] In a fourth aspect, the present application provides a computer program product, comprising a computer program code. When the computer program code is executed, the method of the first aspect or any embodiment of the first aspect of the present application is performed.

[0042] In a fifth aspect, the present application provides a computer-readable storage medium, which stores computer program code. When the computer program code is executed, it performs the method of the first aspect of the present application or any embodiment of the first aspect.

[0043] This application has the following beneficial effects:

[0044] Based on the characteristic that charging pile equipment abnormalities usually manifest as abnormal current fluctuations, this application first analyzes the abnormal current fluctuations in the charging pile to determine the current abnormality parameters to reflect the abnormal equipment status of the charging pile; then, based on the characteristic that abnormal charging operation or current transient pulses generated by frequent charging will aggravate equipment loss, the degree of equipment loss of the charging pile is determined based on the current mutation situation and relevant parameters of the charging process; further, based on the correlation between equipment loss and equipment abnormality and the overall size of current abnormality parameters and equipment loss degree, the operation and maintenance characteristic values ​​that characterize the operation and maintenance requirements of the charging pile in the current operation and maintenance cycle of the charging station are comprehensively determined; thereby, the operation and maintenance characteristic values ​​are introduced into the XGBoost algorithm, so that the division of the operation and maintenance priorities of the charging piles in the charging station is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 A flowchart of a method for adaptive operation and maintenance of a charging station based on the XGBoost algorithm provided by one embodiment of the present invention;

[0047] Figure 2 This is a structural diagram of a charging station adaptive operation and maintenance system based on the XGBoost algorithm provided by one embodiment of the present invention;

[0048] Figure 3 The present invention provides a schematic diagram of a computer device structure according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail the specific implementation method, structure, features and effects of a charging station adaptive operation and maintenance method and system based on the XGBoost algorithm proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment, and the specific features, structures or characteristics in one or more embodiments may be combined in any suitable form. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0051] The following describes in detail a specific scheme of a charging station adaptive operation and maintenance method and system based on the XGBoost algorithm provided by the present invention with reference to the accompanying drawings.

[0052] This application embodiment provides a charging station adaptive operation and maintenance method based on the XGBoost algorithm, please refer to Figure 1 , which shows a flow chart of a charging station adaptive operation and maintenance method based on the XGBoost algorithm provided by one embodiment of the present invention, the method comprising:

[0053] Step S101: in the current operation and maintenance cycle of the charging station, the current data of each charging pile at all sampling moments in each charging time period is obtained.

[0054] In a specific implementation of an embodiment of the present invention, a Hall current sensor is configured at the DC output end of each charging pile of the charging station to collect current data of the charging pile at all sampling moments in each charging time period; wherein, each charging time period corresponds to the time period of a complete charging behavior, and the sampling frequency is set to collect once per second; the operation and maintenance cycle of the charging station is set to one day; it can be adjusted according to the specific implementation environment, and no further limitation or elaboration is made here.

[0055] Step S102: Determine the corresponding current anomaly parameters based on the abnormal fluctuation of the local current data in each charging time period; determine the corresponding transient impact intensity based on the sudden change distribution of the local current data in each charging time period; determine the corresponding equipment loss degree based on the transient impact intensity of each charging time period, the length of the charging time period, and the time interval of the charging time period.

[0056] During charging station operation, brief overloads can prevent heat accumulation and irreversible damage through rapid cooling, preventing major failures after recovery. However, sustained overloads can lead to component insulation aging and fracture, increasing the risk of failure. Long-term underloads reduce efficiency, resulting in insufficient motor activation, low energy conversion efficiency, shortened battery life, or instability, while also accelerating component aging and impacting long-term stability. Therefore, it is necessary to analyze the abnormal current fluctuations manifested by overloads. This analysis can analyze the abnormal current parameters for each charging period and indirectly indicate any abnormal deterioration in the charging state of the charging station during that period.

[0057] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the abnormal current parameters includes:

[0058] Each charging period is divided into at least two time windows; all time windows have the same length. In one specific implementation of the present invention, the time window length is set to 1 minute, which can be adjusted according to the specific implementation environment. First, the entire charging period is divided into time windows, and the local abnormal current state is analyzed in each time window.

