Power transformation equipment parameter processing method and device and electronic equipment

Abstract

The application discloses a power transformation equipment parameter processing method and device and electronic equipment. It relates to the technical field of power equipment detection. The method comprises the following steps: obtaining a target data set of power transformation equipment, wherein the target data set is obtained based on initial parameter data collected at multiple sampling time points; performing clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; determining the average value and the standard deviation corresponding to the N clusters based on the parameter data included in the N clusters; and determining abnormal parameter data based on the average value and the standard deviation corresponding to the N clusters. The application solves the technical problem that the accuracy of determining abnormal parameter data is low due to the fact that the power transformation equipment parameter processing method in the related art does not comprehensively consider various factors.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment detection, in particular to a power transformation equipment parameter processing method and device and electronic equipment. BACKGROUND

[0002] In the traditional power transformation equipment parameter processing method, there is often a key problem that some potential abnormal situations may be ignored, so that problems are not discovered in time or even normal situations are misreported. This is mainly due to the fact that the traditional method can usually only process device parameters from a limited perspective, and it is difficult to fully capture the diversity of abnormalities. For example, some abnormalities may only appear under the interaction of multiple parameters, but the traditional method may only focus on one parameter, thus missing the overall picture of the abnormality.

[0003] At present, no effective solution has been proposed to solve the problem of low accuracy of determining abnormal parameter data due to the fact that the power transformation equipment parameter processing method in the related art does not consider all factors. SUMMARY

[0004] The embodiments of the present application provide a power transformation equipment parameter processing method, device and electronic equipment to at least solve the technical problem of low accuracy of determining abnormal parameter data due to the fact that the power transformation equipment parameter processing method in the related art does not consider all factors.

[0005] According to an aspect of the embodiments of the present application, a power transformation equipment parameter processing method is provided, including: obtaining a target data set of a power transformation equipment, wherein the target data set is obtained based on initial parameter data collected at a plurality of sampling time points; performing clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; determining mean values and standard deviations corresponding to the N clusters respectively based on parameter data included in the N clusters; and determining abnormal parameter data based on the mean values and standard deviations corresponding to the N clusters respectively.

[0006] According to another aspect of the embodiments of the present application, a power transformation equipment parameter processing device is also provided, including: a first obtaining module configured to obtain a target data set of a power transformation equipment, wherein the target data set is obtained based on initial parameter data collected at a plurality of sampling time points; a first clustering module configured to perform clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; a first determining module configured to determine mean values and standard deviations corresponding to the N clusters respectively based on parameter data included in the N clusters; and a second determining module configured to determine abnormal parameter data based on the mean values and standard deviations corresponding to the N clusters respectively.

[0007] According to another aspect of the embodiments of the present application, an electronic device is provided, which includes one or more processors and a memory storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement any of the power transformation device parameter processing methods.

[0008] In the embodiments of the present application, the target data set of the power transformation device is acquired, wherein the target data set is acquired based on the initial parameter data collected at a plurality of sampling time points; the parameter data included in the target data set is clustered to obtain N clusters, wherein N is an integer greater than or equal to 2; the average value and the standard deviation corresponding to the N clusters are determined based on the parameter data included in the N clusters respectively; and the abnormal parameter data is determined based on the average value and the standard deviation corresponding to the N clusters, so as to achieve the purpose of accurately determining the abnormal parameter data based on the average value and the standard deviation corresponding to the target parameter data, thereby realizing the technical effect of improving the accuracy of the power transformation device parameter processing method, and further solving the technical problem that the power transformation device parameter processing method in the related art does not comprehensively consider factors, resulting in low accuracy of determining the abnormal parameter data. BRIEF DESCRIPTION OF DRAWINGS

[0009] The accompanying drawings, which are included to provide a further understanding of the present application and constitute a part of this application, illustrate embodiments of the present application and together with the description serve to explain the present application. In the drawings:

[0010] Figure 1 is a schematic diagram of a power transformation device parameter processing method according to an embodiment of the present application;

[0011] Figure 2 is a schematic diagram of a power transformation device parameter processing device according to an embodiment of the present application;

[0012] Figure 3 is a schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0013] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor should fall within the protection scope of the present application.

[0014] It is to be understood that the terms "first", "second", and the like used in the description and the claims of the present application as well as the above-described drawings do not necessarily have to connote any ordinal, sequential or chronological order, but are merely used to distinguish a different set of objects. It is to be understood that the terms so used in the description and the claims are interchangeable under appropriate circumstances. Furthermore, the terms "comprise", "have", "contain" and any variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, system, product or apparatus that comprises, has, contains one or more of a recited step or element does not include the exclusion of any other steps or elements not expressly recited.

[0015] According to an embodiment of the present application, a method embodiment of substation parameter processing is provided. It is to be understood that the steps shown in the flowcharts of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0016] Figure 1 is a flowchart of a substation parameter processing method according to an embodiment of the present application, as shown in Figure 1 the method comprises the following steps:

[0017] Step S102, obtaining a target data set of the substation, wherein the target data set is obtained based on a plurality of sampling time points respectively collected initial parameter data.

[0018] Optionally, the substation includes a transmission and distribution substation, wherein the transmission and distribution substation: mainly used for the high voltage power generated by the power plant through the transformer to transform and transmit, the power from the power plant to the substation in different areas. The distribution substation: mainly used for the power transmitted by the transmission line to further transform and distribute, the power is transmitted to the end user. Transmission and distribution substation and distribution substation are important components in power system, used for transmission, transformation, distribution and control of electric energy, which plays a key role in ensuring the safety, stability and high efficiency of power system. The related parameter data of the substation includes current, voltage, temperature, humidity, etc. These parameter data can be obtained from sensors, monitoring systems or device control systems.

