CN118473097B - Intelligent line loss detection and alarm method and device for power distribution network - Google Patents

Abstract

The present invention discloses a smart line loss detection and alarm method and device for a distribution network, and relates to the technical field of line loss detection in a distribution network. The method comprises: obtaining line design information, identifying line no-load loss according to the line design information, and generating no-load line loss rate; receiving a preset plan, and collecting power sales information; identifying loss according to the power sales information and the preset plan, and generating a real-time line loss rate; identifying loss rate according to the real-time line loss rate and the no-load line loss rate, and collecting a state data set if it exceeds the standard; performing dynamic loss analysis based on the state data set, and generating a dynamic line loss rate; and generating alarm information by combining the three loss rates for abnormal alarm. The method solves the technical problem that the line loss analysis method based on direct judgment of the threshold value cannot provide targeted early warning due to the variety of causes of line loss, and the auxiliaryness of the subsequent abnormal line loss cause investigation is low, and the accuracy of line loss detection is improved, so that the operation and management of the distribution network are more efficient and accurate.

Description

A smart line loss detection and alarm method and device for distribution network

Technical Field

The present application relates to the technical field of line loss detection in distribution networks, and specifically to an intelligent line loss detection and alarm method and device for distribution networks.

Background Art

With the rapid development of the power industry and the continuous expansion of the power grid, line loss management of distribution networks has become a key link in ensuring the efficient operation of power systems and reducing energy losses. However, with the complexity of the power grid structure and the diversification of load demand, line loss detection and management of distribution networks face increasing challenges. Traditional line loss detection methods often rely on direct analysis of line operation data and threshold judgment, which is inadequate for dealing with complex and changeable line loss problems. Since there are many reasons for line loss, including technical factors such as equipment aging and improper line design, as well as management factors such as improper operation and maintenance and unreasonable scheduling strategies, direct warning based on thresholds often cannot identify specific causes in a targeted manner, resulting in low auxiliary power for subsequent investigation of abnormal line loss causes.

Summary of the invention

The present application provides an intelligent line loss detection and alarm method and device for a distribution network, thereby solving the technical problem that a line loss analysis method based on direct threshold judgment cannot provide targeted warnings due to the various causes of line loss, and has low auxiliary power for subsequent investigation of abnormal line loss causes. This achieves the effect of improving the accuracy of line loss detection and making the operation and management of the distribution network more efficient and accurate.

The present application provides an intelligent line loss detection and alarm method for a distribution network, the method comprising: obtaining multiple line design information of multiple distribution lines of a target distribution network, identifying line no-load losses according to the multiple line design information, and generating multiple no-load line loss rates; receiving a preset distribution plan of the target distribution network in a preset time zone, and collecting power sales information in the preset time zone; identifying power losses of the multiple distribution lines according to the power sales information and the preset distribution plan, and generating multiple real-time line loss rates; identifying variable line loss rates according to the multiple real-time line loss rates and the multiple no-load line loss rates, and if the variable line loss rate is greater than the preset variable line loss rate, collecting multiple line load status data sets of the multiple distribution lines; performing dynamic loss analysis based on the multiple line load status data sets, and generating multiple dynamic line loss rates; generating alarm information in combination with the multiple dynamic line loss rates, the multiple no-load line loss rates, and the multiple real-time line loss rates to issue an abnormal line loss alarm.

The present application also provides an intelligent line loss detection and alarm device for a distribution network, comprising: a no-load loss identification module, the no-load loss identification module is used to obtain multiple line design information of multiple distribution lines of the target distribution network, identify line no-load losses according to the multiple line design information, and generate multiple no-load line loss rates; a data analysis module, the data analysis module is used to receive a preset distribution plan of the target distribution network in a preset time zone, and collect power sales information in the preset time zone; an energy loss identification module, the energy loss identification module is used to identify energy losses of the multiple distribution lines according to the power sales information and the preset distribution plan, and generate multiple real-time power loss rates. a variable line loss rate identification module, the variable line loss rate identification module is used to identify the variable line loss rate according to the multiple real-time line loss rates and the multiple no-load line loss rates, if the variable line loss rate is greater than the preset variable line loss rate, collect multiple line load status data sets of the multiple distribution lines; a dynamic loss analysis module, the dynamic loss analysis module is used to perform dynamic loss analysis based on the multiple line load status data sets to generate multiple dynamic line loss rates; a line loss abnormality alarm module, the line loss abnormality alarm module is used to generate alarm information based on the multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates to issue a line loss abnormality alarm.

