Bus loss analysis system based on balanced dynamic threshold

Abstract

The invention discloses a bus loss analysis system based on a balance dynamic threshold, which belongs to the technical field of power system data and comprises a data acquisition module, a bus analysis module and a balance evaluation module. The data acquisition module is used for collecting operation data of a bus and topological structure information of a power grid where the bus is located in real time, and constructing a topological network structure diagram of the bus in the power grid; the bus balance module is used for carrying out preliminary classification and secondary clustering analysis on buses and configuring different balance thresholds and line loss calculation algorithms for different types of buses; the balance evaluation module is used for monitoring the balance state of the bus electric quantity in the power grid in real time, judging whether the bus electric quantity balance is abnormal or not and analyzing the abnormal phenomenon.

Description

Technical Field

[0001] The invention relates to a bus loss analysis system based on a balanced dynamic threshold value, belonging to the technical field of power system data. Background Art

[0002] The busbar is an important component of the power system, which carries the task of transmitting and distributing electric energy. During the process of power transmission and distribution, the existence of electric current will cause certain losses in the power system, which can be divided into load losses and no-load losses. The existence of busbar losses will not only lead to waste of electric energy and increase the pressure on power load and energy supply, but may also cause busbar overheating, aging or damage due to long-term accumulation of heat, affecting the reliability and safety of the power system. With the development of the power system, the monitoring and analysis of busbar losses has become increasingly important. Real-time monitoring and analysis of busbar losses can help operation and maintenance personnel to promptly discover and deal with potential faults and problems, improve the reliability and safety of the power system, and provide strong technical support and decision-making basis for the optimized operation and energy saving and consumption reduction of the power system.

[0003] The existing invention patent with publication number CN117114494A discloses a method and system for intelligent diagnosis of substation line loss, which obtains daily data of the substation in the time period to be detected and calculates the substation line loss information; determines whether the substation line loss is abnormal, and then determines the type of substation line loss; if the substation line loss is abnormal, conducts intelligent diagnosis and analysis of the substation line loss to determine the influencing factors causing the substation line loss to be abnormal; based on the substation line loss type, conducts a thorough check and technical line loss analysis on the substation electric energy meter, determines the types of various line losses and determines the main influencing factors causing the substation line loss to be abnormal, and obtains the abnormal user details; based on the obtained judgment of various line loss types, the main influencing factors and the abnormal user details, the intelligent diagnosis result of the substation line loss is obtained according to the frequency of occurrence of the influencing factors.

[0004] Although the existing technology has formed a set of intelligent diagnosis methods and processes for substation line losses, which are based on the experience of substation line loss management and have high practicality and convenience, and have been verified to be convenient and reliable in practice, it does not take into account the changes in the operating state of the power system, which causes fluctuations in the busbar power balance state, thereby affecting the applicability of the balance threshold and the calculation results of the line loss.

[0005] Therefore, the present application provides a bus loss analysis system based on a dynamic balance threshold, which performs preliminary classification and secondary clustering analysis on the bus in real time, configures different types of buses with different balance thresholds and line loss calculation algorithms according to the final clustering results, and realizes dynamic update of the bus balance threshold, effectively avoiding the inapplicability of the balance threshold due to changes in clustering results; and monitors the balance state of the bus power in the power grid in real time, and analyzes abnormal conditions of the bus power balance. Summary of the invention

[0006] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a bus loss analysis system based on a balanced dynamic threshold, which performs preliminary classification and secondary clustering analysis on the bus in real time, configures the balanced threshold and line loss calculation algorithm, and monitors the balance state of the bus electricity in the power grid in real time.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A bus loss analysis system based on balanced dynamic threshold, comprising: a data acquisition module, a bus analysis module and a balanced evaluation module;

[0009] The data acquisition module is used to collect the operation data of the busbar and the topological structure information of the power grid where the busbar is located in real time, and to construct a topological network structure diagram of the busbar in the power grid;

[0010] The busbar balancing module is used to perform preliminary classification and secondary clustering analysis on the busbars, and configure different balancing thresholds and line loss calculation algorithms for different types of busbars;

[0011] The balance assessment module is used to monitor the balance state of the busbar electricity in the power grid in real time, determine whether the busbar electricity balance is abnormal, and analyze abnormal phenomena.