[0019] In an optional embodiment, the target data set of the power transformation equipment is acquired, including: acquiring an initial data set of the power transformation equipment, wherein the initial data set includes initial parameter data collected at a plurality of sampling time points; performing first data preprocessing on the parameter data included in the initial data set to obtain a first data set, wherein the first data preprocessing includes at least one of the following: missing value processing, outlier processing; performing screening processing on the parameter data included in the first data set to obtain a second data set, wherein the screening processing is used to determine the parameter data related to the state and performance of the power transformation equipment; and performing second data preprocessing on the parameter data included in the second data set to obtain the target data set, wherein the second data preprocessing includes at least one of the following: data standardization processing, data normalization processing, and data conversion processing.

[0020] Optionally, the initial data set of the power transformation equipment is acquired. Since the parameter data included in the initial data set may have missing values, the missing values can be handled by filling, deleting or interpolation. Since the parameter data included in the initial data set may have outliers, the outliers may be caused by sensor failure, data collection error, etc. These outliers may have a negative impact on subsequent analysis. Therefore, statistical methods (such as Z-score or box plot) can be used to identify and handle outliers, and these outliers can be selected to be rejected, repaired or marked. After the first data preprocessing operation on the parameter data included in the initial data set, the first data set is obtained. The parameter data included in the first data set is screened to select features related to the state and performance of the power transformation equipment, which can help reduce dimensions and eliminate irrelevant information, thereby improving the efficiency and accuracy of subsequent analysis. Feature selection methods such as correlation analysis, variance analysis, mutual information, etc. can be used to determine which features are most relevant to the target. On the basis of feature selection, the features that have the greatest impact on the state and performance of the power transformation equipment can be further selected. The domain knowledge, feature importance ranking (such as feature importance of random forest) or model-based method (such as recursive feature elimination) can be used to complete the selection. On the basis of feature selection and feature screening, new features can be created to capture the relationship between data according to the needs of a specific problem. For example, the power factor can be calculated based on current and voltage, thereby introducing more information. After the screening processing of the parameter data included in the first data set, the second data set is obtained. The parameter data included in the second data set is subjected to data standardization processing, data normalization processing or data conversion processing to ensure that the scales of different features are consistent. The final target data set contains the parameter data most relevant to the state and performance of the power transformation equipment, which is the basis for subsequent steps and can help better perform clustering analysis and anomaly detection.

[0021] Step S104, the parameter data included in the target data set is clustered to obtain N clusters, where N is an integer greater than or equal to 2.

[0022] Optionally, the K-means algorithm can be used for clustering analysis to divide the parameter data into different clusters. K-means is a commonly used clustering algorithm that divides data points into K clusters so that each data point is closest to the center point of the cluster it belongs to. For a given parameter data x i where i = 1, 2,..., N, and the cluster center y k where k = 1, 2,..., K, the goal of the K-means algorithm is to minimize the following loss function:

[0023]

[0024] where w ik is an indicator variable, which is 1 when the i-th parameter data belongs to cluster k, and 0 otherwise; ||x i -y k || 2 is the square of the distance between the i-th parameter data and the cluster center.

[0025] In an optional embodiment, the parameter data included in the target data set is clustered to obtain N clusters, including: determining the number of clusters as N; determining N target parameter data from the parameter data included in the target data set; based on the N target parameter data, determining N initial cluster center points, and repeatedly performing the following operations until a predetermined termination condition is met: calculating the distance from other parameter data in the target data set, except the N target parameter data, to the N initial cluster center points; based on the distance from the other parameter data to the N initial cluster center points, determining the cluster to which the other parameter data belongs; calculating the average value of the parameter data included in the N clusters, respectively; based on the average value, updating the N initial cluster center points, N new cluster center points; based on the N new cluster center points, obtaining N clusters when the predetermined termination condition is met.

[0026] Optionally, the step of clustering the parameter data included in the target data set is as follows:

[0027] Step S1041, the number of clusters N is selected, i.e. the parameter data included in the target data set is divided into N clusters.

[0028] Step S1042, randomly select N parameter data from the parameter data as initial cluster center points.

[0029] Step S1043, for the other parameter data in the target data set except for the N target parameter data, calculate the distance between each initial cluster center point and the parameter data, and assign the parameter data to the cluster to which the nearest initial cluster center point belongs.

[0030] Step S1044, calculate the average value of all parameter data in each cluster, and take it as a new cluster center point.

[0031] Step S1045, determine whether the preset iteration number is reached, if not, execute step S1043, if yes, execute step S1046.

[0032] Step S1046, finally, based on the N new cluster center points, obtain N clusters, each cluster containing a group of similar parameter data.

[0033] In an optional embodiment, new parameter data of the power transformation equipment is obtained; the similarity of the new parameter data with the N clusters is calculated to obtain N similarities; if all the N similarities are greater than or equal to a preset threshold, the maximum similarity in the N similarities is determined, and the new parameter data is divided into the cluster corresponding to the maximum similarity; if any similarity in the N similarities is less than the preset threshold, a new cluster is created, and the new parameter data is divided into the new cluster.

[0034] Optionally, when the power transformation equipment has new parameter data, a clustering algorithm for concept learning (COBWEB algorithm) can be used to monitor the changes of the equipment parameters in real time. When the COBWEB algorithm is used to monitor the changes of the equipment parameters in real time, the algorithm will dynamically update the existing clustering model according to the newly appeared parameter data in the data stream, and assign the new parameter data to the appropriate cluster, which allows the changes of the equipment parameters to be identified in the constantly changing data stream and responded in time. The specific steps of monitoring the changes of the equipment parameters in real time by using the COBWEB algorithm are as follows: according to the existing clustering model, the COBWEB algorithm calculates the similarity of the new parameter data with each cluster, if the similarity of the new parameter data with each cluster is high, the new parameter data is assigned to the cluster with the highest similarity; if the similarity of the new parameter data with each cluster is low, a new cluster is created, and the new parameter data is divided into the new cluster. With the influx of more parameter data, the COBWEB algorithm will gradually adjust the position and size of the cluster according to the new data pattern to adapt to the changes of the data. In this way, the COBWEB algorithm can monitor the changes of the power transformation equipment parameters in real time in the constantly changing data stream, and update the clustering model according to the characteristics of the new parameter data, so as to maintain the real-time perception of the equipment state.

