Method and device for determining fault data label of charging equipment

Abstract

The embodiment of the invention discloses a method and device for determining a fault data label of charging equipment, and the method comprises the steps: obtaining a data sample which comprises at least one charging abnormal code, and the fault weight of each charging abnormal code under different fault conditions, determining fault scores of the corresponding fault situations according to different fault weights corresponding to the fault situations; and determining the fault situation with the highest fault score as the fault label of the data sample. Wherein the fault weight is used for representing the importance of the charging abnormal code on fault condition labeling, and the fault label comprises a gun fault and / or a module fault. Therefore, the score of each fault condition and the fault label of the data sample are determined according to the importance of labeling different fault conditions by using different charging abnormal codes, so that the accuracy of labeling the fault label for the data sample is higher, the labeling is more convenient, and the sample labeling with the labeling accuracy and the labeling efficiency at the same time is realized.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more specifically to a method and apparatus for determining fault data tags for charging equipment. Background Technology

[0002] In the task of fault diagnosis of charging equipment, a high-quality sample set is the key to ensuring model performance, and accurate sample labeling is the key to building a high-quality sample set.

[0003] Existing sample annotation methods include on-site troubleshooting annotation, expert annotation, and rule-based annotation. On-site troubleshooting annotation involves dispatching maintenance personnel to the site to investigate the fault type and ultimately annotate the sample. Expert annotation involves experts annotating based on experience using data from dimensions such as orders and work orders. Rule-based annotation involves extracting deterministic rules from data from dimensions such as historical orders and historical work orders and using these rules for annotation. However, on-site troubleshooting annotation is time-consuming, requiring maintenance personnel to commute to the site for troubleshooting; expert annotation is also time-consuming and cannot be done in batches; rule-based annotation is highly subjective, with the setting of annotation rules greatly affected by human factors, making it difficult to guarantee annotation quality. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method and apparatus for determining fault data tags for charging equipment, so as to achieve sample labeling that simultaneously takes into account both labeling accuracy and labeling efficiency.

[0005] In a first aspect, embodiments of the present invention aim to provide a method for determining fault data tags of charging equipment, the method comprising:

[0006] Obtain a data sample, the data sample including at least one charging error code;

[0007] Obtain the fault weight of each charging error code under different fault conditions. The fault weight is used to characterize the importance of the charging error code to the fault condition labeling.

[0008] The fault score for each fault condition is determined based on the different fault weights corresponding to each fault condition.

[0009] The fault condition with the highest fault score is identified as the fault label of the data sample, and the fault label includes gun fault and / or module fault.

[0010] Furthermore, the method also includes:

[0011] Obtain the set of abnormal codes for the charging device, wherein the set of abnormal codes includes multiple charging abnormal codes;

[0012] An anomaly weight matrix is ​​determined based on the relative importance of each charging anomaly code to the fault label. The anomaly weight matrix includes the label weight of each charging anomaly code, and the label weight is used to characterize the importance of the charging anomaly code to the fault label.

[0013] The fault weight matrix of each charging anomaly code is determined based on the relative importance of each charging anomaly code in different fault conditions and the labeling weight of each charging anomaly code. The fault weight matrix includes the fault weight of the corresponding charging anomaly code under different fault conditions.

[0014] Furthermore, the step of determining an anomaly weight matrix based on the relative importance of each charging anomaly code to the fault labeling, wherein the anomaly weight matrix includes the labeling weights of each charging anomaly code, including:

[0015] Construct a first judgment matrix, which includes quantified values ​​of the relative importance of different charging anomaly codes to fault labeling, the quantified values ​​being determined based on a preset scaling method;

[0016] The first judgment matrix is ​​preprocessed to determine the corresponding initial annotation matrix;

[0017] Based on the initial annotation matrix, the first judgment matrix is ​​subjected to consistency verification to determine the verification result of the first judgment matrix;

[0018] In response to the verification result indicating that the verification is passed, the initial annotation matrix is ​​determined as an anomaly weight matrix, the anomaly weight matrix including the annotation weight of each of the charging anomaly codes;

[0019] In response to the verification result indicating that the verification failed, the first judgment matrix is ​​adjusted until the verification result corresponding to the adjusted first judgment matrix indicates that the verification passed.

[0020] Further, the preprocessing of the first judgment matrix to determine the corresponding initial annotation matrix includes:

[0021] Standardize each column vector in the first judgment matrix to determine the first standard matrix;

[0022] Summing the row vectors in the first standard matrix determines the first weight matrix;

[0023] The first weight matrix is ​​standardized to determine the initial annotation matrix.

[0024] Furthermore, determining the fault weight matrix for each charging anomaly code based on the relative importance of each charging anomaly code for different fault conditions and the labeling weight of each charging anomaly code includes:

[0025] Construct a second judgment matrix corresponding to the target anomaly code, wherein the target anomaly code is any charging anomaly code in the anomaly code set, and the second judgment matrix includes a quantified value of the relative importance of the target anomaly code to different fault conditions, wherein the quantified value is determined based on a preset scaling method;

[0026] The second judgment matrix is ​​preprocessed to determine the initial fault matrix;

[0027] Based on the initial fault matrix, the second judgment matrix is ​​subjected to consistency verification to determine the verification result of the second judgment matrix;

[0028] In response to the verification result indicating that the verification has passed, the initial fault matrix is ​​determined as the fault labeling matrix of the target anomaly code;

[0029] The fault labeling matrix of the target anomaly code is multiplied by the labeling weights to determine the fault weight matrix, which includes the fault weights of the target anomaly code under different fault conditions. Further, the preprocessing of the second judgment matrix to determine the initial fault matrix includes:

[0030] Standardize each column vector in the second judgment matrix to determine the second standard matrix;

[0031] Summing the row vectors in the second standard matrix determines the second weight matrix;

[0032] The second weight matrix is ​​standardized to determine the initial fault matrix.

[0033] Furthermore, determining the fault weight matrix for each charging anomaly code based on the relative importance of each charging anomaly code for different fault conditions and the labeling weight of each charging anomaly code further includes:

[0034] In response to the verification result indicating a failed verification, the second judgment matrix is ​​adjusted until the verification result corresponding to the adjusted second judgment matrix indicates a passed verification. Further, determining the fault score for each fault condition based on the respective fault weights includes:

[0035] The fault weights of each charging abnormal code under the same fault condition are summed to determine the fault score for the corresponding fault condition.

