Electricity meter with tripping current recording and state analysis function

Abstract

The present application relates to the technical field of electric energy meter monitoring, and particularly relates to an electric energy meter with tripping current recording and state analysis function; different dimensions of similarity evaluation values are obtained according to the difference characteristics of current characteristic sequences and voltage characteristic sequences between different tripping events; clustering is performed according to the similarity evaluation values, and the contribution weights of the similarity evaluation values are obtained according to the difference characteristics of the clustering results corresponding to the similarity evaluation values of different dimensions; clustering iteration is performed according to the similarity evaluation values and the contribution weights, and the final clustering results are obtained according to the cluster quantity characteristics and the difference characteristics of the cluster tripping events; the cluster credibility is obtained according to the intra-cluster distribution characteristics of the final clustering results. According to the difference characteristics of the tripping events to be detected and the tripping events of different clusters in the final clustering results, and the cluster credibility, an abnormal evaluation value is obtained; the tripping reason is judged according to the abnormal evaluation value, and the accuracy of the electric energy meter tripping judgment is improved.

Description

Technical Field

[0001] This invention relates to the field of electricity meter monitoring technology, specifically to an electricity meter with trip current recording and status analysis functions. Background Technology

[0002] Energy meters with built-in load switches are widely used in scenarios such as power consumption control, remote tripping and closing, and abnormal power consumption protection. However, when a trip occurs, existing energy meters with built-in load switches typically only record the tripping event or execution result, lacking the recording and analysis of current and voltage data before the trip. This makes it difficult to accurately reflect the circuit condition at the time of the trip, which may lead to maintenance personnel being unable to accurately determine whether the trip is a normal response of the energy meter under overload or current surge, or an erroneous trip caused by abnormal tripping function. This affects the effective determination of normal tripping and erroneous tripping, and is not conducive to the accurate assessment of the reliability of energy meter operation and the service life of load switches. Summary of the Invention

[0003] To address the aforementioned technical problems, the present invention aims to provide an energy meter with trip current recording and status analysis functions. The specific technical solution adopted is as follows:

[0004] The data acquisition module is used to acquire the current characteristic sequence and voltage characteristic sequence when the same type of energy meter trips;

[0005] The first data analysis module is used to obtain similarity evaluation values ​​in different dimensions based on the differences in current and voltage characteristic sequences between different tripping events; to perform clustering based on the similarity evaluation values ​​between different tripping events; and to obtain the contribution weight of the similarity evaluation values ​​in different dimensions based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions.

[0006] The second data analysis module is used to perform clustering iterations based on the similarity evaluation value and the contribution weight; obtain the final clustering result based on the cluster quantity characteristics and the difference characteristics of cluster tripping events in the clustering iteration results; and obtain the cluster credibility based on the intra-cluster distribution characteristics of the final clustering result.

[0007] The tripping detection module is used to obtain an anomaly evaluation value based on the difference characteristics between the tripping event to be detected and the tripping events of different clusters in the final clustering results, and the credibility of the clusters; and to determine the tripping cause of the tripping event to be detected based on the anomaly evaluation value.

[0008] Furthermore, the step of obtaining similarity evaluation values ​​in different dimensions based on the differences in current characteristic sequences and voltage characteristic sequences between different tripping events includes:

[0009] The similarity evaluation values ​​in different dimensions include a first evaluation value, a second evaluation value, a third evaluation value, and a fourth evaluation value. For any two tripping events, the dynamic time warping distance of the current feature sequence is calculated and negatively correlated to obtain the first evaluation value; the dynamic time warping distance of the voltage feature sequence is calculated and negatively correlated to obtain the second evaluation value; the absolute value of the difference between the average values ​​of the current feature sequences is calculated and negatively correlated to obtain the third evaluation value; and the absolute value of the difference between the average values ​​of the voltage feature sequences is calculated and negatively correlated to obtain the fourth evaluation value.

[0010] Furthermore, the step of clustering based on the similarity assessment values ​​between different tripping events, and obtaining the contribution weights of the similarity assessment values ​​in different dimensions based on the difference features of the clustering results corresponding to the similarity assessment values ​​in different dimensions, includes:

[0011] Based on the similarity evaluation values ​​in any dimension, different tripping events are clustered using the HDBSCAN density clustering algorithm to obtain the clustering results corresponding to the similarity evaluation values ​​in any dimension.