[0035] In an optional embodiment, after clustering the parameter data included in the target data set to obtain N clusters, the method further comprises: based on the parameter data respectively included in the N clusters, establishing a parameter prediction model corresponding to each of the N clusters, wherein the parameter prediction model is used to predict the parameter value at a future sampling time.

[0036] Optionally, for the parameter data respectively included in each cluster, a time series analysis can be used to construct the parameter prediction model, and an exponential smoothing model, an autoregressive moving average model, a long short-term memory neural network, etc. can be used to construct the parameter prediction model. For example, a parameter prediction model based on an autoregressive moving average model can be established according to the following formula:

[0037]

[0038] wherein X t is the parameter value of the power transformation device at time t; X t-1 is the parameter value of the power transformation device at time t-1, i.e. at the previous time of time t; and θ1 are parameters of the parameter prediction model; c is a constant term; ε t is an error term at the current time point. The parameter prediction model is constructed so that the future possible parameter value can be predicted according to the trend and pattern of the historical data. In addition to time series analysis, other features such as device operating time, environmental factors, etc. can also be considered. These features can have an impact on the change of device state, and therefore they can also be included in the parameter prediction model.

[0039] In an optional embodiment, the method further comprises: obtaining second parameter data; determining a target cluster to which the second parameter data belongs; determining a second parameter prediction model corresponding to the target cluster from the parameter prediction models corresponding to the N clusters; based on the second parameter data, using the second parameter prediction model to obtain a prediction result; and based on the prediction result, determining a maintenance strategy for the power transformation device.

[0040] Optionally, if the parameter value of the second parameter data at a future time period is to be predicted, the similarity between the second parameter data and the N clusters needs to be calculated to determine the target cluster to which the second parameter data belongs, and a second parameter prediction model corresponding to the target cluster is determined from the parameter prediction models corresponding to the N clusters. The second parameter data is input into the second parameter prediction model to obtain a prediction result. If the predicted value is within the normal range, no special maintenance measures need to be taken. If the predicted value deviates from the normal range, shutdown maintenance or device inspection may be considered to avoid potential failure. Through predictive maintenance, the device state can be managed more intelligently, potential failures can be handled in a timely manner, and the availability and maintenance efficiency of the device can be maximized.

[0041] Step S106, determining the average value and the standard deviation corresponding to each of the N clusters based on the parameter data respectively included in the N clusters.

[0042] Optionally, for each cluster, the average value and the standard deviation of the data in the cluster are calculated. These statistics will serve as a reference of the normal state, because in the normal case, the data in the same cluster should have similar parameter values. The average value can be calculated by the following formula:

[0043]

[0044] The standard deviation can be calculated by the following formula:

[0045]

[0046] wherein C k represents the set of parameter data belonging to cluster k; n k is the number of parameter data included in cluster k; x i is the i-th parameter data in cluster k.

[0047] Step S108, determining the abnormal parameter data based on the average value and the standard deviation corresponding to each of the N clusters.

[0048] In an optional embodiment, determining the abnormal parameter data based on the average value and the standard deviation corresponding to each of the N clusters comprises: determining a target value range based on the average value and the standard deviation corresponding to each of the N clusters; determining, as the abnormal parameter data, the parameter data respectively included in the N clusters that are not within the target value range.

[0049] Optionally, since in the case of sufficient parameter quantity, the parameter data in the same cluster should have similar parameter values, according to the average value and the standard deviation, the target value range of each cluster can be determined, which can be defined as the average value plus or minus several times the standard deviation, for example, the average value plus or minus 2 times the standard deviation. For the parameter data in each cluster, it is checked whether its value is within the target value range, and if not, it is determined as the abnormal parameter data.

[0050] In an optional embodiment, the method further comprises: determining the abnormal score corresponding to the parameter data respectively included in the N clusters based on the target value range by using an isolation forest algorithm, wherein the isolation forest algorithm is used to detect data with large differences from the normal state; judging whether there is first parameter data with an abnormal score greater than or equal to a preset score threshold in the parameter data respectively included in the N clusters; and determining the first parameter data as the abnormal parameter data.

[0051] Optionally, the Isolation Forest algorithm can be used to isolate abnormal data by randomly splitting the data. The algorithm assumes that abnormal data points are more likely to be isolated than normal data points. Isolation Forest randomly assigns data points to different branches of a tree, and identifies anomalies by building an isolation tree, the specific implementation process is as follows:

[0052] Step S1061, input the parameter data included in any one cluster into the algorithm as input.

[0053] Step S1062, construct an isolation tree according to the parameter data, wherein the isolation tree is a binary tree structure, each node is a feature, and the depth of the tree is pre-set.

[0054] Step S1063, randomly select a feature from the parameter data.

[0055] Step S1064, randomly select a split point from the target value range of the selected feature as the split point of the current node.

[0056] Step S1065, divide the parameter data into two parts, left subtree and right subtree, according to the selected feature and split point.

[0057] Step S1066, recursively construct an isolation tree for the left subtree and the right subtree, respectively, until the stop condition is met, wherein the stop condition for constructing the isolation tree can be that the depth of the tree reaches the pre-set maximum depth, or the number of samples in the subtree is less than a threshold.

[0058] Step S1067, repeat steps S1062-S1066 to construct multiple isolation trees until the pre-set execution times are reached.

