A method and device for predicting risk of heavy-load lines based on digital twins
By constructing a digital twin model of high-load lines and combining simulation technology with a risk prediction model, the problem of low accuracy in risk prediction for high-load lines was solved, achieving more efficient and accurate risk prediction.
Patent Information
- Application Number
- CN202311758466.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-20
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2043-12-20
AI Technical Summary
Existing technologies for risk prediction of high-load lines have low accuracy, are time-consuming and labor-intensive, and rely on data processing, which is not accurate enough.
A digital twin model of a high-load power line is constructed, including a line simulation model and a line risk prediction model. By acquiring line modeling data and reference line power datasets, the model is constructed and verified using simulation technology. The simulation index and risk prediction accuracy index are calculated to achieve risk prediction.
It improves the accuracy and efficiency of risk prediction for high-load lines, and enables more accurate simulation performance evaluation and risk prediction performance evaluation.
Smart Images

Figure CN117689205B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of fault detection technology, and in particular to a method and device for predicting high-load line risks based on digital twins. Background Art
[0002] Digital twins are a set of virtual information structures that can fully describe existing or future physical products from the microscopic atomic level to the macroscopic geometric level. In the best state, all information about physical entities can be obtained from digital twin models. Digital twins are mappings of the full spatial scale and life cycle of physical entities. The two are exactly the same and run synchronously in the physical and digital worlds. Based on twins, the real-time status of physical entities can be tracked throughout their life cycle, and the status of physical entities in specific environments can be simulated and predicted, further enhancing the understanding and cognition of physical entities. In the fields of engineering and manufacturing, digital twins can be used to create virtual physical systems, enabling engineers and decision makers to test and optimize designs in a digital environment, predict the performance of equipment, and even conduct real-time monitoring and maintenance. High-load lines usually refer to high-load lines in power systems. These lines may face some potential risks, such as the risk of overload caused by excessive load, the risk of equipment failure due to heat loss, etc. If the risks of high-load lines are not resolved in a timely manner, they will have an impact on the power system.
[0003] In the existing technology, models are often established based on the physical equations and theories of the power system to analyze the performance of the line under different load conditions, or to perform statistical analysis and prediction on historical data. Line risk prediction is also performed with the help of machine learning algorithms.
[0004] For example, the invention patent application with publication number: CN113033989A discloses a method, apparatus and terminal equipment for evaluating the fault risk of a transmission line, comprising: obtaining the location information of a target transmission line; determining the load removal amount of the target transmission line according to the location information, and determining a first fault risk rate of the target transmission line according to the load removal amount; wherein the load removal amount of the target transmission line is the load removal amount required to be removed when the target transmission line fails; determining environmental information corresponding to the target transmission line according to the location information, inputting the environmental information into a preset fault prediction model to obtain a second fault risk rate of the target transmission line; and determining a fault risk index of the target transmission line based on the first fault risk rate and the second fault risk rate.
[0005] For example, the invention patent application with publication number: CN116757465A discloses a line risk assessment method and device based on a dual training weight allocation model, including: obtaining historical load data of the object under different working conditions, preprocessing the historical load data to obtain historical load characteristic data, and inputting the data into a target recurrent neural network model and a target autoregressive moving average model for training to obtain first load forecast data and second load forecast data respectively; weight optimization of the first load forecast data and the second load forecast data is performed through a BP neural network algorithm to obtain a first target weight and a second target weight; and the target load forecast data is calculated based on the first load forecast data, the first target weight, the second load forecast data and the second target weight to determine the line risk index.
[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:
[0007] In the existing technology, building models based on the physical equations and theories of the power system and then using machine learning to make predictions both require a lot of time and effort to process data. Statistical analysis methods also rely on data processing, which leads to inaccurate line risk predictions and low accuracy in risk predictions for heavy-load lines. Summary of the invention
[0008] The embodiments of the present application solve the problem of low accuracy in heavy-load line risk prediction in the prior art by providing a heavy-load line risk prediction method and device based on digital twins, thereby improving the accuracy of heavy-load line risk prediction.
[0009] An embodiment of the present application provides a method for predicting the risk of a heavy-load line based on a digital twin, comprising the following steps: constructing a digital twin model of a heavy-load line, the digital twin model of the heavy-load line comprising a line simulation model and a line risk prediction model, the line simulation model being used to simulate the working condition of the heavy-load line to be tested, and the line risk prediction model being used to predict the line risk index data of the heavy-load line to be tested; obtaining the line risk index data of the heavy-load line to be tested within a monitoring period, and simulating in combination with the line simulation model to obtain a corresponding line risk index, the line risk index being used to describe the probability of a line risk occurring in the heavy-load line to be tested corresponding to the line risk index data; predicting the line risk situation of the heavy-load line to be tested within the monitoring period according to the calculated line risk index, and taking corresponding risk prevention measures in a timely manner.