It is intended to obtain multiple line design information of multiple distribution lines of a target distribution network through an intelligent line loss detection and alarm method and device for a distribution network proposed in this application, identify line no-load losses according to the multiple line design information, and generate multiple no-load line loss rates; receive a preset distribution plan for the target distribution network in a preset time zone, and collect power sales information in the preset time zone; identify power losses of the multiple distribution lines according to the power sales information and the preset distribution plan, and generate multiple real-time line loss rates; identify variable line loss rates according to the multiple real-time line loss rates and the multiple no-load line loss rates, and if the variable line loss rate is greater than the preset variable line loss rate, collect multiple line load status data sets of the multiple distribution lines; perform dynamic loss analysis based on the multiple line load status data sets, and generate multiple dynamic line loss rates; generate alarm information in combination with the multiple dynamic line loss rates, the multiple no-load line loss rates, and the multiple real-time line loss rates to issue an abnormal line loss alarm. The line loss analysis method based on direct judgment of thresholds has solved the technical problem that due to the various causes of line loss, targeted early warning cannot be carried out, and the auxiliary effect on the subsequent investigation of abnormal causes of line loss is low. The accuracy of line loss detection is improved, making the operation and management of the distribution network more efficient and accurate.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the device according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed precisely in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

FIG1 is a schematic diagram of a flow chart of a smart line loss detection and alarm method for a distribution network provided in an embodiment of the present application;

FIG2 is a schematic diagram of the structure of an intelligent line loss detection and alarm device for a distribution network provided in an embodiment of the present application.

Explanation of the accompanying drawings: no-load loss identification module 1, data analysis module 2, power loss identification module 3, variable line loss rate identification module 4, dynamic loss analysis module 5, line loss abnormality alarm module 6.

DETAILED DESCRIPTION

The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

The embodiment of the present application provides a smart line loss detection and alarm method for a distribution network, as shown in FIG1 , the method comprising:

A plurality of line design information of a plurality of distribution lines of a target distribution network is obtained, line no-load loss is identified according to the plurality of line design information, and a plurality of no-load line loss rates are generated.

In the embodiment of the present application, in the line loss management of the distribution network, it is very important to accurately identify and calculate the no-load loss, that is, the fixed loss or basic loss. The no-load loss refers to the fixed energy loss generated by certain equipment or components in the distribution network without considering the load change. These losses mainly include transformer iron loss, high-voltage line corona loss, and electric energy meter coil loss. In order to accurately evaluate the no-load loss of the target distribution network, the system terminal first obtains multiple line design information of multiple distribution lines. This information includes key parameters such as the length, material, conductor cross-section, transformer model and capacity used, and voltage level of the line. Based on these detailed line design information, the system terminal performs cumulative calculation of the no-load loss of the grid components, and calculates the no-load line loss rate of each line design information in combination with the rated distribution power. In this way, the system terminal can generate multiple no-load line loss rates. These no-load line loss rates not only reflect the energy loss of the distribution network in the no-load state, but also provide an important basis for the subsequent investigation of abnormal causes of line loss and energy-saving and consumption-reducing measures.

Furthermore, the present application provides a method for identifying line no-load loss according to the plurality of line design information, and generating a plurality of no-load line loss rates, including:

Based on the multiple line design information, the first line design information of the first distribution line is extracted, wherein the first line design information includes the specifications and quantity of the grid elements in the first distribution line; the no-load losses of the grid elements are cumulatively calculated according to the first line design information to generate a first no-load loss; based on the rated distribution power of the first distribution line and the first no-load loss, a first no-load line loss rate is calculated and obtained, and added to the multiple no-load line loss rates.

Preferably, the system terminal randomly extracts one line design information from a plurality of line design information as the detailed design data of the first distribution line, i.e., the first line design information of the first distribution line. This information includes the specifications and quantity of the grid elements used in the line. The grid elements are the basic components of the grid, such as voltage regulators, transformers, cables, voltage coils, capacitors, etc. Subsequently, based on the extracted first line design information, the system terminal calculates the no-load loss of the grid elements. No-load loss refers to the energy loss generated by the grid elements when there is no load or the load is extremely low. For the transformer, the system terminal directly uses the no-load loss data on its nameplate, or finds the corresponding standard no-load loss data by querying the model and capacity of the transformer. For components such as cables and capacitors, their no-load losses are related to their dielectric loss data per unit length and the actual length of the line. The system terminal obtains the total no-load loss by multiplying the dielectric loss per unit length by the actual length of the line. After completing the no-load loss calculation of each component, the system terminal performs the operation of accumulating the total no-load loss. This means that the system terminal adds up the no-load loss values of all grid elements to obtain the total no-load loss of the entire first distribution line, that is, the first no-load loss. Then, the first no-load line loss rate is calculated by combining the rated distribution power of the first distribution line and the calculated first no-load loss. The no-load line loss rate is the ratio of the no-load loss to the rated distribution power, which reflects the energy loss of the grid under no-load state. Then, the system terminal adds the calculated no-load line loss rate to the set of multiple no-load line loss rates to evaluate and compare the performance of the entire grid system.

Receive a preset power distribution plan of the target power distribution network in a preset time zone, and collect power sales information in the preset time zone.

In one embodiment, in order to accurately evaluate the real-time line loss rate, the system terminal receives the preset distribution plan of the target distribution network within the preset time zone. This plan includes information such as the expected load of the power grid during this period, the expected power distribution strategy, and possible power dispatch plans. This preset time zone is set based on the actual operating needs of the power grid, the characteristics of load changes, and the convenience of management. Subsequently, the system terminal collects the actual power sales information in this preset time zone. The power sales information mainly reflects the actual power sales situation of the power grid during this period, including key data such as power sales volume, power sales time, and electricity price. These data are crucial to understanding the actual operating status and load changes of the power grid.