[0059] Step S1068, for each parameter data, define an anomaly score using the average path length of the data in all isolation trees (i.e., the average value of the path length from the root node to the data point).

[0060] Since the deeper the data point is, the higher the anomaly score is, if there is first parameter data with an anomaly score greater than or equal to a preset score threshold, it means that the first parameter data is more likely to be isolated and is abnormal parameter data. There can be complex nonlinear relationships between the parameters of the power transformation equipment, and Isolation Forest can capture nonlinear anomaly patterns without prior knowledge, which is very useful for the special case of side power transformation equipment. By combining K-means clustering and Isolation Forest anomaly detection, anomaly detection can be performed from different angles, and the combination of the two can detect anomalies from multiple dimensions, quickly identify abnormal parameters, and make maintenance decisions at any time; improve the comprehensiveness of anomaly detection, better adapt to the distribution changes of parameters in different states; and improve the accuracy of abnormal parameter detection.

[0061] In an optional embodiment, after determining the parameter data not in the target value range in the parameter data respectively included in the N clusters as abnormal parameter data, the method further comprises: determining the fault mode corresponding to the abnormal parameter data based on the abnormal parameter data using a pre-trained decision tree fault classifier, wherein the pre-trained decision tree fault classifier is trained based on a historical data set of the power transformation equipment, the historical data set comprising a plurality of historical parameter data collected at a plurality of historical sampling times and a plurality of historical parameter data respectively corresponding to a fault label, and the plurality of historical sampling times correspond one-to-one to the plurality of historical parameter data.

[0062] Optionally, for the abnormal parameter data marked, feature engineering, statistical analysis and domain knowledge can be used to identify potential fault patterns. This may include feature changes related to specific fault types, such as sudden current rise, abnormal temperature fluctuations, etc. By analyzing the feature patterns of abnormal parameter data, a decision tree-based fault classifier can be established. Decision tree is a commonly used classification algorithm that can make classification decisions step by step based on the features of the data. In this step, a decision tree is used to establish a fault classifier to classify abnormal data into different fault types. The decision tree fault classifier can be established in the following way: assuming that there is a historical data set as a training set D, which contains N historical parameter data, and each historical parameter data contains M features (x1, x2, …, x M) and the corresponding label y (representing the fault type) by recursively selecting the optimal feature to split the parameter data and building a tree structure. At each node, a feature is selected to split so that the subsets after splitting are as pure as possible (as many data of the same fault type as possible). For example, if a decision tree is built, the decision rule for a node is "if the current is greater than threshold X and the temperature is less than threshold Y, then the fault type is A". Such a rule can be learned from the data to help classify abnormal parameter data into different fault types. By building a decision tree fault classifier, abnormal parameter data can be classified into different fault types according to the characteristics of the device parameter data. This provides more detailed fault diagnosis information for subsequent predictive maintenance, helping engineers to take appropriate maintenance measures more quickly.

[0063] Through the above steps S102 to S108, the purpose of accurately determining the abnormal parameter data based on the average value and the standard deviation corresponding to the target parameter data can be achieved, thereby realizing the technical effect of improving the accuracy of the power transformation equipment parameter processing method, and further solving the technical problem that the power transformation equipment parameter processing method in the related art does not consider all factors, resulting in low accuracy in determining abnormal parameter data.

[0064] Based on the above embodiments and optional embodiments, the present application proposes an optional implementation, which comprises:

[0065] Step S1, obtaining an initial data set of the power transformation equipment, wherein the initial data set includes initial parameter data collected at a plurality of sampling times.

[0066] Step S2, performing first data preprocessing on the parameter data included in the initial data set to obtain a first data set, wherein the first data preprocessing includes at least one of the following: data cleaning processing, outlier rejection processing.

[0067] Step S3, performing filtering processing on the parameter data included in the first data set to obtain a second data set.

[0068] Step S4, performing second data preprocessing on the parameter data included in the second data set to obtain a target data set, wherein the second data preprocessing includes at least one of the following: data standardization processing, data normalization processing, data conversion processing.

[0069] Step S5, performing clustering processing on the parameter data included in the target data set, and the specific implementation steps are as follows:

[0070] Step S51, determining the number of clusters as N.

[0071] Step S52, determining N target parameter data from the parameter data included in the target data set.

[0072] Step S53, based on the N target parameter data, determine N initial clustering center points.

[0073] Step S54, calculate the distance from other parameter data in the target data set to the N initial clustering center points.

[0074] Step S55, based on the distance from other parameter data to the N initial clustering center points, determine the cluster to which the other parameter data belongs.

[0075] Step S56, calculate the average value of the parameter data included in the N clusters respectively.

[0076] Step S57, based on the average value, update the N initial clustering center points, N new clustering center points.

[0077] Step S58, determine whether the preset termination condition is reached, if the preset termination condition is not reached, execute step S54, if the preset termination condition is reached, execute step S59.

[0078] Step S59, based on the N new clustering center points, obtain N clusters.

[0079] Step S6, obtain new parameter data of the power transformation equipment.

[0080] Step S7, calculate the similarity of the new parameter data to the N clusters respectively, obtain N similarities.

[0081] Step S8, if the N similarities are all greater than or equal to a preset threshold, determine the maximum similarity in the N similarities, and divide the new parameter data into the cluster corresponding to the maximum similarity.

[0082] Step S9, if there is any similarity in the N similarities less than the preset threshold, create a new cluster, and divide the new parameter data into the new cluster.

[0083] Step S10, based on the parameter data included in the N clusters respectively, establish N parameter prediction models respectively corresponding to the N clusters, wherein the parameter prediction model is used to predict the parameter value at a future sampling time.

[0084] Step S11, obtain second parameter data.

[0085] Step S12, determine the target cluster to which the second parameter data belongs.