[0036] Furthermore, the method also includes:

[0037] The data samples are labeled based on the fault labels, and a fault diagnosis model is trained based on the labeled data samples. The fault diagnosis model is used to determine the cause of the fault in the charging device to be identified.

[0038] Secondly, embodiments of the present invention aim to provide a device for determining fault data tags of charging equipment, the device comprising:

[0039] A sample acquisition unit is used to acquire a data sample, the data sample including at least one charging error code;

[0040] The weight acquisition unit is used to acquire the fault weight of each charging abnormal code under different fault conditions. The fault weight is used to characterize the importance of the charging abnormal code to the fault condition labeling.

[0041] The scoring unit is used to determine the fault score of the corresponding fault situation based on the different fault weights corresponding to each of the aforementioned fault situations.

[0042] A label determination unit is used to determine the fault condition with the highest fault score as the fault label of the data sample, and the fault label includes gun fault and / or module fault.

[0043] Thirdly, embodiments of the present invention aim to provide a computer program product, the computer program product including a computer program / instruction, which, when executed by a processor, implements the method described in any of the preceding claims.

[0044] Fourthly, embodiments of the present invention aim to provide an electronic device, including a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method as described in any of the preceding claims.

[0045] Fifthly, embodiments of the present invention aim to provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the preceding claims.

[0046] The technical solution of this invention, after acquiring data samples, uses the fault weights of each charging anomaly code under different fault conditions to determine the fault score of each fault condition, and determines the fault condition with the highest fault score as the fault label of the data sample. This fully considers the importance of different charging anomaly codes to the labeling of different fault conditions, improving the accuracy of fault labeling of data samples. At the same time, since it is only necessary to determine the fault score of each fault condition based on the fault weight of each charging anomaly code and the fault label based on the fault score of each fault condition, the labeling of data samples is more convenient. Thus, while improving the accuracy of data sample labeling, it can also take into account labeling efficiency, achieving sample labeling that simultaneously considers labeling accuracy and labeling efficiency. Attached Figure Description

[0047] The above and other objects, features and advantages of the present invention will become clearer from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0048] Figure 1 This is a flowchart of the charging equipment fault data tag determination method according to an embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram of the hierarchical analysis model according to an embodiment of the present invention;

[0050] Figure 3 This is a flowchart illustrating the determination of fault weights according to an embodiment of the present invention;

[0051] Figure 4 This is a flowchart illustrating the determination of the anomaly weight matrix according to an embodiment of the present invention;

[0052] Figure 5 This is a flowchart of the preprocessing of the first judgment matrix in an embodiment of the present invention;

[0053] Figure 6 This is a flowchart illustrating the determination of the fault weight matrix according to an embodiment of the present invention;

[0054] Figure 7 This is a flowchart of the preprocessing of the second judgment matrix in an embodiment of the present invention;

[0055] Figure 8 This is a charging equipment fault data tag determination device according to an embodiment of the present invention;

[0056] Figure 9 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0057] The present application is described below based on embodiments, but it is not limited to these embodiments. In the detailed description of the present application below, certain specific details are described in detail. Those skilled in the art can fully understand the present application without these details. To avoid obscuring the substance of the present application, well-known methods, processes, flows, elements, and circuits are not described in detail.

[0058] Furthermore, those skilled in the art should understand that the accompanying drawings provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0059] Unless the context explicitly requires it, words such as "including" or "contains" throughout the application should be interpreted as including rather than exclusive or exhaustive; that is, meaning "including but not limited to".

[0060] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0061] The solutions described in this specification and embodiments, if involving information acquisition, will collect data under legal and compliant conditions, ensuring the legality of the data source, and will take appropriate technical and management measures to ensure data security. If involving personal information processing, processing will be carried out under legal grounds (e.g., obtaining the consent of the personal information subject, or being necessary for contract performance), and will only be conducted within the prescribed or agreed scope. A user's refusal to process personal information beyond what is necessary for basic functions will not affect the user's use of basic functions.

[0062] In this embodiment, the labeling of charging equipment data samples in the scenario of charging equipment fault diagnosis is used as an example for illustration. However, it should be noted that the method for determining data labels in this embodiment can also be applied to data labeling in other data labeling scenarios.

[0063] Figure 1 This is a flowchart of a method for determining fault data tags for charging equipment according to an embodiment of the present invention. Figure 1 As shown, the label determination method in this embodiment includes the following steps.

[0064] In step S110, a data sample is obtained, which includes at least one charging error code.

[0065] In this embodiment, considering that some charging orders may report abnormal codes due to improper operation or accidental factors such as weather conditions, but not due to a fault in the charging equipment itself, in order to ensure accurate fault diagnosis of the charging equipment based on different data samples, a data sample in this embodiment includes all charging reporting data of multiple charging orders (such as 10 orders) generated on the same charging equipment.

[0066] Typically, one charging report corresponds to one charging order. Each charging report includes charging anomaly information. Normal charging orders are identified by no charging anomaly code, while abnormal charging orders are identified by a charging anomaly code. During fault analysis, all charging reports from all charging orders in a data sample must include at least one charging anomaly code. For different types of charging anomaly codes within the same or different charging orders, and for the same type of charging anomaly codes within different charging orders, a cumulative count is used. For example, suppose data sample 1 includes charging orders 1, 2, 3, and 4. Charging order 1 includes charging anomaly code 1, charging order 2 has no charging anomaly code, charging order 3 includes charging anomaly code 1, and charging order 4 includes charging anomaly code 2. Then data sample 1 includes three charging anomaly codes: two charging anomaly codes 1 and one charging anomaly code 2.

[0067] Optionally, the total number of charging orders in a data sample in this embodiment can be set according to the actual application scenario. Meanwhile, multiple charging orders in a data sample can be a certain number of charging orders continuously generated on the same charging device, such as 10 consecutive orders; or they can be a certain number of charging orders obtained by sampling multiple consecutive charging orders generated on the same charging device according to a certain sampling rule, such as sampling 20 consecutive charging orders starting from the first order, sampling once for every other order, and finally obtaining 10 orders as a data sample.

[0068] Furthermore, in this embodiment, multiple charging orders on the charging equipment are preferentially used as data samples. This allows for better capture of the fault characteristics of the charging equipment through densely distributed charging orders, which is beneficial to further improve the accuracy of fault labels on the data samples.

