A data and model-based dual-driven distribution network line loss analysis method and device

Abstract

This application provides a data- and model-driven distribution network line loss analysis method and device. The method includes: S1: obtaining raw data collected by SCADA / AMI equipment from the distribution network and preprocessing it into a feature matrix; S2: using a multi-objective algorithm library to calculate theoretical line loss and calculate the residual between the theoretical line loss and the measured line loss; S3: concatenating the feature matrix and the residual into an input vector; S4: inputting the input vector into a pre-trained LSTM network to generate parameter corrections; S5: dynamically adjusting the current algorithm parameters of the multi-objective algorithm library based on the corrections; S6: recalculating the theoretical line loss using the updated parameters to obtain a calibrated line loss value; S7: using a sliding window to evaluate the mean absolute percentage error of the calibrated line loss value. If it exceeds a preset threshold, S2 to S6 are re-executed; otherwise, the line loss value is output. This method achieves high-precision, adaptive analysis, reduces manual intervention, improves efficiency and reliability, and supports intelligent management of power systems.

Description

Technical Field

[0001] The present application relates to the field of line loss analysis, and in particular to a method and device for analyzing line loss in a distribution network based on dual drive of data and models. Background Art

[0002] In the field of distribution network operation and management, line loss analysis, as a key indicator for measuring the economic efficiency, stability, and efficiency of power systems, has always attracted much attention. However, traditional distribution network line loss analysis methods are unable to meet the requirements of intelligent and refined management of modern power systems.

[0003] Traditional line loss analysis methods often rely on single statistical methods, such as estimating line losses through simple averaging or trend analysis of historical data. While this approach is simple to use, it ignores the complexity and dynamic nature of the distribution network's real-time operating conditions, resulting in significant deviations between the estimated line losses and actual values. Meanwhile, some methods use fixed theoretical models to calculate line losses, such as formulas based on parameters such as resistance and inductance. However, the structure of distribution networks is complex and dynamic, and load characteristics constantly change over time. Fixed models struggle to adapt to these changes, resulting in a lack of accuracy and adaptability in the analysis results.

[0004] Furthermore, traditional line loss analysis methods also have significant shortcomings in terms of intelligence and automation. These methods often require extensive manual intervention, such as data collection, processing, and analysis, making them inefficient and error-prone. Furthermore, due to the lack of intelligent tools, traditional methods struggle to achieve real-time monitoring and early warning of distribution network line losses, making it impossible to promptly address potential issues, posing a threat to the safe and stable operation of the power system.

[0005] With the development of intelligent power systems, distribution networks are generating massive amounts of power operation data, including voltage, current, power, temperature, and other types of data. This data contains a wealth of information and is crucial for improving the accuracy and efficiency of distribution network line loss analysis. However, how to fully utilize this data and combine it with advanced analysis models to achieve high-precision, adaptive distribution network line loss analysis has become a pressing issue. Summary of the Invention

[0006] The purpose of this application is to overcome the defects in the above-mentioned prior art and provide a distribution network line loss analysis method and device based on dual drive of data and model.

[0007] This application provides a distribution network line loss analysis method based on dual data and model driving, including:

[0008] S1. Obtaining raw data collected by SCADA / AMI equipment in the distribution network, and preprocessing the raw data to obtain a feature matrix;

[0009] S2. Calculating theoretical line loss using a multi-objective algorithm based on the characteristic matrix, and calculating the residual between the theoretical line loss and the measured line loss;

[0010] S3, concatenating each row element in the feature matrix with the residual to generate an input vector;

[0011] S4, inputting the input vector into a pre-trained LSTM network to generate parameter corrections;

[0012] S5. Dynamically adjust the current algorithm parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters;

[0013] S6. Recalculating the theoretical line loss using the multi-objective algorithm library based on the updated parameters to obtain a calibrated line loss value;

[0014] S7. Obtain a mean absolute percentage error (MAPE) of the calibrated line loss value through a sliding window. When the MAPE exceeds a preset threshold, re-execution of S2 to S6 is triggered; otherwise, the calibrated line loss value is output.

