Distribution network line loss abnormity diagnosis method and device

Abstract

The embodiment of the invention discloses a power distribution network line loss abnormity diagnosis method and device, and relates to the technical field of data processing. The method comprises the following steps: acquiring monitoring state characteristics according to monitoring statistical data of a target line and corresponding state information; sorting at least one sub-state feature in the monitoring state feature according to data acquisition time; respectively inputting each sub-state feature in the monitoring state features into a first feature extraction network and a second feature extraction network of a first target neural network according to a time positive sequence and a time reverse sequence, and respectively carrying out positive sequence calculation and reverse sequence calculation to obtain a first data feature and a second data feature; acquiring a state change feature of the target line based on the first data feature and the second data feature; inputting the state change feature into a second target neural network to extract a depth feature, and obtaining at least one depth state feature; and based on the at least one depth state feature, carrying out line loss abnormity diagnosis and prediction on the target line.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and device for diagnosing abnormal line loss in a distribution network. Background Art

[0002] Since the factors that affect line loss are diverse and their relationships are complex, and the existing technical means for diagnosing line loss anomalies are limited, most of them rely on manual troubleshooting based on the experience of managers. Therefore, when line loss anomalies occur, it is difficult to accurately and effectively diagnose the true cause of the line loss anomaly, which seriously affects the work efficiency of distribution network operation and maintenance.

[0003] Based on this, it is necessary to study a distribution network line loss abnormality diagnosis method and device to improve the accuracy and effectiveness of line loss abnormality diagnosis, thereby improving the work efficiency of distribution network operation and maintenance. Summary of the invention

[0004] To solve the above problem, one aspect of an embodiment of this specification provides a method for diagnosing abnormal line loss in a distribution network, the method comprising:

[0005] According to monitoring statistical data of a target line and status information corresponding to the monitoring statistical data, a monitoring status feature of the target line is obtained, wherein the status information corresponding to the monitoring statistical data is used to indicate an abnormal line loss state of the target line, the monitoring statistical data is time series data generated based on real-time monitoring data of the target line, the time series data includes a data acquisition time when the target line acquires real-time monitoring data each time, and real-time monitoring data corresponding to each data acquisition time, the real-time monitoring data includes power supply data, power sales data, meteorological data, and load data and electrical measurement data of each load point in the target line;

[0006] Sorting at least one sub-state feature in the monitoring state feature according to the order of the data acquisition time;

[0007] Input each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network according to the time positive sequence corresponding to the data acquisition time, and perform positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature, wherein the first data feature is used to characterize the positive evolution law of the monitoring state feature;

[0008] Inputting each sub-state feature in the monitoring state feature into a second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and performing reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature, wherein the second data feature is used to characterize the reverse evolution law of the monitoring state feature;

[0009] Based on the first data feature and the second data feature, acquiring a state change feature of the target line;

[0010] Inputting the state change feature into a second target neural network, and extracting the deep features corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line;

[0011] Based on the at least one deep state feature of the target line, line loss abnormality diagnosis and prediction are performed on the target line.

[0012] In some embodiments, acquiring the monitoring status feature of the target line according to the monitoring statistical data of the target line and the status information corresponding to the monitoring statistical data includes:

[0013] Inputting at least one state information into a multilayer perceptron;

[0014] Filtering out a target line loss abnormal state from at least one line loss abnormal state corresponding to the at least one state information by the multilayer perceptron, and performing weight allocation on the target line loss abnormal state to obtain a feature corresponding to the at least one line loss abnormal state;

[0015] Inputting the real-time monitoring data into a time series analysis model, and extracting the time series features corresponding to each item of data in the real-time monitoring data through the time series analysis model to obtain a monitoring data feature set;

[0016] Inputting the monitoring data feature set into a cross processing network, and performing feature cross processing on each time series feature in the monitoring data feature set through the cross processing network to obtain features corresponding to the real-time monitoring data, wherein the cross processing network includes a deep cross neural network;

[0017] The features corresponding to the at least one abnormal line loss state and the features corresponding to the real-time monitoring data are input into a deep feature extraction network, and the composite features of the real-time monitoring data and the at least one abnormal line loss state are extracted through the deep feature extraction network to obtain the monitoring state features of the target line.

[0018] In some embodiments, inputting each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network in the time positive sequence corresponding to the data acquisition time, and performing positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature, includes: performing positive sequence moving average calculation on the monitoring state feature with a first window size to obtain the first data feature;

[0019] The method comprises: performing reverse moving average calculation on the monitoring state feature with a second window size to obtain the second data feature, wherein the second window size is different from the first window size.

[0020] In some embodiments, acquiring the state change feature of the target line based on the first data feature and the second data feature includes:

[0021] Fusing the first data feature and the second data feature to obtain a third data feature;

[0022] Performing weight distribution on the third data feature to obtain a fourth data feature, wherein the fourth data feature is used to characterize the comprehensive evolution law of the monitoring state feature;

[0023] Based on the third data feature and the fourth data feature, a state change feature of the target line is obtained.

[0024] In some embodiments, performing weight distribution on the third data feature to obtain the fourth data feature includes:

[0025] Processing the third data feature through a first attention model to obtain at least one first weight parameter, wherein each of the first weight parameters is used to represent an importance measure of an item of real-time monitoring data and a corresponding abnormal line loss state;

[0026] performing normalization processing on the first weight parameter to obtain at least one second weight parameter;

[0027] The third data feature is weighted according to the second weight parameter to obtain the fourth data feature.

[0028] In some embodiments, the performing line loss abnormality diagnosis and prediction on the target line based on the at least one deep state feature of the target line includes:

[0029] Performing weight parameter learning based on the second attention model and the at least one deep state feature to obtain at least one third weight parameter, wherein each third weight parameter is used to represent an importance measure of a corresponding deep state feature;

[0030] Performing weight allocation on the at least one depth state feature according to the third weight parameter to obtain a target depth state feature;

[0031] Based on the target depth status characteristics, line loss abnormality diagnosis and prediction are performed on the target line.

[0032] In some embodiments, each layer of the second target neural network outputs one of the deep state features, and the method further comprises:

[0033] The deep state features output by the first processing layer are generated based on the network layer features of the first processing layer in the second target neural network and the deep state features generated by the second processing layer, wherein the network layer features of the first processing layer are used to represent the state of the state change features in the first processing layer, and the second processing layer is the previous processing layer before the first processing layer in the second target neural network.

[0034] In some embodiments, the network layer characteristic of the first processing layer is determined based on the state change characteristic and the network layer characteristic of the second processing layer.

[0035] In some embodiments, the line loss abnormality diagnosis and prediction of the target line based on the target depth state feature includes: inputting the target depth state feature into a trained line loss abnormality recognition model to obtain a prediction result corresponding to the target depth state feature, wherein the prediction result includes a predicted line loss abnormality type and its corresponding predicted abnormality cause;

[0036] The line loss anomaly recognition model is trained based on the following method:

[0037] Acquire a sample depth state feature obtained based on sample monitoring statistical data, and a sample label corresponding to the sample depth state feature, wherein the sample label is used to characterize the line loss anomaly type and its anomaly cause corresponding to the sample monitoring statistical data;

[0038] The sample depth state feature is used as input, and the sample label corresponding to the sample depth state feature is used as output to train the initial line loss anomaly recognition model until the preset conditions are met, thereby obtaining the trained line loss anomaly recognition model.