[0069] Optionally, considering that in addition to the charging error code, the charging report data of each charging order usually includes other charging-related data, such as charging time and charging capacity, in order to improve the annotation efficiency of subsequent annotation of data samples based on error codes, in this embodiment, after obtaining the charging report data of multiple charging orders corresponding to the data sample, the charging report data of each charging order will be preprocessed to filter out other data except for the charging error code, and only the error code in the charging report data of each charging order will be retained, thereby obtaining the corresponding data sample.

[0070] In step S120, the fault weight of each charging error code under different fault conditions is obtained. The fault weight is used to characterize the importance of the charging error code to the fault condition labeling.

[0071] In this embodiment, the purpose of fault condition labeling is to label data samples with fault condition tags. Since charging equipment faults typically include multiple fault conditions (such as charging module malfunction, charging gun malfunction, etc.), each fault condition corresponds to multiple fault indications (such as charging module malfunction including three-phase imbalance, undervoltage output, overcurrent output, etc.), and each fault indication corresponds to a charging anomaly code. This results in different charging anomaly codes having different probabilities of representing the same fault condition, and the same charging anomaly code having different probabilities of representing different fault conditions. In other words, different charging anomaly codes have different importance for labeling the same fault condition, and the same charging anomaly code has different importance for labeling different fault conditions. Therefore, it is necessary to obtain the importance of each charging anomaly code for labeling different fault conditions, so as to label data samples with fault condition tags based on the importance of each charging anomaly code for labeling different fault conditions, thereby improving the accuracy of data labeling.

[0072] Furthermore, in order to quantify the importance of charging error codes to fault condition labeling, this embodiment uses fault weights to characterize the importance of charging error codes to fault condition labeling. The fault weights of the same error code can be different under different fault conditions, and the fault weights of different error codes under the same fault condition can also be different.

[0073] Optionally, in this embodiment, the fault weights of each charging anomaly code under different fault conditions can be predetermined or determined in real time when determining the fault label of the data sample. Specifically, the timing of determining the fault weights can be selected according to the actual application scenario. For example, in scenarios with sufficient computing resources and high requirements for annotation efficiency, predetermined determination can be chosen, while in scenarios with limited computing resources but low requirements for annotation efficiency, real-time determination can be chosen. This allows for convenient selection of a suitable determination method based on the actual application scenario, making the annotation of data samples more convenient.

[0074] Furthermore, in this embodiment, a predetermined method is used to determine the fault weight of each charging anomaly code under different fault conditions, and each fault weight is saved. When it is necessary to determine the fault label of a data sample, the saved fault weights can be used directly, which helps to improve the overall processing efficiency of data sample labeling.

[0075] Optionally, in this embodiment, the fault weight of each charging anomaly code under different fault conditions can be determined based on a multi-level analysis method. Specifically, the fault weight of each charging anomaly code can be determined by qualitative and quantitative analysis of the importance of different charging anomaly codes to fault labeling. Alternatively, other methods can be used, such as experts directly defining the fault weight of each charging anomaly code under different fault conditions based on their experience.

[0076] To further improve the accuracy and efficiency of data sample annotation, this embodiment employs a multi-level analysis method to determine the fault weights of different anomaly codes under different fault conditions. Simultaneously, this embodiment is based on... Figure 2 The hierarchical analysis model shown determines the fault weights. For example... Figure 2 As shown, the hierarchical analysis model includes: a target layer 21, a criterion layer 22, and a solution layer 23. The target layer 21, the highest layer of the hierarchical analysis model, is used to determine the fault weights of each charging anomaly code under different fault conditions, thereby achieving correct labeling of data samples. The criterion layer 22, located in the middle layer, is used to measure and evaluate the contribution of each charging anomaly code to the labeling of fault conditions, i.e., the contribution degree of each charging anomaly code to the labeling of fault conditions. The criterion layer 22 includes multiple charging anomaly codes, each representing a specific evaluation indicator and possessing measurability and comparability. The solution layer 23, the lowest layer of the hierarchical analysis model, includes at least one fault condition of the charging equipment, used to provide feasible options for fault condition labeling, and through comparison and evaluation with the criterion layer, helps determine the fault weights of each anomaly code under different fault conditions.

[0077] The following section explains the method for determining the fault weights of each anomaly code under different fault conditions in the hierarchical analysis model.

[0078] Figure 3 This is a flowchart illustrating the determination of fault weights according to an embodiment of the present invention. Figure 3 As shown, in this embodiment, the fault weight of each charging fault code under different fault conditions is determined based on the following method.

[0079] In step S310, an error code set for the charging device is obtained, which includes multiple charging error codes.

[0080] In this embodiment, the set of exception codes can be obtained through the exception code list provided by the supplier when the charging equipment is delivered, exception codes collected by maintenance personnel based on on-site maintenance, etc.

[0081] Optionally, since the set of abnormal codes for charging devices typically includes various abnormal codes, but not all abnormal situations corresponding to all abnormal codes represent charging device malfunctions, this embodiment only acquires the various types of charging abnormal codes reported when the charging device malfunctions when obtaining the set of abnormal codes for the charging device, and adds all types of abnormal codes to the set of abnormal codes, so as to determine the fault weight of each charging abnormal code under different fault situations based on the set of abnormal codes.

[0082] In step S320, an anomaly weight matrix is ​​determined based on the relative importance of each charging anomaly code to the fault label. The anomaly weight matrix includes the label weight of each charging anomaly code, and the label weight is used to characterize the importance of the charging anomaly code to the fault label.

[0083] In this embodiment, because different charging anomaly codes have varying degrees of impact on the use of the charging equipment—for example, some charging anomaly codes correspond to fault situations that require urgent handling, while others correspond to fault situations that are merely meant to draw attention—the probability of different charging anomaly codes representing charging equipment faults differs, meaning that different charging anomaly codes have different importance in the labeling of charging equipment faults. Therefore, to improve the accuracy of fault situation labeling, this embodiment, when determining the fault weight of each charging anomaly code under different fault situations, first determines the importance of each charging anomaly code in the labeling of charging equipment faults.

[0084] Meanwhile, in order to quantify the importance of charging error codes to the fault labeling of charging equipment, this embodiment uses label weight to characterize the importance of charging error codes to the fault labeling, and different charging error codes have different fault weights when the charging equipment fails.