The power loss of the plurality of power distribution lines is identified according to the power selling information and the preset power distribution plan, and a plurality of real-time line loss rates are generated.

In one embodiment, in order to accurately calculate the real-time line loss rate of the power grid, we need to calculate the actual operating power loss. This process is based on the collected power sales information and the preset power distribution plan, and the power loss of multiple lines in the distribution network is identified through the real-time loss identification model. Specifically, the system terminal collects multiple sets of historical line loss detection records from the historical line loss database. These historical line loss detection records all include historical power sales information and historical power distribution plans, as well as the corresponding historical line loss rates, and then divides the multiple sets of historical line loss detection records into training sets and validation sets. Subsequently, the long short-term memory network model structure is designed, including determining the number of hidden layers, the number of hidden units, the input layer dimension, the output layer dimension, etc. These are all determined according to actual problems and data characteristics. After the model structure is determined, the system terminal initializes the model parameters, such as weights and bias terms. Afterwards, the system terminal uses the training set to train the constructed model structure and updates the model parameters through the back propagation algorithm. At the end of each training cycle, the system terminal uses the validation set to evaluate the performance of the model, determine whether it meets the preset expectations, and record the performance indicators. If not, the system terminal adjusts the model parameters, such as learning rate, number of hidden layers, etc., according to the performance indicators of the validation set, performs hyperparameter tuning until the preset expectations are met, and outputs the trained long short-term memory network model to generate a real-time loss identification model. After the real-time loss identification model is built, the system terminal inputs the obtained power sales information and preset power distribution plan into the real-time loss identification model to identify the energy loss of multiple distribution lines. After receiving the power sales information and the preset power distribution plan, the real-time loss identification model will calculate the real-time line loss rate of each distribution line based on the learned knowledge and generate multiple real-time line loss rates. These real-time line loss rates reflect the power loss of the line in a time period and are important indicators for evaluating the line operation efficiency.

Variable line loss rates are identified based on the multiple real-time line loss rates and the multiple no-load line loss rates. If the variable line loss rate is greater than a preset variable line loss rate, multiple line load status data sets of the multiple distribution lines are collected.

In one embodiment, variable line loss rate identification is an important step in line loss detection, which involves accurate calculation and evaluation of line losses of the power grid under normal operation, that is, with actual load. This calculation process is to use the real-time line loss rate of a distribution line minus the corresponding no-load line loss rate to calculate the variable line loss rate. This variable line loss rate reflects the change in line loss caused by changes in load current. If the variable line loss rate is greater than the preset variable line loss rate threshold, it means that there are some abnormal conditions in the power grid, such as line aging, uneven load distribution, equipment failure, etc. This preset variable line loss rate is set based on historical operating experience and safety standards. In this case, in order to further diagnose the problem and take corresponding measures, the system terminal collects multiple line load status data sets of multiple distribution lines, and each line load status data set corresponds to a distribution line one by one. These line load status data sets include real-time parameters such as current, voltage, power factor, etc. on the line, as well as information such as the temporal and spatial distribution of load, which will be used for further analysis and processing.

Dynamic loss analysis is performed based on the multiple line load status data sets to generate multiple dynamic line loss rates.

In one embodiment, dynamic loss analysis is a key analysis process in line loss detection, which is used to evaluate the line loss caused by variable factors in line operation, such as load changes, ambient temperature changes, etc. These variable factors will cause changes in parameters such as resistance and inductance in the line, which in turn affect the line loss. When performing dynamic loss analysis, the system terminal calls the dynamic loss identification model to calculate the dynamic loss of each line based on the obtained multiple line load status data sets. Dynamic loss refers to the variable line loss caused by variable factors in line operation, except for the no-load line loss itself. Dynamic loss changes in real time and is related to multiple factors such as the load state of the line and the ambient temperature. Through dynamic loss analysis, the system terminal obtains multiple dynamic line loss rates. These dynamic line loss rates reflect the line loss under different load conditions. In theory, the sum of the no-load line loss and the dynamic line loss of each line should be equal to the real-time line loss. Therefore, dynamic loss analysis is an important tool for power grid line loss detection, fault diagnosis, etc. By real-time monitoring and analysis of the dynamic loss of the line, the line loss can be more accurately understood and potential problems can be discovered in time.

Furthermore, the present application provides a method for performing dynamic loss analysis based on the multiple line load status data sets to generate multiple dynamic line loss rates, including:

The multiple line load status data sets are screened and reduced in dimension according to a preset load factor set to generate multiple reduced-dimensional load data; the multiple reduced-dimensional load data are analyzed through a dynamic loss identification model to output the multiple dynamic line loss rates; wherein the dynamic loss identification model includes multiple single-factor identification channels and a factor-weighted fusion channel, and the construction method of the dynamic loss identification model includes.

