A method for monitoring the charging process of a high-power direct current charging device

Abstract

The application relates to the technical field of battery charging of electric vehicles, in particular to a charging process monitoring method of a high-power direct-current charging device, which comprises the following steps: determining normal data frames, mode change frames and abnormal fluctuation frames by analyzing the data fluctuation degree of charging current and voltage and the difference among each voltage data frame and the previous and subsequent voltage data frames; clustering the voltage data frames according to the difference degree of abnormal fluctuation frames and normal data frames and in combination with the abnormal fluctuation degree of the voltage data frames, and adaptively determining the weight of the clustering cluster according to the data abnormality degree of the clustering center in the clustering cluster; weighting and superimposing the data abnormality degrees of adjacent voltage data frames to obtain neighborhood abnormality degrees, and performing overcurrent fault detection through the neighborhood abnormality degrees; the method can avoid the misoperation of the charging device caused by the change of charging power during the charging process, improve the monitoring accuracy of overcurrent faults in the charging process, and improve the safety of electric vehicle battery charging.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery charging of electric vehicles, in particular to a charging process monitoring method of a high-power direct current charging device. BACKGROUND

[0002] The charging time of an electric vehicle is one of the important factors affecting the user experience. The high-power direct current charging device can output high-power and high-current direct current, which can power various battery types of electric vehicles such as lithium ion batteries, and realize fast charging of electric vehicles. Compared with alternating current charging devices, the safety risk of high-power direct current charging devices is greater. Monitoring the charging process of the high-power direct current charging device can perform protection actions such as circuit disconnection in the early stage of failure, thereby improving the safety of the electric vehicle charging process.

[0003] In the charging process of the high-power direct current charging device, as the internal aging of the electric vehicle power battery continues, the internal resistance of the power supply increases, which is prone to overcurrent failure. Overcurrent failure not only causes damage to the high-power direct current charging device, but also causes internal short circuit of the electric vehicle power battery, leading to thermal runaway and inducing serious safety problems. Overcurrent failure usually occurs in the early stage, accompanied by fluctuations in charging data, which is similar to data fluctuations caused by changes in charging process power. Currently, when using outlier detection algorithms to monitor the charging process of the charging device, data fluctuations caused by changes in charging power are easily misjudged as overcurrent failure, which in turn leads to false actions of the charging device, affecting the normal charging of the high-power direct current charging device to the electric vehicle. SUMMARY

[0004] To solve the above technical problems, the present application provides a charging process monitoring method of a high-power direct current charging device to solve the existing problems.

[0005] The charging process monitoring method of the high-power direct current charging device of the present application adopts the following technical scheme:

[0006] One embodiment of the present application provides a charging process monitoring method of a high-power direct current charging device, which comprises the following steps:

[0007] Collecting current and voltage data during the charging process of the electric vehicle, and constructing each current data frame and each voltage data frame;

[0008] Analyzing the data fluctuation degree in each data frame to construct a first feature value of each data frame, and constructing a forward and backward difference vector of each voltage data frame in combination with the data difference between each voltage data frame and the two voltage data frames before and after it;

[0009] In the adjacent voltage data frames before the current voltage data frame, the abnormality degree of the front and back difference vectors of each adjacent voltage data frame and the sign difference of the elements in the front and back difference vectors are analyzed to distinguish normal data frames, mode change frames and abnormal fluctuation frames, and the mode change frames are replaced by normal data frames;

[0010] In the replaced adjacent voltage data frames, the data abnormality degree of each adjacent voltage data frame is determined based on the modulus of the front and back difference vectors of each adjacent voltage data frame and the first characteristic value of each adjacent voltage data frame and its corresponding current data frame, and a feature analysis vector of each adjacent voltage data frame is constructed.

[0011] The neighborhood abnormality degree of the current voltage data frame is determined based on the data abnormality degree difference and the feature analysis vector difference between all adjacent voltage data frames to perform abnormal detection of the high-power direct current charging equipment.

[0012] In one embodiment, the process of obtaining the first characteristic value is as follows:

[0013] Both the current and voltage data are recorded as charging data, stretching processing is performed on each charging data to obtain enhanced data of each charging data, the enhanced data mean of each data frame is calculated, and the cumulative sum of the difference between all enhanced data of each data frame and the enhanced data mean is calculated as the first characteristic value of each data frame.