[0010] Furthermore, the specific construction method of the line simulation model is as follows: obtain line modeling data of the large-load line to be tested, and the line modeling data is used to simulate the topological structure and technical equipment of the large-load line to be tested; obtain a reference line power data set of the large-load line to be tested and divide the data set to obtain a reference line power data training set and a reference line power data verification set, and the reference line power data set is a collection of reference line power data, which is used to simulate the power conditions of the large-load line to be tested, and the reference line power data includes reference load data, reference voltage data, reference current data, reference environmental data and reference motor operating status data; construct a line simulation model in combination with simulation technology, and conduct a line simulation model verification experiment through the reference line power data verification set, and the line simulation model verification experiment is used to verify the working condition of the line simulation model under the reference line power data verification set, and the simulation index of the line simulation model is greater than the reference simulation index, and the simulation index is used to describe the simulation performance of the constructed line simulation model.
[0011] Furthermore, the simulation index is calculated using the following formula:
[0012] Where SI is the simulation index, t is the number of the circuit simulation model verification experiment, t = 1, 2, ..., T, T is the total number of circuit simulation model verification experiments, FL t and ΔFL t are the simulated load data and reference load data of the t-th line simulation model verification experiment, FU t and ΔFU t are the simulation voltage data and reference voltage data of the t-th line simulation model verification experiment, FI t and ΔFI t are the simulation current data and reference current data of the t-th line simulation model experiment, ΔL, ΔU and ΔI are the reference load deviation, reference voltage deviation and reference current deviation, respectively.
[0013] Furthermore, the specific construction method of the line risk prediction model is as follows: obtain reference line risk indicator data of the heavy-load line to be tested within a historical time period; construct a reference line risk indicator data set based on the obtained reference line risk indicator data, and divide the data set to obtain a reference line risk indicator training set and a reference line risk indicator verification set; construct a line risk prediction model, the line risk prediction model is obtained by training and optimizing a preset model based on the reference line risk indicator training set and verifying the model through a reference line risk indicator verification set; through the line risk prediction model, obtain the predicted line risk indicator data corresponding to the reference line risk indicator verification set, and combine the predicted line risk indicator data with the corresponding reference line risk indicator data to obtain the risk prediction accuracy index of the line risk prediction model, the risk prediction accuracy index of the line risk prediction model is greater than the reference risk prediction accuracy index, and the risk prediction accuracy index is used to describe the probability of correctly predicting the risk of the heavy-load line to be tested through the constructed line risk prediction model.
[0014] Furthermore, the risk prediction accuracy index is calculated using the following formula:
[0015] Where, e is a natural constant, h is the line risk type number of the large-load line to be tested, h = 1, 2, 3, 4, h = 1 is the overload risk of the large-load line to be tested, h = 2 is the low-voltage risk of the large-load line to be tested, h = 3 is the heat loss risk of the large-load line to be tested, h = 4 is the instability risk of the large-load line to be tested, PA h is the risk prediction accuracy index of the hth type of line risk of the heavy load line to be tested, FS h is the accuracy score of the hth type of line risk of the heavy-load line to be tested, AR h is the adjusted R-squared score of the h-th line risk of the heavy-load line to be tested, BS h is the Bayesian score of the h-th line risk of the heavy-load line to be tested, CS his the bias score of the hth line risk of the large-load line to be tested, α is the correction factor of the accuracy score, β is the correction factor of the adjusted R-square score, χ is the correction factor of the Bayes score, and δ is the correction factor of the bias score; the overload risk of the large-load line to be tested is used to describe the risk caused by the line load of the large-load line to be tested exceeding the reference first line load, and the line risk indicator data is the peak load data of the large-load line to be tested; the low-voltage risk of the large-load line to be tested is used to describe the risk caused by the line voltage of the large-load line to be tested being less than the reference first line voltage, and the line risk indicator data is the preset node voltage data of the large-load line to be tested; The heat loss risk of the large-load line to be tested is used to describe the risk caused by the heat loss of the large-load line to be tested exceeding the heat loss of the reference first line, and the line risk index data is the technical equipment temperature data of the large-load line to be tested; the instability risk of the large-load line to be tested is used to describe the risk caused by the instability of the corresponding power system of the large-load line to be tested, and the line risk index data is the power system output frequency data of the large-load line to be tested; the accuracy score, adjusted R-square score, Bayesian score and bias score are used to describe the degree of deviation between the predicted line risk index data obtained by the line risk prediction model and the corresponding reference line risk index data.