Preferably, in the process of performing dynamic loss analysis to generate multiple dynamic line loss rates, the system terminal first screens and reduces the dimensions of multiple line load status data sets. The purpose of this step is to select key data with a high correlation with the dynamic line loss rate from a large amount of complex data, so as to more accurately evaluate the line loss situation. The process of screening and reducing dimensions is based on a preset set of load factors. This preset set of load factors is obtained by associating and identifying multiple initial load factor record sequences and line loss rate record sequences, and then judging the identification results, which is used to guide the direction and scope of data screening. By applying these load factors, the system terminal can identify load status data that has a significant impact on dynamic losses and retain them, while excluding those data that are not highly correlated with dynamic losses. After screening and reducing dimensions, the system terminal generates multiple reduced-dimensional load data. These reduced-dimensional load data contain key information closely related to dynamic losses, and the amount of data is more concise than the original data set, which is convenient for subsequent analysis and processing. Subsequently, the system terminal uses a pre-built dynamic loss identification model to analyze multiple reduced-dimensional load data. This model includes multiple single-factor identification channels and a factor weighted fusion channel, which can identify and fuse the dynamic loss characteristics in the load data. By inputting the reduced-dimensional load data, the model can output the corresponding multiple dynamic line loss rates. By real-time monitoring and analysis of the dynamic line loss rate, the loss status of the line can be understood more accurately and early warning can be issued.

The line loss detection log is collected according to the preset load factor set, wherein the line loss detection log includes multiple groups of line loss detection records, and any group of line loss detection records includes multiple preset load factor record data and multiple line loss rate record data; the multiple single factor identification channels are trained using the multiple groups of line loss detection records, and multiple convergence accuracy rates are recorded.

Preferably, the dynamic loss identification model is of great significance in line loss detection and is used to accurately evaluate the line loss under different load conditions. This model consists of multiple single factor identification channels and a factor weighted fusion channel to ensure a comprehensive analysis of line loss from multiple angles. In the process of model construction, the system terminal first collects line loss detection logs from the historical line loss database according to a preset set of load factors. These logs contain multiple sets of line loss detection records, each of which records in detail the multiple load factor data at the time, such as load size, current, voltage, etc., as well as the corresponding line loss rate data. Subsequently, each single factor identification channel in the dynamic loss identification model is trained using the multiple sets of line loss detection records obtained. Each single factor identification channel focuses on extracting information related to the line loss rate from a specific load factor data. Specifically, the system terminal divides the collected multiple sets of line loss detection records into a training set and a validation set to ensure that the division of the data set is representative and non-overlapping. Subsequently, for each single factor identification channel to be constructed, a single load factor data corresponding to the channel is extracted from each set of line loss detection records in the training set as a feature. Afterwards, a neural network structure is constructed for each single-factor identification channel based on a multi-layer perceptron (MLP) to form multiple initial single-factor identification channels, and random numbers are used to initialize the weights and biases of these single-factor identification channels. Then, the system terminal uses the feature data and the corresponding line loss rate data to train each initial single-factor identification channel. The parameters of the initial single-factor identification channel are iteratively updated through the back-propagation algorithm and the gradient descent optimizer, so that the prediction loss of the initial single-factor identification channel on the training set gradually decreases. After each training cycle, the system terminal uses multiple sets of line loss detection records in the validation set to evaluate the performance of the corresponding initial single-factor identification channel. The loss function value and accuracy on the validation set are calculated. As the training progresses, the performance of the initial single-factor identification channel on the validation set will gradually improve and stabilize. At this time, the system terminal will judge the initial single-factor identification channel according to the set early stopping strategy. This early stopping strategy is set based on the accuracy requirements. For example, on the validation set, the performance of the initial single-factor identification channel has not improved for five consecutive cycles. When the initial single-factor identification channel meets the early stopping condition, the system terminal determines that the initial single-factor identification channel has converged, and records the accuracy of each single-factor identification channel at the time of convergence. This accuracy reflects the prediction ability of the channel on a specific load factor. Finally, the system terminal outputs multiple converged initial single-factor identification channels to generate multiple single-factor identification channels. Through this training process, each channel will learn how to predict the corresponding line loss rate from its corresponding load factor data.

The factor weighted fusion channel is constructed by training with the multiple convergence accuracy rates, and the multiple single factor identification channels are connected with the factor weighted fusion channel to obtain the dynamic loss identification model.

Preferably, in the process of constructing a dynamic loss identification model, the training and construction of a factor-weighted fusion channel is an important step. The purpose of this step is to effectively fuse the outputs of multiple single-factor identification channels to obtain a more comprehensive and accurate dynamic loss prediction. The system terminal first partitions the factor-weighted fusion channel, that is, divides the line loss rate storage address output by each single-factor identification channel. Subsequently, the system terminal assigns weights to the storage addresses corresponding to each single-factor identification channel according to the obtained multiple convergence accuracies, and the storage addresses corresponding to the single-factor identification channels with higher convergence accuracy will obtain larger weights. After that, the system terminal inputs each group of line loss detection records into the trained multiple single-factor identification channels in turn, and calculates the line loss rates of multiple preset load factors for each group of line loss detection records. Then, the system terminal inputs the outputs of multiple single-factor identification channels into the factor-weighted fusion channel for weighted summation, and compares the weighted summation result with the sum of multiple line loss rate record data of each group of line loss detection records to calculate the mean square error. Furthermore, the system terminal iteratively updates the weights in the factor-weighted fusion channel through the back-propagation algorithm and the gradient descent optimizer, so that the error between the predicted line loss rate and the actual value gradually decreases. After the training is completed, the system terminal connects the output ends of multiple single-factor identification channels with the input ends of the factor-weighted fusion channel to form a complete dynamic loss identification model. The output of each single-factor identification channel is weighted according to its weight in the factor-weighted fusion channel, and then fused to obtain a final dynamic loss prediction value. Through the above steps, a dynamic loss identification model that can comprehensively consider the influence of multiple load factors and has high prediction accuracy can be constructed. This model can provide strong support for line loss detection.