[0059] Further, based on the deviation between the overall size of the current data in each time window and the preset rated current value, the current abnormality of each time window is determined; in a specific implementation method of an embodiment of the present invention, the process of obtaining the current abnormality includes: taking the mean of the current data at all sampling moments in each time window as the corresponding current effective value; performing positive correlation mapping on the difference between the current effective value and the preset rated current value to determine the current abnormality of each time window; it should be noted that the preset rated current value is the rated current size of each charging pile in the charging station, which is obtained through the factory data provided by the charging pile manufacturer and will not be further elaborated here.

[0060] In one specific implementation of an embodiment of the present invention, a method for positively correlating the difference between the effective current value and a preset rated current value is implemented by calculating the reciprocal of the preset rated current value and multiplying the difference between the effective current value and the preset rated current value by the reciprocal to obtain a positively correlated value, i.e., the current anomaly degree for each time window. Using the preset rated current value as the denominator effectively limits the impact of different rated current values ​​on the current anomaly degree calculation in different implementation environments, thereby improving the robustness of the analysis process.

[0061] Under normal circumstances, the current data should be maintained near the rated current value and remain stable. Therefore, for each time window, the greater the difference between the overall current data size and the current effective value and the preset rated current value, the more abnormal the current data in the time window is, and the corresponding current abnormality such as overload or underload of the current charging pile in the corresponding time window.

[0062] For each charging time period, if the individual time windows exhibit a significant degree of current anomaly as a whole, the charging time period should be divided more finely to more accurately capture the impact of transient current on the equipment load during the operation of the charging pile. Therefore, the average of the current anomaly levels across all time windows is taken as the overall anomaly level. For a charging time period, the greater the current anomaly level across all time windows, the more refined the division is required to more accurately capture the current anomaly. Therefore, the division window needs to be set smaller. Therefore, the product of the negative correlation mapping value of the overall anomaly level and the time length of the time window is further used as the correction time length. Each charging time period is divided into at least two correction windows, and the time length of the correction window is the correction time length.

[0063] It should be noted that, when the embodiment of the present invention divides the charging time period, the divided time windows or correction windows all meet the characteristics of temporal continuity of adjacent time windows, that is, there are no moments outside the windows between the time windows or correction windows; in addition, it should be noted that when dividing the time windows or correction windows, if the length of the charging time period is insufficient, resulting in the incompleteness of the last time window or correction window, the time window or correction window is merged with the previous window for analysis.

[0064] Furthermore, based on the principle of obtaining the current anomaly degree of each time window, the current anomaly degree of each correction window is calculated; that is, after replacing the time window in the calculation process of the current anomaly degree of the time window with the correction window, the current anomaly degree of each correction window is calculated; further, the cumulative value of the current anomaly degrees of all correction windows in each charging time period is used as the corresponding current anomaly parameter, and by combining the current anomaly degrees of all correction windows, the abnormal load accumulated by the current anomaly in a single charging process and the charging time period is quantified; the larger the current anomaly parameter, the greater the impact of the abnormal load on the device, the greater the risk of aging and damage of the internal components, and the worse the health of the corresponding device.

[0065] During operation, users frequently connect and disconnect charging piles, especially when the charging gun or plug connection is unstable, causing transient current fluctuations. In particular, poor plug contact or rapid plugging and unplugging can cause sudden current fluctuations, generating high-frequency pulses that exacerbate device wear. To further quantify device wear, it's necessary to analyze the characteristics of high-frequency pulses and the wear and tear of the charging pile during the charging process. High-frequency pulse characteristics typically correspond to sudden changes in current data. Therefore, the corresponding transient impact intensity is first determined based on the distribution of local current data changes during each charging time period.

[0066] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the transient impact strength includes:

[0067] Based on the similarity of local fluctuations between adjacent sampling moments in chronological order, all similar fluctuation moments in each charging time period are screened out. In a specific implementation of an embodiment of the present invention, the process of obtaining similar fluctuation moments includes: in each charging time period, using the standard deviation of the current data at all sampling moments within a preset neighborhood window of each sampling moment as the corresponding local fluctuation degree; using the difference between the local fluctuation degree at each sampling moment and the local fluctuation degree at the next sampling moment as the fluctuation change reference value at each sampling moment; and using sampling moments with a fluctuation change reference value less than a preset change threshold as similar fluctuation moments. In a specific implementation of an embodiment of the present invention, the preset change threshold is set to 0.5, and the preset neighborhood window is set to a window with a time length of 9 seconds centered on each sampling moment, which can be adjusted according to the specific implementation environment.