[0016] Furthermore, the accuracy score is calculated using the following formula: In the formula, MSE h is the mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, ε is the correction factor of the mean square error data, ΔMSE h is the reference first mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ΔMSE is the reference mean square error data of the predicted line risk index data of the heavy load line to be tested.
[0017] Furthermore, the adjusted R-squared score is calculated using the following formula: In the formula, R2 h is the determination coefficient of the predicted line risk index data of the hth line risk of the heavy load line to be tested, TR h is the adjusted R-square data of the predicted line risk index data of the h-th line risk, μ is the correction factor of the determination coefficient, and η is the correction factor of the adjusted R-square data.
[0018] Furthermore, the Bayesian score is calculated using the following formula: In the formula, BIC his the Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested, ΔBIC is the correction factor for the Bayesian information criterion data for predicting line risk index data. h is the reference Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy load line to be tested, ΔB h It is the reference Bayesian information criterion deviation data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested.
[0019] Furthermore, the bias score is calculated using the following formula: In the formula, MAE h is the mean absolute error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, ΔMAE h is the reference mean absolute error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ζ is the correction factor of the mean absolute error data of the predicted line risk index data.
[0020] The embodiment of the present application provides a high-load line risk prediction device based on digital twins, including a high-load line digital twin model construction module, a line risk index acquisition module and a line risk prediction module; the high-load line digital twin model construction module is used to construct a high-load line digital twin model, and the high-load line digital twin model includes a line simulation model and a line risk prediction model. The line simulation model is used to simulate the working condition of the high-load line to be tested, and the line risk prediction model is used to predict the line risk index data of the high-load line to be tested; the line risk index acquisition module is used to obtain the line risk index data of the high-load line to be tested within the monitoring time period, and simulate in combination with the line simulation model to obtain the corresponding line risk index, and the line risk index is used to describe the probability of line risk occurring in the high-load line to be tested corresponding to the line risk index data; the line risk prediction module is used to predict the line risk situation of the high-load line to be tested within the monitoring time period according to the calculated line risk index, and take corresponding risk prevention measures in time.
[0021] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:
[0022] 1. By obtaining line modeling data and reference line power data sets, and dividing the data sets, then combining simulation technology to build a line simulation model and conduct model verification experiments, then building a reference line risk indicator data set and dividing the data set, and then building a line risk prediction model, and finally combining the line risk prediction model and the line simulation model to predict the line risk situation, thereby achieving accurate prediction of the line risk of the large-load line to be tested, and then achieving the improvement of the accuracy of large-load line risk prediction, effectively solving the problem of low accuracy of large-load line risk prediction in the prior art.
[0023] 2. By obtaining the simulated load data, simulated voltage data and simulated current data of each line simulation model verification experiment, and then obtaining the corresponding reference data and reference deviation, the simulation index of the line simulation model is calculated, and finally the anti-vibration performance of the line simulation model is evaluated in combination with the simulation index, thereby realizing the digitization of the simulation performance of the line simulation model, and then realizing a more accurate evaluation of the simulation performance of the line simulation model.
[0024] 3. By calculating the mean square error data, determination coefficient, adjusted R-square data, Bayesian information criterion data and mean absolute error data of the predicted line risk index data of various line risks, and then obtaining the corresponding reference data, the risk prediction accuracy index of various line risks is calculated, and finally the prediction performance of the constructed risk prediction model is evaluated according to the risk prediction accuracy index, thereby realizing the digitization of the prediction performance of the line risk prediction model, and then realizing a more comprehensive evaluation of the prediction performance of the line risk prediction model. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 A flow chart of a high-load line risk prediction method based on digital twins provided in an embodiment of the present application;
[0026] Figure 2 A flow chart for constructing a circuit simulation model provided in an embodiment of the present application;
[0027] Figure 3 A flowchart for constructing a line risk prediction model provided in an embodiment of the present application;
[0028] Figure 4 A structural diagram of a high-load line risk prediction device based on digital twins provided in an embodiment of the present application. DETAILED DESCRIPTION
[0029] The embodiment of the present application solves the problem of low accuracy of large-load line risk prediction in the prior art by providing a large-load line risk prediction method and device based on digital twins. The method obtains line modeling data and a reference line power data set of the large-load line to be tested, and divides the reference line power data set to obtain a reference line power data training set and a reference line power data verification set. Then, a line simulation model is constructed in combination with simulation technology, and a line simulation model verification experiment is performed through the reference line power data verification set. Then, reference line risk index data of the large-load line to be tested within a historical time period is obtained, a reference line risk index data set is constructed and the data set is divided. Then, a line risk prediction model is constructed, and the prediction performance of the risk prediction accuracy index description model is calculated. Finally, the line risk index data of the large-load line to be tested within a monitoring time period is obtained in combination with the line risk prediction model, and the corresponding line risk index is obtained by simulation in combination with the line simulation model. The line risk situation is predicted according to the calculated line risk index, thereby improving the accuracy of large-load line risk prediction.