Furthermore, the present application provides filtering and reducing the dimensions of the plurality of line load status data sets according to a preset load factor set to generate a plurality of reduced-dimensional load data, including:

An initial load factor set is constructed; and line loss detection record mining is performed using the initial load factor set as a variable to generate multiple initial load factor record sequences and line loss rate record sequences.

Optionally, in the process of building a dynamic loss identification model, a key step is to determine and process the load factor set. The system terminal first builds an initial load factor set. This set contains various factors that affect the line loss of the power grid, such as load size, load type, voltage level, equipment aging, etc. These factors are crucial for understanding and analyzing energy loss in the power grid. Subsequently, the system terminal uses this initial load factor set as a variable to mine the line loss detection records. The purpose of this step is to extract line loss information related to the load factor from a large amount of historical data. Through data mining technology, multiple initial load factor record sequences and corresponding line loss rate record sequences can be generated. The load factor record sequence describes the specific value or state of each load factor at different times or under different conditions. The line loss rate record sequence records the actual energy loss of the power grid under these conditions. There is a one-to-one correspondence between the two sequences, because changes in load factors affect the energy loss of the power grid. By generating these sequences, the relationship between load factors and line losses can be analyzed and understood more deeply. This not only helps to identify the key factors affecting line loss, but also provides strong data support for the subsequent construction of a dynamic loss identification model.

The multiple initial load factor record sequences and the line loss rate record sequence are associated and identified to generate multiple correlation degrees; based on the multiple correlation degrees, initial load factors with correlation degrees greater than or equal to a predetermined correlation degree are extracted to construct the preset load factor set.

Optionally, after obtaining multiple initial load factor record sequences and line loss rate record sequences, the system terminal associates and identifies these sequences. The purpose of this step is to find out which load factors have a strong correlation with the line loss rate. Specifically, the system terminal randomly extracts an initial load factor record sequence from multiple initial load factor record sequences as the first initial load factor record sequence. Subsequently, the system terminal aligns the first initial load factor record sequence and the line loss rate record sequence in time sequence, calculates the absolute difference at the corresponding positions, and extracts the maximum absolute difference and the minimum absolute difference. After that, the system terminal sets a resolution coefficient, which is used to adjust the sensitivity of the correlation coefficient to the absolute difference. In practical applications, it is usually set to 0.5. Then, the system terminal adds the minimum absolute difference to the product of the resolution coefficient and the maximum absolute difference, and calculates the absolute difference between the corresponding positions of the first initial load factor record sequence and the line loss rate record sequence. Then, the calculated absolute difference is added to the product of the resolution coefficient and the maximum absolute difference. Further, the system terminal calculates the ratio of the sum of the first calculation to the sum of the second calculation to obtain the correlation coefficient of the current position. Subsequently, this process is repeated until the correlation coefficients of all positions of the first initial load factor record sequence and the line loss rate record sequence are calculated. After that, the system terminal calculates the mean of all the calculated correlation coefficients to obtain the first correlation degree. Then, the above process is repeated, and the system terminal obtains multiple correlation degrees, each of which corresponds to an initial load factor record sequence. Finally, the system terminal extracts the initial load factors whose correlation degrees are greater than or equal to the predetermined correlation degrees based on the multiple correlation degrees calculated. These load factors are closely related to the line loss rate, so they have important reference value for predicting and analyzing the line loss rate. The system terminal then combines these extracted load factors to construct a preset load factor set. The load factors in this set are factors that have a significant impact on the line loss rate, and they will serve as important input variables in the subsequent construction of the dynamic loss identification model.

The multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates are combined to generate alarm information to issue an abnormal line loss alarm.

In one embodiment, after obtaining multiple dynamic line loss rates, multiple no-load line loss rates, and multiple real-time line loss rates. The system terminal identifies the deviation of the three line loss rates. In theory, the dynamic line loss rate plus the no-load line loss rate should be equal to or close to the real-time line loss rate. If the sum of the dynamic line loss rate and the no-load line loss rate deviates too much from the real-time line loss rate, this indicates that there is management line loss. Management line loss refers to the loss of electric energy caused by errors in metering equipment, mistakes in meter reading, electricity theft, and poor management during the transmission, transformation, distribution, and power supply processes. In order to promptly discover and handle these management line loss problems, the system terminal generates alarm information, issues an abnormal line loss alarm, and sends relevant information to the operator in the form of an alarm message so that timely measures can be taken to reduce management line losses and improve the operating efficiency of the power grid.