[0039] Another aspect of the embodiments of this specification further provides a distribution network line loss abnormality diagnosis device, the device comprising:

[0040] A first acquisition module is used to acquire monitoring status characteristics of the target line according to monitoring statistical data of the target line and status information corresponding to the monitoring statistical data, wherein the status information corresponding to the monitoring statistical data is used to indicate abnormal line loss status of the target line, the monitoring statistical data is time series data generated based on real-time monitoring data of the target line, the time series data includes data acquisition time when the target line acquires real-time monitoring data each time, and real-time monitoring data corresponding to each data acquisition time, the real-time monitoring data includes power supply data, power sales data, meteorological data, and load data and electrical measurement data of each load point in the target line;

[0041] A sorting module, used to sort at least one sub-state feature in the monitoring state feature according to the order of the data acquisition time;

[0042] A first data feature calculation module is used to input each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network according to the time positive sequence corresponding to the data acquisition time, and perform positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature, wherein the first data feature is used to characterize the positive evolution law of the monitoring state feature;

[0043] A second data feature calculation module is used to input each sub-state feature in the monitoring state feature into a second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and perform reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature, wherein the second data feature is used to characterize the reverse evolution law of the monitoring state feature;

[0044] A second acquisition module, configured to acquire a state change feature of the target line based on the first data feature and the second data feature;

[0045] A deep state feature extraction module, used to input the state change feature into a second target neural network, and extract the deep features corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line;

[0046] A line loss anomaly identification module is used to diagnose and predict line loss anomaly of the target line based on the at least one deep state feature of the target line.

[0047] The beneficial effects that the distribution network line loss abnormality diagnosis method and device provided in the embodiments of this specification may bring include at least:

[0048] (1) The monitoring status features are obtained by processing the monitoring statistical data and the status information corresponding to the monitoring statistical data. Compared with the features obtained only based on the monitoring statistical data, the abnormal line loss state reflected by the target line during the detection process can be more accurately reflected, thereby improving the accuracy of the abnormal line loss diagnosis results in the subsequent process;

[0049] (2) The first target neural network is used to perform forward and reverse calculations on the sub-state features of each time series in the monitoring state feature to obtain the first data feature and the second data feature, which can respectively reflect the data change characteristics of the monitoring state feature in different evolution directions and different time scales, thereby improving the accuracy and reliability of the data features to a certain extent;

[0050] (3) The deep state features corresponding to the target line are extracted through the second target neural network, and line loss anomaly diagnosis and prediction of the target line are performed based on the deep state features. Potential line loss problems can be discovered in a timely manner through monitoring statistical data, thereby realizing line loss anomaly diagnosis and prediction at the same time, providing strong technical support for the stable operation of the power grid.

[0051] Additional features will be described in part in the following description. For those skilled in the art, it will become apparent by reviewing the following and accompanying drawings, or it may be understood by the production or operation of the examples. The features of this specification may be implemented and obtained by practicing or using various aspects of the methods, tools, and combinations described in the following detailed examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0053] Figure 1 is an exemplary module diagram of a distribution network line loss abnormality diagnosis device according to some embodiments of this specification;

[0054] Figure 2 is an exemplary flow chart of a method for diagnosing abnormal line loss in a distribution network according to some embodiments of this specification;

[0055] Figure 3 is a flowchart of exemplary sub-steps of a method for diagnosing abnormal line loss in a distribution network according to some embodiments of this specification;

[0056] Figure 4 It is an exemplary structural diagram of a computer processing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0058] It should be understood that the "system", "device", "unit" and / or "module" used in this specification is a method for distinguishing different components, elements, parts, parts or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0059] As shown in this specification and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0060] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0061] The distribution network refers to the power grid that receives electric energy from the transmission network or regional power plants and distributes it locally or step by step according to voltage to various users through distribution facilities. The distribution network is usually composed of overhead lines, towers, cables, distribution transformers, switchgear, reactive power compensation capacitors and other distribution equipment and ancillary facilities. Its main function is to distribute electric energy.

[0062] Distribution network line loss refers to the energy loss caused by various reasons in the process of transmitting electricity from power plants to users. The main causes include technical line loss, management line loss and other line losses. Among them, technical line loss is the energy loss caused by the physical characteristics of power equipment and lines, including resistance loss, transformer loss, capacitor loss, insulation loss, etc.; management line loss is the energy loss caused by poor management or human factors, including meter reading errors, metering errors, electricity theft, data processing errors, etc.; other line losses refer to energy losses caused by unforeseen or special reasons, such as fault losses, energy losses caused by environmental factors, etc.

[0063] Technical line loss can be understood as the energy loss caused by the heat emitted during the operation of the distribution network. This is the active power consumed by the conductivity and resistance in the power grid. Once the technical line loss in the distribution network is too high, the speed of line aging and damage will accelerate, which will have a greater impact on the safe operation of the power grid. Therefore, in actual work, it is necessary to pay full attention to the problem of distribution network line loss, and take effective measures to solve the abnormal causes of distribution network line loss, so as to comprehensively improve the level of line loss management.

[0064] However, since the factors that affect line loss are diverse and their relationships are complex, and the existing technical means for diagnosing line loss anomalies are limited, most of them rely on the experience of managers for manual investigation. Therefore, when line loss anomalies occur, it is difficult to accurately and effectively diagnose the real cause of the line loss anomaly, which not only makes the operation and maintenance tasks heavy, but also inefficient.

[0065] Based on the above technical problems, this specification provides a distribution network line loss abnormality diagnosis method and device based on artificial intelligence technology to improve the accuracy and effectiveness of line loss abnormality diagnosis. The distribution network line loss abnormality diagnosis method and device provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0066] Figure 1 is an exemplary module diagram of a distribution network line loss abnormality diagnosis device provided in an embodiment of the present application. Figure 1 The distribution network line loss abnormality diagnosis device 100 may include a first acquisition module 110, a sorting module 120, a first data feature calculation module 130, a second data feature calculation module 140, a second acquisition module 150, a deep state feature extraction module 160 and a line loss abnormality identification module 170.

[0067] The first acquisition module 110 can be used to acquire the monitoring status characteristics of the target line according to the monitoring statistical data of the target line and the status information corresponding to the monitoring statistical data, wherein the status information corresponding to the monitoring statistical data is used to indicate the abnormal line loss state of the target line, and the monitoring statistical data is time series data generated based on the real-time monitoring data of the target line, and the time series data includes the data acquisition time each time the target line acquires real-time monitoring data, and the real-time monitoring data corresponding to each data acquisition time, and the real-time monitoring data includes power supply data, power sales data, meteorological data, and load data and electrical measurement data of each load point in the target line.

[0068] The sorting module 120 may be used to sort at least one sub-state feature in the monitoring state feature according to the order of the data acquisition time.

[0069] The first data feature calculation module 130 can be used to input each sub-state feature in the monitoring state feature into the first feature extraction network of the first target neural network in the positive time sequence corresponding to the data acquisition time, and perform positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain the first data feature, wherein the first data feature is used to characterize the positive evolution law of the monitoring state feature.

[0070] The second data feature calculation module 140 can be used to input each sub-state feature in the monitoring state feature into the second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and perform reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature, wherein the second data feature is used to characterize the reverse evolution law of the monitoring state feature.

[0071] The second acquisition module 150 may be configured to acquire a state change feature of the target line based on the first data feature and the second data feature.

[0072] The deep state feature extraction module 160 can be used to input the state change feature into the second target neural network, and extract the deep features corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line.

[0073] The line loss anomaly identification module 170 may be configured to diagnose and predict line loss anomalies on the target line based on the at least one deep state feature of the target line.

[0074] For more details about the above modules, please refer to other places in this manual (for example Figure 2~Figure 3 part and its related description), which will not be discussed in detail here.

[0075] It should be understood that Figure 1The distribution network line loss abnormality diagnosis device 100 and its modules shown can be implemented in various ways. For example, in some embodiments, the device and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art can understand that the above methods and systems can be implemented using computer executable instructions and / or included in processor control codes, such as such codes provided on carrier media such as disks, CDs or DVD-ROMs, programmable memories such as read-only memories (firmware), or data carriers such as optical or electronic signal carriers. The device and its modules of this specification can not only be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (for example, firmware).

