Lightning protection device emergency plan planning system based on transformer substation grounding grid fault

Abstract

The invention relates to the technical field of emergency plan planning, and provides a lightning protection device emergency plan planning system based on transformer substation grounding grid faults, a data collection module collects historical data and real-time data of a transformer substation, and a prediction module constructs a grounding grid corrosion prediction model based on soil environment parameters in the historical data. A thunderstorm probability prediction model is generated based on radar meteorological data, a real-time risk assessment module establishes a grounding grid lightning stroke risk association model based on prediction results of the two models, and the lightning stroke risk of the current grounding grid is judged according to real-time data; the emergency plan planning module plans an emergency plan corresponding to the lightning stroke risk based on the lightning stroke risk in combination with a built-in standard plan; the thunderstorm risk association model is constructed by combining the prediction results of the linear regression model and the logistic regression model, and the emergency plan planning module plans the emergency plan according to the thunderstorm risk association model, so that the whole-course control of the lightning stroke risk of the transformer substation from prediction to the emergency plan is realized, and the stable operation of a power grid is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency plan planning, and more specifically, to a lightning protection device emergency plan planning system based on substation grounding grid faults. Background Art

[0002] In modern power systems, substations are key links in power transmission and distribution, and their stable operation is crucial for the reliability of the entire power grid. Frequent lightning activities pose a serious threat to the safety of substations. Faults caused by lightning strikes occur frequently every year, bringing challenges to power supply and equipment safety.

[0003] With the expansion of the scale and the improvement of the intelligence level of power systems, the importance of substation lightning protection grounding grids has become increasingly prominent. As the core channel for lightning current discharge, the reliability of the grounding grid is directly related to equipment safety and the stable operation of the power grid. In recent years, certain progress has been made in grounding grid design, material optimization, and monitoring technologies at home and abroad. For example, technologies such as square-hole grounding grids and unequal-spacing layouts are used to improve the voltage equalization and current dispersion capabilities, and deep-well grounding, electrolytic ion grounding, etc. are combined to reduce the grounding resistance in areas with high soil resistivity. Galvanized steel, copper, and anti-rust coatings are used to extend the service life of grounding electrodes, and the corrosion problem of welding points is reduced through the improvement of construction techniques. Some power grid companies deploy grounding conduction detection and sensor networks, and dynamically adjust the detection plan in combination with meteorological data.

[0004] However, traditional technologies still mainly rely on manual inspections and periodic maintenance, lacking the ability to dynamically respond to complex working conditions. Although some systems have introduced the Internet of Things and big data analysis, there are still the following deficiencies:

[0005] Insufficient prediction of the corrosion faults of the grounding grid itself and thunderstorm prediction. The grounding grid is buried underground for a long time and is affected by various factors in the soil. Corrosion is common, and different corrosion conditions lead to different lightning withstand capabilities of the grounding grid. The protection time limit of conventional galvanized anti-corrosion measures is limited, and there is a lack of a precise corrosion rate prediction model. Once the grounding grid fails due to corrosion, the lightning current cannot be smoothly introduced into the ground during a lightning strike, which will cause overvoltage of in-station equipment and increase the risk of equipment damage. Summary of the Invention

[0006] To solve the above technical problems, the present invention provides a lightning protection device emergency plan planning system based on substation grounding grid faults. By using linear regression and logistic regression models to predict the corrosion rate of the grounding grid and the probability of thunderstorm occurrence respectively, a thunderstorm risk correlation model is constructed in combination with a real-time risk assessment module, and an emergency plan is planned by an emergency plan planning module accordingly, so as to achieve the overall control of substation lightning strike risks from prediction to emergency plan, ensure the stable operation of the power grid, and solve the problems in the prior art.

[0007] A lightning protection device emergency plan planning system based on the grounding grid failure of a substation, comprising:

[0008] A data collection module, which collects historical data and real-time data of the substation. Both the historical data and the real-time data include soil environment parameter data of the grounding grid, substation meteorological data, and operating status data of the grounding grid equipment;

[0009] A prediction module, which is data-connected to the data collection module. Based on the soil environment parameters in the collected historical data, a grounding grid corrosion prediction model is constructed to predict the corrosion rate of the grounding grid. Based on the historical radar meteorological data, a thunderstorm probability prediction model is generated to predict the occurrence probability of thunderstorms;

[0010] A real-time risk assessment module, which is data-connected to the prediction module and the data collection module. Based on the grounding grid corrosion prediction model and the thunderstorm probability prediction model, a grounding grid lightning strike risk association model is established to perform real-time analysis on the real-time data and judge the lightning strike risk of the current grounding grid;

[0011] An emergency plan planning module, which is data-connected to the real-time risk assessment module. Based on the lightning strike risk and combined with the built-in standard plan, an emergency plan corresponding to the lightning strike risk is planned.