[0014] In one embodiment, the process of obtaining the front and back difference vector of each voltage data frame is as follows:

[0015] For voltage data, the difference between the enhanced data mean of the previous frame and the current frame of the current frame is calculated, multiplied by the first characteristic value of the current frame to obtain a first component, the difference between the enhanced data mean of the current frame and the next frame thereof is calculated, multiplied by the first characteristic value of the current frame to obtain a second component, and the first component and the second component are combined to obtain the front and back difference vector of the current frame.

[0016] In one embodiment, the process of distinguishing normal data frames, mode change frames and abnormal fluctuation frames is as follows:

[0017] The front and back difference vectors of all adjacent voltage data frames of the current voltage data frame are taken as the input of the abnormal detection algorithm to obtain normal data frames and abnormal data frames in all adjacent voltage data frames, and the mode change frames and abnormal fluctuation frames in the abnormal data frames are determined by analyzing the sign of the elements in the front and back difference vectors.

[0018] In one embodiment, the mode change frame and the abnormal fluctuation frame are as follows: the abnormal data frame in which the signs of the two elements in the front and back difference vector are the same is recorded as a mode change frame, and the abnormal data frame in which the signs of the two elements are different is recorded as an abnormal fluctuation frame.

[0019] In one embodiment, the data anomaly degree obtaining process is as follows:

[0020] The mean of the first feature value of each adjacent voltage data frame and its corresponding current data frame is recorded as the second feature value of each adjacent voltage data frame; and the modulus of the front-back difference vector of each adjacent voltage data frame is recorded as the third feature value of each adjacent voltage data frame.

[0021] The fusion result of the second feature value and the third feature value of each adjacent voltage data frame is taken as the data anomaly degree of each adjacent voltage data frame.

[0022] In one embodiment, the feature analysis vector is a vector composed of the second feature value and the third feature value of each adjacent voltage data frame.

[0023] In one embodiment, the neighborhood anomaly degree obtaining process of the current voltage data frame is as follows:

[0024] The feature analysis vectors of all adjacent voltage data frames of the current voltage data frame are clustered, the mean of the data anomaly degrees of the adjacent voltage data frames corresponding to the cluster centers of all clusters is calculated, and the normalized result of the difference between the data anomaly degree of each cluster center corresponding to the adjacent voltage data frame and the mean of the data anomaly degrees is taken as the weight of the cluster cluster where the cluster center is located.

[0025] The neighborhood anomaly degree of the current voltage data frame is determined based on the weight and the data anomaly degree of each adjacent voltage data frame.

[0026] In one embodiment, the neighborhood anomaly degree obtaining process is as follows:

[0027] The product of the data anomaly degree of each adjacent voltage data frame in the cluster and the weight of the cluster is recorded as a first product, and the fusion value of the first products of all adjacent voltage data frames of the current voltage data frame is taken as the neighborhood anomaly degree of the current voltage data frame.

[0028] In one embodiment, the anomaly detection of the high-power DC charging device is specifically as follows:

[0029] If the neighborhood anomaly degree of the current voltage data frame exceeds a preset monitoring threshold, it is determined that the high-power DC charging device has an overcurrent fault risk; otherwise, it is determined that the high-power DC charging device does not have an overcurrent fault risk.

[0030] The present application has at least the following beneficial effects:

[0031] The application constructs the first characteristic value of each current and voltage data frame by analyzing the data fluctuation degree of the charging current and voltage, to preliminarily evaluate the abnormal degree of the data frame; the normal data frame, the mode change frame and the abnormal fluctuation frame are determined through the difference of each voltage data frame and the front and rear voltage data frames, combined with the abnormal detection algorithm, the different changes of the voltage data before and after the charging mode change of the power battery and the charging equipment are utilized, the corresponding voltage data frames of the two are distinguished, and the false action of the charging equipment caused by the change of the charging power in the charging process is avoided; the voltage data frame is clustered according to the difference degree of the abnormal fluctuation frame and the normal data frame, and combined with the abnormal fluctuation degree of the voltage data frame, the weight of the clustering cluster is adaptively determined according to the data abnormal degree of the clustering center in the clustering cluster, and the identification ability of the abnormal fluctuation of the charging data is further enhanced; and then the data abnormal degree of the adjacent voltage data frame is weighted and superimposed to obtain the neighborhood abnormal degree, the overcurrent fault detection is performed through the neighborhood abnormal degree, the monitoring accuracy of the overcurrent fault in the charging process is improved, and the safety of the electric vehicle battery charging is improved. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0033] Figure 1 A flow chart of a charging process monitoring method of a high-power direct current charging equipment provided by the present application is shown in the figure.