[0076] It should be noted that the above description of the distribution network line loss abnormality diagnosis device 100 is provided for illustrative purposes only and is not intended to limit the scope of this specification. It is understandable that for those skilled in the art, according to the description of this specification, without departing from this principle, the various modules can be arbitrarily combined, or a subsystem can be formed to connect with other modules. For example, Figure 1 The first acquisition module 110, the sorting module 120, the first data feature calculation module 130, the second data feature calculation module 140, the second acquisition module 150, the deep state feature extraction module 160 and the line loss abnormality identification module 170 described in the above can be different modules in a system, or one module can realize the functions of two or more of the above modules. For another example, the distribution network line loss abnormality diagnosis device 100 can also include a preprocessing module ( Figure 1 (not shown in the figure), the preprocessing module can be used to perform preprocessing such as filtering and denoising on the aforementioned monitoring statistical data. Such variations are within the protection scope of this specification.

[0077] Figure 2 is an exemplary flow chart of a method for diagnosing abnormal line loss in a distribution network according to some embodiments of this specification. Figure 2 The distribution network line loss abnormality diagnosis method 200 may include:

[0078] Step S210 , acquiring monitoring status characteristics of the target line according to monitoring statistical data of the target line and status information corresponding to the monitoring statistical data. In some embodiments, step S210 may be executed by the first acquisition module 110 .

[0079] In an embodiment of the present application, the target line may refer to any power transmission line and distribution line that requires line loss abnormality diagnosis or status monitoring. The monitoring statistical data of the target line may include multiple historical real-time monitoring data obtained from monitoring the target line, wherein each monitoring obtained corresponds to a data acquisition time. In an embodiment of the present application, the monitoring statistical data may be time series data generated based on multiple real-time monitoring data of the target line, and the time series data includes the data acquisition time each time the target line obtains real-time monitoring data, and the real-time monitoring data corresponding to each data acquisition time. Specifically, in an embodiment of the present application, the real-time monitoring data may include power supply data, power sales data, meteorological data, and load data and electrical measurement data of each load point in the target line. Among them, the power supply data and power sales data can be obtained through the power company's metering system and billing system (these data can be collected through smart meters, data acquisition and monitoring control systems (SCADA) or energy management systems (EMS)); meteorological data can be measured by various sensors, and exemplary meteorological data can include weather patterns, temperature, humidity, wind speed, wind direction and other parameters; load data and electrical measurement data of each load point in the target line can be obtained through one or more of smart meters (smart meters can not only record power supply and power sales data, but also measure electrical parameters such as voltage, current, power factor and other electrical parameters of the load points), distribution automation systems (distribution automation systems can monitor and control the operating status of the distribution network in real time and collect load data and electrical measurement data through various sensors and controllers installed in the distribution network), and substation automation systems (substation automation systems can collect the operating status and electrical parameters of each equipment in the substation, including load data and electrical measurement data such as voltage, current, power factor and other electrical measurement data).

[0080] It can be understood that power supply data and power sales data are key data for calculating line losses, and meteorological factors such as temperature, humidity, and wind speed will affect the operating status and power demand of power equipment, thereby affecting the effectiveness and accuracy of line loss calculation. Similarly, load data can reflect the power demand of each load point. By analyzing the load data, we can understand the changing trend of power demand and peak load, help identify time periods and areas with high line losses, and further evaluate the operating efficiency and line loss of power equipment. Electrical measurement data can directly reflect the operating status and power quality of power equipment, help determine whether the power equipment has faults or abnormal operation, and thus find out the potential causes of line losses. In an embodiment of the present application, by combining the above data to diagnose line loss anomalies in the distribution network, the accuracy and reliability of the line loss anomaly diagnosis results can be improved to a certain extent, and the line loss anomaly type and its cause corresponding to the target line can be more accurately diagnosed.

[0081] In the embodiment of the present application, the state information corresponding to the aforementioned monitoring statistical data can be used to indicate whether the target line has a line loss abnormal state. In other words, in the embodiment of the present application, a state identifier can be used to indicate whether the target line has a corresponding line loss abnormal state. Exemplary line loss abnormal states may include but are not limited to line aging (which can be determined based on the service life and life of the line or equipment), abnormal insulation performance, poor contact, overload operation, equipment failure, abnormal voltage, current imbalance, exceeding the power generation rule, bad weather, etc. In the embodiment of the present application, the state identifier may include a first identifier and a second identifier, wherein the first identifier may be used to indicate that the target line has a corresponding abnormal state, and the second identifier may be used to indicate that the target line does not have a corresponding abnormal state. For example, in some embodiments, the state information may be represented in the form of "01...001", wherein 1 represents the first identifier, and 0 represents the second identifier. When any position in the state information (i.e., the digit of the identifier in the state information) has the first identifier, it indicates that the target line has an abnormal state corresponding to the position, and when any position in the state information has the second identifier, it indicates that the target line does not have an abnormal state corresponding to the position. It should be noted that in the embodiments of the present application, other characters may be used to represent the first identifier and the second identifier. As an example, in some embodiments, A may be used to represent the first identifier and B may be used to represent the second identifier. In some embodiments of the present application, the status information may be obtained by setting corresponding judgment rules based on expert experience.

[0082] After each real-time monitoring data acquisition for the target line, the monitoring statistics data corresponding to the current data acquisition time can be obtained. Based on the monitoring statistics data acquired each time, the state information of the target line can be determined. Therefore, in the embodiment of the present application, each state information can correspond to a data acquisition time, and the monitoring statistics data of the target line includes at least one data acquisition time, then the monitoring statistics data of the target line corresponds to at least one state information. Just as an example, in the embodiment of the present application, the state information corresponding to each data acquisition time of the monitoring statistics data can be represented as a state code, and the state code can be composed of 0 and 1, and each digit in the state code can correspond to an abnormal situation. When the data on a certain digit is 0, it means that the target line does not have the abnormal situation corresponding to the digit, and when the data on any digit is 1, it means that the target line has the abnormal situation corresponding to the digit.

[0083] To facilitate subsequent calculations, in some embodiments, data preprocessing may be performed on the monitoring statistics data of the target line and the state information corresponding to the monitoring statistics data so that the monitoring statistics data and the state information corresponding to the monitoring statistics data conform to a corresponding format, and then state feature extraction may be performed on the preprocessed data to obtain the monitoring state feature of the target line. Figure 3 In some possible implementations, step S210 may include the following sub-steps:

[0084] Sub-step S2101, inputting at least one state information into a multi-layer perceptron.

[0085] Sub-step S2102, filtering out a target line loss abnormal state from at least one line loss abnormal state corresponding to the at least one state information through the multi-layer perceptron, and performing weight assignment on the target line loss abnormal state to obtain a feature corresponding to the at least one line loss abnormal state.

[0086] In the embodiment of the present application, since the state information of the target line includes multiple preset states (each preset state corresponds to a line loss abnormality), the target line has some or all of the multiple preset states. In some embodiments, at least one state information corresponding to at least one data acquisition time can be input into a multilayer perceptron for feature screening, and the target line loss abnormal state is determined based on the abnormal state of the target line at each data acquisition time (i.e., the preset state corresponding to the first identifier in the aforementioned state information).

[0087] In some embodiments, a multi-dimensional one-hot vector may be used to represent the state information corresponding to the monitoring statistical data, so as to pre-process the state information corresponding to the monitoring statistical data. The multi-dimensional one-hot vector may be composed of 0 and 1, where 0 represents that the target line does not have an abnormality corresponding to the corresponding digit, and 1 represents that the target line has an abnormality corresponding to the digit.

[0088] In some embodiments, the multidimensional one-hot vector can be input into a multilayer perceptron, and the multilayer perceptron network can filter out the target line loss abnormal state of the target line through the coding matrix, and weight the filtered target line loss abnormal state to obtain the characteristics of the state information corresponding to the monitoring statistical data. It should be pointed out that in the embodiment of the present application, by weighting the filtered target line loss abnormal state, the characteristics of the state information corresponding to the multidimensional one-hot vector can be concentrated in the processed result.