[0012] Specifically, the busbar analysis module includes a classification unit, a cluster analysis unit and a configuration update unit;

[0013] The classification unit obtains characteristic parameters of the busbar in the topological network structure diagram, and performs preliminary classification of the busbar based on the characteristic parameters;

[0014] The cluster analysis unit is used to perform secondary cluster analysis on the busbar data to refine the busbar type;

[0015] The configuration updating unit is used to configure different balancing thresholds and line loss calculation algorithms for different types of buses based on the clustering results.

[0016] Specifically, the classification unit is configured with a classification strategy for preliminary classification of the busbars;

[0017] The classification strategy includes:

[0018] Collect historical bus data, classify and label each sample, and divide the data into training set and test set;

[0019] Extract the key parameters in the training set as feature vectors and construct a feature vector set X = (X1…X a ) T ; Wherein, a is the number of samples in the training set;

[0020] A binary classification SVM model is constructed for each bus category; and the parameters of each binary classification SVM model are initialized; wherein the classification results of each binary classification SVM model are predefined types and other types, and the predefined types of each binary classification SVM model are power plant bus, substation bus, transmission line bus, and distribution line bus;

[0021] Input the training set data into the model, and calculate the inner product between the feature vector of each sample and the current model parameter;

[0022] According to the inner product value, determine the category to which each sample belongs; if Y i >0, the sample belongs to the predefined type; if Y i <0, the sample belongs to other types;

[0023] Calculate the error between the expected classification label and the actual label of the sample And update the model parameters in the opposite direction of the gradient;

[0024] Repeatedly calculate the inner product value and update the model parameters until the number of iterations reaches the preset upper limit;

[0025] The binary SVM model is verified using the test set; if the verification fails, the training set is expanded and the model is continued to be trained; if the verification passes, the binary SVM model is successfully constructed.

[0026] Specifically, the classification strategy also includes:

[0027] Obtain samples of the bus to be classified, calculate the inner product value of the samples using the binary SVM model, and construct a category decision set y=(y1, y2, y3, y4) of the bus to be classified;

[0028] The maximum value max(y1, y2, y3, y4) of the category decision set y is obtained, and the category corresponding to the maximum value is defined as the final category label of the bus to be classified.

[0029] Specifically, the cluster analysis unit is configured with a clustering strategy for performing secondary cluster analysis on the busbar data;

[0030] The clustering strategy includes:

[0031] Obtaining a bus data set and a preliminary classification result corresponding to the bus from the topological network structure diagram to generate a cluster sample set;

[0032] Set clustering parameters, including the number of clusters K, initial cluster center, number of iterations, and distance measurement method;

[0033] According to the selected initial cluster center and the number of clusters K, the cluster sample set is divided into K initial clusters;

[0034] Calculating the distance between each data point in the cluster sample set and each cluster center, and assigning each data point to the nearest cluster center;

[0035] Recalculate the center position of each cluster according to the allocation results, and repeat the iterative update process until the preset number of iterations is reached;

[0036] Calculate the average value of the cluster sample set as the silhouette coefficient of the entire clustering result

[0037] According to the silhouette coefficient Adjust the clustering parameters;

[0038] After adjusting the clustering parameters, re-perform the clustering analysis and calculate the new silhouette coefficient Repeat the above steps until the silhouette coefficient Achieve expected value;

[0039] Generate a busbar clustering results report.

[0040] Specifically, the balance assessment module includes a bus balancing unit and an abnormality analysis unit;

[0041] The bus balancing unit monitors the balance state of the power on the bus in the power grid in real time by calculating the bus power balance value and the bus loss calculation value, and issues a warning signal when the bus power balance exceeds the balance threshold;

[0042] The abnormality analysis unit is used to analyze the bus abnormality in the early warning signal and generate an abnormality analysis report.

[0043] Specifically, the bus balancing unit is configured with a balancing monitoring strategy for monitoring the balancing state of the electric quantity on the bus in the power grid;

[0044] The balance monitoring strategy includes:

[0045] Obtain real-time power data of each bus in the power grid, determine the number of input and output paths of each bus, and record the power data of each path;

[0046] Calculate the input power P of each bus in , output power P out and the unbalance ΔP;

[0047] Calculate the busbar unbalance rate η and line loss rate λ;

[0048] Determine whether there is an abnormal power balance problem on the bus; if η≤1-ηTh And λ≤λ Th , then there is no abnormal power balance problem on the bus; if η>1-η Th or λ>λ Th , then there is an abnormal power balance problem on the busbar, and an early warning signal is generated.