[0068] The standard deviation reflects the degree of discreteness of a set of data. Therefore, the degree of local fluctuation obtained through the standard deviation can reflect the fluctuation characteristics within the local range of each sampling moment to a certain extent. Therefore, for each sampling moment, the larger the corresponding fluctuation change reference value, the more similar the local fluctuation characteristics between it and the next sampling moment. Therefore, the corresponding sampling moment meets the characteristics similar to the neighborhood fluctuation characteristics, and is therefore analyzed as a similar fluctuation moment.

[0069] Traverse all similar fluctuation moments and determine all similar fluctuation time periods; similar fluctuation time periods are all continuous similar fluctuation moments and the first sampling moment before and the first sampling moment after the similar fluctuation time period are not similar fluctuation moments; according to the properties of similar fluctuation moments, since similar fluctuation time periods correspond to continuous similar fluctuation moments, each similar fluctuation time period corresponds to similar fluctuation characteristics, so further determine the local fluctuation characteristic value of the fluctuation characteristic of each similar fluctuation time period based on the average of the fluctuation change reference values ​​of all similar fluctuation moments in each similar fluctuation time period.

[0070] Furthermore, for each similar fluctuation time period, if there is a large deviation in the local fluctuation characteristic value between it and the adjacent similar fluctuation time period, then the current mutation between the corresponding similar fluctuation time period and the adjacent similar fluctuation time period is more obvious, and the higher the probability of corresponding high-frequency pulse characteristics. Therefore, the degree of fluctuation anomaly of each similar fluctuation time period is further determined based on the fluctuation deviation between the local fluctuation characteristic value of each similar fluctuation time period and the local fluctuation characteristic value of the adjacent similar fluctuation time period.

[0071] In a specific implementation of the embodiment of the present invention, the process of obtaining the degree of fluctuation anomaly includes:

[0072] The local fluctuation characteristic value of the previous similar fluctuation time period of each similar fluctuation time period is used as the first reference fluctuation characteristic value; the local fluctuation characteristic value of the next similar fluctuation time period of each similar fluctuation time period is used as the second reference fluctuation characteristic value. Based on the difference between the second reference fluctuation characteristic value and the first reference fluctuation characteristic value, the degree of fluctuation difference between the two sides of each similar fluctuation time period is determined; based on the difference between the local fluctuation characteristic value of each similar fluctuation time period and the second reference fluctuation characteristic value, the reference fluctuation difference degree of each similar fluctuation time period is determined.

[0073] For each similar fluctuation time period, if the local fluctuation characteristic values ​​of the similar fluctuation time periods on either side are similar, and there is a significant difference in the local fluctuation characteristic value between this similar fluctuation time period and the previous similar fluctuation time period, it indicates that a significant current signal characteristic mutation has occurred at the location of this similar fluctuation time period, reflecting the characteristics of transient pulse fluctuations in the signal, and the degree of fluctuation anomaly is greater. Therefore, the degree of fluctuation anomaly for each similar fluctuation time period is further determined based on the ratio between the reference fluctuation difference degree and the fluctuation difference degree on both sides. The greater the degree of fluctuation anomaly, the more consistent it is with the characteristics of transient pulse fluctuations in the signal, and the greater the transient impact intensity at the local location reflected by the corresponding similar fluctuation time period. Furthermore, the corresponding transient impact intensity is determined based on the cumulative value of the fluctuation anomaly degrees of all similar fluctuation time periods in each charging time period. The greater the transient impact intensity, the more frequent current transient pulses occur during charging station operation, the more severe the arc erosion of the charging pile equipment contactor contacts, and the greater the damage to the charging pile equipment.