[0085] In this embodiment, by means of... Figure 4 The method shown determines the anomaly weight matrix, specifically including the following steps.

[0086] In step S410, a first judgment matrix is ​​constructed. The first judgment matrix includes quantified values ​​of the relative importance of different charging abnormal codes to fault labeling. The quantified values ​​are determined based on a preset scaling method.

[0087] In this embodiment, the relative importance of different charging abnormal codes to fault labeling is determined by comparing the importance of different charging abnormal codes to fault labeling among all charging abnormal codes of the charging equipment. A first judgment matrix is ​​constructed based on the quantified value of the relative importance of different charging abnormal codes to fault labeling, and the labeling weight of each charging abnormal code is determined based on the first judgment matrix.

[0088] Optionally, in this embodiment, the relative importance of different charging error codes to fault labeling is determined based on expert experience. The relative importance can be described as charging error code 1 and charging error code 2 being equally important, charging error code 2 being slightly more important than charging error code 1, charging error code 2 being significantly more important than charging error code 1, and so on.

[0089] Furthermore, in this embodiment, a preset scaling method is used to quantify the relative importance of different charging error codes to fault labeling, and to determine the corresponding quantization value. For example, when charging error code 1 and charging error code 2 are equally important, the corresponding quantization value is determined to be 1; when charging error code 2 is significantly more important than charging error code 1, the corresponding quantization value is determined to be 5.

[0090] Optionally, the preset scaling method in this embodiment can be Santy's 1-9 scaling method, or it can be a fuzzy comprehensive evaluation method, entropy weight method, Delphi method, grey relational analysis method, etc., to quantify the relative importance of different anomaly codes to fault labeling. The specific method can be selected according to the actual application scenario to facilitate the rapid determination of the quantified value of the relative importance of different anomaly codes to fault labeling and improve the convenience of constructing the first judgment matrix.

[0091] Furthermore, the first decision matrix is ​​constructed using Santy's 1-9 scaling method in this embodiment. The quantization strategies corresponding to the scaling method are shown in Table (1):

[0092] Table (1)

[0093] Scale (i.e., quantified value) Meaning (i.e., description of relative importance) a1 This indicates that the two charging error codes are of equal importance. a3 This indicates that, compared to the two charging error codes, the former is slightly more important than the latter. a5 This indicates that, compared to the two charging error codes, the former is significantly more important than the latter. a7 This indicates that, compared to the two charging error codes, the former is far more important than the latter. a9 This indicates that, compared to the two charging error codes, the former is significantly more important than the latter. a2, a4, a6, a8 This represents the intermediate value of the above adjacent judgments. The reciprocals of a1 to a9 This indicates the importance of comparing the transformation order of the two corresponding charging error codes.

[0094] In Table (1), a1 < a2 < a3 < a4 < a5 < a6 < a7 < a8 < a9.

[0095] Meanwhile, taking a charging device with four charging error codes as an example, a first judgment matrix M is constructed as shown below. The first judgment matrix M includes 16 data items, and the data items M in the first judgment matrix M... ij This represents the quantification of the relative importance of charging error codes i and j to the fault labeling. The values ​​of i and j can both be 1, 2, 3, or 4. For example, data item M... 21 This represents the quantified value of the relative importance of charging error code 2 and charging error code 1 to the fault labeling, in M 21 When the value is a3, it indicates that charging error code 2 is slightly more important than charging error code 1, that is, charging error code 2 is more important for fault labeling than charging error code 1.

[0096]

[0097] It should be understood that the quantification values ​​a1 to a9 of the relative importance between charging error codes given in this embodiment are only examples, and the specific values ​​can be set according to the actual use scenario.

[0098] In step S420, the first judgment matrix is ​​preprocessed to determine the corresponding initial annotation matrix.

[0099] In this embodiment, by preprocessing the first judgment matrix to determine the corresponding initial annotation matrix, the first judgment matrix can be converted into an initial standard matrix, and the quantization values ​​in the first judgment matrix can be converted to a unified scale, reducing the accumulation of errors in the calculation process and improving the accuracy and reliability of determining the annotation weights of subsequent charging anomaly codes. At the same time, since the preprocessed quantization values ​​have the same comparison basis, the annotation weights between different charging anomaly codes can be directly compared and sorted, thereby more accurately reflecting the importance of each charging anomaly code to the fault situation annotation.

[0100] Figure 5 This is a flowchart illustrating the preprocessing of the first judgment matrix in an embodiment of the present invention. For example... Figure 5 As shown, in this embodiment, the first judgment matrix is ​​preprocessed using the following method to determine the corresponding initial annotation matrix.

[0101] In step S510, the column vectors in the first judgment matrix are standardized to determine the first standard matrix.

[0102] In this embodiment, each column of the first judgment matrix is ​​taken as a standardization processing object, and each column vector in the first judgment matrix is ​​standardized. After the standardization of each column vector is completed, the standardized column vectors are merged and arranged according to their relative positions in the first judgment matrix to generate the first standard matrix.

[0103] Optionally, in this embodiment, a summation standardization method is used to standardize each column vector in the first judgment matrix. Specifically, each data item in the column vector is transformed into the ratio of that data item to the sum of all data items in the column vector, ensuring that the sum of all standardized data items is 1, and that the value of each data item is between 0 and 1.

[0104] It should be understood that other standardization methods can also be used to standardize each column vector in this embodiment, and in order to ensure that the data items in each column vector are standardized under a uniform scale, each column vector adopts the same standardization method.

[0105] In step S520, the row vectors in the first standard matrix are summed to determine the first weight matrix.

[0106] In this embodiment, after determining the first standard matrix corresponding to the first judgment matrix, each row in the first standard matrix is ​​used as a summation processing object, and the data items in each row vector are summed respectively. After the summation of each row vector is completed, the summation results of each row vector are merged and arranged according to the relative position of each row vector in the first standard matrix to generate the first weight matrix.

[0107] In step S530, the first weight matrix is ​​standardized to determine the initial labeling matrix.

[0108] In this embodiment, the standardization method for the first weight matrix can be the same as the standardization method for the first judgment matrix described above, so the standardization process will not be described again here. However, it should be understood that when standardizing the first weight matrix, since the first weight matrix only includes one column of data, only one column vector needs to be standardized, and there is no need to merge and arrange multiple standardized column vectors.

[0109] In step S430, the consistency of the first judgment matrix is ​​checked based on the initial annotation matrix to determine the verification result of the first judgment matrix.