[0049] Specifically, the abnormal analysis unit is configured with a linkage analysis strategy for analyzing a busbar with abnormal power balance problems;

[0050] The linkage analysis strategy includes:

[0051] Obtain the early warning information in the busbar balancing unit and the electrical parameters of the main transformer of the substation;

[0052] Calculate the main transformer loss P loss ;

[0053] Obtain the main transformer loss P of two adjacent busbars in the substation 1,loss , P 2,loss , and calculate the loss difference ΔP loss ;

[0054] Set the seesaw threshold to ΔP Th,loss , determine whether there is a seesaw phenomenon between two adjacent busbars; if ΔP loss <ΔP Th,loss , there is no seesaw phenomenon; if ΔP loss ≥ΔP Th,loss , there is a seesaw phenomenon;

[0055] Set the main transformer loss threshold to P Th , and determine the main transformer loss P loss Is it abnormal? If P loss <P Th , the main transformer loss is normal; if P loss ≥P Th , then there is abnormal loss in the main transformer.

[0056] Specifically, the linkage analysis strategy includes:

[0057] Obtain the electrical parameters of each line under the bus and calculate the line loss rate P of each line under the bus line_loss ;

[0058] Draw the line loss rate of each line into a time series diagram and calculate the correlation coefficient γ between the line loss rates of each line loss ;

[0059] Combine the timing diagram and correlation coefficient to determine whether there is an abnormality in the line loss; if If the time series diagram shows a reverse fluctuation trend, the line loss is abnormal; otherwise, the line loss is not abnormal;

[0060] Generate and output exception analysis reports.

[0061] Beneficial effects of the present invention:

[0062] 1. The busbar analysis module performs real-time preliminary classification and secondary clustering analysis on the busbars, configures different balance thresholds and line loss calculation algorithms for different types of busbars, and identifies busbar groups with similar electrical characteristics and operating conditions, thereby configuring unified balance thresholds and line loss calculation algorithms to improve the efficiency and accuracy of the analysis; when the busbar type changes, the balance threshold can be dynamically updated to ensure the accuracy of the line loss calculation and the accurate judgment and management of the busbar power balance state, which helps to avoid power imbalance and excessive line loss problems in the power grid;

[0063] 2. Real-time monitoring of the busbar power balance in the power grid. When the busbar power balance is abnormal, an alarm is immediately issued and a linkage analysis strategy is quickly initiated. By integrating other relevant data and information in the power grid, the abnormal location is accurately determined, shortening the time for abnormal location, reducing the workload of operation and maintenance personnel, and helping to avoid equipment damage and power outages. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a structural diagram of a bus loss analysis system based on balanced dynamic threshold;

[0065] Figure 2 It is a classification flow chart of bus loss analysis system based on balanced dynamic threshold;

[0066] Figure 3 It is a cluster analysis flow chart of bus loss analysis system based on balanced dynamic threshold;

[0067] Figure 4 The present invention is a flow chart of bus loss analysis system balance monitoring based on balance dynamic threshold.

[0068] Figure 5 The figure is a linkage analysis flow chart of a bus loss analysis system based on a balanced dynamic threshold. DETAILED DESCRIPTION

[0069] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0070] Example 1

[0071] refer to Figures 1 to 5 As shown, this embodiment introduces a bus loss analysis system based on a balanced dynamic threshold, including: a data acquisition module, a bus analysis module and a balanced evaluation module;

[0072] The data acquisition module is used to collect real-time bus operation data, such as voltage, current, active power and reactive power, as well as topological information of the power grid where the bus is located, including: collecting detailed information of bus connection equipment, such as equipment number, rated power, type, recording the connection method between equipment, such as direct connection, connection through switch equipment, and recording the layout information of the power grid, such as the direction of the bus and the location of the equipment; cleaning, integrating and verifying the collected data to ensure the accuracy and completeness of the data; and constructing a topological network structure diagram of the bus in the power grid, in which the equipment connected to the bus is regarded as a node in the topological network, and the bus is regarded as an edge in the topological network. In order to enhance the readability and aesthetics of the structure diagram, different colors, line thicknesses and shapes are used to distinguish different types of equipment and connection methods, and text labels and annotations are added to the equipment, connecting lines and key nodes to explain the important information of the equipment name, number and connection method;