[0086] Step S13, determine the second parameter prediction model corresponding to the target cluster from the N parameter prediction models respectively corresponding to the N clusters.

[0087] Step S14, based on the second parameter data, use the second parameter prediction model to obtain a prediction result.

[0088] In step S15, a maintenance strategy of the power transformation equipment is determined based on the prediction result.

[0089] In the embodiment, a power transformation equipment parameter processing apparatus is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. As used below, the term "module" "apparatus" can be a combination of software and / or hardware that implements a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, implementation of hardware, or a combination of software and hardware, is also possible and is contemplated.

[0090] According to the embodiment of the present application, a device embodiment for implementing the above-mentioned power transformation equipment parameter processing method is also provided, Figure 2 is a structural schematic diagram of a power transformation equipment parameter processing apparatus according to an embodiment of the present application, as Figure 2 shown, the power transformation equipment parameter processing apparatus comprises a first acquisition module 202, a first clustering module 204, a first determination module 206, and a second determination module 208, wherein:

[0091] The first acquisition module 202 is configured to acquire a target data set of the power transformation equipment, wherein the target data set is acquired based on initial parameter data collected at a plurality of sampling time points;

[0092] The first clustering module 204 is connected to the first acquisition module 202 and is configured to perform clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2;

[0093] The first determination module 206 is connected to the first clustering module 204 and is configured to determine mean values and standard deviations corresponding to the N clusters based on parameter data included in the N clusters, respectively;

[0094] The second determination module 208 is connected to the first determination module 206 and is configured to determine abnormal parameter data based on the mean values and the standard deviations corresponding to the N clusters, respectively.

[0095] By setting the first obtaining module 202, the target data set of the power transformation equipment is obtained, wherein the target data set is obtained based on the initial parameter data collected at the plurality of sampling moments; the first clustering module 204 is used for clustering the parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; the first determining module 206 is used for determining the average value and the standard deviation corresponding to the N clusters based on the parameter data included in the N clusters; the second determining module 208 is used for determining the abnormal parameter data based on the average value and the standard deviation corresponding to the N clusters, thereby achieving the purpose of accurately determining the abnormal parameter data based on the average value and the standard deviation corresponding to the target parameter data, thereby realizing the technical effect of improving the accuracy of the power transformation equipment parameter processing method, and further solving the technical problem that the power transformation equipment parameter processing method in the related art does not comprehensively consider the factors, resulting in low accuracy of determining the abnormal parameter data.

[0096] In an optional embodiment, the first obtaining module described above includes: a first obtaining submodule, configured to obtain an initial data set of the power transformation equipment, wherein the initial data set includes initial parameter data collected at a plurality of sampling moments; a first processing submodule, configured to perform first data preprocessing on the parameter data included in the initial data set to obtain a first data set, wherein the first data preprocessing includes at least one of the following: missing value processing, abnormal value processing; a second processing submodule, configured to perform screening processing on the parameter data included in the first data set to obtain a second data set, wherein the screening processing is used to determine the parameter data related to the state and performance of the power transformation equipment; and a third processing submodule, configured to perform second data preprocessing on the parameter data included in the second data set to obtain a target data set, wherein the second data preprocessing includes at least one of the following: data standardization processing, data normalization processing, and data conversion processing.

[0097] In an optional embodiment, the first clustering module comprises: a first determining submodule, configured to determine the number of clusters as N; a second determining submodule, configured to determine N pieces of target parameter data from the parameter data included in the target data set; a third determining submodule, configured to determine N initial clustering center points based on the N pieces of target parameter data, and repeatedly perform the following operations until a preset termination condition is reached: a first calculating submodule, configured to calculate distances from other parameter data in the target data set, except the N pieces of target parameter data, to the N initial clustering center points; a fourth determining submodule, configured to determine clusters to which the other parameter data belong based on the distances from the other parameter data to the N initial clustering center points; a second calculating submodule, configured to calculate average values of the parameter data included in the N clusters, respectively; a first updating submodule, configured to update the N initial clustering center points based on the average values to obtain N new clustering center points; and a fifth determining submodule, configured to obtain the N clusters based on the N new clustering center points in a case where the preset termination condition is met.

[0098] In an optional embodiment, the second determining module comprises: a sixth determining submodule, configured to determine a target value range based on the average values and the standard deviations corresponding to the N clusters, respectively; and a seventh determining submodule, configured to determine, as abnormal parameter data, parameter data that is not within the target value range from among the parameter data included in the N clusters, respectively.

[0099] In an optional embodiment, the apparatus further comprises: an eighth determining submodule, configured to determine, based on the target value range, abnormal scores corresponding to the parameter data included in the N clusters, respectively, by using an isolation forest algorithm, wherein the isolation forest algorithm is used to detect data that is greatly different from a normal state; a first judging submodule, configured to determine whether there is first parameter data whose abnormal score is greater than or equal to a preset score threshold from among the parameter data included in the N clusters, respectively; and a ninth determining submodule, configured to determine the first parameter data as abnormal parameter data.

[0100] In an optional embodiment, the apparatus further comprises: a tenth determining submodule, configured to determine a fault mode corresponding to the abnormal parameter data by using a pre-trained decision tree fault classifier based on the abnormal parameter data, wherein the pre-trained decision tree fault classifier is trained based on a historical data set of the power transformation equipment, the historical data set comprising a plurality of pieces of historical parameter data collected at a plurality of historical sampling moments and fault labels corresponding to the plurality of pieces of historical parameter data, respectively, the plurality of historical sampling moments corresponding to the plurality of pieces of historical parameter data in one-to-one correspondence.