[0110] In this embodiment, to ensure the accuracy of the labeling weights of each charging anomaly code and avoid situations such as "charging anomaly code 1 is more important than charging anomaly code 2, charging anomaly code 2 is more important than charging anomaly code 3, but charging anomaly code 3 is more important than charging anomaly code 1", a consistency check is performed on the first judgment matrix based on the initial labeling matrix to determine the check result of the first judgment matrix. This can evaluate the rationality of the logical relationship of the relative importance of each charging anomaly code in the first judgment matrix, ensuring the accuracy of the labeling weights of each charging anomaly code, so that the fault weight of each charging anomaly code can be determined subsequently based on the labeling weights of each charging anomaly code.

[0111] Optionally, in this embodiment, the consistency verification result of the initial annotation matrix is ​​determined based on the consistency index (CI) and the consistency ratio (CR).

[0112] CI (Conformity Index) measures whether the judgment matrix satisfies the consistency requirement, that is, whether the elements in the judgment matrix satisfy transitivity and consistency. The value of CI typically ranges from 0 to 1. The closer the CI value is to 1, the higher the consistency of the evaluated object; while the closer the CI value is to 0, the lower the consistency. The formula for calculating CI is: CI = (λ...) max -n) / (n-1), where λ max To determine the largest eigenvalue of the matrix, n is the order of the matrix.

[0113] CR (Conformity Ratio) is used to further evaluate the consistency of the judgment matrix, specifically to determine whether the consistency of the judgment matrix is ​​acceptable. CR is the ratio of CI (Conformity Index) to the Random Index (RI), i.e., CR = CI / RI. RI is a constant related to the matrix order n and can be obtained by looking up a table. Generally, when CR is less than 0.1, the consistency of the first judgment matrix is ​​considered acceptable; when CR is between 0.1 and 0.2, the consistency of the judgment matrix is ​​still acceptable, but further review may be needed; when CR is greater than 0.2, the consistency of the judgment matrix is ​​poor, and the judgment matrix needs to be corrected.

[0114] Furthermore, in this embodiment, when performing consistency verification on the first judgment matrix based on the initial annotation matrix, the largest eigenvalue λ of the first judgment matrix is ​​first determined according to the initial annotation matrix. max The formula for calculating the largest eigenvalue is: Where n is the order of the first judgment matrix, A is the first judgment matrix, and W is the initial annotation matrix; then, the CI value is determined based on the maximum eigenvalue and the above CI calculation formula, and the final consistency verification result CR is determined according to the CI value and the above CR calculation formula.

[0115] In step S440, the verification result is used to determine whether the verification has passed.

[0116] In this embodiment, after determining the consistency check result CR, the first judgment matrix corresponding to the check result is compared with a preset threshold to determine whether the check passes. Specifically, when the check result CR is less than the preset threshold (e.g., 0.1), it indicates that the consistency level of the first judgment matrix is ​​considered to be within the acceptable range. In this case, the check passes, the initial annotation matrix is ​​usable, and subsequent processing can be performed based on the initial annotation matrix. Conversely, if the check result CR is greater than or equal to the preset threshold, it indicates that a logical error occurred when constructing the judgment matrix. In this case, the check fails, the initial annotation matrix is ​​unusable, and the judgment matrix needs to be adjusted to determine a usable initial annotation matrix.

[0117] Furthermore, in this embodiment, if the verification result indicates that the verification has passed, step S450 is executed; if the verification result indicates that the verification has failed, step S460 is executed.

[0118] In step S450, in response to the verification result indicating that the verification is passed, the initial annotation matrix is ​​determined as the anomaly weight matrix, which includes the annotation weight of each charging anomaly code.

[0119] In this embodiment, when the first judgment matrix passes the verification, it indicates that the initial annotation matrix determined based on the current first judgment matrix is ​​usable. At this time, the initial annotation matrix is ​​determined as the anomaly weight matrix, and each data item in the anomaly weight matrix is ​​the annotation weight of each charging anomaly code.

[0120] In step S460, in response to the verification result indicating that the verification failed, the first judgment matrix is ​​adjusted until the verification result corresponding to the adjusted first judgment matrix indicates that the verification passed.

[0121] In this embodiment, when the first judgment matrix fails verification, the quantization values ​​in the current first judgment matrix are adjusted based on expert experience. After adjustment, the above steps are continued to analyze the adjusted first judgment matrix to determine the corresponding verification result, until the verification result corresponding to the adjusted first judgment matrix represents the verification passed, so that the initial annotation matrix corresponding to the verified first judgment matrix is ​​determined as the abnormal weight matrix.

[0122] Therefore, in this embodiment, by combining expert experience and the hierarchical analysis method to analyze all charging anomaly codes of charging equipment faults, multiple charging anomaly codes can be processed in batches, and the labeling weight of each charging anomaly code representing the charging equipment fault can be determined. This improves the efficiency of labeling weight determination while ensuring the accuracy of labeling weight.

[0123] In step S330, a fault weight matrix for each charging abnormal code is determined based on the relative importance of each charging abnormal code for different fault conditions and the labeling weight of each charging abnormal code. The fault weight matrix includes the fault weight of the corresponding charging abnormal code under different fault conditions.

[0124] In this embodiment, after determining the labeling weight of each charging abnormal code, a fault weight matrix for each charging abnormal code is determined based on the relative importance of each charging abnormal code to different fault conditions and the labeling weight of each charging abnormal code. The fault weight matrix is ​​used to identify the fault weight of each charging abnormal code under different fault conditions by referring to the data in the fault weight matrix.

[0125] Furthermore, this embodiment takes a charging error code as an example to illustrate the process of determining the fault weight matrix for the same charging error code under different fault conditions.

[0126] Figure 6 This is a flowchart illustrating the determination of the fault weight matrix according to an embodiment of the present invention. Figure 6 As shown, this embodiment uses the following steps to determine the fault weight matrix.

[0127] In step S610, a second judgment matrix corresponding to the target anomaly code is constructed. The target anomaly code is any charging anomaly code in the anomaly code set. The second judgment matrix includes a quantified value of the relative importance of the target anomaly code to different fault conditions. The quantified value is determined based on a preset scaling method.

[0128] In this embodiment, the second judgment matrix corresponding to different target anomaly codes is usually different, but the method for determining the second judgment matrix corresponding to each target anomaly code is the same.