[0073] The busbar analysis module is used to perform real-time preliminary classification and secondary clustering analysis on the busbars, and configure different balance thresholds and line loss calculation algorithms for different types of busbars. Since the electrical characteristics and operating conditions of the busbars are affected by many factors, the classification and secondary clustering results of the busbars may change. The topological network structure diagram of the busbars in the power grid is used for real-time analysis to generate busbar clustering results, determine whether the busbar type has changed, and set a new balance threshold and corresponding line loss calculation algorithm for the busbar whose type has changed, and overwrite the original balance threshold of the busbar to achieve dynamic update of the busbar balance threshold. For example, a new busbar is added to the system, or the electrical characteristics of the existing busbar change, resulting in the original balance threshold no longer being applicable. By continuously updating the balance threshold, it helps to improve the accuracy of the line loss calculation, ensure that the busbar power balance state is accurately judged and managed, and thus ensure the safe and stable operation of the power system.

[0074] The balance assessment module is used to monitor the balance state of busbar electricity in the power grid in real time, and judge whether the busbar electricity balance is abnormal based on the configured threshold. Once an abnormal phenomenon is found, it will be analyzed in time, and the electricity balance problem at the bus level and the line loss abnormality at the line level will be analyzed separately. The abnormal location can be accurately judged through the linkage analysis strategy, reducing the workload of operation and maintenance personnel, avoiding equipment damage and power outages, and thus reducing operation and maintenance costs.

[0075] Specifically, the busbar analysis module includes a classification unit, a cluster analysis unit, and a configuration update unit;

[0076] The classification unit obtains characteristic parameters of the bus in the topological network structure diagram, such as location, function, connection equipment, voltage level, and current capacity, and preliminarily classifies the bus based on the characteristic parameters. The classification results include but are not limited to power plant bus, substation bus, transmission line bus, and distribution line bus;

[0077] The cluster analysis unit uses clustering algorithms, such as the K-means algorithm, to perform secondary clustering analysis on bus data, refine bus types, such as subdividing substation buses into high-voltage side buses and low-voltage side buses, and generate bus clustering results for subsequent configuration of balancing thresholds and line loss algorithms. At the same time, it enables staff to better understand the characteristics and distribution patterns of the bus, providing strong support for the optimized operation of the power system;

[0078] The configuration update unit is used to configure different balance thresholds and line loss calculation algorithms for different types of buses based on the clustering results; analyze the bus clustering results to determine whether the bus type has changed. Once a change occurs, the corresponding balance threshold and line loss calculation algorithm (adjustment of line loss calculation parameters) are scheduled from the preset information library to achieve dynamic update of the balance threshold to ensure the accuracy and reliability of the line loss calculation results; and record the change time and the bus type before and after the change, and update the scheduled balance threshold and line loss calculation algorithm to the bus balance monitoring to achieve real-time monitoring and adjustment of the bus status.

[0079] Specifically, the classification unit configures the classification strategy for preliminary classification of the mother line;

[0080] The specific steps of the classification strategy include:

[0081] Collect a large amount of historical bus data, such as the bus voltage level, current capacity, location, function, type and number of connected devices, and topological structure information, and classify and label each sample so that each sample has a corresponding category label, including power plant bus, substation bus, transmission line bus, and distribution line bus; and divide the data into training set and test set;

[0082] Extract features from the sample data in the training set, select key parameters as feature vectors, such as the voltage level, current capacity, location information, type and number of connected devices, and construct a feature vector set Among them, a is the number of samples in the training set, b is the number of feature vectors, and X a is the ath sample, x a (b) is the value of the bth eigenvector in the ath sample;