[0034] Figure 2 A schematic diagram of the acquisition process of the front and rear difference vector is shown in the figure. DETAILED DESCRIPTION

[0035] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific implementation, structure, features and effects of the charging process monitoring method of a high-power direct current charging equipment according to the present application are described in detail as follows by combining with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0037] The application provides a charging process monitoring method of a high-power direct current charging device.

[0038] The application provides a charging process monitoring method of a high-power direct current charging device.

[0039] Specifically, the application provides a charging process monitoring method of a high-power direct current charging device, please refer to Figure 1 The method comprises the following steps:

[0040] In step S1, current and voltage data in the charging process of the electric vehicle are collected, and each current data frame and each voltage data frame are constructed.

[0041] The application collects voltage data and current data in the charging process of the high-power direct current charging device through a voltmeter and an ammeter installed in the high-power direct current charging device, and the sampling frequency of both is 1 kHz. For the convenience of description, the voltage data and the current data are collectively referred to as charging data below.

[0042] The application performs the same frame processing on the collected voltage data and current data, and the data frames corresponding to the voltage data and the current data are respectively recorded as voltage data frames and current data frames, wherein the number of corresponding data contained in each data frame is N, and N is 100 in the embodiment. For the convenience of description, the voltage data frames and the current data frames are collectively referred to as charging data frames below.

[0043] In step S2, the first characteristic value of each data frame is constructed by analyzing the data fluctuation degree in each data frame, and the front and rear difference vectors of each voltage data frame are constructed by combining the data difference between each voltage data frame and the two voltage data frames before and after it.

[0044] The aging of the power battery of the electric vehicle causes abnormal data fluctuation in the charging process, and then causes overcurrent failure. In the early stage of overcurrent failure, the data change caused by the aging of the power battery is relatively rapid, and the instantaneous value of the current change is relatively large.

[0045] (1) Due to the introduction of V2G technology, the charging mode of the electric vehicle changes frequently. The data fluctuation caused by the change of the charging power of the electric vehicle lasts for a long time, and the numerical value fluctuation range in a short time is small. In order to improve the identification ability of the abnormal charging data, the application adopts a nonlinear function to stretch the charging data. For the stretching processing of the charging data, preferably, a nonlinear function for stretching processing given in the embodiment of the application is:

[0046]

[0047] wherein, enhanced data representing the current charging data, representing the charging data after stretching processing of the current charging data; current charging data, i.e. the charging data before stretching; max, representing the maximum value of the charging data in the current charging data frame; is a preset index, in the embodiments of the present application, the value of is set to 10 to enhance the difference between the abnormal charging data and the normal charging data. In other embodiments of the present application, the implementer can set the value of according to the actual situation.

[0048] It should be noted that for the stretching processing of the charging data, the present application only provides a stretching processing method, there are many existing stretching processing methods, and the implementer can also use other functions to perform stretching processing on the charging data, which is not specifically limited in the present application.

[0049] To further quantify the degree of data fluctuation in the charging process, the present application calculates the first eigenvalue of any charging data frame. First, the mean value of the enhanced data of all charging data in the charging data frame is calculated. Second, the absolute value of the difference between each enhanced data in the charging data frame and the mean value is calculated, and all the absolute values of the difference of the charging data frame are accumulated. Finally, the accumulation result is recorded as the first eigenvalue of the charging data frame, reflecting the degree of data fluctuation in the charging data frame caused by the aging of the power battery or the change of the power of the charging device.

[0050] When the charging power remains stable, the output voltage of the high-power direct current charging device does not change much, and the probability of current fluctuation is low, the difference between the enhanced data in the charging data frame and the mean value is small, and the first eigenvalue is small.

[0051] The electric vehicle battery is charged in "constant current mode", and as the battery power changes, the charging power also needs to be adjusted, at this time the current does not change and the voltage changes. That is, the voltage changes when the charging power changes, and under normal circumstances the current does not change. Compared with when the power does not change, when the power changes, the probability of current fluctuation increases, the absolute value of the difference between the enhanced data in the charging data frame and the mean value is large, and the first eigenvalue is large.