[0101] In an optional embodiment, the apparatus further comprises a second obtaining module, configured to obtain new parameter data of the power transformation equipment; a third calculating module, configured to calculate similarities between the new parameter data and the N clusters respectively, to obtain N similarities; an eleventh determining module, configured to determine a maximum similarity in the N similarities if all the N similarities are greater than or equal to a preset threshold, and divide the new parameter data into a cluster corresponding to the maximum similarity; a twelfth determining module, configured to create a new cluster and divide the new parameter data into the new cluster if any of the N similarities is less than the preset threshold.

[0102] In an optional embodiment, the apparatus further comprises a first establishing module, configured to establish parameter prediction models corresponding to the N clusters respectively based on parameter data included in the N clusters respectively, wherein the parameter prediction models are used to predict parameter values at a future sampling time.

[0103] In an optional embodiment, the apparatus further comprises a third obtaining module, configured to obtain second parameter data; a thirteenth determining module, configured to determine a target cluster to which the second parameter data belongs; a fourteenth determining module, configured to determine a second parameter prediction model corresponding to the target cluster from parameter prediction models corresponding to the N clusters respectively; a fifteenth determining module, configured to obtain a prediction result by using the second parameter prediction model based on the second parameter data; and a sixteenth determining module, configured to determine a maintenance strategy of the power transformation equipment based on the prediction result.

[0104] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, the modules can be located in the same processor or in different processors in any combination.

[0105] It should be noted that the first obtaining module 202, the first clustering module 204, the first determining module 206, and the second determining module 208 correspond to steps S102-S108 in the embodiments, and have the same instances and application scenarios as the corresponding steps, but are not limited to the disclosed contents in the above embodiments. It should be noted that the modules as part of the apparatus can run in a computer terminal.

[0106] It should be noted that the optional or preferred embodiments of the present embodiment can refer to the related descriptions in the embodiments, which will not be repeated here.

[0107] The transformer equipment parameter processing apparatus can further include a processor and a memory, and the first acquisition module 202, the first clustering module 204, the first determination module 206, the second determination module 208, and the like are stored in the memory as program modules, and the processor executes the program modules stored in the memory to realize the corresponding functions.

[0108] The processor includes a core, and the core calls the corresponding program modules from the memory, and the core can be one or more. The memory can include a non-permanent memory in a computer readable medium, a random access memory (RAM), and / or a non-volatile memory such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one memory chip.

[0109] According to the embodiments of the present application, an embodiment of a non-volatile storage medium is further provided. Optionally, in the embodiment, the non-volatile storage medium includes a stored program, and when the program runs, the non-volatile storage medium controls a device where the non-volatile storage medium is located to execute any of the transformer equipment parameter processing methods.

[0110] Optionally, in the embodiment, the non-volatile storage medium can be located in any of computer terminals in a computer terminal group in a computer network or in any of mobile terminals in a mobile terminal group, and the non-volatile storage medium includes a stored program.

[0111] Optionally, when the program runs, the non-volatile storage medium controls a device where the non-volatile storage medium is located to execute the following functions: acquiring a target data set of transformer equipment, where the target data set is acquired based on initial parameter data collected at a plurality of sampling moments; performing clustering processing on parameter data included in the target data set to obtain N clusters, where N is an integer greater than or equal to 2; determining mean values and standard deviations corresponding to the N clusters based on parameter data included in the N clusters, respectively; and determining abnormal parameter data based on the mean values and the standard deviations corresponding to the N clusters, respectively.

[0112] According to the embodiments of the present application, an embodiment of a processor is further provided. Optionally, in the embodiment, the processor is used to run a program, and when the program runs, the processor executes any of the transformer equipment parameter processing methods.

[0113] According to the embodiments of the present application, an embodiment of a computer program product is further provided, which is adapted to execute a program initialized with steps of any of the transformer equipment parameter processing methods when executed on a data processing device.

[0114] Optionally, the computer program product described above, when executed on the data processing device, is adapted to execute a program that is initialized with the following method steps: obtaining a target data set of the power transformation device, wherein the target data set is obtained based on initial parameter data collected at a plurality of sampling time points respectively; performing clustering processing on the parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; determining the mean value and the standard deviation corresponding to each of the N clusters based on the parameter data included in each of the N clusters; and determining abnormal parameter data based on the mean value and the standard deviation corresponding to each of the N clusters.

[0115] Optionally, the computer program product described above is further adapted to execute a program that is initialized with the following method steps: obtaining an initial data set of the power transformation device, wherein the initial data set includes initial parameter data collected at a plurality of sampling time points respectively; performing first data preprocessing on the parameter data included in the initial data set to obtain a first data set, wherein the first data preprocessing includes at least one of the following: missing value processing, outlier processing; performing screening processing on the parameter data included in the first data set to obtain a second data set, wherein the screening processing is used to determine parameter data related to the state and performance of the power transformation device; and performing second data preprocessing on the parameter data included in the second data set to obtain a target data set, wherein the second data preprocessing includes at least one of the following: data standardization processing, data normalization processing, data conversion processing.

[0116] Optionally, the computer program product described above is further adapted to execute a program that is initialized with the following method steps: determining the number of clusters to be N; determining N target parameter data from the parameter data included in the target data set; determining N initial clustering center points based on the N target parameter data, and repeatedly performing the following operations until a preset termination condition is met: calculating the distance from other parameter data in the target data set, other than the N target parameter data, to the N initial clustering center points; determining the cluster to which the other parameter data belongs based on the distance from the other parameter data to the N initial clustering center points; calculating the mean value of the parameter data included in each of the N clusters; updating the N initial clustering center points based on the mean value to obtain N new clustering center points; and based on the N new clustering center points, obtaining the N clusters if the preset termination condition is met.

[0117] Optionally, the computer program product described above is further adapted to execute a program that is initialized with the following method steps: determining a target value range based on the mean value and the standard deviation corresponding to each of the N clusters; and determining parameter data that is not within the target value range from the parameter data included in each of the N clusters as abnormal parameter data.