[0083] A binary classification SVM model is constructed for each bus category, wherein the classification results of each binary classification SVM model are predefined types and other types, and the predefined types of each binary classification SVM model are power plant bus, substation bus, transmission line bus, and distribution line bus. For example, a model is constructed for the power plant bus category to distinguish the power plant bus samples from all other category samples; and each binary classification SVM model parameter is initialized, such as the penalty parameter C and the kernel function; in this embodiment, a linear kernel is selected as the kernel function of the model;

[0084] Taking the binary classification SVM model of the power plant bus as an example, the training set data is input into the model, and the inner product between the feature vector of each sample and the current model parameter is calculated. The expression is as follows:

[0085] Y i =W·X i +B

[0086] Where W is the weight vector, X i is the i-th sample, B is the bias term, Y i is the inner product value of the i-th sample, i = 1, 2, …, a; the weight vector and bias term are the model parameters to be solved, which are solved and optimized through training;

[0087] According to the inner product value, determine the category to which each sample belongs; if Y i >0, the sample belongs to the power plant bus type; if Y i <0, the sample belongs to other types;

[0088] Calculate the error between the expected classification label and the actual label of the sample As well as the gradient of the error function with respect to the model parameters, the model parameters are updated in the opposite direction of the gradient to gradually reduce the value of the loss function; the expression is as follows:

[0089]

[0090] In the formula, is the true label value, if The real label is the power plant bus. The true label is other; θ is the model parameter, including the weight vector W and the bias term B, θ t is the model parameter value before updating, θ t+1 is the updated model parameter value, ε is the learning rate, which is used to control the step size of parameter update, is the error function Regarding the model parameter θ at time point θ t The gradient value of

[0091] Repeatedly calculate the inner product value and update the model parameters until the number of iterations reaches the preset upper limit;

[0092] The binary classification SVM model of the power plant bus is verified using the test set; if the verification fails, the training set is expanded and the model is further trained; if the verification passes, the binary classification SVM model of the power plant bus is successfully constructed;

[0093] Obtain samples of the busbar to be classified, use 4 trained binary SVM models to calculate the inner product value of the samples, record the inner product value of each model, and retain the sign (positive or negative) and absolute value, and construct the category decision set y = (y1, y2, y3, y4) of the busbar to be classified;

[0094] Get the maximum value max(y1,y2,y3,y4) of the category decision set y, and define the category corresponding to the maximum value as the final category label of the bus to be classified.

[0095] Specifically, the cluster analysis unit is configured with a clustering strategy for performing secondary cluster analysis on the busbar data;

[0096] The specific steps of the clustering strategy include:

[0097] Obtain the bus data set and the preliminary classification results corresponding to the bus from the topological network structure diagram, and preprocess the collected data to generate a cluster sample set, such as checking whether there are missing values, outliers and duplicate data in the data set, using interpolation to process missing values, using the 3σ principle to detect and eliminate outliers, and directly deleting duplicate data. Z-score standardization converts the data into a distribution with a mean of 0 and a standard deviation of 1, and standardizes the data to eliminate the dimensional differences between different indicators;

[0098] Set clustering parameters, including the number of clusters K, initial cluster centers, number of iterations, and distance measurement methods. For example, use the silhouette coefficient method to determine the K value, and use the K-means++ method to select the initial cluster center by calculating the distance between data points in the cluster sample set to accelerate the convergence of the clustering process.

[0099] According to the selected initial cluster center and the number of clusters K, the cluster sample set is divided into K initial clusters;

[0100] For each data point in the cluster sample set, calculate the distance between each data point and each cluster center, and assign each data point to the nearest cluster center;

[0101] The center position of each cluster is recalculated according to the allocation results, and the iterative update process is repeated until the preset number of iterations is reached. The expression is as follows:

[0102]

[0103] In the formula, C k is the kth cluster center, S k is the sample point set of the kth cluster, d j For S k The jth sample point in |S k | is the number of sample points in the kth cluster, j = 1, 2, ..., |S k |;

[0104] Calculate the silhouette coefficient S(l) of each data point and calculate the average value of all sample points as the silhouette coefficient of the entire clustering result The expression is as follows:

[0105]

[0106] In the formula, r(l) is the average distance from sample point l to other sample points in its cluster, t(l) is the minimum value of the average distance from sample point l to all sample points in adjacent clusters, and e is the total number of sample points.