[0012] Preferably, the data collection module collects grounding grid-related data through various sensors;

[0013] The soil environment parameter data includes data linearly related to the grounding grid corrosion rate, including soil temperature, soil humidity, soil resistivity, and the exposed area of the grounding grid material; the substation meteorological data includes lightning activity frequency, intensity, and thunderstorm warning information; the operating status data of the grounding grid equipment includes whether the grounding grid is faulty information / fault type information and fault node information.

[0014] Preferably, the prediction module includes a data preprocessing unit, a grounding grid corrosion prediction unit, and a thunderstorm probability prediction unit;

[0015] The data preprocessing unit is used to remove noise and outliers from the collected data and provide effective data samples for the subsequent grounding grid corrosion prediction unit and thunderstorm probability prediction unit;

[0016] The grounding grid corrosion prediction unit, based on the data samples linearly related to the grounding grid corrosion rate after data preprocessing, uses a linear regression model to construct a linear regression equation and adopts the least squares method to solve the coefficients to minimize the sum of squared errors, so as to realize the prediction of the grounding grid corrosion rate;

[0017] The thunderstorm probability prediction unit uses the logistic regression model to establish a calculation method for the probability of thunderstorm occurrence based on the pre-processed substation meteorological data, and uses the maximum likelihood estimation method to solve the coefficients to maximize the log-likelihood function, so as to realize the prediction of the probability of thunderstorm occurrence.

[0018] Preferably, the specific process for the grounding grid corrosion prediction unit to establish a grounding grid corrosion prediction model is as follows:

[0019] First, select n data samples with a linear relationship with the grounding grid corrosion rate from historical data, and use the linear regression model to construct a grounding grid corrosion prediction model to predict the grounding grid corrosion rate y; assume that there are m independent variables (x1, x2, x3,... x m ) that are nearly linear factors affecting the corrosion rate in the data samples, and construct a linear regression equation:

[0020] y = α0 + α1x1 + α2x2 +... + α m x m + ∈

[0021] Among them, α0 is the intercept, α i (i = 1, 2, 3,..., m) is the regression coefficient, and ∈ is the error term;

[0022] Model training, solve the α i coefficient by the least squares method to minimize the sum of squared errors S:

[0023]

[0024] Among them, i represents the index of the data sample, j represents the index of the independent variable in the data sample, y i is the actual corrosion rate of the i-th sample, x ij is the value of the j-th characteristic independent variable related to the grounding grid corrosion rate in the i-th sample, and n represents the total number of samples.

[0025] Preferably, the specific process for the thunderstorm probability prediction unit to establish a thunderstorm probability prediction model is as follows:

[0026] First, based on n pre-processed substation meteorological data samples, use the logistic regression model to construct a thunderstorm probability prediction model. Assume that the independent variable characteristics with a near-linear relationship with the thunderstorm probability in the substation meteorological data samples are (y1, u2, u3,... u c ), and the logistic regression model outputs the probability of thunderstorm occurrence P(v = 1|u), where v represents whether a thunderstorm occurs, 1 represents occurrence, and 0 represents non-occurrence;

[0027]

[0028] Among them, c is the number of input features, β0 is the intercept in the thunderstorm probability prediction unit, and β i (i = 1, 2, 3,..., c) are the logistic coefficients;

[0029] Model training, solve for β i coefficient by the maximum likelihood estimation method to maximize the log-likelihood function:

[0030]

[0031] Among them, v i is the thunderstorm occurrence situation of the i-th sample, 0 or 1, and u i is the meteorological feature related to the thunderstorm occurrence in the i-th data sample.