[0030] The technical solution in the embodiment of the present application is to solve the problem of low accuracy of risk prediction of heavy load lines mentioned above, and the overall idea is as follows:
[0031] By obtaining the line modeling data of the large-load line to be tested and the reference line power data set, and dividing the data sets, the line simulation model is constructed by combining simulation technology and a line simulation model verification experiment is carried out. Then, a reference line risk indicator data set is constructed and divided, and then a line risk prediction model is constructed. Finally, the line risk prediction model and the line simulation model are combined to predict the line risk situation, thereby achieving the effect of improving the accuracy of large-load line risk prediction.
[0032] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0033] like Figure 1As shown, a flow chart of a high-load line risk prediction method based on digital twins provided in an embodiment of the present application is applied to a high-load line risk prediction device based on digital twins, and the method includes the following steps: Construction of a high-load line digital twin model: constructing a high-load line digital twin model, the high-load line digital twin model including a line simulation model and a line risk prediction model, the line simulation model is used to simulate the working condition of the high-load line to be tested, and the line risk prediction model is used to predict the line risk index data of the high-load line to be tested; line risk index acquisition: obtaining the line risk index data of the high-load line to be tested within the monitoring time period, and simulating it in combination with the line simulation model to obtain the corresponding line risk index, the line risk index is used to describe the probability of line risk occurring in the high-load line to be tested corresponding to the line risk index data; line risk prediction: predicting the line risk situation of the high-load line to be tested within the monitoring time period according to the calculated line risk index, and taking corresponding risk prevention measures in a timely manner.
[0034] In this embodiment, heavy-load lines are prone to line risks due to their high loads. If there is no accurate prediction and intervention, it will affect the power of the entire area; the line simulation model performs simulation based on the data obtained by the line risk prediction model, and judges the line risk probability corresponding to the obtained data based on the simulation results, which requires that the accuracy of the line simulation model and the line risk prediction model meet the expected requirements; the heavy-load line digital twin model based on the digital twin model can quickly and accurately predict line risks and intervene based on the simulation results, thereby improving the accuracy of heavy-load line risk prediction.
[0035] Further, such as Figure 2 As shown, it is a flow chart for constructing a line simulation model provided in an embodiment of the present application, and the specific construction method of the line simulation model is as follows: obtain line modeling data: obtain line modeling data of the heavy load line to be tested, and the line modeling data is used to simulate the topological structure and technical equipment of the heavy load line to be tested; obtain a reference line power data set and divide it: obtain a reference line power data set of the heavy load line to be tested and divide the data set to obtain a reference line power data training set and a reference line power data verification set, the reference line power data set is a collection of reference line power data, used to simulate the power conditions of the heavy load line to be tested, the reference line power data includes reference load data, reference voltage data, reference current data, reference environment data and reference motor operating status data; construct a line simulation model and verify it: construct a line simulation model in combination with simulation technology, and conduct a line simulation model verification experiment through the reference line power data verification set, the line simulation model verification experiment is used to verify the working condition of the line simulation model under the reference line power data verification set, the simulation index of the line simulation model is greater than the reference simulation index, and the simulation index is used to describe the simulation performance of the constructed line simulation model.
[0036] In this embodiment, the simulation of the line topology structure generally requires obtaining information such as nodes, branches, transformers, etc. of the line, and constructing a corresponding line model according to the topology structure; the simulation of technical equipment generally requires obtaining technical parameters of each device on the line, such as transformer capacity, resistance, reactance, etc., and a correction process may also be involved after the simulation; the reference simulation index limits the reference simulation performance of the constructed line simulation model, which is equivalent to having certain requirements for the training and optimization of the line simulation model; and a more accurate construction of a large-load line simulation model is achieved.