[0107] According to the silhouette coefficient Adjust the clustering parameters; when the value of the silhouette coefficient is low or does not meet expectations, adjust the clustering parameters, for example, increase or decrease the number of clusters K, change the selection method of the initial cluster center; when the silhouette coefficient is large, the clustering effect is better, and the clustering parameters are not adjusted;

[0108] After adjusting the clustering parameters, re-perform the clustering analysis and calculate the new silhouette coefficient If the adjusted silhouette coefficient If the expected value is still not reached, repeat the above steps, that is, adjust the parameters according to the evaluation results, re-perform cluster analysis, and calculate the evaluation index value; if the adjusted silhouette coefficient When the expected value is reached, the clustering parameter adjustment is completed;

[0109] Generate a busbar clustering result report, which provides detailed clustering information, such as the final category label, cluster label, cluster center location, and number of clusters to which each busbar belongs.

[0110] Specifically, the balance assessment module includes a busbar balance unit and an abnormality analysis unit;

[0111] The bus balancing unit calculates the bus power balance value and bus loss calculation value according to the actual situation of the power grid and the principle of bus balancing, and monitors the balance state of the power on the bus in the power grid in real time. When the bus power balance exceeds the balance threshold, an early warning signal is issued;

[0112] The abnormality analysis unit is used to analyze the bus abnormality in the early warning signal. When the bus line loss is abnormal, it triggers the linkage analysis with the main transformer loss and the line line loss, and generates an abnormality analysis report.

[0113] Specifically, a balancing monitoring strategy for monitoring the balancing state of the electric quantity on the busbar in the power grid is configured in the busbar balancing unit;

[0114] Specific steps in a balanced monitoring strategy include:

[0115] Through the power grid monitoring system or sensor network, the real-time power data of each bus in the power grid, including voltage, current, active power and reactive power, is obtained in real time; the number of input paths and output paths of each bus is determined, and the power data of each path is recorded;

[0116] According to the input path and output path data of the bus, calculate the input power P of each bus in , output power P out and the imbalance ΔP, the expression is as follows:

[0117]

[0118] ΔP=P in -P out

[0119] Where N is the number of input paths of bus x, P n,in is the input power of the nth input path of bus x, n=1,2,…,N, M is the number of output paths of bus x, P m,in is the output power of the mth output path of bus x, n=1,2,…,M;

[0120] Calculate the busbar unbalance rate η and line loss rate λ, the expressions are as follows:

[0121]

[0122] Where ΔP load The power change of the load carried by the bus;

[0123] The calculated imbalance rate and line loss rate are compared with the preset balance threshold η Th and line loss threshold λ Th Compare and judge whether there is abnormal power balance problem on the bus; if η≤1-η Th And λ≤λ Th , then there is no abnormal power balance problem on the bus; if η>1-η Th or λ>λ ThIf the bus has an abnormal power balance problem, an early warning signal is generated and transmitted to the abnormality analysis unit. The early warning signal includes the real-time power data, location information, and clustering information of the bus with the abnormal balance.

[0124] Specifically, a linkage analysis strategy for analyzing a bus with abnormal power balance problems is configured in the abnormal analysis unit;

[0125] The specific steps of the linkage analysis strategy include:

[0126] Obtain early warning information in the bus balancing unit and electrical parameters of the substation main transformer, such as current, voltage, and power, to ensure the accuracy and integrity of the data;

[0127] Calculate the main transformer loss P loss , the expression is as follows:

[0128] P loss =I 2 R+P fe +P var

[0129] Where I is the current flowing through the main transformer, R is the resistance of the main transformer winding, and P is fe P is the iron loss, which is the loss caused by the change of magnetic flux in the iron core when the transformer is unloaded or lightly loaded. var is the reactive power loss;

[0130] Obtain the main transformer loss P of two adjacent busbars in the substation 1,loss , P 2,loss , and calculate the loss difference ΔP loss =|P 1,loss -P 2,loss |;

[0131] Set the seesaw threshold to ΔP Th,loss , determine whether there is a seesaw phenomenon between two adjacent busbars; if ΔP loss <ΔP Th,loss , there is no seesaw phenomenon; if ΔP loss ≥ΔP Th,loss , there is a seesaw phenomenon. When the loss of one bus increases, the loss of the other bus decreases accordingly. There is an abnormal loss of the main transformer or a bus connection problem between the two adjacent busbars;