[0015] Optionally, concatenating each row element in the feature matrix with the residual to generate an input vector includes:

[0016] The voltage, current, resistance, and temperature feature vectors of each time step in the feature matrix are concatenated with the residual values ​​at the corresponding time points in a column direction to generate an input vector containing device state features and error feedback.

[0017] Optionally, based on the characteristic matrix, in calculating the theoretical line loss using a multi-objective algorithm, dynamically selecting an optimal algorithm using a multi-objective algorithm library includes:

[0018] According to the historical error standard deviation of each algorithm and the residual between the current theoretical line loss and the measured line loss, the weight coefficients α and β are used to perform weight calculation, and the algorithm with the smallest weight value is selected as the current optimal algorithm.

[0019] Optionally, dynamically adjusting the current parameters of the multi-objective algorithm library based on the parameter correction amount includes:

[0020] According to the comparison between the absolute value of the residual obtained in real time and a preset threshold, when the residual exceeds the preset threshold, the parameter adjustment amplitude is reduced, and when the residual is lower than the preset threshold, the adjustment amplitude is increased.

[0021] Optionally, obtaining a mean absolute percentage error of the calibrated line loss value through a sliding window includes:

[0022] Within the set sliding window, the absolute percentage error between the theoretical line loss and the measured line loss is calculated point by point, and the average value of the error within the window is obtained as an evaluation indicator.

[0023] The present application also provides a distribution network line loss analysis device based on dual drive of data and model, comprising:

[0024] An acquisition module acquires raw data collected by SCADA / AMI equipment in the distribution network and preprocesses the raw data to obtain a feature matrix;

[0025] a calculation module, which calculates theoretical line loss based on the characteristic matrix using a multi-objective algorithm, and calculates the residual between the theoretical line loss and the measured line loss;

[0026] A concatenation module concatenates each row element in the feature matrix with the residual to generate an input vector;

[0027] A model module, inputting the input vector into a pre-trained LSTM network to generate parameter corrections;

[0028] An adjustment module dynamically adjusts the current algorithm parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters;

[0029] The calculation module recalculates the theoretical line loss based on the updated parameters using the multi-objective algorithm library to obtain a calibrated line loss value;

[0030] The output module obtains the mean absolute percentage error of the calibrated line loss value through a sliding window, and when the mean absolute percentage error exceeds a preset threshold, triggers the re-execution of the above module, otherwise outputs the calibrated line loss value.

[0031] Optionally, the concatenation module concatenates each row element in the feature matrix with the residual to generate an input vector, including:

[0032] The voltage, current, resistance, and temperature feature vectors of each time step in the feature matrix are concatenated with the residual values ​​at the corresponding time points in a column direction to generate an input vector containing device state features and error feedback.

[0033] Optionally, the calculation module calculates the theoretical line loss using a multi-objective algorithm based on the characteristic matrix, and dynamically selects an optimal algorithm through a multi-objective algorithm library, including:

[0034] According to the historical error standard deviation of each algorithm and the residual between the current theoretical line loss and the measured line loss, the weight coefficients α and β are used to perform weight calculation, and the algorithm with the smallest weight value is selected as the current optimal algorithm.

[0035] Optionally, the adjustment module dynamically adjusts the current parameters of the multi-objective algorithm library based on the parameter correction amount, including:

[0036] According to the comparison between the absolute value of the residual obtained in real time and a preset threshold, when the residual exceeds the preset threshold, the parameter adjustment amplitude is reduced, and when the residual is lower than the preset threshold, the adjustment amplitude is increased.

[0037] Optionally, the output module obtains a mean absolute percentage error (MAPE) of the calibrated line loss value through a sliding window, including:

[0038] Within the set sliding window, the absolute percentage error between the theoretical line loss and the measured line loss is calculated point by point, and the average value of the error within the window is obtained as an evaluation indicator.