[0052] At the same time, when the power supply produces abnormal data fluctuation due to aging, the first eigenvalue of the corresponding data frame is large.

[0053] (2) The degree of data fluctuation caused by the aging of the power battery and the change of the power of the charging device is close, and it is difficult to distinguish the charging data frame caused by the aging of the battery through the first eigenvalue.

[0054] The high-power DC charging device supplies power to the electric vehicle in a constant current mode. The voltage and current data change obviously before and after the charging mode changes. The output power of the charging device differs before and after the charging mode changes, and the corresponding output voltage or current changes obviously before and after the charging mode changes. The data fluctuation caused by the aging of the power battery of the electric vehicle usually appears in the form of pulses, the fluctuation time is short, and the voltage and current data can recover to the previous value in a short time. Therefore, the charging data frames corresponding to the aging of the power battery and the different changes of the data before and after the charging mode changes of the charging device are distinguished.

[0055] Firstly, any charging data frame is taken as a current frame, the difference between the enhanced data mean of the previous frame of the current frame and the enhanced data mean of the current frame is recorded as a first forward difference value, the difference between the enhanced data mean of the current frame and the enhanced data mean of the next frame of the current frame is recorded as a first backward difference value, and the forward difference value and the backward difference value are multiplied by the first feature value of the current frame respectively, and the products are recorded as a second forward difference value and a second backward difference value respectively.

[0056] The second forward difference value and the second backward difference value are taken as a first component and a second component respectively to form a forward-backward difference vector of the current frame, which is used to distinguish different charging data frames. In this application, the purpose of weighting the two components of the forward-backward difference vector by the first feature value is to increase the discrimination degree of the charging data frames in normal charging and abnormal charging, and thus to improve the accuracy of overcurrent fault monitoring.

[0057] When the charging mode of the charging device remains unchanged, the output voltage or current remains basically stable, the enhanced data means of adjacent data frames differ little, and the absolute values of the forward difference value and the backward difference value of the current frame are small.

[0058] When the output voltage or current changes due to the change of the charging mode of the charging device, the enhanced data means of the current frame and its previous and next frames differ greatly because of the obvious difference between the output voltage or current before and after the change of the charging mode, and the absolute values of the forward difference value and the backward difference value of the current frame are large and the two values are of the same sign.

[0059] When the power battery of the electric vehicle appears abnormal fluctuation of charging data due to aging, the voltage or current values before and after the data fluctuation remain basically unchanged, the enhanced data means of the current frame and its previous and next frames differ greatly, and the absolute values of the forward difference value and the backward difference value of the current frame are large and the two values are of different signs.

[0060] Step S3, in the adjacent voltage data frames before the current voltage data frame, analyze the abnormal degree of the front and rear difference vectors of each adjacent voltage data frame and the sign difference of the values of the elements in the front and rear difference vectors, to distinguish normal data frames, mode change frames and abnormal fluctuation frames, and replace the mode change frames with normal data frames.

[0061] To capture the charging data fluctuation before the overcurrent fault and improve the accuracy of overcurrent fault monitoring of the high-power DC charging equipment during the charging process, the application takes any voltage data frame as an example, takes the voltage data frame as the starting frame, and takes M-1 voltage data frames forward to form an observation window of the voltage data frame, expands the monitoring range of the data, and obtains effective charging data features. Wherein, all voltage data frames in the observation window of any voltage data frame are adjacent voltage data frames of the voltage data frame. The observation window contains M voltage data frames, and M is 100 in the present application. In other embodiments of the application, the implementer can set the size of the observation window according to the actual situation.

[0062] Take any voltage data frame as the current voltage data frame, and for all adjacent voltage data frames in the observation window thereof, take the front and rear difference vectors of all adjacent voltage data frames as input, adopt an anomaly detection algorithm, and divide the input front and rear difference vectors into normal vectors and abnormal vectors. Since the abnormal fluctuation duration of the charging data is short, the normal voltage data frames account for the majority in the observation window, so the adjacent voltage data frames corresponding to the normal vectors are recorded as normal data frames, and the adjacent voltage data frames corresponding to the abnormal vectors are recorded as abnormal data frames.

[0063] Wherein, the anomaly detection algorithm adopted in the embodiments of the application is the LOF algorithm. There are many existing anomaly detection algorithms, and the implementer can also adopt other anomaly detection algorithms to detect the abnormality of the front and rear difference vectors. The application does not make specific limitations.