[0089] Specifically, in the embodiment of the present application, each network node in the multi-layer perception network can calculate the data in the multi-dimensional one-hot vector by the following formula:

[0090]

[0091] in, is a feature vector of the state information corresponding to the j-th data acquisition time, the feature vector represents the feature of the state information corresponding to the j-th data acquisition time, and j is an integer greater than or equal to 1; Represents a processing function; is a feature coding matrix, which may be a pre-trained matrix or a matrix trained by the multi-layer perception network in the process of calculating the features of the state information; Get the multidimensional one-hot vector of the state information corresponding to the time for the jth data; Represents the deviation term.

[0092] For example, when j represents the total number of times real-time monitoring data is acquired for the target line, after calculating all the state information corresponding to the monitoring statistical data through the above multi-layer perception network, the feature G of the state information corresponding to the monitoring statistical data can be obtained: j ]. It should be noted that, in the embodiments of the present application, in addition to calculating the characteristics of the aforementioned state information through a multi-layer perception network, the characteristics of the state information can also be calculated through other neural networks. For example, in some embodiments, a fully connected neural network can also be used.

[0093] It can be understood that, since one state information is used to indicate at least one state of the target line at a data acquisition time in the embodiment of the present application, when the state information indicates more states, the dimension of the multi-dimensional one-hot vector may be higher, which will increase the complexity of subsequent calculations. Based on this, in the embodiment of the present application, by screening out the target line loss abnormal state from multiple state information corresponding to multiple data acquisition times, the feature dimension of the state information can be reduced compared to the dimension of the initial multi-dimensional one-hot vector, thereby achieving data dimensionality reduction for subsequent calculations.

[0094] Sub-step S2103: input the real-time monitoring data into a time series analysis model, and extract the time series features corresponding to each item of data in the real-time monitoring data through the time series analysis model to obtain a monitoring data feature set.

[0095] In an embodiment of the present application, the time series analysis model can be a recurrent neural network model based on deep learning, such as a long short-term memory network (LSTM) or a gated recurrent unit (GRU), which can capture the time dependency in the data sequence, thereby effectively identifying and predicting patterns and trends in the time series data. By training the model, it can be ensured that it has a high recognition accuracy rate for the time series features of a specific type of monitoring data. In addition, the time series analysis model can also optimize its performance by continuously learning new monitoring data to adapt to changes in the monitoring environment.

[0096] In some embodiments, each item of the aforementioned real-time monitoring data may be processed separately to obtain the features corresponding to each item of data. As an example only, in the embodiments of the present application, the time series features corresponding to each item of data may include but are not limited to features such as data distribution, entropy, and scaling properties representing this type of data, which may represent the inherent correlation of this type of data.

[0097] For example, when the monitoring statistical data acquired at the jth data acquisition time is input into the time series analysis model, the time series analysis model can extract the features of each type of data in the jth monitoring statistical data based on the preset feature extraction rules. .in, It represents the time series characteristics of the kth data type (i.e., the kth item of data) in the monitoring statistical data obtained at the jth data acquisition time. z is an integer greater than 1, and k is the number of the data type.

[0098] After processing the monitoring statistical data corresponding to all data acquisition times through the time series analysis model, the time series analysis model can store the extracted features of the monitoring statistical data of each data acquisition time in a monitoring data feature set. , finally, the time series analysis model can output the monitoring data feature set.

[0099] Since the features in the aforementioned monitoring data feature set can only represent the intrinsic correlation of these types of data, and cannot reflect the features of all monitoring statistical data, in the embodiment of the present application, it is also necessary to obtain the features of the monitoring statistical data by executing the following steps.

[0100] Sub-step S2104, inputting the monitoring data feature set into a cross-processing network, and performing feature cross-processing on each time series feature in the monitoring data feature set through the cross-processing network to obtain features corresponding to the real-time monitoring data, wherein the cross-processing network includes a deep cross-neural network.

[0101] In the embodiment of the present application, the cross-processing network can be understood as a deep learning model, which extracts and learns features of input data through nonlinear transformation of multiple layers of neurons. During the feature cross-processing process, the network can automatically identify and combine the correlations between different features, thereby enhancing the model's ability to capture the inherent laws of the data. In this way, the cross-processing network can generate richer and more abstract feature representations, providing more accurate data support for subsequent line loss anomaly diagnosis and prediction tasks.

[0102] In some embodiments, the cross-processing network may include a deep cross neural network, which is composed of a cross network and a deep network. By inputting the aforementioned monitoring data feature set into the cross network and the deep network respectively, multi-layer abstraction and feature extraction of the data can be achieved. Among them, the cross network part can capture the complex relationship between the input features through multi-layer cross operations, while the deep network can learn the deep representation of the data through its deep structure. Specifically, multiple time series features in the monitoring data feature set can be crossed through the cross network, and the crossed features are output, and the deep features of all features in the monitoring data feature set can be extracted through the deep network. Finally, by combining the crossed features with the deep features extracted by the deep network, the features of the monitoring statistical data are obtained.

[0103] Sub-step S2105, inputting the features corresponding to the at least one abnormal line loss state and the features corresponding to the real-time monitoring data into a deep feature extraction network, and extracting the composite features of the real-time monitoring data and the at least one abnormal line loss state through the deep feature extraction network to obtain the monitoring state features of the target line.

[0104] In some embodiments, the features of the status information and the features of the monitoring statistical data can be fused first to obtain a fused feature, and then the fused feature can be input into the deep feature extraction network to extract the composite features of the real-time monitoring data and the at least one abnormal line loss state to obtain the monitoring status features of the target line.

[0105] For example, in some embodiments, the concat function can be used to combine the features of the state information corresponding to the j-th data acquisition time and the characteristics of the monitoring statistical data obtained at the jth data acquisition time ( = ) is merged into ,but =concat[ , ], then Input the deep feature extraction network. Each node in the deep feature extraction network can extract the input Processing is performed to extract the composite features of the real-time monitoring data and the at least one abnormal line loss state, and obtain the monitoring state features of the target line.

[0106]

[0107] in, represents the monitoring status characteristics of the target line at the jth data acquisition time; is the first weight parameter matrix, is the first bias term. It should be noted that due to Each weight parameter in is used to represent The importance measure corresponding to each element in is Can achieve Each element in the equation is weighted, and then integrated The ReLU function can better mine the relevant features between data. Therefore, in the embodiment of the present application, the ReLU function is used to Processing to obtain monitoring status characteristics , which can better express the composite characteristics of the real-time monitoring data and the at least one abnormal line loss state.

[0108] Continue to refer to Figure 2 After step S210, the distribution network line loss abnormality diagnosis method 200 further includes:

[0109] Step S220 , sorting at least one sub-state feature in the monitoring state feature according to the order of the data acquisition time. In some embodiments, step S220 may be performed by the aforementioned sorting module 120 .

[0110] In an embodiment of the present application, the features of the state information corresponding to each data acquisition time and the features of the monitoring statistical data are fused, and then the monitoring state features at each data acquisition time can be extracted through the aforementioned deep feature extraction network. For ease of description, the monitoring state features at each data acquisition time are referred to as sub-state features in this specification, and therefore, the monitoring state features ultimately output by the deep feature extraction network may include at least one sub-state feature.

[0111] In the embodiment of the present application, at least one sub-state feature in the monitoring state feature can be sorted according to the order of the data acquisition time, and the at least one sub-state feature can be stored in the aforementioned monitoring state feature according to the order, so as to obtain the monitoring state feature W=[W1, W2, ..., W j ], where W j It represents the j-th sub-state feature, that is, the monitoring state feature corresponding to the j-th data acquisition time.

[0112] In the embodiment of the present application, since the characteristics of the monitoring statistical data and the characteristics of the status information include the characteristics of various types of data in each time series, the monitoring status characteristics have multiple dimensions and multiple modalities, and the processing process of the monitoring status characteristics can be regarded as a multimodal feature extraction process.