[0012] In the formula, R represents the contribution weight of the similarity evaluation value in any dimension, and G represents the number of clusters in the clustering results of the similarity evaluation value in any dimension. The number of tripping events in the g-th cluster is represented by M, and M represents the number of other dimensions of the similarity evaluation value. This represents an exponential function with the natural constant as its base. This represents the number of tripping events that occur most frequently in any cluster class among the clustering results corresponding to the g-th cluster and other dimensions in the m-th dimension. This represents the number of tripping events in any cluster class within the clustering results corresponding to the m-th other dimension. This represents the intersection rate of events.

[0013] Further, the step of performing clustering iterations based on the similarity evaluation value and the contribution weight includes:

[0014] In the formula Let H represent the initial weights of the similarity assessment values ​​for the h-th dimension, where H represents the number of dimensions. The initial weight represents the contribution weight of the similarity evaluation value of the h-th dimension; the weighted sum of the similarity evaluation values ​​among the tripping events is calculated based on the initial weight to obtain the comprehensive similarity; the comprehensive similarity is used to perform clustering iteration on all tripping events to obtain different clustering iteration results; the initial weights corresponding to each dimension are iteratively adjusted according to the particle swarm optimization algorithm.

[0015] Furthermore, the step of obtaining the final clustering result based on the cluster quantity characteristics and the difference characteristics of cluster tripping events from the clustering iteration results includes:

[0016] In the formula This represents the clustering result value of the t-th clustering iteration. This represents the average comprehensive similarity among all cluster centers in the clustering results of the t-th clustering iteration. This represents the number of clusters in the clustering result of the t-th clustering iteration. This represents the average comprehensive similarity among all tripping events in the f-th cluster.

[0017] An effect curve is constructed based on the clustering effect values ​​of different clustering iterations, and the optimal number of clustering iterations is obtained by using the elbow method. The clustering iteration result corresponding to the optimal number of clustering iterations is taken as the final clustering result.

[0018] Furthermore, the step of obtaining cluster confidence based on the intra-cluster distribution characteristics of the final clustering result includes:

[0019] In the formula This represents the cluster confidence of the nth cluster in the final clustering result. This represents the number of tripping events in the nth cluster. This represents an exponential function with the natural constant as its base. This represents the number of different energy meters in the nth cluster, and N represents the number of clusters in the final clustering result.

[0020] Furthermore, the step of obtaining anomaly evaluation values ​​based on the difference features between the tripping event to be detected and the tripping events of different clusters in the final clustering results, and the cluster confidence level, includes:

[0021] In the formula, K represents the anomaly assessment value of the tripping event to be detected. This indicates linear normalization, and N represents the number of clusters in the final clustering result. This represents the cluster confidence level of the nth cluster. This represents the minimum value of the overall similarity among tripping events in the nth cluster. This represents the minimum value of the overall similarity between the tripping event to be detected and the tripping events in the nth cluster.

[0022] Furthermore, the step of determining the tripping cause of the tripping event to be detected based on the abnormal evaluation value includes:

[0023] When the abnormal assessment value of the tripping event to be detected exceeds the preset percentage of the abnormal assessment value of historical tripping events, a false tripping alert will be issued.

[0024] The present invention has the following beneficial effects:

[0025] In this invention, obtaining similarity evaluation values ​​in different dimensions can measure the similarity of circuit operating states before different tripping events, thus enabling the classification of different tripping events based on these similarity evaluation values. Clustering is performed based on the similarity evaluation values ​​between different tripping events, and the contribution weights of the similarity evaluation values ​​in different dimensions are obtained based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions. This measures the importance of the similarity evaluation values ​​in classifying tripping events, allowing for more accurate classification of tripping events according to the actual scenario. Iterative clustering continuously adjusts the weights of the similarity evaluation values ​​in each dimension, thereby selecting a more accurate classification result for the tripping events. Obtaining the final clustering result determines the final classification result of the tripping events, thus improving the accuracy of subsequent tripping cause determination. Obtaining cluster credibility characterizes the probability that the tripping event in the cluster is caused by a real overload or short circuit, thus enabling the determination of the causative factors of the tripping event to be detected. Finally, obtaining the anomaly evaluation value accurately determines the tripping cause of the tripping event to be detected, improving the accuracy of the tripping factor determination for the electricity meter. Attached Figure Description

[0026] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a system block diagram of an energy meter with trip current recording and status analysis functions, provided as an embodiment of the present invention. Detailed Implementation

[0028] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an energy meter with trip current recording and state analysis functions proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0029] 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 this invention pertains.