[0032] Preferably, the real-time risk assessment module constructs a lightning strike risk association model based on the grounding grid fault and the thunderstorm probability. The specific construction process of the model is as follows:

[0033] Data preparation: Input the real-time data collected by the data collection module into the established grounding grid corrosion prediction model and thunderstorm probability prediction model, and calculate the current grounding grid corrosion rate prediction result and thunderstorm occurrence probability prediction result as the input data of the lightning strike risk association model;

[0034] At the same time, classify the lightning strike event data, that is, the damage degree to the substation according to certain rules to obtain the lightning strike risk level R, and R represents the degree index of the lightning strike risk;

[0035] Use the Bayesian network to construct the lightning strike risk association model. Let the grounding grid corrosion rate prediction results based on n historical data samples be (y1, y2, y i ,...y n ). The Bayesian formula:

[0036]

[0037] Among them, P(R|y i ) represents the occurrence probability of lightning strike risk R when the corrosion rate is y i in the historical data samples. y i represents the different corrosion rates of the grounding grid calculated from different data samples. P(y i |R) represents the probability of the corrosion rate y i in the historical data when the lightning strike risk is R;

[0038] P(R) represents the prior probability when the lightning strike risk is R, that is, the probability that the lightning strike risk is at level R in all data samples according to historical data statistics;

[0039] P(yi ) is the prior probability of the occurrence of event y, that is, the probability that the corrosion rate of the grounding grid is y obtained according to historical data statistics. i Then, the conditional probabilities between each node are learned through historical data, and a complete Bayesian network model is trained and constructed.

[0040] Preferably, the lightning strike risk association model also establishes a lightning strike risk association model based on multiple fault types and a thunderstorm probability prediction model;

[0041] First, based on the operation status data of the grounding grid equipment collected by the data collection module;

[0042] Let the fault type of the grounding grid be F z , the total number of fault types be μ, where z represents the index of the fault type. For the case of multiple grounding grid fault types, the joint probability formula is:

[0043] P(R,F1,F2,F3,...F μ ) = P(R|F1,F2,F3,...F μ )P(F1)P(F2)...P(F μ )

[0044] Among them, P(R,F1,F2,F3,...F μ ) is the joint probability of the lightning strike risk level R and multiple grounding grid fault types F1,F2,F3,...F μ occurring simultaneously. P(R|F1,F2,F3,...F μ ) is the conditional probability when the lightning strike risk level is R under the condition that the grounding grid has a fault F z . P(F z )(z = 1,2,3,...μ) is the prior probability that the grounding grid has the z-th fault type F z in the historical data samples;

[0045] Then, the conditional probabilities between each node are learned through historical data, and a complete Bayesian network model is constructed.

[0046] Preferably, the specific process of the emergency plan planning module for emergency plan planning is as follows:

[0047] Obtain the lightning strike risk level R, the corrosion rate y of the grounding grid, and the thunderstorm occurrence probability P(v = 1|u) from the real-time risk assessment module. Set three goals: minimizing the emergency cost, minimizing the fault repair time, and minimizing the equipment damage loss. Let h be the index of the three goals, and the value of h is 1, 2, or 3. 1 represents the emergency cost, 2 represents the fault repair time, and 3 represents the equipment damage loss;

[0048] Generate a set of pre - plan solutions. According to different combinations of lightning strike risk levels, grounding grid corrosion rates, and thunderstorm occurrence probabilities, initially generate multiple emergency plans. Let the set of emergency plans be (g1, g2,..., g T ), where T is the total number of solutions;

[0049] Construct a decision matrix. For each emergency plan solution g, calculate its values under each objective function

[0050]

[0051] Emergency cost objective: Considering the cost of personnel allocation and the cost of material allocation, the emergency cost k of solution g g1 is obtained by multiplying the number of repair personnel by the per - capita cost plus the number of required materials by the unit material cost;

[0052] Fault repair time objective: Based on experience and simulation, calculate the time k from the start of the emergency to the completion of fault repair for solution g g2 ;

[0053] Equipment damage loss objective: According to the lightning strike risk level R, grounding grid corrosion rate y, and thunderstorm occurrence probability, combined with the importance and vulnerability of the equipment, calculate the equipment damage loss amount k under solution g g3 ;

[0054] Normalize the decision matrix X to obtain the normalized decision matrix:

[0055]

[0056] Where:

[0057]

[0058] Determine the positive and negative ideal solutions. The positive ideal solution Where:

[0059]

[0060] Among them, the emergency cost minimization objective is the minimum value of the normalized emergency cost, the fault repair time minimization objective is the minimum value of the normalized fault repair time, and the equipment damage loss minimization objective is the minimum value of the normalized equipment damage loss.