[0039] The beneficial effects of this application are:

[0040] The present application provides a distribution network line loss analysis method based on dual-drive of data and model, including: S1, obtaining raw data collected by SCADA / AMI equipment in the distribution network, and preprocessing the raw data to obtain a feature matrix; S2, based on the feature matrix, calculating the theoretical line loss through a multi-objective algorithm, and calculating the residual between the theoretical line loss and the measured line loss; S3, splicing each row element in the feature matrix with the residual to generate an input vector; S4, inputting the input vector into a pre-trained LSTM network to generate a parameter correction; S5, dynamically adjusting the current algorithm parameters of the multi-objective algorithm library based on the parameter correction to obtain updated parameters; S6, based on the updated parameters, recalculating the theoretical line loss through the multi-objective algorithm library to obtain a calibrated line loss value; S7, obtaining the average absolute percentage error of the calibrated line loss value through a sliding window, and when the average absolute percentage error exceeds a preset threshold, triggering re-execution of S2 to S6, otherwise outputting the calibrated line loss value. This application uses data and model dual drivers to achieve high-precision, adaptive analysis of distribution network line losses, reduce manual intervention, improve analysis efficiency and reliability, and provide strong support for intelligent management of power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of the distribution network line loss analysis process based on data and model dual drive in this application;

[0042] Figure 2 This is a schematic diagram of a distribution network line loss analysis device based on dual drive of data and model in this application. DETAILED DESCRIPTION

[0043] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it is understood that various forms of implementing the present disclosure are not limited by the embodiments set forth herein. Rather, the embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0044] The present invention discloses a method for analyzing distribution network line losses based on dual-drive of data and model, which adopts a dual-drive architecture, with the data layer and the model layer working together, as follows:

[0045] Data layer:

[0046] Acquisition module: Through the SCADA (Supervisory Control and Data Acquisition) / AMI (Advanced Measurement Infrastructure) data acquisition equipment deployed in the distribution network, data such as voltage U, current I, line resistance R, and temperature T are collected in real time at a frequency of seconds.

[0047] Feature Engineering Module: This module preprocesses the collected data, including data cleaning (removing outliers and missing values) and normalization (mapping the data to the interval [0, 1]). It then combines the voltage U, current I, line resistance R, temperature T, and the corresponding timestamp t in chronological order to construct a spatiotemporal feature matrix X = [U, I, R, T, t]. This feature matrix integrates multi-source data, converting data of different types and time scales into a unified structured data format. This provides standardized and effective data input for the model layer's algorithm calculations, facilitates the extraction of effective features from the data, and explores potential relationships between the data, thereby improving the accuracy of line loss analysis.

[0048] Model layer:

[0049] The model layer consists of a multi-objective algorithm library and a dynamic calibration engine. The multi-objective algorithm library includes various algorithms suitable for line loss analysis, including the root mean square current method, the average current method, and the maximum current method. The dynamic calibration engine consists of a data interface module, a line loss difference analysis module, a parameter adjustment module, and a feedback evaluation module.

[0050] Multi-objective algorithm library:

[0051] Data interface module: This module is responsible for interacting with the data layer and obtaining the feature matrix , providing a data foundation for the entire dynamic calibration engine, ensuring that the engine can work based on the latest and most accurate power operation information.

[0052] Line loss difference analysis module: The line loss difference analysis module calculates the residual between the theoretical line loss and the measured value.

[0053] Parameter adjustment module: The parameter adjustment module predicts the parameter correction amount based on the LSTM network (long short-term memory network) and adjusts the algorithm parameters. The feedback evaluation module evaluates the calibration results.

[0054] Feedback evaluation module: Visually displays the calibrated line loss analysis results in the form of charts (such as line charts, bar charts), reports, etc., and sets warning thresholds for indicators such as line loss rate and error.

[0055] Please refer to Figure 1 As shown, the present application provides a distribution network line loss analysis method based on dual drive of data and model, including:

[0056] S1. Obtaining raw data collected by SCADA / AMI equipment in the distribution network, and preprocessing the raw data to obtain a feature matrix;

[0057] The data layer collects data such as voltage U, current I, line resistance R, and temperature T in real time at a frequency of seconds through SCADA (Supervisory Control and Data Acquisition) / AMI (Advanced Measurement Infrastructure) data acquisition equipment deployed in the distribution network.

[0058] The collected data are preprocessed, including data cleaning (removing outliers and missing values) and normalization (mapping the data to the [0, 1] interval).