Furthermore, the present application provides a method for generating alarm information by combining the multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates to perform line loss abnormality alarm, including:

The multiple dynamic line loss rates and the multiple no-load line loss rates are superimposed and calculated to generate multiple reference line loss rates; determine whether the line loss rate deviation between the multiple reference line loss rates and the multiple real-time line loss rates meets a preset deviation threshold, and if so, generate a first alarm information to issue a line loss abnormality alarm.

Preferably, in order to determine whether the line loss is abnormal, the system terminal performs superposition calculations on multiple dynamic line loss rates and no-load line loss rates to generate multiple reference line loss rates. These reference line loss rates represent the expected line losses under theoretical conditions, that is, without management problems. Subsequently, the system terminal compares these reference line loss rates with multiple real-time line loss rates measured in real time to determine whether there is a significant line loss rate deviation. If the deviation between the reference line loss rate and the real-time line loss rate does not exceed the preset deviation threshold, the system terminal will generate a first alarm message to issue an abnormal line loss alarm. This alarm message indicates that no abnormality has been found in the management at this time, that is, the line loss abnormality is not caused by poor management. Therefore, the line loss abnormality is likely to be caused by abnormal line operation, for example, the operating voltage and current are too large, or there are physical reasons such as leakage. This alarm mechanism helps operation and maintenance personnel to quickly locate and handle line operation abnormalities, eliminate safety hazards in a timely manner, and ensure the stable and efficient operation of the power grid.

Further, the present application provides a method for determining whether a line loss rate deviation between the multiple reference line loss rates and the multiple real-time line loss rates meets a preset deviation threshold, including:

If the line loss rate deviation does not meet the preset deviation threshold, locate the abnormal management line; locate the abnormal management node of the abnormal management line, and generate second alarm information with the abnormal management node to issue a distribution management abnormal alarm.

Optionally, when the line loss rate deviation exceeds a preset deviation threshold, this indicates that there is a management anomaly. In order to solve this problem, the system terminal first locates the lines whose line loss rate deviation exceeds the preset deviation threshold. And these lines are defined as management abnormality lines. Once the management abnormality lines are determined, the system terminal collects the historical electric energy metering timing information of the electric energy metering nodes in the management abnormality lines. Subsequently, based on the collected historical electric energy metering timing information, the metering cycle anomaly is identified, and a second alarm message is generated to alert the operator to the anomaly in the distribution management. This alarm message is intended to help operators quickly locate the problem and take effective measures to repair it in order to restore the normal operation of the power grid and reduce power loss.

Furthermore, the present application provides a method for locating management abnormal nodes on the management abnormal line, including:

Acquire multiple electric energy metering nodes in the abnormal management line; collect multiple historical electric energy metering time series information of the multiple electric energy metering nodes, and the deadline of any historical electric energy metering time series information is the current time.

Optionally, when an abnormal management line is detected, in order to further analyze and locate the root cause of the problem, the system terminal first obtains multiple power metering nodes on these abnormal lines. These power metering nodes are directly connected to the user end and are responsible for real-time monitoring and recording of the user's power usage. Subsequently, multiple historical power metering time series information of these metering nodes are collected. This information records in detail the power usage data of each metering node from a certain point in the past until the current moment. These time series data are crucial for analyzing power loss and anomalies because they can reflect the user's power usage behavior in various time periods. By collecting these historical power metering time series information, the system terminal can more accurately understand which users or areas have abnormal power usage, and then locate possible management problems, such as power theft, equipment failure or human error.

Metering cycle anomalies are identified on the multiple historical electric energy metering time series information to obtain cycle anomaly nodes; and the second alarm information is generated based on the cycle anomaly nodes.

Optionally, after the power management abnormal line is found, in order to further analyze the specific cause of the problem, the system terminal conducts in-depth metering cycle anomaly identification on the historical power metering time series information of multiple power metering nodes. Since the metering node is directly connected to the user, the user's power consumption usually shows certain regular changes over a period of time, such as the daily peak power consumption period, the difference between weekend and weekday power consumption, etc. The system terminal detects whether abnormal changes in power consumption have occurred in the regular cycle by comparing and analyzing these historical power metering time series information. If the power consumption of a metering node shows a significant deviation or abnormal fluctuation in the regular cycle, the system terminal determines that the metering node has a periodic anomaly and marks it as a periodic abnormal node. Once the periodic abnormal node is identified, the system terminal generates a second alarm message based on these abnormal nodes. These alarm messages will clearly indicate which metering nodes have periodic abnormalities, and include detailed information such as the time period when the abnormality occurs and the specific manifestations of the abnormality. After receiving these alarm messages, the operator can quickly locate the specific abnormal metering node, and then inspect and repair the metering equipment to ensure the normal operation and accurate metering of the power grid.

In the above, a smart line loss detection and alarm method for a distribution network according to an embodiment of the present invention is described in detail with reference to Fig. 1. Next, a smart line loss detection and alarm device for a distribution network according to an embodiment of the present invention will be described with reference to Fig. 2.