[0061] The negative ideal solution Where:

[0062]

[0063] That is, the maximum value of each target value after normalization. Since it is to minimize the target, the maximum value is taken as the negative ideal solution;

[0064] Calculate the distance from the emergency plan to the positive ideal solution:

[0065]

[0066] Calculate the distance to the negative ideal solution of the emergency plan

[0067]

[0068] Calculate the relative closeness B g :

[0069]

[0070] Among them, B g The closer it is to 1, the better the solution g; according to the relative closeness B g Select the emergency plan with the largest relative closeness as the final plan; combine this plan with the built-in plan template, personnel allocation, material allocation and emergency measure execution steps as the output of the emergency plan planning module.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] 1. The present invention uses a linear regression model through the grounding grid corrosion prediction unit and a logistic regression model through the thunderstorm probability prediction unit to provide effective data samples, and the real-time risk assessment module constructs a lightning strike risk association model based on the Bayesian network using the prediction results of the former two, realizing the mutual association and collaborative operation of multiple algorithms, achieving a full-process accurate analysis from data processing to single-factor prediction to comprehensive risk assessment, and providing comprehensive protection for the lightning protection safety of substations.

[0073] 2. The present invention constructs a lightning strike risk association model based on grounding grid faults and thunderstorm probabilities through the real-time risk assessment module, and the emergency plan planning module plans the emergency plan according to the lightning strike risk level, grounding grid corrosion rate and thunderstorm occurrence probability, realizing the real-time analysis and dynamic response to the lightning strike risk of the substation, improving the substation's ability to cope with lightning strike disasters, reducing the risk of equipment damage caused by lightning strikes, and ensuring the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 is a schematic diagram of the working process of the modules of the present invention;

[0075] Figure 2 is a schematic diagram of the specific working process of the prediction module and the real-time risk assessment module of the present invention;

[0076] Figure 3 It is a schematic diagram of the specific working process of the emergency plan planning module of the present invention. Specific embodiments

[0077] The following further describes the embodiments of the present invention in detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0078] The present invention provides a lightning protection device emergency plan planning system based on substation grounding grid faults, including:

[0079] A data collection module, which collects historical data and real-time data of the substation. Both the historical data and real-time data include soil environment parameter data of the grounding grid, substation meteorological data, and operating status data of the grounding grid equipment;

[0080] A prediction module, which is data-connected to the data collection module. Based on the soil environment parameters in the collected historical data, a grounding grid corrosion prediction model is constructed to predict the corrosion rate of the grounding grid. Based on the historical radar meteorological data, a thunderstorm probability prediction model is generated to predict the occurrence probability of thunderstorms;

[0081] A real-time risk assessment module, which is data-connected to the prediction module and the data collection module. Based on the grounding grid corrosion prediction model and the thunderstorm probability prediction model, a grounding grid lightning strike risk association model is established to perform real-time analysis on the real-time data and judge the current lightning strike risk of the grounding grid;

[0082] An emergency plan planning module, which is data-connected to the real-time risk assessment module. Based on the lightning strike risk, an emergency plan corresponding to the lightning strike risk is planned in combination with the built-in standard plan.

[0083] Example 1:

[0084] As Figures 1-3 shown, in this embodiment, for a substation in a certain area, due to the relatively high soil humidity and salt content in the local area, the grounding grid has been in an environment prone to corrosion for a long time. At the same time, lightning activities are frequent in this area in summer, bringing great challenges to the safe operation of the substation. The traditional manual inspection and periodic maintenance methods are difficult to cope with the complex and changeable working conditions, and it often happens that the grounding grid corrosion faults are not discovered in time, resulting in equipment damage during thunderstorm weather. Therefore, the present invention is introduced.

[0085] The data collection module collects the soil environmental parameter data of the grounding grid of this substation in the past year through various sensors, such as soil temperature, humidity, resistivity, the exposed area of the grounding grid material, the meteorological data of the substation, the lightning activity frequency, intensity, and thunderstorm warning information, as well as the operation status data of the grounding grid equipment. The data preprocessing unit processes these data to remove noise and outliers, providing effective data samples for subsequent model construction.

[0086] First, the grounding grid corrosion prediction unit selects n data samples with a linear relationship with the grounding grid corrosion rate from historical data and uses a linear regression model to construct a grounding grid corrosion prediction model to predict the grounding grid corrosion rate y; assume that there are m independent variables (x1, x2, x3,... x m ) that are nearly linear factors affecting the corrosion rate in the data samples, and construct a linear regression equation:

[0087] y = α0 + α1x1 + α2x2 +... + α m x m + ∈

[0088] Among them, α0 is the intercept, α i (i = 1, 2, 3,..., m) are the regression coefficients, and ∈ is the error term;

[0089] Model training, solve for the α i coefficients by the least squares method to minimize the sum of squared errors S:

[0090]

[0091] Among them, i represents the index of the data sample, j represents the index of the independent variable in the data sample, y i is the actual corrosion rate of the i-th sample, x ij is the value of the j-th characteristic independent variable related to the grounding grid corrosion rate in the i-th sample, and n represents the total number of samples.