[0064] Further, the two components of the abnormal vector are judged. The adjacent voltage data frames corresponding to the abnormal vectors with the same sign of the two components are recorded as mode change frames; and the adjacent voltage data frames corresponding to the abnormal vectors with different signs of the two components are recorded as abnormal fluctuation frames.

[0065] At this point, for the current voltage data frame, all adjacent voltage data frames in the observation window thereof are divided into normal data frames, mode change frames and abnormal fluctuation frames, further, all mode change frames in the observation window thereof are deleted, and the nearest normal data frames are used to fill the vacant positions.

[0066] The voltage data frame with data fluctuation caused by the charging mode change of the charging device is screened out through the acquired mode change frame, so as to avoid the false operation of the charging device caused by the charging power change during the charging process, reduce the influence caused by the charging mode change, and improve the monitoring accuracy for the overcurrent fault.

[0067] In step S4, the data abnormality degree of each adjacent voltage data frame is determined based on the modulus of the front-back difference vector of each adjacent voltage data frame and the first characteristic value of each adjacent voltage data frame and its corresponding current data frame, and a feature analysis vector of each adjacent voltage data frame is constructed.

[0068] The current data frame with the same sampling time as each voltage data frame is taken as the current data frame corresponding to each voltage data frame.

[0069] For each adjacent voltage data frame in the modified observation window of the current voltage data frame, the mean of the first characteristic value of the adjacent voltage data frame and its corresponding current data frame is taken as the second characteristic value of the adjacent voltage data frame, reflecting the overall fluctuation degree of the charging data.

[0070] Meanwhile, the difference in the modulus of the front-back difference vector of different voltage data frames can enhance the difference between the abnormal fluctuation frame and the normal data frame. Therefore, the modulus of the front-back difference vector of each adjacent voltage data frame is taken as the third characteristic value of each adjacent voltage data frame.

[0071] The second characteristic value and the third characteristic value of each adjacent voltage data frame are taken as the first component and the second component, respectively, to form the feature analysis vector of each adjacent voltage data frame.

[0072] Any adjacent voltage data frame is taken as the current adjacent voltage data frame. Taking the current adjacent voltage data frame as an example, the data abnormality degree thereof is calculated according to the second characteristic value and the third characteristic value thereof, and the expression is as follows:

[0073]

[0074] wherein, represents the data abnormality degree of the current adjacent voltage data frame, and is used to measure the data abnormality degree of the current adjacent voltage data frame; and respectively represent the second characteristic value and the third characteristic value of the current adjacent voltage data frame; and respectively represent the maximum value of the second characteristic value and the maximum value of the third characteristic value of all adjacent voltage data frames in the observation window.

[0075] The third characteristic value represents the difference degree of the abnormal fluctuation frame and the normal data frame, and the second characteristic value reflects the overall fluctuation degree of the charging data. The greater the values of the two, the greater the abnormal degree of the corresponding charging data, and the greater the data abnormality degree obtained by calculation. and The purpose is to control the calculation range of T.

[0076] In step S5, the neighborhood abnormality degree of the current voltage data frame is determined based on the data abnormality degree difference and the feature analysis vector difference between all adjacent voltage data frames, so as to perform abnormality detection of the high-power direct current charging equipment.

[0077] The more serious the internal aging of the power battery of the electric vehicle, the greater the risk of overcurrent failure, and the more frequent the short-time data fluctuation before the overcurrent failure occurs.

[0078] The frequency of data abnormal fluctuation in different observation windows is different. In order to further enhance the recognition ability of the charging data abnormal fluctuation, the voltage data frames in the observation window of any voltage data frame are clustered according to the data abnormal fluctuation degree in the embodiment of the application. Specifically, the feature analysis vectors of all adjacent voltage data frames of the current voltage data frame are taken as input, and the FCM clustering algorithm is used to output the clustered clusters. In this embodiment, the number of clustering clusters is set to 5, and the maximum iteration number is set to 20. In other embodiments of the application, the implementer can set the number of clustering clusters and the maximum iteration number according to the actual situation. The FCM clustering algorithm is a known technology, and the specific process will not be described here.