[0037] Furthermore, the simulation index is calculated using the following formula:
[0038] Where SI is the simulation index, t is the number of the circuit simulation model verification experiment, t = 1, 2, ..., T, T is the total number of circuit simulation model verification experiments, FL t and ΔFL t are the simulated load data and reference load data of the t-th line simulation model verification experiment, FU t and ΔFU t are the simulation voltage data and reference voltage data of the t-th line simulation model verification experiment, FI t and ΔFI t are the simulation current data and reference current data of the t-th line simulation model experiment, ΔL, ΔU and ΔI are the reference load deviation, reference voltage deviation and reference current deviation, respectively.
[0039] In this embodiment, the smaller the actual deviation of the simulated load, voltage and current in the circuit simulation model verification experiment, the larger the corresponding simulation index. When the deviation between the simulated load data and the reference load data exceeds the corresponding reference load deviation, it indicates that the simulation performance of the model is very poor, and the corresponding simulation index is directly 0; when the actual deviation of the simulated load, voltage and current is 0, that is, The corresponding simulation index is 1; when the actual deviation of the simulated load, voltage and current is the same as the corresponding reference deviation, that is, The corresponding simulation index is 0.5, which realizes the quantification of the simulation performance of the circuit simulation model.
[0040] Further, such as Figure 3As shown, it is a flow chart for constructing a line risk prediction model provided in an embodiment of the present application, and the specific construction method of the line risk prediction model is as follows: obtain reference line risk indicator data: obtain reference line risk indicator data of the heavy-load line to be tested within a historical time period; construct a reference line risk indicator data set and divide it: construct a reference line risk indicator data set according to the obtained reference line risk indicator data, and divide the data set to obtain a reference line risk indicator training set and a reference line risk indicator verification set; construct a line risk prediction model and perform model verification: construct a line risk prediction model, and the line risk prediction model is obtained by training and optimizing a preset model according to the reference line risk indicator training set and performing model verification through a reference line risk indicator verification set; obtain predicted line risk indicator data, and calculate a risk prediction accuracy index: through the line risk prediction model, obtain predicted line risk indicator data corresponding to the reference line risk indicator verification set, and combine the predicted line risk indicator data and the corresponding reference line risk indicator data to obtain the risk prediction accuracy index of the line risk prediction model, the risk prediction accuracy index of the line risk prediction model is greater than the reference risk prediction accuracy index, and the risk prediction accuracy index is used to describe the probability of correctly predicting the risk of the heavy-load line to be tested through the constructed line risk prediction model.
[0041] In this embodiment, the preset model refers to a common regression model that predicts future data based on historical data; the reference risk prediction accuracy index also gives the reference accuracy of the line risk prediction model, which requires that the training and optimization of the model must meet certain requirements. The parameters of the model can be obtained through automatic parameter search and other methods, and the data set can also be preprocessed, such as data enhancement, etc.; a more accurate construction of a line risk prediction model is achieved.
[0042] Furthermore, the risk prediction accuracy index is calculated using the following formula:
[0043] Where, e is a natural constant, h is the line risk type number of the large-load line to be tested, h = 1, 2, 3, 4, h = 1 is the overload risk of the large-load line to be tested, h = 2 is the low-voltage risk of the large-load line to be tested, h = 3 is the heat loss risk of the large-load line to be tested, h = 4 is the instability risk of the large-load line to be tested, PA h is the risk prediction accuracy index of the hth type of line risk of the heavy load line to be tested, FS h is the accuracy score of the hth type of line risk of the heavy-load line to be tested, AR h is the adjusted R-squared score of the h-th line risk of the heavy-load line to be tested, BS h is the Bayesian score of the h-th line risk of the heavy-load line to be tested, CS his the bias score of the hth line risk of the large-load line to be tested, α is the correction factor of the accuracy score, β is the correction factor of the adjusted R-square score, χ is the correction factor of the Bayes score, and δ is the correction factor of the bias score; the overload risk of the large-load line to be tested is used to describe the risk caused by the line load of the large-load line to be tested exceeding the reference first line load, and the line risk indicator data is the peak load data of the large-load line to be tested; the low-voltage risk of the large-load line to be tested is used to describe the risk caused by the line voltage of the large-load line to be tested being less than the reference first line voltage, and the line risk indicator data is the preset node voltage data of the large-load line to be tested ; The heat loss risk of the large-load line to be tested is used to describe the risk caused by the heat loss of the large-load line to be tested exceeding the heat loss of the reference first line. The line risk index data is the technical equipment temperature data of the large-load line to be tested; The instability risk of the large-load line to be tested is used to describe the risk caused by the instability of the corresponding power system of the large-load line to be tested. The line risk index data is the power system output frequency data of the large-load line to be tested; The accuracy score, adjusted R-square score, Bayesian score and bias score are used to describe the degree of deviation between the predicted line risk index data obtained by the line risk prediction model and the corresponding reference line risk index data.