[0074] It should be noted that in order to ensure that the calculation results are meaningful, when performing fractional operations in the embodiments of the present invention, when encountering a situation where the denominator is 0, it is necessary to add a parameter adjustment factor greater than 0 to the denominator to prevent the denominator from being 0. The value of the parameter adjustment factor is set by the implementer according to the actual situation, and is set to 1 in this application.

[0075] It should be noted that, considering that the degree of fluctuation anomaly cannot be calculated for similar fluctuation time periods ending at the boundary, the embodiment of the present invention calculates the degree of fluctuation anomaly of the second similar fluctuation time period as the degree of fluctuation anomaly of the first similar fluctuation time period; and calculates the degree of fluctuation anomaly of the second-to-last similar fluctuation time period as the degree of fluctuation anomaly of the last similar fluctuation time period to ensure the completeness of the embodiment. It can be adjusted according to the specific implementation environment and will not be further elaborated here.

[0076] Typically, when charging, current flowing through conductors (such as cables and contactors) generates Joule heating, causing the charging station equipment to heat up. During the charging interval, the equipment releases this heat into the environment through heat dissipation (convection and radiation). If the interval is too short, the heat cannot be fully dissipated, resulting in a cumulative temperature rise. In this case, if the charging station experiences a high transient impulse intensity during a single charge, this will result in more arc erosion on the equipment contactor contacts, exacerbating equipment wear. Transient impulse intensity can also indicate wear on the charging station equipment. Therefore, the corresponding equipment wear level is further determined based on the transient impulse intensity of each charging period, the duration of the charging period, and the time interval between charging periods.

[0077] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the degree of device wear includes:

[0078] Long-term charging also leads to a large amount of accumulated temperature rise. Therefore, the longer the charging period, the more severe the accumulated temperature rise will lead to the loss of charging pile equipment. The shorter the charging interval, the more severe the temperature rise accumulation due to the inability to fully dissipate heat, and the higher the degree of equipment loss. Therefore, the time interval between each charging period and the previous charging period is further used as the reference charging interval. The degree of equipment loss in each charging period is determined by multiplying the negative correlation mapping value of the reference charging interval, the length of each charging period, and the transient impact intensity. The greater the degree of equipment loss, the more severe the charging pile equipment loss in the corresponding charging period.

[0079] In one specific implementation of an embodiment of the present invention, a negative correlation mapping method for the reference charging interval is performed by calculating a normalized value after linear normalization of the reference charging interval, and then subtracting the normalized value from the real number 1 to obtain the negative correlation mapping value for the reference charging interval. Implementers may adopt other negative correlation mapping methods, such as the inverse method, depending on the specific implementation environment, and further description is omitted here.

[0080] Step S103: Determine the corresponding operation and maintenance characteristic value based on the correlation between the current abnormality parameters and the degree of equipment loss corresponding to each charging time period in the current charging station operation and maintenance cycle of each charging pile, as well as the overall size; and perform charging station operation and maintenance classification based on the operation and maintenance characteristic value combined with the XGBoost algorithm.

[0081] For each charging pile, the larger the overall current anomaly parameter during each charging time period of the charging station's operation and maintenance cycle, the worse the current health condition. The greater the degree of equipment loss during each charging time period, the more severe the charging pile equipment loss. Therefore, both the current anomaly parameter and the degree of equipment loss can characterize the operation and maintenance requirements of each charging pile. Furthermore, if the deterioration of the charging pile's health condition is highly correlated with equipment loss, this indicates that there is a long-term cumulative effect between health risks and equipment loss, and the charging pile is under the dual pressure of "short-term anomalies and long-term aging," resulting in a greater corresponding operation and maintenance requirement. If the correlation between the deterioration of the charging pile's health condition and equipment loss is poor, this indicates that the health problem may be an occasional failure unrelated to equipment aging, and the operation and maintenance requirement should be relatively small. Therefore, the corresponding operation and maintenance characteristic value is further determined based on the correlation and overall magnitude between the current anomaly parameter and the degree of equipment loss corresponding to each charging time period in the current charging station's operation and maintenance cycle.