[0030] The following description, in conjunction with the accompanying drawings, details a specific scheme for an energy meter with trip current recording and status analysis functions provided by the present invention.

[0031] Please see Figure 1 The diagram illustrates a system block diagram of an energy meter with trip current recording and status analysis functions according to an embodiment of the present invention. The energy meter includes the following modules:

[0032] The data acquisition module S1 is used to acquire the current characteristic sequence and voltage characteristic sequence when the same type of energy meter trips.

[0033] In this embodiment of the invention, the implementation scenario is to detect the cause of tripping events in electricity meters and improve the accuracy of electricity meter reliability assessment. During the operation of electricity meters, the normal triggering types of tripping events mainly include command-type tripping, overload-type tripping, and short-circuit-type tripping. Among them, command-type tripping is triggered by a command issued by the remote control terminal system, and can be accurately traced through the control log. Therefore, this embodiment of the invention does not analyze command-type tripping events. This embodiment of the invention mainly focuses on non-command-type tripping events that occur in the normal operation of electricity meters, analyzing whether their tripping behavior meets the normal protection triggering conditions such as overload or short circuit, thereby determining whether there is a false tripping phenomenon in the electricity meter. First, the current characteristic sequence and voltage characteristic sequence of the same model of electricity meter when tripping are obtained. When a tripping occurs, the voltage sequence and current sequence 10 seconds before the tripping are obtained and normalized respectively to obtain the corresponding current characteristic sequence and voltage characteristic sequence. The implementer can determine the acquisition length according to the implementation scenario.

[0034] The first data analysis module S2 is used to obtain similarity evaluation values ​​in different dimensions based on the differences in current and voltage characteristic sequences between different tripping events; to perform clustering based on the similarity evaluation values ​​between different tripping events; and to obtain the contribution weight of the similarity evaluation values ​​in different dimensions based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions.

[0035] For energy meters of the same model and specifications, under normal tripping conditions, their protection actions are usually triggered by overload or short-circuit conditions. False tripping often occurs when electrical parameters such as voltage and current do not meet the normal tripping trigger conditions. Therefore, similarity assessment can be performed based on the current and voltage data collected before the trip to determine the cause of the trip. For different tripping events of the same type, the overall trend of voltage and current changes before the trip is usually similar. For example, in overload tripping, the current gradually increases; in short-circuit tripping, the current suddenly increases and is accompanied by a sharp drop in voltage. However, due to differences in the time of tripping and the load at the time of tripping, although the changes in the collected voltage and current data may be similar, the phase of the data may have a certain degree of offset. If a point-to-point alignment method is used for comparison, it is difficult to accurately reflect the true similarity of the operating state before the trip. Therefore, this embodiment of the invention introduces a dynamic time warping algorithm to match and calculate the changes in voltage and voltage timing before the trip, thereby eliminating the influence of time offset on the similarity assessment and quantifying the consistency of different tripping events in the trend of electrical parameter changes. On the other hand, judging similarity solely based on the morphology of the events is insufficient to reflect the differences in the overall electrical operation level across various tripping events. Therefore, it is necessary to further extract the average amplitude characteristics of the voltage and voltage timing before the tripping event to characterize the load level and energy characteristics before the tripping event. Thus, similarity assessment values ​​in different dimensions are obtained based on the differences in current and voltage characteristic sequences among different tripping events.