[0059] Then, the voltage U, current I, line resistance R, temperature T and the corresponding timestamp t are combined in chronological order to construct the spatiotemporal feature matrix X = [U, I, R, T, t].

[0060] The data interface module in the model layer interacts with the data layer to obtain the feature matrix and sends it to the multi-line loss difference analysis module.

[0061] S2. Calculating theoretical line loss using a multi-objective algorithm based on the characteristic matrix, and calculating the residual between the theoretical line loss and the measured line loss;

[0062] Multi-objective algorithm library: Contains a variety of algorithms suitable for line loss analysis, including the root mean square current method, average current method, maximum current method, etc.

[0063] The optimal algorithm is dynamically selected through the entropy weight method, and the line loss calculation algorithm that best suits the current distribution network operation status is dynamically selected and sent to the line loss difference analysis module.

[0064] First, based on the historical data samples, for each algorithm, according to the historical calculation results and the measured line loss data, the historical error standard deviation σ of each algorithm is calculated according to the following formula:

[0065]

[0066] Among them, m is the number of historical data samples, The algorithm calculates the theoretical line loss for the jth sample, is the measured line loss of the jth sample.

[0067] Then, in the current calculation, the theoretical line loss is calculated using the root mean square current method, the average current method, and the maximum current method. .

[0068] Obtaining measured line loss .

[0069] Next, calculate the weight of each algorithm according to the formula :

[0070]

[0071] in and is the weight coefficient ( The value is between 0.6 and 0.8. The value range is 0.2 - 0.4).

[0072] Finally, choose the weights The smallest algorithm is taken as the current optimal algorithm.

[0073] This mechanism can dynamically select the line loss calculation algorithm that best suits the current distribution network operating status based on actual data and algorithm historical performance, avoiding the limitations brought by fixed algorithms and improving the accuracy and adaptability of line loss calculation.

[0074] Line loss difference analysis module:

[0075] Obtain the theoretical line loss calculated by the optimal algorithm in the multi-objective algorithm library and measured line loss

[0076] Calculate the residual ε between the two:

[0077]

[0078] By calculating the residual, the difference between the theoretical calculation results and the actual measurement results can be quantified, intuitively reflecting the error size of the current line loss calculation, and providing a clear basis for subsequent parameter adjustments, so as to optimize the algorithm parameters in a targeted manner.

[0079] S3, concatenating each row element in the feature matrix with the residual to generate an input vector;

[0080] The parameter combines the feature matrix X constructed by the feature engineering module and the residual calculated by the difference analysis module Splice to form a new input vector [X, ].

[0081] S4, inputting the input vector into a pre-trained LSTM network to generate parameter corrections;

[0082] The input vector is input into the LSTM network, which calculates the parameter correction value by learning the time series characteristics and error relationship in the historical data. :

[0083]

[0084] in, is the weight matrix, b is the bias vector, is the activation function. Based on the current data characteristics and calculation error, the correction amount required to adjust the algorithm parameters is predicted.

[0085] S5. Dynamically adjust the current algorithm parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters;

[0086] This correction is applied to the relevant algorithm parameters in the multi-objective algorithm library to dynamically adjust the algorithm and further improve the accuracy of line loss calculation.

[0087] S6. Recalculating the theoretical line loss using the multi-objective algorithm library based on the updated parameters to obtain a calibrated line loss value;

[0088] Specifically, the above step S2 is executed again.

[0089] S7. Obtain a mean absolute percentage error (MAPE) of the calibrated line loss value through a sliding window. When the MAPE exceeds a preset threshold, re-execution of S2 to S6 is triggered; otherwise, the calibrated line loss value is output.

[0090] Establish a sliding window of length n (n is between 10 and 20). As new data is continuously generated, the window slides forward in the time series.

[0091] In each window, according to the formula:

[0092]

[0093] Calculate the mean absolute percentage error.

[0094] By calculation , quantitatively evaluate the calibrated line loss calculation results, and monitor the accuracy and reliability of line loss analysis in real time. When the threshold is exceeded (e.g., 10%), a recalibration mechanism is triggered, re-analyzing the residuals and recalibrating the parameters to ensure that the line loss calculation results remain highly accurate. If the threshold is not exceeded, the calibrated line loss analysis results are visualized. If the results exceed warning thresholds such as line loss rate or error, the system issues an abnormality warning.