[0132] Set the main transformer loss threshold to P Th , the calculated main transformer loss P loss and the preset loss threshold P Th Compare and determine the main transformer loss P loss Is it abnormal? If P loss <P Th, the main transformer loss is normal; if P loss ≥P Th , then there is abnormal loss in the main transformer;

[0133] Obtain the electrical parameters of each line under the bus and calculate the line loss rate P of each line under the bus line_loss , the expression is as follows:

[0134]

[0135] Where P send is the transmission power, P receive is the received power;

[0136] Draw the line loss rate of each line into a time series diagram and calculate the correlation coefficient γ between the line loss rates of each line loss ;

[0137] Combined with the time series diagram and correlation coefficient, analyze whether there is a seesaw phenomenon between the line loss data of the two lines, and judge whether there is an abnormality in the line loss; if Moreover, the time series diagram shows a reverse fluctuation trend, with an obvious seesaw phenomenon. There is a correlation between the two lines, and the line loss is abnormal; otherwise, there is no abnormality in the line loss;

[0138] Based on the above analysis, an abnormality analysis report is generated and output, which includes the existing abnormal phenomena, causes, and abnormal locations.

[0139] In summary, the present invention performs real-time preliminary classification and secondary clustering analysis on the bus, and configures different balance thresholds and line loss calculation algorithms for different types of buses according to the bus clustering results; applies the configured parameters to the bus power balance evaluation, obtains the bus operation data in real time, monitors the balance state of the bus power in the power grid in real time, and determines whether the bus power balance is abnormal based on the configured threshold, and analyzes the bus abnormality.

[0140] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A bus loss analysis system based on balanced dynamic threshold, characterized in that: include: Data acquisition module, busbar analysis module and balance assessment module; The data acquisition module is used to collect the operation data of the busbar and the topological structure information of the power grid where the busbar is located in real time, and to construct a topological network structure diagram of the busbar in the power grid; The busbar balancing module is used to perform preliminary classification and secondary clustering analysis on the busbars, and configure different balancing thresholds and line loss calculation algorithms for different types of busbars; The balance assessment module is used to monitor the balance state of the busbar electricity in the power grid in real time, determine whether the busbar electricity balance is abnormal, and analyze abnormal phenomena.

2. A bus loss analysis system based on balanced dynamic threshold according to claim 1, characterized in that: The busbar analysis module includes a classification unit, a cluster analysis unit and a configuration update unit; The classification unit obtains characteristic parameters of the busbar in the topological network structure diagram, and performs preliminary classification of the busbar based on the characteristic parameters; The cluster analysis unit is used to perform secondary cluster analysis on the busbar data to refine the busbar type; The configuration updating unit is used to configure different balancing thresholds and line loss calculation algorithms for different types of buses based on the clustering results.

3. A bus loss analysis system based on balanced dynamic threshold according to claim 2, characterized in that: The classification unit is configured with a classification strategy for preliminary classification of the mother line; The classification strategy includes: Collect historical bus data, classify and label each sample, and divide the data into training set and test set; Extract the key parameters in the training set as feature vectors and construct a feature vector set X = (X1…X a ) T ; Wherein, a is the number of samples in the training set; A binary classification SVM model is constructed for each bus category; and the parameters of each binary classification SVM model are initialized; wherein the classification results of each binary classification SVM model are predefined types and other types, and the predefined types of each binary classification SVM model are power plant bus, substation bus, transmission line bus, and distribution line bus; Input the training set data into the model, and calculate the inner product between the feature vector of each sample and the current model parameter; According to the inner product value, determine the category to which each sample belongs; if Y i >0, the sample belongs to the predefined type; if Y i <0, the sample belongs to other types; Calculate the error between the expected classification label and the actual label of the sample And update the model parameters in the opposite direction of the gradient; Repeatedly calculate the inner product value and update the model parameters until the number of iterations reaches the preset upper limit; The binary SVM model is verified using the test set; if the verification fails, the training set is expanded and the model is continued to be trained; if the verification passes, the binary SVM model is successfully constructed.