[0036] Preferably, in this embodiment of the invention, the step of obtaining similarity evaluation values ​​includes: similarity evaluation values ​​in different dimensions, including a first evaluation value, a second evaluation value, a third evaluation value, and a fourth evaluation value; for any two tripping events, calculating the dynamic time warping distance of the current feature sequence and negatively correlated mapping to obtain the first evaluation value; calculating the dynamic time warping distance of the voltage feature sequence and negatively correlated mapping to obtain the second evaluation value; it should be noted that the dynamic time warping distance is calculated using an existing dynamic time warping algorithm. The more similar the two sequences are, the smaller the dynamic time warping distance. This algorithm is existing technology, and the specific steps will not be elaborated further. Calculating the absolute value of the difference between the average values ​​of the current feature sequences and negatively correlated mapping to obtain the third evaluation value; calculating the absolute value of the difference between the average values ​​of the voltage feature sequences and negatively correlated mapping to obtain the fourth evaluation value. In this embodiment of the invention, through Perform negative correlation mapping, where 'a' represents the mapping object. This represents an exponential function with the natural constant as its base. A higher similarity score indicates greater similarity in the data characteristics preceding the two tripping events.

[0037] Furthermore, similarity assessment values ​​across different dimensions can evaluate the degree of similarity between voltage or current characteristics prior to a tripping event. However, accurately setting the contribution weights when combining similarity assessment values ​​from different dimensions is difficult. Accurately setting these contribution weights can improve the accuracy of tripping event classification; therefore, further analysis based on the actual characteristics prior to the tripping event is necessary. Since the clustering effects of similarity assessment values ​​across different dimensions on tripping events vary, if the clustering results for a certain dimension differ little from those for other dimensions, it means that this dimension has limited ability to distinguish tripping events, and its contribution weight should be set relatively low. Conversely, if the clustering results for a certain dimension differ significantly from those for other dimensions, it indicates that this dimension has strong ability to distinguish tripping events, and its contribution weight should be set relatively high. Therefore, clustering can be performed based on the similarity evaluation values ​​between different tripping events. The contribution weights of the similarity evaluation values ​​in different dimensions can be obtained based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions. Preferably, in this embodiment of the invention, the steps of clustering and obtaining contribution weights include: clustering different tripping events using the HDBSCAN density clustering algorithm based on the similarity evaluation values ​​in any dimension, obtaining the clustering results corresponding to the similarity evaluation values ​​in any dimension. Thus, all tripping events have been clustered using the similarity evaluation values ​​in each dimension, resulting in different clustering results. It should be noted that the HDBSCAN density clustering algorithm is an improved version of the existing DBSCAN density clustering algorithm, and clustering can be performed directly without preset parameters; the specific steps are not described in detail here.

[0038]

[0039] In the formula, R represents the contribution weight of the similarity evaluation value in any dimension, and G represents the number of clusters in the clustering results of the similarity evaluation value in any dimension. This represents the number of tripping events in the g-th cluster. The larger the value, the larger the cluster size, and the greater its weight when measuring cluster differences. M represents the number of other dimensions in the similarity assessment value. This represents an exponential function with the natural constant as its base. This represents the number of tripping events that occur most frequently in any cluster class among the clustering results corresponding to the g-th cluster and other dimensions in the m-th dimension. The larger the value, the more identical tripping events occur between the two clusters in the two dimensions, and the more similar the two clusters are. This represents the number of tripping events in any cluster within the clustering results corresponding to the m-th other dimension. This represents the event intersection rate. A higher event intersection rate means a greater proportion of identical tripping events, indicating a more similarity between the two clusters. Therefore, a higher event intersection rate between all clusters in this dimension and all clusters in all other dimensions means a more similar clustering result for this dimension compared to other dimensions. The smaller the value, the smaller the contribution weight of that dimension to the similarity evaluation value. Conversely, the smaller the event intersection rate between that dimension and all other dimensions, the greater the difference between the clustering results of that dimension and other dimensions. The larger the value, the greater its contribution to the similarity assessment value of that dimension.

[0040] The second data analysis module S3 is used to perform clustering iteration based on similarity evaluation values ​​and contribution weights; obtain the final clustering result based on the cluster quantity characteristics and cluster tripping event difference characteristics of the clustering iteration results; and obtain the cluster credibility based on the intra-cluster distribution characteristics of the final clustering result.