[0082] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the operation and maintenance characteristic value includes:

[0083] Arrange the equipment loss levels for all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an equipment loss level sequence; arrange the current anomaly parameters for all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine a current anomaly parameter sequence; and normalize the Pearson correlation coefficient between the equipment loss level sequence and the current anomaly parameter sequence to determine a loss anomaly correlation. In a specific implementation of an embodiment of the present invention, the method for normalizing the Pearson correlation coefficient between the equipment loss level sequence and the current anomaly parameter sequence is as follows: linearly normalize the Pearson correlation coefficient between the equipment loss level sequence and the current anomaly parameter sequence to obtain a corresponding normalized result, i.e., a loss anomaly correlation.

[0084] Health risks and equipment loss have a long-term cumulative effect. That is, when there is a positive feedback effect, the corresponding equipment loss level and current anomaly parameters should show a positive correlation. Therefore, the Pearson correlation coefficient is normalized to both characterize the positive correlation and avoid the impact of negative values ​​on the calculation process. It should be noted that the calculation of the Pearson correlation coefficient is a technical means well known to those skilled in the art and will not be further defined or elaborated here.

[0085] During the current charging station operation and maintenance cycle, the accumulated values ​​of the equipment wear levels for all charging time periods are normalized to determine the cumulative wear level; and the accumulated values ​​of the overload capacity parameters for all charging time periods are normalized to determine the cumulative overload parameters. In a specific implementation of an embodiment of the present invention, both the method for normalizing the accumulated values ​​of the equipment wear levels for all charging time periods and the method for normalizing the accumulated values ​​of the overload capacity parameters for all charging time periods employ linear normalization to reduce the impact of different dimensions on the calculation process and improve the accuracy of the calculation results.

[0086] Since the greater the degree of equipment loss that characterizes equipment loss in all charging time periods in the current charging station operation and maintenance cycle, the greater the current abnormality parameter that characterizes health deterioration, and the greater the correlation between the degree of equipment loss and the current abnormality parameter, the higher the operation and maintenance demand of the corresponding charging pile in the current charging operation and maintenance cycle, normalization is performed based on the product of the cumulative loss degree, the cumulative overload parameter, and the loss abnormality correlation to determine the operation and maintenance characteristic value of each charging pile in the current charging station operation and maintenance cycle; wherein, the method for normalizing the product of the cumulative loss degree, the cumulative overload parameter, and the loss abnormality correlation adopts linear normalization. So that when the operation and maintenance characteristic value is larger, the operation and maintenance demand of the corresponding charging pile is higher, and the divided operation and maintenance priority should be higher. The operation and maintenance characteristic value that characterizes the operation and maintenance demand is further introduced as a feature into the XGBoost algorithm, so as to more accurately divide the operation and maintenance priority of the charging pile.

[0087] Preferably, in a specific implementation of an embodiment of the present invention, the process of grading charging station operation and maintenance based on operation and maintenance feature values ​​in conjunction with the XGBoost algorithm includes: inputting the operation and maintenance feature values ​​as features into an XGBoost model, and determining the operation and maintenance risk level of each charging pile in the current charging station operation and maintenance cycle based on the output of the XGBoost model. In a specific implementation of an embodiment of the present invention, the operation and maintenance risk levels include low operation and maintenance priority, medium operation and maintenance priority, and high operation and maintenance priority. The higher the operation and maintenance priority, the higher the operation and maintenance demand. Charging piles with low operation and maintenance priority exhibit stable equipment operation and low failure rate; charging piles with medium operation and maintenance priority exhibit average equipment operation and high failure rate; charging piles with high operation and maintenance priority exhibit severe equipment aging and require more maintenance and attention. After prioritizing each charging pile in the current charging pile operation and maintenance cycle, charging piles with high operation and maintenance priority are prioritized for maintenance and repair, and charging piles with medium operation and maintenance priority are screened for equipment risks. Charging piles with low operation and maintenance priority are not processed.