[0044] In this embodiment, overload risk may cause overheating of cables or equipment, or even cause a fire; low voltage risk may have an adverse effect on equipment and systems connected to the line; heat loss risk may shorten the life of equipment and require more frequent maintenance; instability risk may cause oscillation or collapse of the power system; other risks of the high-load line to be tested can also be predicted, and it is only necessary to build a prediction model based on the corresponding reference data; when the corresponding accuracy score is 0, the risk prediction accuracy index is directly 0; when the actual accuracy score and adjusted R-square score of this type of line risk are both 1, the corresponding Bayesian score is 1, and the bias score is 0, the risk prediction accuracy index of this type of line risk is 1; a more accurate evaluation of the prediction performance of the line risk prediction model is achieved.
[0045] Furthermore, the accuracy score is calculated using the following formula: In the formula, MSE h is the mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, ε is the correction factor of the mean square error data, ΔMSE h is the reference first mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ΔMSE is the reference mean square error data of the predicted line risk index data of the heavy load line to be tested.
[0046] In this embodiment, the mean square error (MSE) refers to the expected value of the square of the difference between the parameter estimate and the true value of the parameter; MSE can evaluate the degree of change of the data. The smaller the value of MSE, the better the accuracy of the prediction model in describing the experimental data. The mean square error data is a number not less than 0. When the actual mean square error data is 0, that is, The corresponding accuracy score is 1; when the actual mean square error data is equal to the reference mean square error data, that is, The corresponding accuracy score is When the actual mean square error data is greater than the reference mean square error data, the corresponding accuracy score is 0; this achieves more efficient calculation of the accuracy scores of various line risks.
[0047] Furthermore, the adjusted R-squared score is calculated using the following formula: In the formula, R2 h is the determination coefficient of the predicted line risk index data of the hth line risk of the heavy load line to be tested, TR h is the adjusted R-square data of the predicted line risk index data of the h-th line risk, μ is the correction factor of the determination coefficient, and η is the correction factor of the adjusted R-square data.
[0048] In this embodiment, R2 (R-Square), the coefficient of determination ranges from [0,1], and is usually used to describe how well the data fits the model: if the result is 0, it means that the model fits very poorly, and if the result is 1, it means that the model has no errors. Generally speaking, the larger the value, the better the model fits. The adjusted R-square is generally smaller than the coefficient of determination. When the actual coefficient of determination is 0, the adjusted R-square score is directly 0. The adjusted R-square score increases with the increase of the adjusted R-square data. This achieves a more accurate evaluation of the fitting performance of the route risk prediction model.
[0049] Furthermore, the Bayesian score is calculated using the following formula: In the formula, BIC h is the Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested, ΔBIC is the correction factor for the Bayesian information criterion data for predicting line risk index data. h is the reference Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy load line to be tested, ΔB h It is the reference Bayesian information criterion deviation data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested.
[0050] In this embodiment, the Bayesian information criterion (BIC) is a method in statistics for selecting the best model from a limited set of models, which helps to avoid overfitting; within a certain range, the smaller the Bayesian information accuracy data, the better the fitting performance of the model; a more comprehensive evaluation of the fitting performance of the route risk prediction model is achieved.
[0051] Furthermore, the bias score was calculated using the following formula: In the formula, MAE h is the mean absolute error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, ΔMAE h is the reference mean absolute error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ζ is the correction factor of the mean absolute error data of the predicted line risk index data.
[0052] In this embodiment, the mean absolute error (MAE) is the average value of the absolute error, which is a number not less than 0. The mean absolute error can better reflect the actual situation of the prediction value error; when the actual mean absolute error data obtained is not less than the corresponding reference mean absolute error data, the bias score is directly 1, indicating that the deviation of the predicted line risk index data of this type of line risk is large; when the actual mean absolute error data obtained is 0, that is, The corresponding bias score is 0, indicating that the prediction effect is very good at this time; a more accurate assessment of the deviation of the risk indicator data of the predicted line is achieved.