[0041] After obtaining the contribution weights corresponding to the similarity evaluation values ​​for all dimensions, clustering iterations can be performed based on the similarity evaluation values ​​and contribution weights; preferably, in this embodiment of the invention, the clustering iteration steps include: In the formula Let H represent the initial weights of the similarity assessment values ​​for the h-th dimension, where H represents the number of dimensions. This represents the contribution weight of the similarity assessment value in the h-th dimension; the initial weight is the normalized value of the contribution weight. A weighted sum of the similarity assessment values ​​between tripping events is calculated based on the initial weight to obtain the comprehensive similarity. The comprehensive similarity combines the similarity assessment values ​​of all dimensions to measure the differences between two tripping events. The initial weight allows for adaptive assignment of more accurate weights to the similarity assessment values ​​of different dimensions according to the actual scenario, resulting in higher accuracy in subsequent tripping cause determination. A higher comprehensive similarity indicates a higher consistency in the operating states of the two tripping events before the trip, making the segmentation of tripping events more accurate. All tripping events are clustered iteratively based on the comprehensive similarity to obtain different clustering iteration results. The initial weights corresponding to each dimension are iteratively adjusted using the particle swarm optimization algorithm. It should be noted that the particle swarm optimization algorithm is an existing technology, and its specific steps are not detailed here. After each output of the initial weights corresponding to each dimension, the HDBSCAN density clustering algorithm performs a clustering iteration on the tripping events based on the updated comprehensive similarity. After clustering iteration, the effects of different clusters need to be evaluated to determine a more accurate clustering result. Therefore, the final clustering result is obtained based on the cluster quantity characteristics and the difference characteristics of cluster tripping events in the clustering iteration results. Preferably, in this embodiment of the invention, the step of obtaining the final clustering result includes:

[0042]

[0043] In the formula, This represents the clustering result value of the t-th clustering iteration. This represents the average comprehensive similarity among all cluster centers in the clustering results of the t-th clustering iteration. The smaller the comprehensive similarity among cluster centers of different clusters, the greater the difference between clusters and the better the clustering effect. This represents the number of clusters in the clustering result of the t-th clustering iteration. To avoid excessive splitting, it is necessary to constrain the number of clusters. Therefore, the fewer the number of clusters, the better the clustering effect. This represents the average comprehensive similarity among all tripping events in the f-th cluster. A higher average comprehensive similarity within a cluster indicates more similar tripping events and better clustering performance. Therefore, a higher clustering effect value signifies better clustering results for that iteration and more accurate event segmentation. Furthermore, an effect curve is constructed based on the clustering effect values ​​of different iterations, and the optimal number of iterations is obtained using the elbow method. The clustering iteration result corresponding to the optimal number of iterations is taken as the final clustering result. It should be noted that the elbow method is existing technology, and its specific steps will not be elaborated further. The final clustering result accurately segments tripping events under the same meter model, improving the accuracy of subsequent tripping event judgments.

[0044] After obtaining the final clustering results, the distribution characteristics of tripping events within different clusters are different, and the cluster confidence levels will also differ. Therefore, the cluster confidence level is obtained based on the intra-cluster distribution characteristics of the final clustering results. Preferably, in this embodiment of the invention, the step of obtaining the cluster confidence level includes:

[0045]

[0046] In the formula, This represents the cluster confidence of the nth cluster in the final clustering result. This represents the number of tripping events in the nth cluster. This represents the number of different energy meters in the nth cluster, and N represents the number of clusters in the final clustering result. Since abnormal tripping of energy meters is less frequent than normal overload or short-circuit tripping, the larger the cluster size, the more likely the tripping event in that cluster is a normal trip. On the other hand, normal tripping usually originates from different energy meters, while abnormal tripping usually originates from or is concentrated in a single energy meter. Therefore, if... The smaller This means that the greater the likelihood that multiple tripping events in a cluster originate from individual electricity meters, the more likely the cluster is to represent an abnormal false tripping, and the lower its reliability. Therefore, the higher the reliability of a cluster, the more likely the tripping events within that cluster are genuine tripping events.

[0047] The trip detection module S4 is used to obtain an anomaly evaluation value based on the difference characteristics between the trip event to be detected and the trip events of different clusters in the final clustering results, and the cluster confidence level; and to determine the tripping cause of the trip event to be detected based on the anomaly evaluation value.