[0113] Since, in the embodiment of the present application, the aforementioned monitoring status feature is obtained based on the monitoring statistical data and the status information corresponding to the monitoring statistical data, it can more accurately reflect the abnormal line loss state reflected by the target line during the detection process compared to the feature obtained only based on the monitoring statistical data. In addition, the monitoring statistical data is the actual monitored data, which can be used as an objective basis to make the acquired monitoring status feature interpretable, and the status information corresponding to the monitoring statistical data is the result obtained based on expert experience judgment. Therefore, the monitoring status feature is obtained based on the status information and the monitoring statistical data, and subsequent calculations are performed based on the monitoring status feature to realize the abnormal diagnosis of the distribution network line loss, which can improve the accuracy of the diagnosis result to a certain extent.

[0114] Step S230, input each sub-state feature in the monitoring state feature into the first feature extraction network of the first target neural network according to the time sequence corresponding to the data acquisition time, and perform a positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature. In some embodiments, step S230 can be performed by the aforementioned first data feature calculation module 130.

[0115] Step S240, input each sub-state feature in the monitoring state feature into the second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and perform reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature. In some embodiments, step S240 can be performed by the aforementioned second data feature calculation module 140.

[0116] In an embodiment of the present application, the first data feature can be used to characterize the positive evolution law of the monitoring status feature, and the second data feature is used to characterize the reverse evolution law of the monitoring status feature.

[0117] Specifically, in an embodiment of the present application, the monitoring state feature obtained by the aforementioned process can be input into the first target neural network, and the sub-state features on each time series in the monitoring state feature are respectively calculated in forward order and reverse order through the first target neural network to obtain the first data feature and the second data feature, wherein the first data feature is used to represent the forward evolution law of the monitoring state feature, and the second data feature is used to represent the reverse evolution law of the monitoring state feature. In an embodiment of the present application, the first target neural network can be a bidirectional RNN (Bidirectional Recurrent Neural Networks, BiRNN) with an attention mechanism, and the bidirectional RNN can be composed of a first feature extraction network and a second feature extraction network, wherein the first feature extraction network is used to obtain the aforementioned first data feature, and the second feature extraction network is used to obtain the aforementioned second data feature.

[0118] In an embodiment of the present application, the monitoring status feature can be input into the first feature extraction network of the first target neural network in the positive time sequence corresponding to the data acquisition time, and the monitoring status feature can be calculated in positive order through the first feature extraction network to obtain the first data feature; similarly, the monitoring status feature can be input into the second feature extraction network of the first target neural network in the reverse time sequence corresponding to the data acquisition time, and the monitoring status feature can be calculated in reverse order through the first feature extraction network to obtain the second data feature.

[0119] As an example only, in some embodiments of the present application, the monitoring state characteristic can be subjected to a forward moving average calculation with a first window size to obtain the first data characteristic, and the monitoring state characteristic can be subjected to a reverse moving average calculation with a second window size to obtain the second data characteristic, wherein the second window size is different from the first window size. For example, the first window size can be 5 data acquisition cycles, and the second window size can be 10 data acquisition cycles. In this way, two different data characteristics can be obtained, which respectively reflect the change characteristics of the monitoring state characteristics in different time scales and different evolution directions, thereby improving the accuracy and reliability of the data characteristics. The specific implementation method of the moving average calculation can be regarded as the prior art and will not be described in detail in this specification.

[0120] In some embodiments, other data processing techniques, such as filtering, normalization, etc., can also be combined to further optimize the aforementioned first data feature and second data feature. In addition, in some embodiments, by comparing the first data feature and the second data feature, the mutation point of the monitoring status feature can be identified, so as to promptly discover potential problems or anomalies. It should be pointed out that in the embodiments of the present application, by using this dual-window moving average calculation method, not only the long-term trend of the monitoring status can be captured, but also short-term fluctuations can be sensitively reflected, providing strong technical support for real-time monitoring and abnormal diagnosis of distribution network line losses.

[0121] It can be understood that, since in the embodiment of the present application, at least one sub-state feature in the monitoring state feature is sorted in the order of data acquisition time, each sub-state feature in the monitoring state feature can be input into the first feature extraction network in a forward time sequence to obtain the first data feature. Similarly, each sub-state feature in the monitoring state feature can be input into the second feature extraction network in a reverse time sequence to obtain the second data feature.

[0122] Specifically, in some embodiments of the present application, the monitoring state feature W=[W1, W2, ..., W j ] is sequentially input into the nodes of the input layer of the first feature extraction network. In other words, W1 is input into the first node of the input layer of the first feature extraction network, W2 is input into the second node of the input layer of the first feature extraction network, and so on. After the monitoring state feature is input into the first feature extraction network, the first feature extraction network can calculate each sub-state feature in the monitoring state feature based on the calculation rules preset in the first feature extraction network, and finally output the first data feature Y Z ={ ,……, }, wherein z represents positive order, and m represents the number of features contained in the first data feature (in some embodiments, the number depends on the value of the first window size).

[0123] Similarly, in some embodiments, the monitoring state feature W=[W1, W2, ..., W j Each sub-state feature in ] is sequentially input into the nodes of the input layer in the second feature extraction network. For example, W j Input the first node of the input layer of the second feature extraction network and convert W j-1 Input the second node of the input layer of the second feature extraction network, and so on... After the monitoring state feature is input to the second feature extraction network, the second feature extraction network can calculate each sub-state feature in the monitoring state feature based on the calculation rules preset in the second feature extraction network, and finally output the second data feature Y F ={ ,……, }, wherein F represents reverse order, and n represents the number of features contained in the second data feature (in some embodiments, the number depends on the value of the second window size).

[0124] Step S250, based on the first data feature and the second data feature, obtain the state change feature of the target line. In some embodiments, step S250 can be performed by the second acquisition module 150 described above.

[0125] In the embodiment of the present application, in order to obtain a more accurate state change characteristic of the target line, the state change characteristic of the target line is obtained based on the first data characteristic and the second data characteristic. Specifically, a data characteristic representing the comprehensive evolution law of the monitoring state characteristic can be first obtained based on the first data characteristic and the second data characteristic, and then the state change characteristic of the target line is obtained based on the data characteristic.

[0126] For example, in some embodiments, the first data feature and the second data feature may be fused to obtain a third data feature Y={Y1, Y2, . . . , Y l}, wherein l represents the number of features included in the third data feature (in some embodiments, l=m+n).

[0127] Furthermore, the third data feature may be weighted to obtain a fourth data feature. In the embodiment of the present application, the fourth data feature may be used to represent the comprehensive evolution law of the monitoring state feature.

[0128] Specifically, in an embodiment of the present application, the third data feature can be weighted by the first attention model in the first target neural network. For example, in one possible implementation, weight parameter learning can be performed based on the first attention model and the third data feature to obtain at least one first weight parameter, and the first weight parameter is used to indicate the importance of the monitoring statistical data and the status information corresponding to the monitoring statistical data. In other words, each of the first weight parameters is used to represent the importance measure of a real-time monitoring data and the corresponding abnormal line loss state. Among them, the first attention model is any attention model within the first target neural network. In an embodiment of the present application, weight parameter learning can be performed based on the weight parameter learning strategy in the first attention model. In some embodiments of the present application, the weight parameter learning strategy can be a position-based attention weight parameter learning strategy, and the learning strategy can be expressed as:

[0129]

[0130] in, The first weight parameter corresponding to the jth data acquisition time; B τ represents the second weight parameter vector; represents the jth feature element in the third data feature; b τ represents the second bias term.

[0131] It should be noted that in the embodiment of the present application, the weight parameter learning strategy in the first attention model may also be other attention weight parameter learning strategies, and the embodiment of the present application does not specifically limit the weight parameter learning strategy in the first attention model.

[0132] Furthermore, in the embodiment of the present application, the at least one first weight parameter may be standardized to obtain at least one second weight parameter.