[0092] Then, the thunderstorm probability prediction unit constructs a thunderstorm probability prediction model based on the n substation meteorological data samples after data preprocessing using a logistic regression model. Assume that the independent variable characteristics in the substation meteorological data samples that have a nearly linear relationship with the thunderstorm probability are (u1, u2, u3,... u c ), and the logistic regression model outputs the thunderstorm occurrence probability P(v = 1|u), where v represents whether a thunderstorm occurs, 1 represents occurrence, and 0 represents non-occurrence;

[0093]

[0094] Among them, c is the number of input features, β0 is the intercept in the thunderstorm probability prediction unit, β i(i = 1, 2, 3, ..., c) is a logical coefficient;

[0095] Model training, solve β by the maximum likelihood estimation method i Coefficient to maximize the log-likelihood function:

[0096]

[0097] where v i is the thunderstorm occurrence situation of the i-th sample, 0 or 1, and u i is the meteorological characteristics related to the thunderstorm occurrence in the i-th data sample.

[0098] The real-time risk assessment module constructs a lightning strike risk association model based on the grounding grid fault and thunderstorm probability. The specific construction process of the model is as follows:

[0099] Data preparation: Input the real-time data collected by the data collection module into the established grounding grid corrosion prediction model and thunderstorm probability prediction model, and calculate the current grounding grid corrosion rate prediction result and thunderstorm occurrence probability prediction result as the input data of the lightning strike risk association model;

[0100] At the same time, classify the lightning strike event data, that is, the damage degree to the substation according to certain rules to obtain the lightning strike risk level R, and R represents the degree index of the lightning strike risk;

[0101] Use the Bayesian network to construct the lightning strike risk association model. Let the grounding grid corrosion rate prediction results based on n historical data samples be (y1, y2, y i ,... y n ), and the Bayesian formula:

[0102]

[0103] where P(R|y i ) represents the occurrence probability of lightning strike risk R when the corrosion rate is y i in the historical data samples, and y i represents different corrosion rates of the grounding grid calculated from different data samples. P(y i |R) represents the probability of the corrosion rate being y i in the historical data when the lightning strike risk is R;

[0104] P(R) represents the prior probability when the lightning strike risk is R, that is, the probability that the lightning strike risk is at level R in all data samples according to historical data statistics;

[0105] P(y i ) is the prior probability of the event y occurring, that is, the probability that the grounding grid has a corrosion rate of y according to historical data statisticsi The probability, and then, learn the conditional probabilities between each node through historical data, and train and construct a perfect Bayesian network model.

[0106] Meanwhile, the lightning strike risk association model also establishes a lightning strike risk association model based on multiple fault types and the thunderstorm probability prediction model;

[0107] First, based on the operation status data of the grounding grid equipment collected in the data collection module;

[0108] Let the fault type of the grounding grid be F z , and the total number of fault types be μ. Among them, z represents the index of the fault type. For the case of multiple grounding grid fault types, the joint probability formula is:

[0109] P(R,F1,F2,F3,...F μ )=P(R|F1,F2,F3,...F μ )P(F1)P(F2)...P(F μ )

[0110] Among them, P(R,F1,F2,F3,...F μ ) is the joint probability of the lightning strike risk level R and multiple grounding grid fault types F1,F2,F3,...F μ occurring simultaneously. P(R|F1,F2,F3,...F μ ) is the conditional probability when the lightning strike risk level is R under the condition that the grounding grid has a fault F z . P(F z )(z = 1,2,3,...μ) is the prior probability that the grounding grid has the z-th fault type F z in the historical data samples;

[0111] Then, learn the conditional probabilities between each node through historical data, and construct a complete Bayesian network model.

[0112] By combining the three algorithms of linear regression, logistic regression and Bayesian network, a full-process analysis from single-factor prediction to comprehensive risk assessment is realized. Compared with the traditional method, it can more accurately predict the corrosion condition of the grounding grid and the thunderstorm occurrence probability, as well as the lightning strike risk under the combined action of the two. This joint prediction model provides a more scientific decision-making basis for the operation and maintenance personnel of the substation, avoids equipment damage and power outages caused by grounding grid corrosion faults and thunderstorm weather, and improves the operation reliability of the substation.