[0053] like Figure 4As shown, it is a structural diagram of a heavy-load line risk prediction device based on digital twin provided in an embodiment of the present application. A heavy-load line risk prediction device based on digital twin provided in an embodiment of the present application includes: a heavy-load line digital twin model construction module, a line risk index acquisition module and a line risk prediction module; a heavy-load line digital twin model construction module: used to construct a heavy-load line digital twin model, the heavy-load line digital twin model includes a line simulation model and a line risk prediction model, the line simulation model is used to simulate the working condition of the heavy-load line to be tested, and the line risk prediction model is used to predict the line risk index data of the heavy-load line to be tested; a line risk index acquisition module: used to obtain the line risk index data of the heavy-load line to be tested within the monitoring time period, and simulate in combination with the line simulation model to obtain the corresponding line risk index, the line risk index is used to describe the probability of line risk occurring in the heavy-load line to be tested corresponding to the line risk index data; a line risk prediction module: used to predict the line risk situation of the heavy-load line to be tested during the monitoring time period according to the calculated line risk index, and take corresponding risk prevention measures in time.
[0054] In this embodiment, the line simulation model and the line risk prediction model can be further trained and optimized based on real-time data. In theory, as the usage time increases, the prediction performance of the model will become better and better, achieving more efficient prediction of high-load line risks.
[0055] The technical solution in the above-mentioned embodiment of the present application has at least the following technical effects or advantages: relative to a method, device and terminal equipment for fault risk assessment of a power transmission line disclosed in the application with publication number: CN113033989A, the embodiment of the present application obtains the simulated load data, simulated voltage data and simulated current data of each line simulation model verification experiment, and then obtains the corresponding reference data and reference deviation, and then calculates the simulation index of the line simulation model, and finally evaluates the anti-seismic performance of the line simulation model in combination with the simulation index, thereby realizing the digitization of the simulation performance of the line simulation model, and then realizing a more accurate evaluation of the simulation performance of the line simulation model; relative to The disclosure of the application with publication number: CN116757465A discloses a line risk assessment method and device based on a dual training weight allocation model. The embodiment of the application calculates the mean square error data, determination coefficient, adjusted R-square data, Bayesian information criterion data and mean absolute error data of the predicted line risk index data of various line risks, and then obtains the corresponding reference data, and then calculates the risk prediction accuracy index of various line risks, and finally evaluates the prediction performance of the constructed risk prediction model according to the risk prediction accuracy index, thereby realizing the digitization of the prediction performance of the line risk prediction model, and further realizing a more comprehensive evaluation of the prediction performance of the line risk prediction model.
[0056] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0057] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0058] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0060] Although preferred embodiments of the present invention have been described, additional changes and modifications may be made to these embodiments by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A high-load line risk prediction method based on digital twins, characterized in that: The following steps are involved: Constructing a digital twin model of a heavy-load line, wherein the digital twin model of the heavy-load line includes a line simulation model and a line risk prediction model, wherein the line simulation model is used to simulate the working condition of the heavy-load line to be tested, and the line risk prediction model is used to predict line risk index data of the heavy-load line to be tested; Acquire the line risk index data of the heavy-load line to be tested within the monitoring period, and simulate in combination with the line simulation model to obtain the corresponding line risk index, wherein the line risk index is used to describe the probability of occurrence of line risk in the heavy-load line to be tested corresponding to the line risk index data; Predict the line risk situation of the heavy-load line to be tested during the monitoring period based on the calculated line risk index, and take corresponding risk prevention measures in a timely manner; The simulation index of the circuit simulation model is greater than the reference simulation index, and the simulation index is used to describe the simulation performance of the constructed circuit simulation model; The simulation index is calculated using the following formula: Where SI is the simulation index, t is the number of the circuit simulation model verification experiment, t = 1, 2, ..., T, T is the total number of circuit simulation model verification experiments, FL t and ΔFL t are the simulated load data and reference load data of the t-th line simulation model verification experiment, FU t and ΔFU t are the simulation voltage data and reference voltage data of the t-th line simulation model verification experiment, FI t and ΔFI t are respectively the simulation current data and the reference current data of the t-th line simulation model experiment, ΔL, ΔU and ΔI are respectively the reference load deviation, the reference voltage deviation and the reference current deviation; The risk prediction accuracy index of the line risk prediction model is greater than the reference risk prediction accuracy index, and the risk prediction accuracy index is used to describe the probability of correctly predicting the risk of the heavy-load line to be tested by the constructed line risk prediction model; The risk prediction accuracy index is calculated using the following formula: Where, e is a natural constant, h is the line risk type number of the large-load line to be tested, h = 1, 2, 3, 4, h = 1 is the overload risk of the large-load line to be tested, h = 2 is the low-voltage risk of the large-load line to be tested, h = 3 is the heat loss risk of the large-load line to be tested, h = 4 is the instability risk of the large-load line to be tested, PA h is the risk prediction accuracy index of the hth type of line risk of the heavy load line to be tested, FS h is the accuracy score of the hth type of line risk of the heavy-load line to be tested, AR h is the adjusted R-squared score of the h-th line risk of the heavy-load line to be tested, BS h is the Bayesian score of the h-th line risk of the heavy-load line to be tested, CS h is the bias score of the h-th line risk of the heavy-load line to be tested, α is the correction factor of the accuracy score, β is the correction factor of the adjusted R-square score, χ is the correction factor of the Bayesian score, and δ is the correction factor of the bias score.