[0079] It should be noted that for clustering of the feature analysis vectors of all adjacent voltage data frames, only one clustering method is provided in the embodiment of the application, and there are many existing clustering methods. The implementer can also use other clustering algorithms to cluster the feature analysis vectors of all adjacent voltage data frames, and the application does not make specific limitations.

[0080] The data abnormality degree mean of the adjacent voltage data frame corresponding to the clustering center of all clustering clusters is calculated, the normalized result of the difference between the data abnormality degree of the adjacent voltage data frame corresponding to the clustering center and the data abnormality degree mean is calculated as the weight of the clustering cluster where the clustering center is located, reflecting the abnormal degree of the voltage data frame in different clustering clusters. The normalization method used in this embodiment is arctangent normalization, and the normalized weight value interval is (0, 1). The stronger the abnormal degree of the data change of the abnormal fluctuation frame in the clustering cluster, the greater the weight.

[0081] Further, the data abnormality degree of each adjacent voltage data frame in the cluster and the weight of the cluster are multiplied to obtain a first product. Further, the sum of the first products of all adjacent voltage data frames of the current voltage data frame is calculated as the neighborhood abnormality degree of the current voltage data frame, reflecting the abnormality degree of the charging data in the observation window. The stronger the abnormality degree of the data change of the abnormal fluctuation frame in the observation window, the higher the proportion of the abnormal fluctuation frame, the stronger the abnormality degree of the charging data in the observation window, and the greater the risk of overcurrent fault of the current voltage data frame.

[0082] All voltage data frames and current data frames of Q consecutive complete charging processes without overcurrent fault are collected, and the neighborhood abnormality degrees of all voltage data frames are calculated by using the above method, and the maximum value is recorded as a monitoring threshold. If the neighborhood abnormality degree of the current collected voltage data frame exceeds the monitoring threshold, it is determined that the high-power direct current charging device has an overcurrent fault risk, and the device is controlled to perform a protection action. Otherwise, it is determined that the high-power direct current charging device does not have an overcurrent fault risk. Preferably, in the embodiment of the application, the value of Q is set to 500. In other embodiments of the application, the implementer can set the value of Q according to the actual situation.

[0083] The acquisition process of the front-back difference vector is shown in the schematic diagram of FIG. 1. Figure 2

[0084] In summary, the embodiments of the application construct the first feature value of each current and voltage data frame by analyzing the data fluctuation degree of the charging current and voltage, to preliminarily evaluate the abnormality degree of the data frame. By determining the normal data frame, the mode change frame and the abnormal fluctuation frame through the difference of each voltage data frame and the front and back voltage data frames, and combining the abnormality detection algorithm, the different changes of the voltage data before and after the change of the charging mode of the power battery and the charging device are utilized to distinguish the corresponding voltage data frames, so as to avoid the false action of the charging device due to the change of the charging power during the charging process. According to the difference degree of the abnormal fluctuation frame and the normal data frame, and combining the abnormal fluctuation degree of the voltage data frame, the voltage data frame is clustered, the weight of the cluster is adaptively determined according to the data abnormality degree of the cluster center in the cluster, and the recognition ability of the abnormal fluctuation of the charging data is further enhanced. Further, the data abnormality degrees of the adjacent voltage data frames are weighted and superimposed to obtain the neighborhood abnormality degree, and the overcurrent fault is detected through the neighborhood abnormality degree, so as to improve the monitoring accuracy of the overcurrent fault in the charging process and improve the safety of the electric vehicle battery charging.

[0085] ​It should be noted that the above-mentioned order of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes the specific embodiments of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0086] Each of the embodiments in the present application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments.