[0129] Optionally, the method for constructing the second judgment matrix in this embodiment is the same as the method for constructing the first judgment matrix described above, and will not be repeated here.

[0130] In step S620, the second judgment matrix is ​​preprocessed to determine the initial fault matrix.

[0131] In this embodiment, the method for preprocessing the second judgment matrix to determine the initial fault matrix is ​​the same as the method for preprocessing the first judgment matrix to determine the abnormal weight matrix.

[0132] Figure 7 This is a flowchart illustrating the preprocessing of the second judgment matrix in an embodiment of the present invention. For example... Figure 7 As shown, in this embodiment, the second judgment matrix is ​​preprocessed using the following method to determine the corresponding initial fault matrix.

[0133] In step S710, the column vectors in the second judgment matrix are standardized to determine the second standard matrix.

[0134] In step S720, the row vectors in the second standard matrix are summed to determine the second weight matrix.

[0135] In step S730, the second weight matrix is ​​standardized to determine the initial fault matrix.

[0136] It should be understood that the processing method of steps S710-S730 when preprocessing the second judgment matrix in this embodiment corresponds to steps S510-S530 in the aforementioned preprocessing method of the first judgment matrix. The same processing method is used in each step as in the corresponding step, which will not be repeated here.

[0137] In step S630, the consistency of the second judgment matrix is ​​checked based on the initial fault matrix to determine the verification result of the second judgment matrix.

[0138] In this embodiment, after determining the corresponding initial fault matrix based on the second judgment matrix, a consistency check is performed on the second judgment matrix based on the initial fault matrix, and the check result of the second judgment matrix is ​​determined. Specifically, the consistency check method is the same as described above, and will not be repeated here.

[0139] In step S640, the verification result is used to determine whether the verification has passed.

[0140] In this embodiment, if the verification result indicates that the second judgment matrix has passed the verification, step S650 is executed; if the verification result indicates that the second judgment matrix has failed the verification, step S670 is executed.

[0141] In step S650, in response to the verification result indicating that the verification has passed, the initial fault matrix is ​​determined as the fault labeling matrix of the target anomaly code.

[0142] In this embodiment, when the second judgment matrix passes the verification, it indicates that the initial fault matrix determined based on the current second judgment matrix is ​​available. At this time, the initial fault matrix is ​​determined as the fault labeling matrix of the target anomaly code.

[0143] In step S660, the fault labeling matrix of the target anomaly code is multiplied by the labeling weight to determine the fault weight matrix, which includes the fault weight of the target anomaly code under different fault conditions.

[0144] In this embodiment, after determining the fault labeling matrix of the target anomaly code, the fault labeling matrix includes only one column of data (i.e., the fault labeling matrix is ​​a column vector), while the labeling weight is a single numerical value. The fault weight matrix can be determined by multiplying the fault labeling matrix of the target anomaly code by the labeling weight.

[0145] In step S670, in response to the verification result indicating that the verification failed, the second judgment matrix is ​​adjusted until the initial fault matrix corresponding to the adjusted second judgment matrix passes the verification.

[0146] In this embodiment, when the second judgment matrix fails verification, the quantization values ​​in the current second judgment matrix are adjusted based on expert experience. After adjustment, the above steps are continued to analyze the adjusted second judgment matrix to determine the corresponding verification result, until the verification result corresponding to the adjusted second judgment matrix passes the verification, so that the fault labeling matrix corresponding to the verified second judgment matrix is ​​determined as the abnormal weight matrix.

[0147] Furthermore, in this embodiment, after determining the fault weight matrix of a charging error code under different fault conditions, the fault weights of other charging error codes are determined based on the same method for determining the fault weight matrix, thereby determining the fault weight matrix of each charging error code under different fault conditions.

[0148] Therefore, this embodiment analyzes all fault scenarios of charging equipment by combining expert experience and hierarchical analysis method, which can process multiple fault scenarios in batch and determine the fault weight of each charging abnormal code in multiple fault scenarios, thereby improving the efficiency of fault weight determination while ensuring the accuracy of fault weight.

[0149] It should be noted that, in order to improve the accuracy of the charging anomaly code in labeling different fault conditions, this embodiment determines the corresponding fault weight matrix based on the labeling weight of the charging anomaly code and the fault labeling matrix. However, in scenarios where the accuracy requirement for fault condition labeling is low but the labeling efficiency requirement is high, this embodiment can also achieve fault condition labeling based only on the fault labeling matrix of each charging anomaly code. In this case, it is only necessary to perform the above steps S610-S650 and S670 for each charging anomaly code and directly determine the fault labeling matrix as the fault weight matrix, without having to perform step S660.

[0150] Furthermore, when determining the fault weight matrix of each charging abnormal code under different fault conditions based on the above method, and when it is necessary to label the data samples with fault conditions, this embodiment will obtain the fault weight matrix of each charging abnormal code in the data samples to determine the fault weight of each charging abnormal code in the data samples under different fault conditions.

[0151] For example, assuming data sample 1 includes charging error code 1 and charging error code 2, it is necessary to obtain the fault weight matrix A1 for charging error code 1 and the fault weight matrix A2 for charging error code 2. The fault weight matrix A1 for charging error code 1 includes fault weights A1, A2, and A3. 11 and fault weight A 12 A 11 Charging error code 1 indicates the importance of fault condition 1, A 12 Charging error code 1 represents the importance of fault condition 2. The fault weight matrix A2 of charging error code 2 includes fault weights A1, A2, and B2. 21 and fault weight A 22 A 21 Charging error code 2 indicates the importance of fault condition 1, A 22 The charging error code 2 indicates the importance of fault condition 2.

[0152] In step S130, the fault score for each fault condition is determined based on the different fault weights corresponding to each fault condition.

[0153] In this embodiment, after obtaining the fault weights of each charging abnormal code in the data sample under different fault conditions, the obtained fault weights are classified according to different fault conditions to determine the fault weights of each charging abnormal code under the same fault condition. The fault weights of each charging abnormal code under the same fault condition are then summed to determine the fault score for the corresponding fault condition.