[0133] Considering that the first weight parameter calculated by the above process is a value obtained by mathematical calculation, it may be too large or too small. Therefore, in an embodiment of the present application, in order to facilitate calculation, the at least one first weight parameter can be standardized so that the size of each second weight parameter obtained after processing is moderate. In an embodiment of the present application, the standardization process may include normalization. It can be understood that in an embodiment of the present application, since each second weight parameter is only the result obtained by standardizing a first weight parameter, the second weight parameter has the same function as the first weight parameter, and is used to indicate a monitoring statistic and the importance measure of the state information corresponding to the monitoring statistic. Based on this, in some embodiments, the aforementioned third data feature can also be weighted based on the at least one second weight parameter to obtain the fourth data feature.

[0134] Specifically, the at least one second weight parameter [ , ,……, , ,]Substitute into the formula , and then use the output of the formula as the fourth data feature to achieve weight allocation for the third data feature. Where c is the fourth data feature, The second weight parameter corresponding to the j-th data acquisition time, is the jth feature element in the third data feature, and l is the number of features contained in the third data feature.

[0135] It can be understood that since a second weight parameter is used to represent the importance measure of a monitoring statistic and the status information corresponding to the monitoring statistic, weighting the third data feature through at least one second weight parameter can make the third data feature have a more accurate feature expression. Therefore, in an embodiment of the present application, by fusing the first data feature and the second data feature to obtain the third data feature, and then weighting the third data feature to obtain the fourth data feature, the evolution law of the aforementioned monitoring status feature can be more accurately reflected.

[0136] Since the first data feature and the second data feature are weighted, the first data feature and the second data feature can be comprehensively represented by the fourth data feature, so the fourth data feature can reflect both the positive evolution law represented by the first data feature and the reverse evolution law represented by the second data feature. In other words, the fourth data feature can represent the comprehensive evolution law of the aforementioned monitoring state features.

[0137] In some embodiments of the present application, the state change feature can be obtained based on the third data feature and the fourth data feature. Specifically, since the correlation between adjacent monitoring statistical data is the highest, and the correlation between adjacent state information is the highest, in order to further predict the state feature of the target line when it is detected next time, the data feature corresponding to the last data acquisition time in the third data feature and the fourth data feature can be substituted into the formula = , and the output As the state change characteristics. is the data feature corresponding to the last data acquisition time of the target line in the third data feature, for The vector after fusion with c, B d is the third weight parameter matrix, is the third bias term.

[0138] Since the fourth data feature represents the comprehensive evolution law of the monitoring status feature, the first data feature can represent the forward evolution law of the monitoring status feature, and the second data feature can represent the reverse evolution law of the monitoring status feature, by assigning weights to these three data features, the result obtained can more accurately represent the state change characteristics of the target line.

[0139] Step S260, input the state change feature into the second target neural network, and extract the depth feature corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line. In some embodiments, step S260 can be performed by the aforementioned deep state feature extraction module 160.

[0140] In an embodiment of the present application, the second target neural network can be a hierarchical and multi-label classification network, such as HMCN (Hierarchical Multi-label Classification Networks). Since the state change characteristics obtained by the aforementioned process cannot reflect the deep characteristics of the target line, in some embodiments of the present application, the deep characteristics corresponding to the target line can be extracted through the second target neural network.

[0141] Specifically, in some embodiments of the present application, the second target neural network can extract the deep features of the target line layer by layer, so that the final extracted detail features can meet the needs for state information prediction. Each layer of the processing network of the second target neural network can output a deep state feature. After the state change feature is input into the second target neural network, it can be input from the input layer to the output layer of the second target neural network. The second target neural network can calculate the state change feature layer by layer. When calculating, the first processing layer of the second target neural network can calculate the network layer features of the first processing layer and the deep state features of the target line in the first processing layer based on the output data of the second processing layer, wherein the first processing layer can refer to any layer of the second target neural network, and the second processing layer is the previous processing layer in the second target neural network that is located before the first processing layer. The network layer feature is used to represent the state of the state change feature in the network layer of the second target neural network, and the network layer feature of the first processing layer is determined by the state change feature and the network layer feature of the second processing layer.

[0142] After the second processing layer of the second target neural network generates the network layer features of the second processing layer and the deep state features of the target circuit at the second processing layer, the second processing layer can output the network layer features of the second processing layer and the state change features (i.e., the output data of the second processing layer) to the first processing layer, so that the first processing layer can receive the network layer features of the second processing layer and the state change features. In this way, as the state change features are output by each network layer of the second target neural network, the state change features can be input to each network layer of the second target neural network.

[0143] Since the network layer characteristics of the first processing layer are determined by the state change characteristics and the network layer characteristics of the second processing layer, when the first processing layer receives the network layer characteristics of the second processing layer and the state change characteristics, the first processing layer can calculate the network layer characteristics of the first processing layer based on the network layer characteristics of the second processing layer and the state change characteristics. Specifically, in some embodiments, the network layer characteristics of the i-th layer in the second target neural network It can be expressed as: , where G stands for global, is the fourth weight parameter matrix, is the network layer feature of the i-1th processing layer (the i-th layer can be regarded as the first processing layer), represents the state change characteristics obtained by the above process, represents the fourth bias term.

[0144] In the embodiment of the present application, the node on the first processing layer can obtain the deep state feature of the target line on the first processing layer based on the network layer feature of the first processing layer and the state change feature. As an example only, in some embodiments, the deep state feature of the target line at the i layer It can be expressed as: . Where L represents the network layer, is the fifth weight parameter matrix, is the fifth bias term.

[0145] It can be understood that in the embodiment of the present application, since each layer of the second target neural network is calculated based on the network layer characteristics and state change characteristics of the previous layer, the deep state characteristics of the target line on each layer of the second target neural network are affected by the deep state characteristics of the previous layer. In addition, since the hierarchical expression of each layer is determined by the network layer characteristics of this layer, the deep state characteristics generated by any network layer in the second target neural network can be used as the parent of the deep state characteristics generated by the next network layer. Based on this, the second target neural network can realize the extraction of the deep characteristics of the target line layer by layer.

[0146] Step S270 , based on the at least one deep state feature of the target line, diagnose and predict line loss anomaly for the target line. In some embodiments, step S270 may be performed by the aforementioned line loss anomaly identification module 170 .

[0147] From the above content, it can be known that in the embodiments of the present application, the deep state feature can reflect the details of the monitoring statistics of the target line at different levels. Considering that there may be many details involved in the above process, in some embodiments of the present application, these details at different levels can be processed centrally to obtain more detailed deep state features, and then line loss abnormality diagnosis and prediction can be performed based on this more detailed deep state feature.

[0148] In some embodiments, at least one depth state feature of the target line can be weighted to obtain a target depth state feature. The target depth state feature is also a more detailed depth state feature. Specifically, in some embodiments, the depth state feature of the target line can be weighted by a second attention model in a second target neural network. Specifically, weight parameter learning can be performed based on the second attention model and the at least one depth state feature to obtain at least one third weight parameter, wherein the third weight parameter is used to represent the importance measure of the corresponding depth state feature.

[0149] In an embodiment of the present application, the second attention model may be any attention model in the second target neural network. In some embodiments, weight parameter learning may be performed based on the second attention model and the at least one deep state feature. Specifically, weight parameters may be learned based on a weight parameter learning strategy in the second attention model. The weight parameter learning strategy in the second attention model may be expressed as:

[0150]

[0151]

[0152] in, represents the i-th third weight parameter, represents the exponential function with base e, and Respectively represent the reference weight parameters corresponding to the i-th and j-th data acquisition times, represents the sixth weight parameter matrix, is the sixth bias term, represents the hyperbolic tangent function.

[0153] It should be noted that the weight parameter learning strategy in the second attention model can also adopt other learning strategies, and the embodiment of the present application does not specifically limit the weight parameter learning strategy in the second attention model. It should also be noted that in the embodiment of the present application, the weight parameter matrix and bias terms involved in the aforementioned content can be obtained through training, and during the training process, the gradient descent method or other optimization algorithms can be used to continuously adjust these parameters to minimize the model error.