[0048] After obtaining the cluster confidence levels of different clusters, anomaly assessment values ​​can be obtained based on the differences between the tripping events to be detected and the tripping events of different clusters in the final clustering results, and the cluster confidence levels. Preferably, in this embodiment of the invention, the step of obtaining anomaly assessment values ​​includes:

[0049]

[0050] In the formula, K represents the anomaly assessment value of the tripping event to be detected. This indicates linear normalization, and N represents the number of clusters in the final clustering result. This represents the cluster confidence level of the nth cluster. This represents the minimum overall similarity among tripping events in the nth cluster. This represents the minimum comprehensive similarity between the tripping event to be detected and the tripping events in the nth cluster. This comprehensive similarity is calculated based on the weights corresponding to the similarity evaluation values ​​of each dimension in the final clustering result. The smaller This means that the less similar the tripping event to be detected is to the tripping events of this cluster, and if... The larger the value, the greater the cluster class reliability, and the more likely the cluster class is to reflect normal tripping events. Therefore, when... The larger the value, the more likely the tripping event to be detected is an abnormal erroneous trip. Therefore, the larger the anomaly assessment value, the more likely the tripping event to be detected is an abnormal erroneous trip.

[0051] Furthermore, the cause of the tripping event to be detected can be determined based on the abnormal assessment value. Preferably, in this embodiment of the invention, when the abnormal assessment value of the tripping event to be detected exceeds a preset percentage of the abnormal assessment values ​​of historical tripping events, a false tripping alert is issued. The implementer can determine the preset percentage according to the implementation scenario. For example, if the abnormal assessment value exceeds 90% of the historical tripping events, the tripping event to be detected is highly likely to be an abnormal false trip, which can remind relevant personnel to check and replace the electricity meter, reduce the time spent on line inspection, improve maintenance efficiency, and prevent similar tripping problems from recurring. The implementer can determine the method for tripping judgment based on the abnormal assessment value according to the implementation scenario, which is not limited here.

[0052] In summary, this invention provides an energy meter with trip current recording and status analysis functions. It obtains similarity evaluation values ​​in different dimensions based on the differences in current and voltage characteristic sequences between different trip events. Clustering is performed based on these similarity evaluation values, and the contribution weight of each similarity evaluation value is obtained based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions. Clustering iteration is performed based on the similarity evaluation values ​​and contribution weights, and the final clustering result is obtained based on the number of clusters and the differences in trip events within each cluster. The cluster reliability is obtained based on the intra-cluster distribution characteristics of the final clustering result. This invention obtains anomaly evaluation values ​​based on the differences in trip events between the detected trip event and the trip events in different clusters in the final clustering result, as well as the cluster reliability. The cause of the trip is determined based on the anomaly evaluation values, improving the accuracy of energy meter trip judgment.

[0053] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An energy meter with trip current recording and status analysis functions, characterized in that, The electricity meter includes the following modules: The data acquisition module is used to acquire the current characteristic sequence and voltage characteristic sequence when the same type of energy meter trips; The first data analysis module is used to obtain similarity evaluation values ​​in different dimensions based on the differences in current and voltage characteristic sequences between different tripping events; to perform clustering based on the similarity evaluation values ​​between different tripping events; and to obtain the contribution weight of the similarity evaluation values ​​in different dimensions based on the differences in the clustering results corresponding to the similarity evaluation values ​​in different dimensions. The second data analysis module is used to perform clustering iteration based on the similarity evaluation value and the contribution weight; and to obtain the final clustering result based on the cluster quantity characteristics and the difference characteristics of cluster tripping events in the clustering iteration results. The cluster confidence level is obtained based on the intra-cluster distribution characteristics of the final clustering results; The trip detection module is used to obtain an anomaly evaluation value based on the difference characteristics between the trip event to be detected and the trip events of different clusters in the final clustering results, and the confidence level of the clusters; and to determine the tripping cause of the trip event to be detected based on the anomaly evaluation value. The step of clustering based on the similarity assessment values ​​between different tripping events, and obtaining the contribution weights of the similarity assessment values ​​in different dimensions based on the difference features of the clustering results corresponding to the similarity assessment values ​​in different dimensions, includes: Based on the similarity evaluation values ​​in any dimension, different tripping events are clustered using the HDBSCAN density clustering algorithm to obtain the clustering results corresponding to the similarity evaluation values ​​in any dimension. In the formula, R represents the contribution weight of the similarity evaluation value in any dimension, and G represents the number of clusters in the clustering results of the similarity evaluation value in any dimension. The number of tripping events in the g-th cluster is represented by M, and M represents the number of other dimensions of the similarity evaluation value. This represents an exponential function with the natural constant as its base. This represents the number of tripping events that occur most frequently in any cluster class among the clustering results corresponding to the g-th cluster and other dimensions in the m-th dimension. This represents the number of tripping events in any cluster class within the clustering results corresponding to the m-th other dimension. This represents the intersection rate of events.