[0095] like Figure 2 As shown, the present application also provides a distribution network line loss analysis device based on dual drive of data and model, comprising:

[0096] An acquisition module 201 acquires raw data collected by SCADA / AMI equipment in the distribution network and preprocesses the raw data to obtain a feature matrix;

[0097] A calculation module 202 calculates theoretical line loss based on the characteristic matrix using a multi-objective algorithm, and calculates the residual between the theoretical line loss and the measured line loss;

[0098] A concatenation module 203 concatenates each row element in the feature matrix with the residual to generate an input vector;

[0099] Model module 204, inputs the input vector into a pre-trained LSTM network to generate parameter corrections;

[0100] An adjustment module 205 dynamically adjusts the current algorithm parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters;

[0101] The calculation module 202 recalculates the theoretical line loss based on the updated parameters using the multi-objective algorithm library to obtain a calibrated line loss value;

[0102] The output module 206 obtains the mean absolute percentage error of the calibrated line loss value through a sliding window. When the mean absolute percentage error exceeds a preset threshold, the above module is triggered to be re-executed; otherwise, the calibrated line loss value is output.

[0103] Furthermore, each row element in the feature matrix is ​​concatenated with the residual to generate an input vector, including:

[0104] The voltage, current, resistance, and temperature feature vectors of each time step in the feature matrix are concatenated with the residual values ​​at the corresponding time points in a column direction to generate an input vector containing device state features and error feedback.

[0105] Furthermore, the calculation module calculates the theoretical line loss using a multi-objective algorithm based on the characteristic matrix, and dynamically selects the optimal algorithm through a multi-objective algorithm library, including:

[0106] According to the historical error standard deviation of each algorithm and the residual between the current theoretical line loss and the measured line loss, the weight coefficients α and β are used to perform weight calculation, and the algorithm with the smallest weight value is selected as the current optimal algorithm.

[0107] Furthermore, dynamically adjusting the current parameters of the multi-objective algorithm library based on the parameter correction amount includes:

[0108] According to the comparison between the absolute value of the residual obtained in real time and a preset threshold, when the residual exceeds the preset threshold, the parameter adjustment amplitude is reduced, and when the residual is lower than the preset threshold, the adjustment amplitude is increased.

[0109] Furthermore, obtaining a mean absolute percentage error of the calibrated line loss value through a sliding window includes:

[0110] Within the set sliding window, the absolute percentage error between the theoretical line loss and the measured line loss is calculated point by point, and the average value of the error within the window is obtained as an evaluation indicator.

[0111] The above description of the embodiments is intended to facilitate understanding and application of this application by those skilled in the art. It will be readily apparent to those skilled in the art that various modifications to the above embodiments can be made, and the general principles described herein can be applied to other embodiments without requiring creative effort. Therefore, this application is not limited to the above embodiments. Any improvements or modifications made to this application by those skilled in the art based on the disclosure of this application should fall within the scope of protection of this application.

Claims

1. A distribution network line loss analysis method based on dual drive of data and model, characterized in that: include: S1. Obtaining raw data collected by SCADA / AMI equipment in the distribution network, and preprocessing the raw data to obtain a feature matrix; S2. Calculating theoretical line loss using a multi-objective algorithm based on the characteristic matrix, and calculating the residual between the theoretical line loss and the measured line loss; S3, concatenating each row element in the feature matrix with the residual to generate an input vector; S4, inputting the input vector into a pre-trained LSTM network to generate parameter corrections; S5. Dynamically adjusting the current parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters; S6. Recalculating the theoretical line loss using the multi-objective algorithm library based on the updated parameters to obtain a calibrated line loss value; S7. Obtain a mean absolute percentage error (MAPE) of the calibrated line loss value through a sliding window. When the MAPE exceeds a preset threshold, re-execute steps S2 to S6; otherwise, output the calibrated line loss value.