[0154] Furthermore, based on the at least one third weight parameter, a weight can be assigned to at least one depth state feature of the target line to obtain the target depth state feature. Specifically, in an embodiment of the present application, at least one third weight parameter and at least one depth state feature of the target line can be substituted into the formula: , and then output it as the target deep state feature to achieve weight allocation of at least one deep state feature of the target line. is the deep state feature corresponding to the monitoring statistical data corresponding to the current time series, is the number of layers of the second target neural network.

[0155] It can be understood that since a third weight parameter in the embodiment of the present application represents an importance measure of a depth state feature, weighting of at least one depth state feature is performed by using at least one third weight parameter to obtain a target depth state feature, which can make the obtained target depth state feature more detailed.

[0156] After obtaining the target depth state feature through the above steps, the target line can be diagnosed and predicted based on the target depth state feature. Specifically, in some embodiments, the target depth state feature can be input into a trained line loss anomaly recognition model to obtain a prediction result corresponding to the target depth state feature, wherein the prediction result includes a predicted line loss anomaly type and its corresponding predicted anomaly cause.

[0157] In the embodiment of the present application, the line loss anomaly recognition model can be trained based on the following method:

[0158] First, obtain the sample depth state characteristics obtained based on the sample monitoring statistical data processing, and the sample label corresponding to the sample depth state characteristics, wherein the sample label is used to characterize the line loss anomaly type and its anomaly cause corresponding to the sample monitoring statistical data. The sample monitoring statistical data can be obtained through experiments. For example, in some embodiments, the line loss anomaly can be artificially created or selected, and then the corresponding monitoring statistical data is collected as the sample monitoring statistical data. It should be noted that in the embodiment of the present application, the sample monitoring statistical data is the time series data collected within a certain period of time.

[0159] Then, the sample monitoring statistical data can be processed by the method described above to obtain the sample depth state characteristics corresponding to the sample monitoring statistical data. The specific implementation details of the process can be referred to above and will not be repeated here. At the same time, the corresponding sample label can be determined for each sample depth state characteristic based on the actual line loss anomaly corresponding to each sample monitoring statistical data and its anomaly cause. The sample label can reflect the line loss anomaly type and its anomaly cause corresponding to each sample monitoring statistical data.

[0160] Finally, the sample deep state feature can be used as input, and the sample label corresponding to the sample deep state feature can be used as output to train the initial line loss anomaly recognition model until the preset conditions are met to obtain the trained line loss anomaly recognition model. In an embodiment of the present application, the initial line loss anomaly recognition model may include one or more neural network layers, such as a convolutional neural network (CNN) layer, a recurrent neural network (RNN) layer, or a fully connected layer. Among them, the convolutional neural network layer can be used to extract features, the recurrent neural network layer can be used to process time series data, and the fully connected layer is used to integrate the relationship between different features. Through such a multi-layer network structure, the line loss anomaly recognition model can learn the complex patterns and correlations in the sample deep state features corresponding to the sample monitoring statistical data.

[0161] During the training process, the back-propagation algorithm can be used to adjust the weights of the model to minimize the difference between the prediction results and the sample labels. In an embodiment of the present application, the difference can be measured by a loss function such as the mean square error (MSE). Through continuous iterative optimization, the model can learn complex patterns and relationships in the sample data, thereby improving the accuracy of identifying line loss anomalies. In addition, in some embodiments of the present application, in order to prevent the model from overfitting, regularization techniques such as L1 or L2 regularization can be used, or the dropout method can be used to randomly discard some neurons to enhance the generalization ability of the model.

[0162] When the aforementioned loss function reaches a preset threshold, the training can be terminated to obtain a trained line loss anomaly recognition model. After the training is completed, the model can be tested by comparing the predicted results on the test set with the actual labels to evaluate the performance of the model. If the line loss anomaly recognition model trained through the above steps performs well on the test set, the trained line loss anomaly recognition model can be used in the actual line loss anomaly detection task. More details about the model training can be regarded as prior art and will not be repeated in this specification.

[0163] It can be understood that in the embodiments of the present application, after obtaining the above-mentioned trained line loss anomaly recognition model, the trained line loss anomaly recognition model can be used to diagnose the real-time monitoring data of any target line, thereby identifying or predicting the corresponding line loss anomaly type, and giving the relevant abnormal reasons, thereby realizing real-time monitoring and management of the power grid operation status. In this way, power companies can promptly discover and deal with potential line loss problems, reduce economic losses, and improve the operation efficiency and reliability of the power grid. In addition, in some embodiments of the present application, the line loss anomaly recognition model can also be continuously optimized and adjusted based on the comparative analysis of historical data and real-time data to adapt to changes in the power grid operation environment and ensure the accuracy and timeliness of the diagnostic results.

[0164] Figure 4 is an exemplary structural diagram of a computer processing device provided in an embodiment of the present application. Figure 4 , wherein the computer processing device 400 may include a memory and a processor. The memory and the processor are electrically connected directly or indirectly to achieve data transmission or interaction. For example, the two may be electrically connected via one or more communication buses or signal lines. The memory may store at least one software function module (computer program) that may exist in the form of software or firmware. The processor may be used to execute the executable computer program stored in the memory, thereby implementing the distribution network line loss abnormality diagnosis method provided in the embodiment of the present application.

[0165] Optionally, in the embodiment of the present application, the memory may be, but not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on chip (SoC), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0166] It should be noted that Figure 4 The structure shown is for illustration only. The computer processing device may also include Figure 4 More or fewer components as shown, or with Figure 4 The different configurations shown, for example, may include a communication unit for information exchange with other devices. In an alternative example, the computer processing device may be a server with data processing capabilities. In some embodiments, the aforementioned distribution network line loss abnormality diagnosis device 100 may be a part of the computer processing device 400.

[0167] In summary, the beneficial effects that may be brought about by the embodiments of the present specification include but are not limited to: (1) In the distribution network line loss abnormality diagnosis method and device provided in some embodiments of the present specification, the monitoring state characteristics are obtained by processing the monitoring statistical data and the state information corresponding to the monitoring statistical data. Compared with the characteristics obtained only based on the monitoring statistical data, it can more accurately reflect the abnormal line loss state reflected by the target line during the detection process, thereby improving the accuracy of the line loss abnormality diagnosis results in the subsequent process; (2) In the distribution network line loss abnormality diagnosis method and device provided in some embodiments of the present specification, the sub-state characteristics of each time series in the monitoring state characteristics are respectively processed by the first target neural network. The forward and reverse calculations can obtain the first data feature and the second data feature, which can respectively reflect the data change characteristics of the monitoring status features in different evolution directions and different time scales, thereby improving the accuracy and reliability of the data features to a certain extent; (3) In the distribution network line loss abnormality diagnosis method and device provided in some embodiments of the present specification, the deep state features corresponding to the target line are extracted through the second target neural network, and the line loss abnormality diagnosis and prediction of the target line are performed based on the deep state features. Potential line loss problems can be discovered in time through monitoring statistical data, thereby realizing line loss abnormality diagnosis and line loss abnormality prediction at the same time, providing strong technical support for the stable operation of the power grid.

[0168] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or may be any other beneficial effects that may be obtained.

[0169] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0170] At the same time, this specification uses specific words to describe the embodiments of this specification. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures or characteristics in one or more embodiments of this specification can be appropriately combined.