2. The energy meter with trip current recording and status analysis function according to claim 1, characterized in that, The steps for obtaining similarity evaluation values ​​in different dimensions based on the differences in current and voltage characteristic sequences between different tripping events include: The similarity evaluation values ​​in different dimensions include a first evaluation value, a second evaluation value, a third evaluation value, and a fourth evaluation value. For any two tripping events, the dynamic time warping distance of the current feature sequence is calculated and negatively correlated to obtain the first evaluation value; the dynamic time warping distance of the voltage feature sequence is calculated and negatively correlated to obtain the second evaluation value; the absolute value of the difference between the average values ​​of the current feature sequences is calculated and negatively correlated to obtain the third evaluation value; and the absolute value of the difference between the average values ​​of the voltage feature sequences is calculated and negatively correlated to obtain the fourth evaluation value.

3. An energy meter with trip current recording and status analysis functions according to claim 1, characterized in that, The step of performing clustering iterations based on the similarity evaluation value and the contribution weight includes: In the formula Let H represent the initial weights of the similarity assessment values ​​for the h-th dimension, where H represents the number of dimensions. The initial weight represents the contribution weight of the similarity evaluation value of the h-th dimension; the weighted sum of the similarity evaluation values ​​among the tripping events is calculated based on the initial weight to obtain the comprehensive similarity; the comprehensive similarity is used to perform clustering iteration on all tripping events to obtain different clustering iteration results; the initial weights corresponding to each dimension are iteratively adjusted according to the particle swarm optimization algorithm.

4. An energy meter with trip current recording and status analysis function according to claim 3, characterized in that, The steps for obtaining the final clustering result based on the cluster quantity characteristics and the difference characteristics of cluster tripping events in the clustering iteration results include: In the formula This represents the clustering result value of the t-th clustering iteration. This represents the average comprehensive similarity among all cluster centers in the clustering results of the t-th clustering iteration. This represents the number of clusters in the clustering result of the t-th clustering iteration. This represents the average comprehensive similarity among all tripping events in the f-th cluster. An effect curve is constructed based on the clustering effect values ​​of different clustering iterations, and the optimal number of clustering iterations is obtained by using the elbow method. The clustering iteration result corresponding to the optimal number of clustering iterations is taken as the final clustering result.

5. An energy meter with trip current recording and status analysis function according to claim 1, characterized in that, The step of obtaining cluster confidence based on the intra-cluster distribution characteristics of the final clustering result includes: In the formula This represents the cluster confidence of the nth cluster in the final clustering result. This represents the number of tripping events in the nth cluster. This represents an exponential function with the natural constant as its base. This represents the number of different energy meters in the nth cluster, and N represents the number of clusters in the final clustering result.

6. An energy meter with trip current recording and status analysis function according to claim 3, characterized in that, The step of obtaining anomaly evaluation values ​​based on the difference characteristics between the tripping event to be detected and the tripping events of different clusters in the final clustering results, and the cluster confidence level, includes: In the formula, K represents the anomaly assessment value of the tripping event to be detected. This indicates linear normalization, and N represents the number of clusters in the final clustering result. This represents the cluster confidence level of the nth cluster class. This represents the minimum value of the overall similarity among tripping events in the nth cluster. This represents the minimum value of the overall similarity between the tripping event to be detected and the tripping events in the nth cluster.

7. An energy meter with trip current recording and status analysis function according to claim 1, characterized in that, The step of determining the tripping cause of the tripping event to be detected based on the abnormal evaluation value includes: When the abnormal assessment value of the tripping event to be detected exceeds the preset percentage of the abnormal assessment value of historical tripping events, a false tripping alert will be issued.

Images