A power distribution network switch layout optimization method

Abstract

The present application relates to power distribution network planning and operation optimization technical field, disclose a kind of switch layout optimization method of power distribution network, comprising: obtaining power distribution network basic data;Different switch point parameters are simulated for fault scenario, obtain the fault index data reflecting power loss scale, switch point parameters include the spacing of automation switch section and the interconnection density of tie switch interconnection;Statistical analysis is carried out to fault index data, determine the nonlinear relationship between section spacing and power loss scale and the coupling interaction relationship between section spacing and the interconnection density of tie switch interconnection;Regression equation containing quadratic term of section spacing and its interaction term with the interconnection density of tie switch interconnection is constructed as quantitative correlation model;N-k fault reconstruction optimization model considering new energy fluctuation is established;Resilience evaluation system is constructed, and the layout scheme meeting the requirement of resilience is determined by calculating resilience evaluation system based on the above two models.The present application realizes the quantitative correlation of switch point and power loss scale, and provides scientific basis for switch layout.

Description

Technical Field

[0001] This invention relates to the field of distribution network planning and operation optimization technology, specifically to a method for optimizing the layout of distribution network switches. Background Technology

[0002] With the large-scale integration of high-proportion distributed power sources (photovoltaics, wind power, etc.) into the smart grid, the randomness and volatility of their output have led to more complex distribution network fault scenarios. The traditional Nk fault reconstruction model has not fully considered the impact of the volatility of new energy sources, and has not established a quantitative correlation between switch placement and the scale of power outages and the resilience of the distribution network. This has resulted in an expansion of the power outage range after a fault and low recovery efficiency, making it difficult to meet the safe, reliable, resilient and disturbance-resistant operation requirements of the new distribution network.

[0003] The current distribution network switch layout relies heavily on engineering experience, using deterministic load and renewable energy output modeling without distinguishing the random fluctuation characteristics of photovoltaic and wind power output; there are no quantitative calculation formulas for switch segment spacing and interconnection density, relying on empirical values ​​from guidelines; fault analysis only uses N-1 single fault verification, lacking N-2 double fault verification; fault reconfiguration uses a fixed power flow model, without considering the islanding power supply capability of distributed generation sources; distribution network resilience is only qualitatively evaluated, without a quantitative indicator system, making it impossible to quantify the correlation between switch layout and power outages and resilience.

[0004] Existing technologies (such as CN121480313A) have proposed a two-layer planning framework for distribution network switch configuration, but they still have the following shortcomings: (1) they do not consider the impact of high-proportion distributed power source volatility on the optimization effect of switch placement; (2) they do not establish switch placement parameters (segment spacing L, interconnection density). ) and the scale of power outage (power outage load rate) Power outage node rate (3) The direct quantitative correlation of the Nk fault reconstruction model cannot provide a specific placement threshold; (4) The volatility of new energy sources is not taken into account in the Nk fault reconstruction model. Summary of the Invention

[0005] This invention provides a method for optimizing the layout of distribution network switches. It overcomes the shortcomings of existing switch placement methods, such as reliance on experience, failure to account for random fluctuations in renewable energy sources, inadequate fault reconfiguration constraints, and a lack of quantitative correlation between switch placement and power outage / resilience. Based on four original fault propagation laws, a quantitative placement system is established to achieve precise switch deployment in different regions (urban areas / industrial parks / rural areas), reducing the scale of power outages and improving fault reconfiguration efficiency and the distribution network's resilience against disturbances.

[0006] This invention is achieved through the following technical solution:

[0007] A method for optimizing the layout of switches in a power distribution network includes the following steps:

[0008] Acquire basic data of the distribution network, including distribution network topology and operating status data, line fault rate data, load data, and historical output data of distributed power sources;

[0009] Based on the aforementioned fundamental data, fault scenario simulations were conducted for different switch placement parameters to obtain fault indicator data reflecting the scale of system power loss under each switch placement scheme. The fault scenarios included single-line and dual-line fault scenarios. The switch placement parameters included the automatic switch segment spacing L and the interconnection density of tie switches. ;

[0010] Statistical analysis was performed on the fault index data to determine the nonlinear relationship between the segment spacing L and the power outage scale, as well as the relationship between the segment spacing L and the interconnection density of the tie switch. The coupling and interaction relationships between them;

[0011] Based on the nonlinear relationship and the coupled interaction relationship, a quadratic term containing the piecewise spacing is constructed. and the interaction term between the segment spacing and the interconnection density of the tie switch. The regression equation serves as a quantitative correlation model between the switch placement parameters and the system power loss scale.