[0171] In addition, it will be understood by those skilled in the art that various aspects of this specification may be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any new and useful improvements thereto. Accordingly, various aspects of this specification may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as "data blocks", "modules", "engines", "units", "components" or "systems". In addition, various aspects of this specification may be represented as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0172] A computer storage medium may include a propagated data signal containing computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may be in a variety of forms, including electromagnetic, optical, etc., or a suitable combination. A computer storage medium may be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit the program for use. The program code on the computer storage medium may be transmitted via any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0173] The computer program codes required for the operation of each part of this specification can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C language, Visual Basic, Fortran2003, Perl, COBOL2002, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby and Groovy, or other programming languages, etc. The program code can be run entirely on the user's computer, or run on the user's computer as an independent software package, or run partly on the user's computer and partly on a remote computer, or run entirely on a remote computer or processing device. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0174] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this specification, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing processing device or mobile device.

[0175] Similarly, it should be noted that in order to simplify the description disclosed in this specification and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this specification, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the subject matter of this specification are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0176] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining the digits. Although the numerical domains and parameters used to confirm the breadth of the range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0177] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, documents, etc., cited in this specification is hereby incorporated by reference in its entirety. Except for application history documents that are inconsistent with or conflicting with the contents of this specification, documents that limit the broadest scope of the claims of this specification (currently or later attached to this specification) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the materials attached to this specification are inconsistent or conflicting with the contents described in this specification, the descriptions, definitions, and / or use of terms in this specification shall prevail.

[0178] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A method for diagnosing abnormal line loss in a distribution network, characterized in that: include: Acquire monitoring status characteristics of the target line according to monitoring statistical data of the target line and status information corresponding to the monitoring statistical data; Sorting at least one sub-state feature in the monitoring state feature in the order of data acquisition time; Input each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network according to the time positive sequence corresponding to the data acquisition time, and perform positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature; Input each sub-state feature in the monitoring state feature into a second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and perform reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature; Based on the first data feature and the second data feature, acquiring a state change feature of a target line; Inputting the state change feature into a second target neural network, and extracting the deep features corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line; Based on the at least one deep state feature of the target line, line loss abnormality diagnosis and prediction are performed on the target line.

2. The method according to claim 1, characterized in that The acquiring the monitoring status feature of the target line according to the monitoring statistical data of the target line and the status information corresponding to the monitoring statistical data includes: Inputting at least one state information into a multilayer perceptron; Filtering out a target line loss abnormal state from at least one line loss abnormal state corresponding to the at least one state information by the multilayer perceptron, and performing weight allocation on the target line loss abnormal state to obtain a feature corresponding to the at least one line loss abnormal state; Inputting the real-time monitoring data in the monitoring statistical data into a time series analysis model, and extracting the time series features corresponding to each item of data in the real-time monitoring data through the time series analysis model to obtain a monitoring data feature set; Inputting the monitoring data feature set into a cross processing network, and performing feature cross processing on each time series feature in the monitoring data feature set through the cross processing network to obtain features corresponding to the real-time monitoring data, wherein the cross processing network includes a deep cross neural network; The features corresponding to the at least one abnormal line loss state and the features corresponding to the real-time monitoring data are input into a deep feature extraction network, and the composite features of the real-time monitoring data and the at least one abnormal line loss state are extracted through the deep feature extraction network to obtain the monitoring state features of the target line.

3. The method according to claim 1, characterized in that The step of inputting each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network in the time sequence corresponding to the data acquisition time, and performing a positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature comprises: performing a positive sequence moving average calculation on the monitoring state feature with a first window size to obtain the first data feature; The method comprises: performing reverse moving average calculation on the monitoring state feature with a second window size to obtain the second data feature, wherein the second window size is different from the first window size.

4. The method according to claim 1, characterized in that The acquiring, based on the first data feature and the second data feature, a state change feature of the target line includes: Fusing the first data feature and the second data feature to obtain a third data feature; Performing weight distribution on the third data feature to obtain a fourth data feature, wherein the fourth data feature is used to characterize the comprehensive evolution law of the monitoring state feature; Based on the third data feature and the fourth data feature, a state change feature of the target line is obtained.

5. The method according to claim 4, characterized in that The step of performing weight distribution on the third data feature to obtain a fourth data feature includes: Processing the third data feature through a first attention model to obtain at least one first weight parameter, wherein each of the first weight parameters is used to represent an importance measure of an item of real-time monitoring data and a corresponding abnormal line loss state; performing normalization processing on the first weight parameter to obtain at least one second weight parameter; The third data feature is weighted according to the second weight parameter to obtain the fourth data feature.

6. The method according to claim 1, characterized in that The performing line loss abnormality diagnosis and prediction on the target line based on the at least one deep state feature of the target line includes: Performing weight parameter learning based on the second attention model and the at least one deep state feature to obtain at least one third weight parameter, wherein each third weight parameter is used to represent an importance measure of a corresponding deep state feature; Performing weight allocation on the at least one depth state feature according to the third weight parameter to obtain a target depth state feature; Based on the target depth status characteristics, line loss abnormality diagnosis and prediction are performed on the target line.

7. The method according to claim 6, characterized in that Each layer of the second target neural network outputs one of the deep state features, and the method further includes: The deep state features output by the first processing layer are generated based on the network layer features of the first processing layer in the second target neural network and the deep state features generated by the second processing layer, wherein the network layer features of the first processing layer are used to represent the state of the state change features in the first processing layer, and the second processing layer is the previous processing layer before the first processing layer in the second target neural network.

8. The method according to claim 7, characterized in that The network layer characteristic of the first processing layer is determined based on the state change characteristic and the network layer characteristic of the second processing layer.

9. The method according to claim 6, characterized in that The method of diagnosing and predicting line loss anomalies on the target line based on the target depth state feature comprises: inputting the target depth state feature into a trained line loss anomaly recognition model to obtain a prediction result corresponding to the target depth state feature, wherein the prediction result includes a predicted line loss anomaly type and a corresponding predicted anomaly cause; The line loss anomaly recognition model is trained based on the following method: Acquire a sample depth state feature obtained based on sample monitoring statistical data, and a sample label corresponding to the sample depth state feature, wherein the sample label is used to characterize the line loss anomaly type and its anomaly cause corresponding to the sample monitoring statistical data; The sample depth state feature is used as input, and the sample label corresponding to the sample depth state feature is used as output to train the initial line loss anomaly recognition model until the preset conditions are met, thereby obtaining the trained line loss anomaly recognition model.

10. A distribution network line loss abnormality diagnosis device, characterized in that: include: A first acquisition module is used to acquire monitoring status characteristics of the target line according to monitoring statistical data of the target line and status information corresponding to the monitoring statistical data, wherein the status information corresponding to the monitoring statistical data is used to indicate abnormal line loss status of the target line, the monitoring statistical data is time series data generated based on real-time monitoring data of the target line, the time series data includes data acquisition time when the target line acquires real-time monitoring data each time, and real-time monitoring data corresponding to each data acquisition time, the real-time monitoring data includes power supply data, power sales data, meteorological data, and load data and electrical measurement data of each load point in the target line; A sorting module, used to sort at least one sub-state feature in the monitoring state feature according to the order of the data acquisition time; A first data feature calculation module is used to input each sub-state feature in the monitoring state feature into a first feature extraction network of a first target neural network according to the time positive sequence corresponding to the data acquisition time, and perform positive sequence calculation on the monitoring state feature through the first feature extraction network to obtain a first data feature, wherein the first data feature is used to characterize the positive evolution law of the monitoring state feature; A second data feature calculation module is used to input each sub-state feature in the monitoring state feature into a second feature extraction network of the first target neural network in reverse order of the time corresponding to the data acquisition time, and perform reverse calculation on the monitoring state feature through the second feature extraction network to obtain a second data feature, wherein the second data feature is used to characterize the reverse evolution law of the monitoring state feature; A second acquisition module, configured to acquire a state change feature of the target line based on the first data feature and the second data feature; A deep state feature extraction module, used to input the state change feature into a second target neural network, and extract the deep features corresponding to the target line layer by layer through the second target neural network to obtain at least one deep state feature of the target line; A line loss anomaly identification module is used to diagnose and predict line loss anomaly of the target line based on the at least one deep state feature of the target line.

Images