2. A method for predicting high-load line risks based on digital twins as claimed in claim 1, characterized in that: The overload risk of the heavy load line to be tested is used to describe the risk of the heavy load line to be tested due to the line load exceeding the reference first line load, and the line risk indicator data is the peak load data of the heavy load line to be tested; The low voltage risk of the large load line to be tested is used to describe the risk caused by the line voltage of the large load line to be tested being lower than the reference first line voltage, and the line risk indicator data is the preset node voltage data of the large load line to be tested; The heat loss risk of the large-load line to be tested is used to describe the risk of the large-load line to be tested due to the line heat loss exceeding the reference first line heat loss, and the line risk indicator data is the technical equipment temperature data of the large-load line to be tested; The instability risk of the large-load line to be tested is used to describe the risk of the large-load line to be tested caused by the instability of the corresponding power system, and the line risk indicator data is the power system output frequency data of the large-load line to be tested; The accuracy score, adjusted R-square score, Bayesian score and bias score are used to describe the degree of deviation between the predicted line risk index data obtained through the line risk prediction model and the corresponding reference line risk index data.
3. A method for predicting high-load line risks based on digital twins as claimed in claim 2, characterized in that: The accuracy score is calculated using the following formula: In the formula, MSE h is the mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, ε is the correction factor of the mean square error data, ΔMSE h is the reference first mean square error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ΔMSE is the reference mean square error data of the predicted line risk index data of the heavy load line to be tested.
4. A method for predicting high-load line risks based on digital twins as claimed in claim 2, characterized in that: The adjusted R-squared score is calculated using the following formula: In the formula, R2 h is the determination coefficient of the predicted line risk index data of the hth line risk of the heavy load line to be tested, TR h is the adjusted R-square data of the predicted line risk index data of the h-th line risk, μ is the correction factor of the determination coefficient, and η is the correction factor of the adjusted R-square data.
5. A method for predicting high-load line risks based on digital twins as claimed in claim 2, characterized in that: The Bayesian score is calculated using the following formula: In the formula, BIC h is the Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested, ΔBIC is the correction factor for the Bayesian information criterion data for predicting line risk index data. h is the reference Bayesian information criterion data of the predicted line risk index data of the h-th line risk of the heavy load line to be tested, ΔB h It is the reference Bayesian information criterion deviation data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested.
6. A method for predicting high-load line risks based on digital twins as claimed in claim 2, characterized in that: The bias score is calculated using the following formula: In the formula, MAE h is the mean absolute error data of the predicted line risk index data of the h-th line risk of the heavy-load line to be tested, ΔMAE h is the reference mean absolute error data of the predicted line risk index data of the hth line risk of the heavy load line to be tested, and ζ is the correction factor of the mean absolute error data of the predicted line risk index data.
7. A high-load line risk prediction device based on digital twins, characterized in that: The digital twin-based heavy-load line risk prediction device is used to execute the digital twin-based heavy-load line risk prediction method according to any one of claims 1 to 6, and the digital twin-based heavy-load line risk prediction device includes a heavy-load line digital twin model construction module, a line risk index acquisition module, and a line risk prediction module; The heavy-load line digital twin model construction module is used to construct a heavy-load line digital twin model, wherein the heavy-load line digital twin model includes a line simulation model and a line risk prediction model, wherein the line simulation model is used to simulate the working condition of the heavy-load line to be tested, and the line risk prediction model is used to predict the line risk index data of the heavy-load line to be tested; The line risk index acquisition module is used to acquire line risk index data of the heavy load line to be tested within the monitoring period, and simulate with the line simulation model to acquire the corresponding line risk index, wherein the line risk index is used to describe the probability of occurrence of line risk in the heavy load line to be tested corresponding to the line risk index data; The line risk prediction module is used to predict the line risk situation of the heavy-load line to be tested during the monitoring period according to the calculated line risk index, and to take corresponding risk prevention measures in time.
Citation Information
Cited By
Risk assessment method for heavy-load line
CN120410206A