2. A distribution network line loss analysis method based on dual drive of data and model according to claim 1, characterized in that: Each row element in the feature matrix is ​​concatenated with the residual to generate an input vector, including: The voltage, current, resistance, and temperature feature vectors of each time step in the feature matrix are concatenated with the residual values ​​at the corresponding time points in a column direction to generate an input vector containing device state features and error feedback.

3. The method for analyzing distribution network line loss based on dual drive of data and model according to claim 1, characterized in that: Based on the characteristic matrix, the theoretical line loss is calculated using a multi-objective algorithm, and the optimal algorithm is dynamically selected through the multi-objective algorithm library, including: According to the historical error standard deviation of each algorithm and the residual between the current theoretical line loss and the measured line loss, the weight coefficients α and β are used to calculate the weight value, and the algorithm with the smallest weight value is selected as the current optimal algorithm.

4. The method for analyzing distribution network line loss based on dual drive of data and model according to claim 1, characterized in that: Dynamically adjusting the current parameters of the multi-objective algorithm library based on the parameter correction amount includes: According to the comparison between the absolute value of the residual obtained in real time and a preset threshold, when the residual exceeds the preset threshold, the parameter adjustment amplitude is reduced, and when the residual is lower than the preset threshold, the adjustment amplitude is increased.

5. The method for analyzing distribution network line loss based on dual drive of data and model according to claim 1, characterized in that: Obtaining a mean absolute percentage error of the calibrated line loss value through a sliding window includes: Within the set sliding window, the absolute percentage error between the theoretical line loss and the measured line loss is calculated point by point, and the average value of the error within the window is obtained as an evaluation indicator.

6. A distribution network line loss analysis device based on dual drive of data and model, characterized in that: include: An acquisition module acquires raw data collected by SCADA / AMI equipment in the distribution network and preprocesses the raw data to obtain a feature matrix; a calculation module, which calculates theoretical line loss based on the characteristic matrix using a multi-objective algorithm, and calculates the residual between the theoretical line loss and the measured line loss; A concatenation module concatenates each row element in the feature matrix with the residual to generate an input vector; A model module, inputting the input vector into a pre-trained LSTM network to generate parameter corrections; An adjustment module dynamically adjusts the current algorithm parameters of the multi-objective algorithm library based on the parameter correction amount to obtain updated parameters; The calculation module recalculates the theoretical line loss based on the updated parameters using the multi-objective algorithm library to obtain a calibrated line loss value; The output module obtains the mean absolute percentage error of the calibrated line loss value through a sliding window, and when the mean absolute percentage error exceeds a preset threshold, triggers the re-execution of the above module, otherwise outputs the calibrated line loss value.

7. The data and model-based dual-driven distribution network line loss analysis device according to claim 6, characterized in that: The concatenation module concatenates each row element in the feature matrix with the residual to generate an input vector, including: The voltage, current, resistance, and temperature feature vectors of each time step in the feature matrix are concatenated with the residual values ​​at the corresponding time points in a column direction to generate an input vector containing device state features and error feedback.

8. The data and model-based dual-driven distribution network line loss analysis device according to claim 6, characterized in that: The calculation module calculates the theoretical line loss using a multi-objective algorithm based on the characteristic matrix and dynamically selects the optimal algorithm through a multi-objective algorithm library, including: According to the historical error standard deviation of each algorithm and the residual between the current theoretical line loss and the measured line loss, the weight coefficients α and β are used to perform weight calculation, and the algorithm with the smallest weight value is selected as the current optimal algorithm.

9. The data and model-based dual-driven distribution network line loss analysis device according to claim 6, characterized in that: The adjustment module dynamically adjusts the current parameters of the multi-objective algorithm library based on the parameter correction amount, including: According to the comparison between the absolute value of the residual obtained in real time and a preset threshold, when the residual exceeds the preset threshold, the parameter adjustment amplitude is reduced, and when the residual is lower than the preset threshold, the adjustment amplitude is increased.

10. The data and model-based dual-driven distribution network line loss analysis device according to claim 6, characterized in that: The output module obtains a mean absolute percentage error (MAPE) of the calibrated line loss value through a sliding window, including: Within the set sliding window, the absolute percentage error between the theoretical line loss and the measured line loss is calculated point by point, and the average value of the error within the window is obtained as an evaluation indicator.

Images