According to an embodiment of the present invention, a smart line loss detection and alarm device for a distribution network is used to solve the technical problem that the line loss analysis method based on direct judgment of thresholds cannot provide targeted warnings due to the various causes of line loss, and has low auxiliary power for subsequent investigation of abnormal line loss causes, thereby achieving the effect of improving the accuracy of line loss detection and making the operation and management of the distribution network more efficient and accurate. A smart line loss detection and alarm device for a distribution network includes: a no-load loss identification module 1, a data analysis module 2, an energy loss identification module 3, a variable line loss rate identification module 4, a dynamic loss analysis module 5, and a line loss abnormality alarm module 6.

No-load loss identification module 1: The no-load loss identification module 1 is used to obtain multiple line design information of multiple distribution lines of the target distribution network, identify line no-load losses according to the multiple line design information, and generate multiple no-load line loss rates;

Data analysis module 2: the data analysis module 2 is used to receive the preset power distribution plan of the target distribution network in the preset time zone, and collect the power sales information in the preset time zone;

Power loss identification module 3: the power loss identification module 3 is used to identify power loss of the plurality of power distribution lines according to the power sales information and the preset power distribution plan, and generate a plurality of real-time line loss rates;

Variable line loss rate identification module 4: the variable line loss rate identification module 4 is used to identify the variable line loss rate according to the multiple real-time line loss rates and the multiple no-load line loss rates, and if the variable line loss rate is greater than the preset variable line loss rate, collect multiple line load status data sets of the multiple distribution lines;

Dynamic loss analysis module 5: the dynamic loss analysis module 5 is used to perform dynamic loss analysis based on the multiple line load status data sets to generate multiple dynamic line loss rates;

Line loss abnormality alarm module 6: The line loss abnormality alarm module 6 is used to generate alarm information by combining the multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates to issue a line loss abnormality alarm.

Furthermore, the no-load loss identification module 1 further includes:

Based on the multiple line design information, the first line design information of the first distribution line is extracted, wherein the first line design information includes the specifications and quantity of the grid elements in the first distribution line; the no-load losses of the grid elements are cumulatively calculated according to the first line design information to generate a first no-load loss; based on the rated distribution power of the first distribution line and the first no-load loss, a first no-load line loss rate is calculated and obtained, and added to the multiple no-load line loss rates.

Furthermore, the dynamic loss analysis module 5 also includes:

The multiple line load status data sets are screened and reduced in dimension according to a preset load factor set to generate multiple reduced-dimensional load data; the multiple reduced-dimensional load data are analyzed by a dynamic loss identification model to output the multiple dynamic line loss rates; wherein the dynamic loss identification model includes multiple single-factor identification channels and a factor-weighted fusion channel, and the method for constructing the dynamic loss identification model includes: collecting line loss detection logs according to the preset load factor set, wherein the line loss detection logs include multiple groups of line loss detection records, and any group of line loss detection records includes multiple preset load factor record data and multiple line loss rate record data; using the multiple groups of line loss detection records to train the multiple single-factor identification channels, and recording multiple convergence accuracy rates; using the multiple convergence accuracy rates to train and construct the factor-weighted fusion channel, connecting the multiple single-factor identification channels with the factor-weighted fusion channel, and obtaining the dynamic loss identification model.

Furthermore, the dynamic loss analysis module 5 also includes:

Construct an initial load factor set; use the initial load factor set as a variable to perform line loss detection record mining to generate multiple initial load factor record sequences and line loss rate record sequences; perform association identification on the multiple initial load factor record sequences and the line loss rate record sequences to generate multiple association degrees; based on the multiple association degrees, extract the initial load factors whose association degrees are greater than or equal to a predetermined association degree to construct the preset load factor set.

Furthermore, the line loss abnormality alarm module 6 further includes:

The multiple dynamic line loss rates and the multiple no-load line loss rates are superimposed and calculated to generate multiple reference line loss rates; determine whether the line loss rate deviation between the multiple reference line loss rates and the multiple real-time line loss rates meets a preset deviation threshold, and if so, generate a first alarm information to issue a line loss abnormality alarm.

Furthermore, the line loss abnormality alarm module 6 further includes:

If the line loss rate deviation does not meet the preset deviation threshold, locate the abnormal management line; locate the abnormal management node of the abnormal management line, and generate second alarm information with the abnormal management node to issue a distribution management abnormal alarm.

Furthermore, the line loss abnormality alarm module 6 further includes:

Acquire multiple electric energy metering nodes in the management abnormal line; collect multiple historical electric energy metering timing information of the multiple electric energy metering nodes, and the deadline of any historical electric energy metering timing information is the current time; identify the metering cycle abnormality of the multiple historical electric energy metering timing information to obtain the cycle abnormality node; generate the second alarm information with the cycle abnormality node.