[0154] For example, continuing with the above example, when data sample 1 includes charging error code 1 and charging error code 2, the total fault weights corresponding to fault condition 1 are determined, including fault weight A. 11 and fault weight A 21 The total fault weights corresponding to fault scenario 2 include fault weight A. 12 and fault weight A 22 Then for fault weight A 11 and fault weight A 21 The fault score for fault case 1 can be determined by summing the results, with fault weight A. 12 and fault weight A 22 The fault score for fault scenario 2 can be determined by summing the results; that is, the fault score for fault scenario 1 is (A). 11 +A 21 The fault score for fault scenario 2 is (A). 12 +A 22 Therefore, the fault scores of the data samples under various fault conditions are determined using the above method.

[0155] Optionally, considering that the occurrence time of each charging anomaly code in the data sample may differ from the time the charging device detects the fault, the reference value of different charging anomaly codes for fault diagnosis of the charging device will vary. To further improve the accuracy of fault condition labeling, this embodiment sets corresponding time weights for charging anomaly codes with different occurrence times. The earlier the occurrence time of the charging anomaly code, the lower its reference value for fault diagnosis, and the smaller its corresponding time weight. For example, if the occurrence time of charging anomaly code 1 is earlier than that of charging anomaly code 2, then the time weight w corresponding to charging anomaly code 1 is... t1 The time weight w is less than the time weight w corresponding to charging error code 2 t2 .

[0156] Meanwhile, when determining the score for a fault condition, the fault weight of each charging fault code is first weighted using the time weight corresponding to each charging fault code. Then, the weighted fault weights are summed to determine the fault score for that fault condition. For example, suppose the time weight corresponding to charging fault code 1 is w. t1 The time weight corresponding to charging error code 2 is w. t2 Then the fault score for charging error code 1 corresponding to fault condition 1 is (A 11 *w t1 +A21 *w t2 The fault score for fault scenario 2 is (A). 12 *w t1 +A 22 *w t2 ).

[0157] In step S140, the fault condition with the highest fault score is determined as the fault label of the data sample. The fault label includes gun fault and / or module fault.

[0158] In this embodiment, after determining the fault score for each fault condition, the fault condition with the highest fault score is determined by sorting the fault scores, and the fault condition with the highest fault score is determined as the fault label of the corresponding data sample.

[0159] The technical solution of this embodiment determines the fault score for each fault condition by using the fault weight of each charging anomaly code under different fault scenarios after acquiring data samples, and identifies the fault condition with the highest fault score as the fault label for the data sample. This approach fully considers the importance of different charging anomaly codes to different fault conditions when determining and labeling data samples, thus improving the accuracy of fault labeling. Furthermore, since determining the fault score for each fault condition only requires considering the fault weight of each charging anomaly code, and determining the fault label based on the fault score, data sample labeling becomes more convenient. This improves both the accuracy and efficiency of data sample labeling, achieving a balance between accuracy and efficiency. Moreover, this method enables batch processing of multiple data samples, ensuring data labeling accuracy while achieving batch processing of data sample labeling, thereby further improving data labeling efficiency.

[0160] Optionally, after determining the fault label of the data sample, the method of this embodiment further includes: labeling the data sample based on the fault label, and training a fault diagnosis model based on the labeled data sample. The fault diagnosis model is used to determine the cause of the fault of the charging device to be identified.

[0161] Furthermore, in this embodiment, when training the fault diagnosis model, multiple data samples are acquired, and the fault label of each data sample is determined based on the above-mentioned label determination method. Then, training samples are constructed using each data sample with a fault label, and the fault diagnosis model with a preset structure is trained using the training samples. The model parameters in the fault diagnosis model are determined, and the trained fault diagnosis model is obtained.

[0162] Optionally, the fault diagnosis model can be the XGBOOST model. The training process of the fault diagnosis model adopts a general model training method, which will not be described in detail here.

[0163] Therefore, in this embodiment, the fault label of each data sample is determined by the above method, and a fault diagnosis model is trained based on each data sample with fault label to perform fault diagnosis of the charging equipment. This is beneficial to improve the training efficiency of the fault diagnosis model and the accuracy of the fault diagnosis results output by the fault diagnosis model.

[0164] Figure 8 This is a charging equipment fault data tag determination device according to an embodiment of the present invention. For example... Figure 8 As shown, the label determination device in this embodiment includes a sample acquisition unit 1, a weight acquisition unit 2, a score determination unit 3, and a label determination unit 4. The sample acquisition unit 1 acquires data samples, which include at least one charging anomaly code. The weight acquisition unit 2 acquires the fault weight of each charging anomaly code under different fault conditions; the fault weight characterizes the importance of the charging anomaly code to the fault condition labeling. The score determination unit 3 determines the fault score for each fault condition based on its corresponding fault weight. The label determination unit 4 identifies the fault condition with the highest fault score as the fault label for the data sample; the fault label includes gun fault and / or module fault.

[0165] Optionally, the scoring determination unit 3 in this embodiment is also used to sum the fault weights of each charging abnormal code under the same fault condition to determine the fault score of the corresponding fault condition.

[0166] Optionally, the label determination unit 4 in this embodiment is further used to label the data samples based on the fault labels, so as to train a fault diagnosis model based on the labeled data samples. The fault diagnosis model is used to determine the cause of the fault of the charging device to be identified.

[0167] Optionally, the label determination device in this embodiment further includes a weight determination unit. The weight determination unit is used to determine the fault weight of each charging anomaly code under different fault conditions based on a multi-layer word analysis method. Specifically, it is used to obtain a set of anomaly codes for the charging device, which includes multiple charging anomaly codes; determine an anomaly weight matrix based on the relative importance of each charging anomaly code to the fault labeling, the anomaly weight matrix including the labeling weight of each charging anomaly code, the labeling weight being used to characterize the importance of the charging anomaly code to the fault labeling; and determine a fault weight matrix for each charging anomaly code based on the relative importance of each charging anomaly code to the labeling of different fault conditions and the labeling weight of each charging anomaly code, the fault weight matrix including the fault weight of the corresponding charging anomaly code under different fault conditions.

[0168] Figure 9 This is a schematic diagram of an electronic device according to an embodiment of the present invention. (For example...) Figure 9 As shown, Figure 9The illustrated electronic device is a general-purpose data processing device, comprising a general-purpose computer hardware architecture, including at least a processor 91 and a memory 92. The processor 91 and memory 92 are connected via a bus 93. The memory 92 is adapted to store instructions or programs executable by the processor 91. The processor 91 can be a standalone microprocessor or a collection of one or more microprocessors. Thus, the processor 91 executes the instructions stored in the memory 92, thereby performing the method flow of the embodiments of the present invention as described above to process data and control other devices. The bus 93 connects the aforementioned components together, and also connects these components to a display controller 94, a display device, and an input / output (I / O) device 95. The input / output (I / O) device 95 can be a mouse, keyboard, modem, network interface, touch input device, motion-sensing input device, printer, and other devices known in the art. Typically, the input / output device 95 is connected to the system via an input / output (I / O) controller 96.