[0088] In the XGBoost classification model, the model iteratively constructs multiple decision trees with the goal of minimizing the loss function. It integrates the nonlinear relationships and interactions between features to ultimately output the operation and maintenance risk level of a charging station. In this embodiment of the present invention, the operation and maintenance feature value is used as a key feature, directly participating in the node splitting process of each decision tree. Its value influences the selection of the feature splitting threshold, which in turn determines the flow of charging station samples within the tree structure and the final classification result. Each time a decision tree splits a node, it traverses all features (including the operation and maintenance feature value) and outputs the gain value (Gain) after the node split. The gain value represents the decrease in the objective function (loss function + regularization term) after the split. If the splitting gain of the operation and maintenance feature value is the largest, the current node is preferentially split based on this feature. In particular, when the samples to be split have significant differences (high-level nodes), traversing all splitting features will result in a higher splitting gain. In this case, preferentially selecting the operation and maintenance feature value as the core splitting basis can guide the samples to be split at high-level nodes to be split based on the differences in the operation and maintenance risk of the charging station, which can help to prevent multi-dimensional failure and excessive splitting. For example, if the over-splitting condition "operation and maintenance eigenvalue greater than 0.6" can significantly separate high-risk from medium- and low-risk charging stations (resulting in a higher gain value), then this splitting condition is preferentially used at this node. This allows the operation and maintenance eigenvalue to participate in the splitting of high-level nodes and reap a higher gain, ensuring the accuracy of the splitting direction while effectively reducing the tree depth, promoting the convergence of the decision tree, and avoiding overfitting of the XGBoost model, which could cause the charging station classification results to collapse. It should be noted that the XGBoost algorithm is a technical means well known to those skilled in the art and will not be further defined or elaborated upon here.

[0089] In summary, an adaptive operation and maintenance method for charging stations based on the XGBoost algorithm is based on the characteristic that charging pile equipment abnormalities usually manifest as abnormal current fluctuations. First, the abnormal current fluctuations in the charging pile are analyzed to determine the current abnormality parameters to reflect the abnormal equipment status of the charging pile; then, based on the characteristic that abnormal charging operation or frequent charging generates current transient pulses that aggravate equipment loss, the degree of equipment loss of the charging pile is determined based on the current mutation situation and relevant parameters of the charging process; further, based on the correlation between equipment loss and equipment abnormality and the overall size of current abnormality parameters and equipment loss degree, the operation and maintenance characteristic values ​​that characterize the operation and maintenance requirements of the charging pile in the current operation and maintenance cycle of the charging station are comprehensively determined; thereby, the operation and maintenance characteristic values ​​are introduced into the XGBoost algorithm, which makes the division of the operation and maintenance priority of the charging piles in the charging station more accurate.

[0090] This application also provides a charging station adaptive operation and maintenance system based on the XGBoost algorithm, please refer to Figure 2, which shows a structural diagram of a charging station adaptive operation and maintenance system based on the XGBoost algorithm provided by an embodiment of the present invention. The system includes: a data acquisition and preprocessing module 201, a parameter determination module 202 and a charging station operation and maintenance classification module 203.

[0091] The data acquisition module 201 is used to obtain the current data of each charging pile at all sampling times in each charging time period during the current charging station operation and maintenance cycle;

[0092] Parameter determination module 202 is configured to determine corresponding current anomaly parameters based on abnormal fluctuations in local current data during each charging time period; determine corresponding transient impact intensity based on the distribution of sudden changes in local current data during each charging time period; and determine corresponding equipment loss levels based on the transient impact intensity, duration, and time interval of each charging time period.

[0093] The charging station operation and maintenance classification module 203 is used to determine the corresponding operation and maintenance characteristic values ​​based on the correlation between the current abnormality parameters and the degree of equipment loss corresponding to each charging time period of each charging pile in the current charging station operation and maintenance cycle, as well as the overall size; and perform charging station operation and maintenance classification based on the operation and maintenance characteristic values ​​combined with the XGBoost algorithm.

[0094] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the above embodiment provides an adaptive operation and maintenance system for charging stations based on the XGBoost algorithm and an adaptive operation and maintenance method for charging stations based on the XGBoost algorithm. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0095] The present application also provides a computer device. Figure 3 , which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, the computer device includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302, wherein when the processor 302 executes the computer program 303, the computer device can execute any one of the above-mentioned adaptive operation and maintenance methods of charging stations based on the XGBoost algorithm.

[0096] An embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute any one of the above-mentioned adaptive operation and maintenance methods for charging stations based on the XGBoost algorithm.