4. A bus loss analysis system based on balanced dynamic threshold according to claim 3, characterized in that: The classification strategy also includes: Obtain samples of the bus to be classified, calculate the inner product value of the samples using the binary SVM model, and construct a category decision set y=(y1, y2, y3, y4) of the bus to be classified; The maximum value max(y1, y2, y3, y4) of the category decision set y is obtained, and the category corresponding to the maximum value is defined as the final category label of the bus to be classified.

5. A bus loss analysis system based on balanced dynamic threshold according to claim 4, characterized in that: The cluster analysis unit is configured with a clustering strategy for performing secondary cluster analysis on the busbar data; The clustering strategy includes: Obtaining a bus data set and a preliminary classification result corresponding to the bus from the topological network structure diagram to generate a cluster sample set; Set clustering parameters, including the number of clusters K, initial cluster center, number of iterations, and distance measurement method; According to the selected initial cluster center and the number of clusters K, the cluster sample set is divided into K initial clusters; Calculating the distance between each data point in the cluster sample set and each cluster center, and assigning each data point to the nearest cluster center; Recalculate the center position of each cluster according to the allocation results, and repeat the iterative update process until the preset number of iterations is reached; Calculate the average value of the cluster sample set as the silhouette coefficient of the entire clustering result According to the silhouette coefficient Adjust the clustering parameters; After adjusting the clustering parameters, re-perform the clustering analysis and calculate the new silhouette coefficient Repeat the above steps until the silhouette coefficient Achieve expected value; Generate a busbar clustering results report.

6. A bus loss analysis system based on balanced dynamic threshold according to claim 5, characterized in that: The balance assessment module includes a busbar balance unit and an abnormality analysis unit; The bus balancing unit monitors the balance state of the power on the bus in the power grid in real time by calculating the bus power balance value and the bus loss calculation value, and issues a warning signal when the bus power balance exceeds the balance threshold; The abnormality analysis unit is used to analyze the bus abnormality in the early warning signal and generate an abnormality analysis report.

7. A bus loss analysis system based on balanced dynamic threshold according to claim 6, characterized in that: The bus balancing unit is configured with a balancing monitoring strategy for monitoring the balancing state of the electric quantity on the bus in the power grid; The balance monitoring strategy includes: Obtain real-time power data of each bus in the power grid, determine the number of input and output paths of each bus, and record the power data of each path; Calculate the input power P of each bus in , output power P out and the unbalance ΔP; Calculate the busbar unbalance rate η and line loss rate λ; Determine whether there is an abnormal power balance problem on the bus; if η≤1-η Th And λ≤λ Th , then there is no abnormal power balance problem on the bus; if η>1-η Th or λ>λ Th , then there is an abnormal power balance problem on the busbar, and an early warning signal is generated.

8. The bus loss analysis system based on balanced dynamic threshold according to claim 7 is characterized in that: The abnormal analysis unit is configured with a linkage analysis strategy for analyzing a busbar with abnormal power balance problems; The linkage analysis strategy includes: Obtain the early warning information in the busbar balancing unit and the electrical parameters of the main transformer of the substation; Calculate the main transformer loss P loss ; Obtain the main transformer loss P of two adjacent busbars in the substation 1,loss , P 2,loss , and calculate the loss difference ΔP loss ; Set the seesaw threshold to ΔP Th,loss , determine whether there is a seesaw phenomenon between two adjacent busbars; if ΔP loss <ΔP Th,loss , there is no seesaw phenomenon; if ΔP loss ≥ΔP Th,loss , there is a seesaw phenomenon; Set the main transformer loss threshold to P Th , and determine the main transformer loss P loss Is it abnormal? If P loss <P Th , the main transformer loss is normal; if P loss ≥P Th , then there is abnormal loss in the main transformer.

9. A bus loss analysis system based on balanced dynamic threshold according to claim 8, characterized in that: The linkage analysis strategy includes: Obtain the electrical parameters of each line under the bus and calculate the line loss rate P of each line under the bus line_loss ; Draw the line loss rate of each line into a time series diagram and calculate the correlation coefficient γ between the line loss rates of each line loss ; Combine the timing diagram and correlation coefficient to determine whether there is an abnormality in the line loss; if If the time series diagram shows a reverse fluctuation trend, the line loss is abnormal; otherwise, the line loss is not abnormal; Generate and output exception analysis reports.

Images