[0118] Optionally, the computer program product is further adapted to execute a program that is initialized with the following method steps: determining, based on the target value range, N clusters respectively including parameter data respectively corresponding to an abnormal score by using an isolation forest algorithm, wherein the isolation forest algorithm is used to detect data that is greatly different from a normal state; determining whether there is first parameter data with an abnormal score greater than or equal to a preset score threshold in the parameter data respectively included in the N clusters; and determining the first parameter data as abnormal parameter data.

[0119] Optionally, the computer program product is further adapted to execute a program that is initialized with the following method steps: determining, based on the abnormal parameter data, a fault mode corresponding to the abnormal parameter data by using a pre-trained decision tree fault classifier, wherein the pre-trained decision tree fault classifier is trained based on a historical data set of the power transformation equipment, the historical data set including a plurality of pieces of historical parameter data collected at a plurality of historical sampling moments and a plurality of historical parameter data respectively corresponding to a fault label, and the plurality of historical sampling moments correspond one-to-one to the plurality of pieces of historical parameter data.

[0120] Optionally, the computer program product is further adapted to execute a program that is initialized with the following method steps: obtaining new parameter data of the power transformation equipment; calculating a similarity between the new parameter data and each of the N clusters to obtain N similarities; if all the N similarities are greater than or equal to a preset threshold, determining a maximum value of the N similarities, and dividing the new parameter data into a cluster corresponding to the maximum value; and if any of the N similarities is less than the preset threshold, creating a new cluster and dividing the new parameter data into the new cluster.

[0121] Optionally, the computer program product is further adapted to execute a program that is initialized with the following method steps: based on the parameter data respectively included in the N clusters, establishing a parameter prediction model corresponding to each of the N clusters, wherein the parameter prediction model is used to predict a parameter value at a future sampling moment.

[0122] Optionally, the computer program product is further adapted to execute a program that is initialized with the following method steps: obtaining second parameter data; determining a target cluster to which the second parameter data belongs; determining a second parameter prediction model corresponding to the target cluster from the parameter prediction models corresponding to the N clusters; obtaining a prediction result by using the second parameter prediction model based on the second parameter data; and determining a maintenance strategy of the power transformation equipment based on the prediction result.

[0123] As Figure 3As shown, the embodiment of the present application provides an electronic device 10, which comprises a processor, a memory, and a program stored on the memory and executable on the processor, and the processor implements the following steps when executing the program: obtaining a target data set of a power transformation device, wherein the target data set is obtained based on initial parameter data collected at a plurality of sampling moments; performing clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; determining mean values and standard deviations corresponding to the N clusters based on parameter data included in the N clusters respectively; and determining abnormal parameter data based on the mean values and the standard deviations corresponding to the N clusters respectively.

[0124] In the above-described embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0125] In several embodiments provided in the present application, it should be understood that the disclosed technical contents can be implemented by other ways. Among them, the above-described device embodiments are only schematic, for example, the division of the above-described modules can be a logical function division, and actual implementation can have another division way, for example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interfaces, indirect coupling or communication connection between modules or modules, which can be electrical or other forms.

[0126] The above-described modules explained as separate components can be or can not be physically separated, and the components displayed as modules can be or can not be physical modules, that is, can be located in one place, or can be distributed to a plurality of modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

[0127] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically, or two or more modules can be integrated in one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of software functional module.

[0128] If the above-mentioned integrated modules are realized in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable nonvolatile storage medium. Based on this understanding, the technical solutions of the present application, essentially or the part that contributes to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the embodiments of the present application. The aforementioned non-volatile storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0129] The above is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method of processing parameters of a power transformation device, characterized by, The method comprises the following steps: acquiring a target data set of a power transformation device, wherein the target data set is acquired based on initial parameter data collected at a plurality of sampling time points, and the initial parameter data includes current, voltage, temperature, and humidity; performing clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; determining average values and standard deviations corresponding to the N clusters based on parameter data included in the N clusters, respectively; determining abnormal parameter data based on the average values and the standard deviations corresponding to the N clusters, including: determining a target value range based on the average values and the standard deviations corresponding to the N clusters; determining abnormal scores corresponding to parameter data included in the N clusters based on the target value range by using an isolation forest algorithm, wherein the isolation forest algorithm is used to detect data that is greatly different from a normal state; determining whether there is first parameter data with an abnormal score greater than or equal to a preset score threshold in the parameter data included in the N clusters; and determining the first parameter data as the abnormal parameter data; determining a fault mode corresponding to the abnormal parameter data by using a pre-trained decision tree fault classifier based on the abnormal parameter data, wherein the pre-trained decision tree fault classifier is trained based on a historical data set of the power transformation device, a plurality of historical parameter data collected at a plurality of historical sampling time points included in the historical data set, and fault labels corresponding to the plurality of historical parameter data, respectively, wherein the plurality of historical sampling time points correspond to the plurality of historical parameter data one by one; determining abnormal scores corresponding to parameter data included in the N clusters based on the target value range by using an isolation forest algorithm, including: taking parameter data included in any one of the N clusters as input, and constructing an isolated tree according to the parameter data included in the any one of the N clusters, wherein the isolated tree is a binary tree structure, each node is a feature, and the depth of the tree is pre-set; randomly selecting a feature from the parameter data included in the any one of the N clusters; randomly selecting a split point from the target value range of the selected feature as a split point of a current node; dividing the parameter data included in the any one of the N clusters into two parts according to the selected feature and the split point to obtain a left sub-tree and a right sub-tree; recursively constructing isolated trees for the left sub-tree and the right sub-tree, respectively, until a stop condition is met, wherein the stop condition for stopping the construction of the isolated tree is that the depth of the tree reaches a pre-set maximum depth, or the number of samples in the sub-tree is less than a pre-set threshold; repeatedly constructing a plurality of isolated trees until a pre-set execution number is reached; and determining the abnormal scores corresponding to each parameter data in the N clusters based on the average path lengths of the parameter data in all isolated trees; The method further comprises: based on the parameter data respectively included in the N clusters, establishing a parameter prediction model corresponding to each of the N clusters, wherein the parameter prediction model is used to predict a parameter value at a future sampling time; obtaining second parameter data; determining a target cluster to which the second parameter data belongs; determining a second parameter prediction model corresponding to the target cluster from the parameter prediction models corresponding to the N clusters; based on the second parameter data, obtaining a prediction result by using the second parameter prediction model; and based on the prediction result, determining a maintenance strategy of the power transformation equipment.