[0097] An embodiment of the present application also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer device, the computer device can execute any one of the above-mentioned adaptive operation and maintenance methods for charging stations based on the XGBoost algorithm.

[0098] In the embodiments provided in the present application, it should be understood that the provided computer devices, computer program products and computer-readable storage media are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0099] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0100] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A charging station adaptive operation and maintenance method based on XGBoost algorithm, characterized in that: The method comprises: During the current operation and maintenance cycle of the charging station, the current data of each charging pile at all sampling times in each charging time period is obtained; Determine the corresponding current anomaly parameters based on the abnormal fluctuations in the local current data in each charging time period; determine the corresponding transient impact intensity based on the distribution of the sudden changes in the local current data in each charging time period; and determine the corresponding equipment loss level based on the transient impact intensity, duration, and time interval of each charging time period. Determine the corresponding operation and maintenance characteristic value based on the correlation between the abnormal current parameters and the degree of equipment loss corresponding to each charging time period in the current charging station operation and maintenance cycle of each charging pile, as well as the overall size; and perform charging station operation and maintenance classification based on the operation and maintenance characteristic value combined with the XGBoost algorithm; The process of obtaining the operation and maintenance characteristic value includes: Arrange the equipment loss degrees of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an equipment loss degree sequence; arrange the current abnormality parameters of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an abnormal current parameter sequence; normalize the Pearson correlation coefficient between the equipment loss degree sequence and the current abnormality parameter sequence to determine the loss abnormality correlation; During the current operation and maintenance cycle of the charging station, the accumulated values ​​of the equipment loss degree in all charging time periods are normalized to determine the cumulative loss degree; the accumulated values ​​of the overload capacity parameters in all charging time periods are normalized to determine the cumulative overload parameter; Normalization is performed based on the product of the cumulative loss degree, the cumulative overload parameter, and the loss anomaly correlation to determine the operation and maintenance characteristic value of each charging pile in the current operation and maintenance cycle of the charging station.

2. The adaptive operation and maintenance method for charging stations based on the XGBoost algorithm according to claim 1, characterized in that: The process of obtaining the abnormal current parameters includes: Divide each charging time period into at least two time windows; wherein all time windows have the same length; Determine the current abnormality of each time window based on the deviation between the overall magnitude of the current data in each time window and the preset rated current value; The average value of the current abnormality of all time windows is used as the overall abnormality; the product of the negative correlation mapping value of the overall abnormality and the time length of the time window is used as the correction time length; Each charging time period is divided into at least two correction windows, where the time length of the correction window is the correction time length; based on the principle of obtaining the current anomaly degree of each time window, the current anomaly degree of each correction window is calculated; and the accumulated value of the current anomaly degrees of all correction windows in each charging time period is used as the corresponding current anomaly parameter.

3. The adaptive operation and maintenance method of a charging station based on the XGBoost algorithm according to claim 2, characterized in that: The process of obtaining the current anomaly degree includes: The mean value of the current data at all sampling moments in each time window is taken as the corresponding current effective value; the difference between the current effective value and the preset rated current value is positively correlated to determine the current abnormality of each time window.

4. The adaptive operation and maintenance method for charging stations based on the XGBoost algorithm according to claim 1, characterized in that: The process of obtaining the transient impact strength includes: Based on the local fluctuation similarity between adjacent sampling moments in time sequence, all similar fluctuation moments in each charging time period are screened out; all similar fluctuation moments are traversed to determine all similar fluctuation time periods; the similar fluctuation time periods are all continuous similar fluctuation moments and the first sampling moment before and the first sampling moment after the similar fluctuation time period are not similar fluctuation moments; Determine the corresponding local fluctuation characteristic value based on the mean of the fluctuation change reference values ​​of all similar fluctuation moments in each similar fluctuation time period; determine the degree of fluctuation anomaly of each similar fluctuation time period based on the fluctuation deviation between the local fluctuation characteristic value of each similar fluctuation time period and the local fluctuation characteristic value of the adjacent similar fluctuation time period; The corresponding transient impact intensity is determined according to the accumulated value of the fluctuation abnormality degree of all similar fluctuation time periods in each charging time period.