[0113] Embodiment 2:

[0114] Such as Figures 1-3As shown, in this embodiment, during a thunderstorm-prone period in a certain summer in the substation of Embodiment 1, affected by the recent continuous alternation of high temperature and heavy rain, the soil humidity and temperature changed violently, further accelerating the corrosion of the grounding grid. At this time, it is difficult for the traditional operation and maintenance method to judge the lightning strike risk faced by the substation according to the complex and changeable real-time situation, and it is also impossible to quickly formulate a targeted emergency plan that suits the actual situation.

[0115] The real-time risk assessment module analyzes the real-time data of the substation according to the established grounding grid corrosion prediction model and thunderstorm probability prediction model, and then combines the current lightning strike risk level calculated through the lightning strike risk correlation model.

[0116] Then, the emergency plan planning module obtains the lightning strike risk level R, the grounding grid corrosion rate y, and the thunderstorm occurrence probability P(v = 1|u) from the real-time risk assessment module, and sets three goals: minimizing the emergency cost, minimizing the fault repair time, and minimizing the equipment damage loss. Let h be the index of the three goals, and the value of h is 1, 2, or 3. 1 represents the emergency cost, 2 represents the fault repair time, and 3 represents the equipment damage loss;

[0117] Generate a set of plan schemes. According to different combinations of the lightning strike risk level, the grounding grid corrosion rate, and the thunderstorm occurrence probability, initially generate multiple emergency plans. Let the emergency plan set be (g1, g2,..., g T ), where T is the total number of schemes;

[0118] Construct a decision matrix. For each emergency plan scheme g, calculate its value under each objective function

[0119]

[0120] Emergency cost objective: Considering the personnel deployment cost and the material deployment cost, the emergency cost k of scheme g g1 is obtained by multiplying the number of repair personnel by the per capita cost plus the number of required materials by the unit material cost;

[0121] Fault repair time objective: Based on experience and simulation, calculate the time k from the start of the emergency to the completion of the fault repair for scheme g g2 ;

[0122] Equipment damage loss objective: According to the lightning strike risk level R, the grounding grid corrosion rate y, and the thunderstorm occurrence probability, combined with the importance and vulnerability of the equipment, calculate the equipment damage loss amount k under scheme g g3 ;

[0123] Normalize the decision matrix X to obtain the normalized decision matrix:

[0124]

[0125] Among them:

[0126]

[0127] Determine the positive and negative ideal solutions. The positive ideal solution Among them:

[0128]

[0129] Among them, the goal of minimizing emergency costs is the minimum value of the normalized emergency costs. The goal of minimizing the fault repair time is the minimum value of the normalized fault repair time. The goal of minimizing the loss of equipment damage is the minimum value of the normalized loss of equipment damage.

[0130] The negative ideal solution Among them:

[0131]

[0132] That is, the maximum value of each normalized target value. Since it is a minimization goal, the maximum value is taken as the negative ideal solution;

[0133] Calculate the distance from the emergency plan to the positive ideal solution:

[0134]

[0135] Calculate the distance from the emergency plan to the negative ideal solution

[0136]

[0137] Calculate the relative closeness B g :

[0138]

[0139] Among them, B g The closer it is to 1, the better the solution g; According to the relative closeness B g Select the emergency plan with the largest relative closeness as the final plan; Combine this plan with the built-in plan template, personnel allocation, material allocation, and emergency measure execution steps as the output of the emergency plan planning module.

[0140] By adopting the multi-objective decision-making method, considering multiple factors such as emergency costs, fault repair time, and equipment damage loss, it is possible to formulate the optimal emergency plan according to different lightning strike risk situations. Compared with the traditional emergency plan, it is more targeted and scientific.

[0141] By means of real-time risk assessment and precise emergency plan planning, the risk of damage to the substation during thunderstorm weather has been effectively reduced, ensuring the stability of power supply.

[0142] The embodiments of the present invention are given for the purposes of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A lightning protection device emergency plan planning system based on substation grounding grid faults, characterized in that, Including: A data collection module that collects historical data and real-time data of a substation. Both the historical data and the real-time data include soil environmental parameter data of the grounding grid, substation meteorological data, and operation status data of the grounding grid equipment; A prediction module that is data-connected to the data collection module. Based on the soil environmental parameters in the collected historical data, it constructs a grounding grid corrosion prediction model to predict the corrosion rate of the grounding grid. Based on the historical radar meteorological data, it generates a thunderstorm probability prediction model to predict the occurrence probability of thunderstorms; A real-time risk assessment module that is data-connected to the prediction module and the data collection module. Based on the grounding grid corrosion prediction model and the thunderstorm probability prediction model, it establishes a grounding grid lightning strike risk association model, conducts real-time analysis on the real-time data, and judges the current lightning strike risk of the grounding grid; An emergency plan planning module that is data-connected to the real-time risk assessment module. Based on the lightning strike risk and combined with the built-in standard plan, it plans an emergency plan corresponding to the lightning strike risk.

2. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 1, wherein: The data collection module collects grounding grid-related data through various sensors; The soil environmental parameter data includes data linearly related to the grounding grid corrosion rate, including soil temperature, soil humidity, soil resistivity, and the exposed area of the grounding grid material. The substation meteorological data includes lightning activity frequency, intensity, and thunderstorm warning information. The operation status data of the grounding grid equipment includes whether the grounding grid is faulty information, fault type information, and fault node information.

3. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 1, wherein: The prediction module includes a data preprocessing unit, a grounding grid corrosion prediction unit, and a thunderstorm probability prediction unit; The data preprocessing unit is used to remove noise and outliers from the collected data, and provide effective data samples for the subsequent grounding grid corrosion prediction unit and thunderstorm probability prediction unit; The grounding grid corrosion prediction unit, based on the data samples linearly related to the grounding grid corrosion rate after data preprocessing, uses a linear regression model to construct a linear regression equation, and adopts the least squares method to solve the coefficients to minimize the sum of squared errors, so as to realize the prediction of the grounding grid corrosion rate; The thunderstorm probability prediction unit, based on the data preprocessed and the substation meteorological data, uses a logistic regression model to establish a calculation method for the occurrence probability of thunderstorms, and adopts the maximum likelihood estimation method to solve the coefficients to maximize the log-likelihood function, so as to realize the prediction of the occurrence probability of thunderstorms.

4. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 3, wherein: The specific process for the grounding grid corrosion prediction unit to establish a grounding grid corrosion prediction model is as follows: First, select n data samples in the historical data that have a linear relationship with the corrosion rate of the grounding grid, and use a linear regression model to construct a grounding grid corrosion prediction model to predict the grounding grid corrosion rate y; assume that there are m independent variables (x1, x2, x3,... x m ) that affect the corrosion rate in the data samples, and construct a linear regression equation: y = α0 + α1x1 + α2x2 +... + α m x m + ∈ where α0 is the intercept, and α i (i = 1, 2, 3, ..., m) are the regression coefficients, and ∈ is the error term; Model training, solving for α by the least squares method i Coefficient, minimizing the sum of squared errors S: where, i represents the index of the data sample, j represents the index of the independent variable in the data sample, and y i is the actual corrosion rate of the i-th sample, and x ij is the value of the j-th characteristic independent variable in the i-th sample that has a linear relationship with the corrosion rate of the grounding grid, and n represents the total number of samples.

5. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 3, wherein: The specific process for the thunderstorm probability prediction unit to establish a thunderstorm probability prediction model is as follows: First, based on the n meteorological data samples of substations after data preprocessing, a thunderstorm probability prediction model is constructed using a logistic regression model. Let the independent variable features in the meteorological data samples of substations that have a near-linear relationship with the thunderstorm probability be (u1, u2, u3,... u c ), and the logistic regression model outputs the thunderstorm occurrence probability P(v = 1|u), where v represents whether a thunderstorm occurs, 1 represents occurrence, and 0 represents non-occurrence; where c is the number of input features, β0 is the intercept in the thunderstorm probability prediction unit, and β i (i = 1, 2, 3, ..., c) are the logistic coefficients; Model training, solving for β by the maximum likelihood estimation method i Coefficient to maximize the log-likelihood function: where v i is the thunderstorm occurrence of the i-th sample, 0 or 1, and u i are the meteorological characteristics related to the thunderstorm occurrence in the i-th data sample.

6. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 1, characterized in that: The specific process for the real-time risk assessment module to construct a lightning strike risk association model based on grounding grid faults and thunderstorm probabilities is as follows: Data preparation: Input the real-time data collected by the data collection module into the established grounding grid corrosion prediction model and thunderstorm probability prediction model, calculate the current grounding grid corrosion rate prediction result and thunderstorm occurrence probability prediction result, and use them as the input data of the lightning strike risk association model; Meanwhile, the lightning strike event data, that is, the degree of damage to the substation is classified according to certain rules to obtain the lightning strike risk level R, where R represents the degree index of the lightning strike risk; Use a Bayesian network to construct a lightning strike risk association model. Let the prediction results of the grounding grid corrosion rate based on n historical data samples be (y1, y2, y i ,...y n ). Bayes' formula: Among them, P(R|y i ) represents the occurrence probability of lightning strike risk R when the corrosion rate is y i in the historical data samples. y i represents different corrosion rates corresponding to the grounding grid calculated from different data samples. P(y i |R) represents the probability that the corrosion rate is y i in the historical data when the lightning strike risk is R; P(R) represents the prior probability when the lightning strike risk is R, that is, the probability that the lightning strike risk is at level R in all data samples according to historical data statistics; P(y i ) is the prior probability of the occurrence of event y, that is, the probability that the corrosion rate of the grounding grid is y obtained by statistical analysis of historical data. i Then, the conditional probabilities between each node are learned from historical data, and a complete Bayesian network model is trained and constructed.

7. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 6, characterized in that: The lightning strike risk association model also establishes a lightning strike risk association model based on multiple fault types and thunderstorm probability prediction models; First, based on the operation status data of the grounding grid equipment collected by the data collection module; Let the fault type of the grounding grid be F z , and the total number of fault types be μ. Among them, z represents the index of the fault type. For the case of multiple grounding grid fault types, the joint probability formula is as follows: P(R,F1,F2,F3,...F μ ) = P(R|F1,F2,F3,...F μ )P(F1)P(F2)...P(F μ ) Among them, P(R, F1, F2, F3,...F μ ) is the joint probability of the lightning strike risk level R and multiple grounding grid fault types F1, F2, F3,...F μ occurring simultaneously. P(R|F1, F2, F3,...F μ ) is the conditional probability when the lightning strike risk level is R under the condition that the grounding grid fails F z . P(F z )(z = 1, 2, 3,...μ) is the prior probability of the z-th fault type F z of the grounding grid appearing in the historical data sample; Next, the conditional probabilities between each node are learned through historical data to construct a complete Bayesian network model.

8. The lightning protection device emergency plan planning system based on the substation grounding grid fault according to claim 1, characterized in that: The specific process of the emergency plan planning module for emergency plan planning is as follows: Obtain the lightning strike risk level R, the grounding grid corrosion rate y, and the thunderstorm occurrence probability P(v = 1|u) from the real-time risk assessment module, and set three goals: minimizing the emergency cost, minimizing the fault repair time, and minimizing the equipment damage loss. Let h be the index of the three goals, and the value of h is 1, 2, or 3. 1 represents the emergency cost, 2 represents the fault repair time, and 3 represents the equipment damage loss; Generate a set of pre - plan solutions. According to different combinations of lightning strike risk levels, grounding grid corrosion rates, and thunderstorm occurrence probabilities, initially generate multiple emergency plans. Let the emergency plan set be (g1, g2,..., g T ), where T is the total number of solutions; Construct a decision matrix and calculate the values of each emergency response plan \(g\) under various objective functions Emergency cost target: Considering the personnel deployment cost and the material deployment cost, the emergency cost k of Plan g g1 is obtained by multiplying the number of emergency repair personnel by the per capita cost and adding the number of required materials multiplied by the unit material cost; Fault repair time target: Based on experience and simulation, calculate the time k from the start of the emergency response to the completion of fault repair for solution g g2 ; Equipment damage loss target: According to the lightning strike risk level R, the grounding grid corrosion rate y, and the probability of thunderstorm occurrence, combined with the importance and vulnerability of the equipment, calculate the equipment damage loss amount k under the plan g g3 ; Normalize the decision matrix X to obtain the normalized decision matrix: Where: Determine the positive and negative ideal solutions, positive ideal solution Wherein: Among them, the emergency cost minimization objective is the minimum value of the normalized emergency cost, and the fault repair time minimization objective is the minimum value of the normalized fault repair time, and the equipment damage loss minimization objective is the minimum value of the normalized equipment damage loss; Negative ideal solution Wherein: That is, the maximum value of each target value after normalization. Since it is a minimization target, the maximum value is taken as the negative ideal solution; Calculate the distance from the emergency plan to the positive ideal solution: Calculate the distance from the emergency plan to the negative ideal solution Calculate the relative proximity degree B g : Among them, B g The closer it is to 1, the better the solution g is; according to the relative closeness degree B g Select the emergency plan with the largest relative closeness degree as the final plan according to the size of; combine this plan with the built-in plan template, personnel allocation, material allocation and emergency measure execution steps as the output of the emergency plan planning module.

Images