[0169] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus (devices), or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0170] This application is described with reference to flowchart illustrations of methods, apparatus (devices), and computer program products according to embodiments of this application. It should be understood that each step in the flowchart can be implemented by computer program instructions.

[0171] These computer program instructions may be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including an instruction means, the implementation process of which is described in the instruction means. Figure 1 The function specified in one or more processes.

[0172] These computer program instructions may also be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, produce instructions for implementing processes. Figure 1 A device for a function specified in one or more processes.

[0173] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program for use by a computer to execute some or all of the above-described method embodiments.

[0174] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program specifying the relevant hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0175] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for determining fault data tags of charging equipment, characterized in that, The method includes: Obtain a data sample, the data sample including at least one charging error code; Obtain the fault weight of each charging error code under different fault conditions. The fault weight is used to characterize the importance of the charging error code to the fault condition labeling. The fault score for each fault condition is determined based on the different fault weights corresponding to each fault condition. The fault condition with the highest fault score is identified as the fault label of the data sample, and the fault label includes gun fault and / or module fault.

2. The method according to claim 1, characterized in that, The method further includes: Obtain the set of abnormal codes for the charging device, wherein the set of abnormal codes includes multiple charging abnormal codes; An anomaly weight matrix is ​​determined based on the relative importance of each charging anomaly code to the fault label. The anomaly weight matrix includes the label weight of each charging anomaly code, and the label weight is used to characterize the importance of the charging anomaly code to the fault label. The fault weight matrix of each charging anomaly code is determined based on the relative importance of each charging anomaly code in different fault conditions and the labeling weight of each charging anomaly code. The fault weight matrix includes the fault weight of the corresponding charging anomaly code under different fault conditions.

3. The method according to claim 2, characterized in that, The step of determining an anomaly weight matrix based on the relative importance of each charging anomaly code to the fault labeling, wherein the anomaly weight matrix includes the labeling weights of each charging anomaly code, including: Construct a first judgment matrix, which includes quantified values ​​of the relative importance of different charging anomaly codes to fault labeling, the quantified values ​​being determined based on a preset scaling method; The first judgment matrix is ​​preprocessed to determine the corresponding initial annotation matrix; Based on the initial annotation matrix, the first judgment matrix is ​​subjected to consistency verification to determine the verification result of the first judgment matrix; In response to the verification result indicating that the verification is passed, the initial annotation matrix is ​​determined as an anomaly weight matrix, the anomaly weight matrix including the annotation weight of each of the charging anomaly codes; In response to the verification result indicating that the verification failed, the first judgment matrix is ​​adjusted until the verification result corresponding to the adjusted first judgment matrix indicates that the verification passed.

4. The method according to claim 3, characterized in that, The step of preprocessing the first judgment matrix to determine the corresponding initial annotation matrix includes: Standardize each column vector in the first judgment matrix to determine the first standard matrix; Summing the row vectors in the first standard matrix determines the first weight matrix; The first weight matrix is ​​standardized to determine the initial annotation matrix.

5. The method according to claim 2, characterized in that, The step of determining the fault weight matrix for each charging anomaly code based on the relative importance of each charging anomaly code for different fault conditions and the labeling weight of each charging anomaly code includes: Construct a second judgment matrix corresponding to the target anomaly code, wherein the target anomaly code is any charging anomaly code in the anomaly code set, and the second judgment matrix includes a quantified value of the relative importance of the target anomaly code to different fault conditions, wherein the quantified value is determined based on a preset scaling method; The second judgment matrix is ​​preprocessed to determine the initial fault matrix; Based on the initial fault matrix, the second judgment matrix is ​​subjected to consistency verification to determine the verification result of the second judgment matrix; In response to the verification result indicating that the verification has passed, the initial fault matrix is ​​determined as the fault labeling matrix of the target anomaly code; The fault labeling matrix of the target anomaly code is multiplied by the labeling weights to determine the fault weight matrix, which includes the fault weights of the target anomaly code under different fault conditions.

6. The method according to claim 5, characterized in that, The step of preprocessing the second judgment matrix to determine the initial fault matrix includes: Standardize each column vector in the second judgment matrix to determine the second standard matrix; Summing the row vectors in the second standard matrix determines the second weight matrix; The second weight matrix is ​​standardized to determine the initial fault matrix.

7. The method according to claim 5, characterized in that, The step of determining the fault weight matrix for each charging anomaly code based on the relative importance of each charging anomaly code for different fault conditions and the labeling weight of each charging anomaly code further includes: In response to the verification result indicating that the verification failed, the second judgment matrix is ​​adjusted until the verification result corresponding to the adjusted second judgment matrix indicates that the verification passed.

8. The method according to claim 1, characterized in that, The process of determining the fault score for each fault condition based on the respective fault weights includes: The fault weights of each charging abnormal code under the same fault condition are summed to determine the fault score for the corresponding fault condition.

9. The method according to claim 1, characterized in that, The method further includes: The data samples are labeled based on the fault labels, and a fault diagnosis model is trained based on the labeled data samples. The fault diagnosis model is used to determine the cause of the fault in the charging device to be identified.

10. A device for determining fault data tags of charging equipment, characterized in that, The device includes: A sample acquisition unit is used to acquire a data sample, the data sample including at least one charging error code; The weight acquisition unit is used to acquire the fault weight of each charging abnormal code under different fault conditions. The fault weight is used to characterize the importance of the charging abnormal code to the fault condition labeling. The scoring unit is used to determine the fault score of the corresponding fault situation based on the different fault weights corresponding to each of the aforementioned fault situations. A label determination unit is used to determine the fault condition with the highest fault score as the fault label of the data sample, and the fault label includes gun fault and / or module fault.

11. A computer program product, characterized in that, The computer program product includes a computer program / instruction that, when executed by a processor, implements the method of any one of claims 1-9.

12. An electronic device comprising a memory and a processor, characterized in that, The memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method of any one of claims 1-9.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1-9.

Images