5. The adaptive operation and maintenance method of a charging station based on the XGBoost algorithm according to claim 4 is characterized in that: The process of obtaining the similar fluctuation moments includes: In each charging time period, the standard deviation of the current data at all sampling moments within the preset neighborhood window of each sampling moment is taken as the corresponding local fluctuation degree; the difference between the local fluctuation degree at each sampling moment and the local fluctuation degree at the next sampling moment is taken as the fluctuation change reference value at each sampling moment; the sampling moment at which the fluctuation change reference value is less than the preset change threshold is taken as the similar fluctuation moment.

6. The adaptive operation and maintenance method for charging stations based on the XGBoost algorithm according to claim 4, characterized in that: The process of obtaining the degree of fluctuation anomaly includes: The local fluctuation characteristic value of the previous similar fluctuation time period of each similar fluctuation time period is used as the first reference fluctuation characteristic value; the local fluctuation characteristic value of the next similar fluctuation time period of each similar fluctuation time period is used as the second reference fluctuation characteristic value; determining a degree of difference in fluctuations on both sides of each similar fluctuation time period based on a difference between the second reference fluctuation characteristic value and the first reference fluctuation characteristic value; determining a reference fluctuation difference degree for each similar fluctuation time period according to a difference between the local fluctuation characteristic value for each similar fluctuation time period and the second reference fluctuation characteristic value; The degree of fluctuation anomaly in each similar fluctuation time period is determined according to the ratio between the reference fluctuation difference degree and the fluctuation difference degree on both sides.

7. The adaptive operation and maintenance method of a charging station based on the XGBoost algorithm according to claim 1, characterized in that: The process of obtaining the degree of equipment loss includes: The time interval between each charging time period and the previous charging time period is used as a reference charging interval; The device loss degree of each charging time period is determined according to the product of the negative correlation mapping value of the reference charging interval, the time length of each charging time period, and the transient impact intensity.

8. The adaptive operation and maintenance method for charging stations based on the XGBoost algorithm according to claim 1, characterized in that: The process of performing charging station operation and maintenance classification based on the operation and maintenance characteristic values ​​in combination with the XGBoost algorithm includes: The operation and maintenance feature value is input as a feature into the XGBoost model, and based on the output of the XGBoost model, the operation and maintenance risk level of each charging pile in the current operation and maintenance cycle of the charging station is determined.

9. An adaptive operation and maintenance system for charging stations based on the XGBoost algorithm, characterized in that: The system comprises: The data acquisition module is used to obtain the current data of each charging pile at all sampling times in each charging time period during the current charging station operation and maintenance cycle; A parameter determination module is used to determine the corresponding current anomaly parameter based on the abnormal fluctuation of the local current data in each charging time period; determine the corresponding transient impact intensity based on the sudden change distribution of the local current data in each charging time period; and determine the corresponding equipment loss degree based on the transient impact intensity, charging time period duration, and charging time interval of each charging time period; The charging station operation and maintenance grading module is used to determine the corresponding operation and maintenance characteristic values ​​based on the correlation between the current abnormality parameters and the degree of equipment loss corresponding to each charging time period in the current charging station operation and maintenance cycle, as well as the overall size; and perform charging station operation and maintenance grading based on the operation and maintenance characteristic values ​​combined with the XGBoost algorithm; The process of obtaining the operation and maintenance characteristic value includes: Arrange the equipment loss degrees of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an equipment loss degree sequence; arrange the current abnormality parameters of all charging time periods in the current charging station operation and maintenance cycle in chronological order to determine an abnormal current parameter sequence; normalize the Pearson correlation coefficient between the equipment loss degree sequence and the current abnormality parameter sequence to determine the loss abnormality correlation; During the current operation and maintenance cycle of the charging station, the accumulated values ​​of the equipment loss degree in all charging time periods are normalized to determine the cumulative loss degree; the accumulated values ​​of the overload capacity parameters in all charging time periods are normalized to determine the cumulative overload parameter; Normalization is performed based on the product of the cumulative loss degree, the cumulative overload parameter, and the loss anomaly correlation to determine the operation and maintenance characteristic value of each charging pile in the current operation and maintenance cycle of the charging station.

Images