[0087] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; modifying the technical solutions described in the above embodiments, or equivalently replacing some technical features, does not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for monitoring the charging process of a high-power DC charging device, characterized in that, The method includes the following steps: Collect current and voltage data during the electric vehicle charging process, and construct each current data frame and each voltage data frame; The first feature value of each data frame is constructed by analyzing the degree of data fluctuation in each data frame. The difference vector between each voltage data frame and the two voltage data frames before and after it is constructed by combining the data difference between each voltage data frame and the two voltage data frames before and after it. In the adjacent voltage data frames preceding the current voltage data frame, analyze the degree of anomaly of the differential vectors before and after each adjacent voltage data frame and the difference in the positive and negative signs of the elements in the differential vectors before and after, in order to distinguish normal data frames, mode change frames and abnormal fluctuation frames, and replace the mode change frames with normal data frames. In the replaced adjacent voltage data frames, based on the magnitude of the difference vector between the preceding and following adjacent voltage data frames, and the first feature value of each adjacent voltage data frame and its corresponding current data frame, the data anomaly degree of each adjacent voltage data frame is determined, and the feature analysis vector of each adjacent voltage data frame is constructed. The neighborhood anomaly degree of the current voltage data frame is determined based on the differences in data anomaly degree and feature analysis vectors among all neighboring voltage data frames, in order to detect anomalies in high-power DC charging equipment. The process of obtaining the first feature value is as follows: Both current and voltage data are recorded as charging data; each charging data is stretched to obtain enhanced data; the mean of the enhanced data for each data frame is calculated; the sum of the differences between all enhanced data in each data frame and the mean of the enhanced data is calculated and used as the first feature value of each data frame. The process of obtaining the data anomaly degree is as follows: The mean of the first characteristic values ​​of each neighboring voltage data frame and its corresponding current data frame is recorded as the second characteristic value of each neighboring voltage data frame; the modulus of the difference vector between the front and back of each neighboring voltage data frame is recorded as the third characteristic value of each neighboring voltage data frame; the data anomaly degree of each neighboring voltage data frame is calculated based on the second and third characteristic values ​​of each neighboring voltage data frame.

2. The method for monitoring the charging process of a high-power DC charging device as described in claim 1, characterized in that, The process of obtaining the difference vectors between the preceding and following voltage data frames is as follows: For voltage data, calculate the difference between the mean of enhanced data of the previous frame and the current frame, multiply it by the first feature value of the current frame to obtain the first component; calculate the difference between the mean of enhanced data of the current frame and the next frame, multiply it by the first feature value of the current frame to obtain the second component. The first component and the second component are combined to obtain the difference vector between the front and back of the current frame.

3. The method for monitoring the charging process of a high-power DC charging device as described in claim 1, characterized in that, The process of distinguishing between normal data frames, mode change frames, and abnormal fluctuation frames is as follows: The difference vectors between the current voltage data frame and all its neighboring voltage data frames are used as input to the anomaly detection algorithm to obtain normal and abnormal data frames among all neighboring voltage data frames. Analyzing the signs of the elements in the difference vectors before and after the data frame determines the pattern change frames and abnormal fluctuation frames in the abnormal data frame.

4. The method for monitoring the charging process of a high-power DC charging device as described in claim 3, characterized in that, The mode change frame and the abnormal fluctuation frame are respectively: the abnormal data frame in which the two elements in the difference vector before and after are of the same sign is recorded as the mode change frame, and the abnormal data frame in which the two elements are of different signs is recorded as the abnormal fluctuation frame.

5. The method for monitoring the charging process of a high-power DC charging device as described in claim 1, characterized in that, The feature analysis vector is a vector composed of the second and third feature values ​​of each adjacent voltage data frame.

6. The method for monitoring the charging process of a high-power DC charging device as described in claim 1, characterized in that, The process of obtaining the neighborhood anomaly degree of the current voltage data frame is as follows: Cluster the feature analysis vectors of all neighboring voltage data frames of the current voltage data frame, calculate the mean data anomaly of the neighboring voltage data frames corresponding to the cluster centers of all clusters, and use the normalized result of the difference between the data anomaly of the neighboring voltage data frames corresponding to each cluster center and the mean data anomaly as the weight of the cluster to which each cluster center belongs. The neighborhood anomaly of the current voltage data frame is determined based on the weights and the data anomalies of each neighboring voltage data frame.

7. The method for monitoring the charging process of a high-power DC charging device as described in claim 6, characterized in that, The process of obtaining the neighborhood anomaly degree is as follows: The product of the data anomaly degree of each neighboring voltage data frame in the cluster and the weight of the cluster is recorded as the first product. The sum of the first products of all neighboring voltage data frames of the current voltage data frame is used as the neighborhood anomaly degree of the current voltage data frame.

8. The method for monitoring the charging process of a high-power DC charging device as described in claim 1, characterized in that, The specific steps for performing anomaly detection on high-power DC charging equipment are as follows: If the neighborhood anomaly of the current voltage data frame exceeds the preset monitoring threshold, it is determined that the high-power DC charging equipment has an overcurrent fault risk; otherwise, it is determined that the high-power DC charging equipment does not have an overcurrent fault risk.

Images