An intelligent line loss detection and alarm device for a distribution network provided in an embodiment of the present invention can execute an intelligent line loss detection and alarm method for a distribution network provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

Although the present application makes various references to certain modules in the device according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and/or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims (8)

1. A smart line loss detection and alarm method for a distribution network, characterized in that the method comprises: Acquire multiple line design information of multiple distribution lines of the target distribution network, identify line no-load losses according to the multiple line design information, and generate multiple no-load line loss rates; Receiving a preset power distribution plan of the target distribution network in a preset time zone, and collecting power sales information in the preset time zone; Identifying power loss of the plurality of power distribution lines according to the power sales information and the preset power distribution plan, and generating a plurality of real-time line loss rates; Identify a variable line loss rate according to the multiple real-time line loss rates and the multiple no-load line loss rates, and if the variable line loss rate is greater than a preset variable line loss rate, collect multiple line load status data sets of the multiple distribution lines; Performing dynamic loss analysis based on the multiple line load status data sets to generate multiple dynamic line loss rates; The multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates are combined to generate alarm information to issue an abnormal line loss alarm. 2. The method according to claim 1, characterized in that the step of identifying line no-load loss according to the plurality of line design information and generating a plurality of no-load line loss rates comprises: Extracting first line design information of a first distribution line according to the plurality of line design information, wherein the first line design information includes specifications and quantity of grid elements in the first distribution line; Accumulate and calculate the no-load loss of the power grid components according to the first line design information to generate a first no-load loss; The first no-load line loss rate is calculated and obtained by combining the rated distribution power of the first distribution line and the first no-load loss, and is added to the multiple no-load line loss rates. 3. The method according to claim 1, characterized in that the generating of alarm information for abnormal line loss alarm by combining the multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates comprises: Superimposing and calculating the multiple dynamic line loss rates and the multiple no-load line loss rates to generate multiple reference line loss rates; Determine whether the line loss rate deviation between the multiple reference line loss rates and the multiple real-time line loss rates meets a preset deviation threshold. If so, generate a first alarm message to issue a line loss abnormality alarm. 4. The method according to claim 3, characterized in that judging whether the line loss rate deviations between the multiple reference line loss rates and the multiple real-time line loss rates meet a preset deviation threshold further comprises: If the line loss rate deviation does not meet the preset deviation threshold, locate and manage the abnormal line; The abnormal management node is located for the abnormal management line, and the abnormal management node is used to generate second alarm information to issue a power distribution management abnormality alarm. 5. The method according to claim 4, characterized in that locating the abnormal management node of the abnormal management line comprises: Acquire multiple electric energy metering nodes in the abnormal management line; Collecting multiple historical electric energy metering time series information of the multiple electric energy metering nodes, and the end time of any historical electric energy metering time series information is the current time; Performing measurement cycle anomaly identification on the plurality of historical electric energy measurement time series information to obtain a cycle anomaly node; The second alarm information is generated using the periodic abnormal node. 6. The method according to claim 1, wherein the step of performing dynamic loss analysis based on the plurality of line load status data sets to generate a plurality of dynamic line loss rates comprises: Screening and reducing the dimensions of the plurality of line load status data sets according to a preset load factor set to generate a plurality of reduced-dimensional load data; Analyze the multiple reduced-dimensional load data by using a dynamic loss identification model, and output the multiple dynamic line loss rates; The dynamic loss identification model includes multiple single factor identification channels and a factor weighted fusion channel, and the construction method of the dynamic loss identification model includes: Collecting a line loss detection log according to the preset load factor set, wherein the line loss detection log includes multiple groups of line loss detection records, and any group of line loss detection records includes multiple preset load factor record data and multiple line loss rate record data; Using the multiple groups of line loss detection records to train the multiple single factor recognition channels, and recording multiple convergence accuracy rates; The factor weighted fusion channel is constructed by training with the multiple convergence accuracy rates, and the multiple single factor identification channels are connected with the factor weighted fusion channel to obtain the dynamic loss identification model. 7. The method according to claim 6, characterized in that the plurality of line load status data sets are screened and reduced in dimension according to a preset load factor set to generate a plurality of reduced-dimensional load data, comprising: Construct an initial set of load factors; Using the initial load factor set as a variable, line loss detection record mining is performed to generate multiple initial load factor record sequences and line loss rate record sequences; Associating and identifying the multiple initial load factor record sequences and the line loss rate record sequence to generate multiple association degrees; According to the multiple association degrees, initial load factors having an association degree greater than or equal to a predetermined association degree are extracted to construct the preset load factor set. 8. An intelligent line loss detection and alarm device for a distribution network, characterized in that the device is used to implement an intelligent line loss detection and alarm method for a distribution network according to any one of claims 1 to 7, comprising: No-load loss identification module: obtains multiple line design information of multiple distribution lines of the target distribution network, identifies line no-load losses according to the multiple line design information, and generates multiple no-load line loss rates; Data analysis module: receiving a preset power distribution plan of the target distribution network in a preset time zone, and collecting power sales information in the preset time zone; An energy loss identification module: identifies energy losses of the plurality of distribution lines according to the electricity sales information and the preset power distribution plan, and generates a plurality of real-time line loss rates; A variable line loss rate identification module: performs variable line loss rate identification according to the multiple real-time line loss rates and the multiple no-load line loss rates, and if the variable line loss rate is greater than a preset variable line loss rate, collects multiple line load status data sets of the multiple distribution lines; Dynamic loss analysis module: performs dynamic loss analysis based on the multiple line load status data sets to generate multiple dynamic line loss rates; Line loss abnormality alarm module: generates alarm information by combining the multiple dynamic line loss rates, the multiple no-load line loss rates and the multiple real-time line loss rates to issue a line loss abnormality alarm.