2. The method of claim 1, wherein, The target data set of the power transformation equipment is obtained by: obtaining an initial data set of the power transformation equipment, wherein the initial data set includes initial parameter data collected at the plurality of sampling time points; performing first data preprocessing on the parameter data included in the initial data set to obtain a first data set, wherein the first data preprocessing includes at least one of missing value processing and abnormal value processing; performing screening processing on the parameter data included in the first data set to obtain a second data set, wherein the screening processing is used to determine parameter data related to the state and performance of the power transformation equipment; performing second data preprocessing on the parameter data included in the second data set to obtain the target data set, wherein the second data preprocessing includes at least one of data standardization processing, data normalization processing, and data conversion processing.

3. The method of claim 1, wherein, The parameter data included in the target data set is clustered to obtain N clusters, including: determining the number of clusters to be N; determining N target parameter data from the parameter data included in the target data set; based on the N target parameter data, determining N initial clustering center points, and repeatedly performing the following operations until a preset termination condition is met: calculating the distances from other parameter data in the target data set, except the N target parameter data, to the N initial clustering center points; based on the distances from the other parameter data to the N initial clustering center points, determining the clusters to which the other parameter data belong; calculating the average values of the parameter data included in the N clusters; based on the average values, updating the N initial clustering center points to obtain N new clustering center points; if the preset termination condition is met, obtaining the N clusters based on the N new clustering center points.

4. The method of claim 3, wherein, The method further comprises: obtaining new parameter data of the power transformation equipment; calculating the similarities between the new parameter data and the N clusters to obtain N similarities; if all the N similarities are greater than or equal to a preset threshold, determining a maximum similarity in the N similarities, and dividing the new parameter data into a cluster corresponding to the maximum similarity; if any of the N similarities is less than the preset threshold, creating a new cluster and dividing the new parameter data into the new cluster.

5. A power transformation equipment parameter processing apparatus characterized by comprising: including: The first obtaining module is configured to obtain a target data set of a power transformation device, wherein the target data set is obtained based on initial parameter data collected at a plurality of sampling time points, and the initial parameter data includes current, voltage, temperature, and humidity; The first clustering module is configured to perform clustering processing on parameter data included in the target data set to obtain N clusters, wherein N is an integer greater than or equal to 2; The first determining module is configured to determine, based on parameter data included in the N clusters respectively, average values and standard deviations corresponding to the N clusters respectively; The second determining module is configured to determine, based on the average values and the standard deviations corresponding to the N clusters respectively, abnormal parameter data; The second determining module is further configured to determine, based on the average values and the standard deviations corresponding to the N clusters respectively, a target value range; determine, based on the target value range, abnormal scores corresponding to parameter data included in the N clusters respectively by using an isolation forest algorithm, wherein the isolation forest algorithm is used to detect data that is greatly different from a normal state; determine, among the parameter data included in the N clusters respectively, whether there is first parameter data with an abnormal score greater than or equal to a preset score threshold; and determine the first parameter data as the abnormal parameter data; The device is further configured to determine, based on the abnormal parameter data, a fault mode corresponding to the abnormal parameter data by using a pre-trained decision tree fault classifier, wherein the pre-trained decision tree fault classifier is trained based on a historical data set of the power transformation device, a plurality of historical parameter data collected at a plurality of historical sampling time points included in the historical data set, and fault labels corresponding to the plurality of historical parameter data respectively, and the plurality of historical sampling time points correspond one-to-one to the plurality of historical parameter data; The device is further configured to take, as input, parameter data included in any one of the N clusters, and construct an isolated tree from the parameter data included in the any one of the N clusters, wherein the isolated tree is a binary tree structure, each node is a feature, and the depth of the tree is pre-set; randomly select a feature from the parameter data included in the any one of the N clusters; randomly select, as a split point of a current node, a split point in a target value range of the selected feature; divide the parameter data included in the any one of the N clusters into two parts according to the selected feature and the split point, to obtain a left sub-tree and a right sub-tree; recursively construct isolated trees for the left sub-tree and the right sub-tree respectively until a stop condition is met, wherein the stop condition for stopping the construction of the isolated trees is that the depth of the tree reaches a preset maximum depth, or the number of samples in the sub-tree is less than a preset threshold; repeatedly perform the construction of a plurality of isolated trees until a preset execution number is reached; and determine, as a corresponding abnormal score, an average path length of each parameter data in the N clusters in all isolated trees. The apparatus is further configured to: based on the parameter data respectively included in the N clusters, establish parameter prediction models respectively corresponding to the N clusters, wherein the parameter prediction models are used to predict parameter values at future sampling times; obtain second parameter data; determine a target cluster to which the second parameter data belongs; determine a second parameter prediction model corresponding to the target cluster from the parameter prediction models respectively corresponding to the N clusters; based on the second parameter data, obtain a prediction result by using the second parameter prediction model; and based on the prediction result, determine a maintenance strategy for the power transformation equipment.

6. An electronic device, comprising: The apparatus includes one or more processors and a memory configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the power transformation equipment parameter processing method of any one of claims 1 to 4.

Images