[0012] Based on the historical output data of the distributed power source, a deep convolutional generative adversarial network is used to generate time-series output data of new energy sources, and Kalman filtering is used to correct the generated data. The corrected data is then embedded into the fault reconstruction model to establish an Nk fault reconstruction optimization model that takes into account the volatility of new energy sources.

[0013] A resilience assessment system is constructed. The resilience assessment system is calculated based on the quantitative correlation model and the fault reconstruction optimization model. The switch placement parameters are comprehensively evaluated based on the calculation results to determine the switch layout scheme that meets the system resilience requirements.

[0014] As an optimization, the fault scenario simulation uses the Monte Carlo simulation method to generate Nk fault scenarios, where k=1,2, and the number of simulations is no less than 10,000; the fault index data includes the power outage rate. and power outage node rate .

[0015] As an optimization, the nonlinear relationship is determined using the controlled variable method: the interconnection density of the tie switches is fixed. Change the segment spacing L, and fit the segment spacing L with the power outage rate. The nonlinear curve between the segments, which exhibits exponential growth after the segment spacing L exceeds a threshold; the coupling interaction relationship is determined through a full-factor experiment: the interconnection density of the tie switches is calculated under different values ​​of segment spacing L. Changes in power outage rate The marginal impact, which increases with the increase of the segment spacing L.

[0016] As an optimization, the regression equation is:

[0017] ;

[0018] ;

[0019] in, Power outage rate, The power outage node rate, , , , , , , , , , For regression coefficients, , The error term is solved using the least squares method to obtain the regression coefficients.

[0020] As an optimization, based on the regression equation and the requirements for safe operation of the distribution network, the optimal threshold range for the segment spacing L and the interconnection density of the tie switches are determined. The optimal threshold range; wherein, the safe upper limit of the power failure load rate is 5%, and the safe upper limit of the power failure node rate is 3%.

[0021] As an optimization, the optimal threshold range of the segment spacing L and the interconnection density of the interconnection switches are... The optimal threshold range is set differently based on the type of distribution network:

[0022] For urban power distribution networks, the optimal threshold range for the segment spacing L is 1-2 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5;

[0023] For industrial park power distribution networks, the optimal threshold range for the segment spacing L is 1.5-2.5 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5;

[0024] For rural power distribution networks, the optimal threshold range for the segment spacing L is 2-3 km, and the interconnection density of the tie switches... The optimal threshold range is 0.3-0.4.

[0025] As an optimization, the objective function of the Nk fault reconstruction optimization model is:

[0026] ;

[0027] in, This represents the power supply status of node i, with a value of 1 indicating power supply and 0 indicating power failure. Let i be the load of node i. Let be the resistance of line ij. Let be the current in line ij. is the weighting coefficient, and n is the number of nodes; the constraints of the Nk fault reconfiguration optimization model include: active power balance constraints, reactive power balance constraints, node voltage safety constraints, line current carrying constraints, operating time constraints of automatic switches and tie switches, and new energy output constraints.

[0028] As an optimization, the action time constraint includes a timing coordination constraint: the nearest automatic switch upstream of the fault point takes priority in action, the tie switch takes a delay after the automatic switch has taken action, and the time difference between the automatic switch and the tie switch meets a preset threshold; the new energy output constraint includes an island support constraint: when the output of the distributed power source meets the downstream load demand, the distributed power source is put into operation as a black start power source.

[0029] As an optimization, the resilience assessment system includes four dimensions of assessment indicators: power outage scale indicators. Recovery efficiency indicators Voltage safety indicators and disturbance resistance index The weights of each evaluation indicator were determined using the entropy weight method, and the comprehensive resilience index was calculated using... Calculation, where The weight of the j-th indicator is given; the Pearson correlation coefficient is used to quantify the segment spacing L and the interconnection density of the tie switches. Based on the correlation with the comprehensive resilience index R, a switch layout scheme that meets the system resilience requirements is determined.

[0030] As an optimization, the specific process for determining the switch layout scheme that meets the system resilience requirements based on the correlation calculation results is as follows:

[0031] Calculate the correlation coefficient between the segment spacing L and the comprehensive toughness index R respectively. and the interconnection density of the communication switch Correlation coefficient with the comprehensive toughness index R :

[0032] When the correlation coefficient satisfy When it is determined that the segment spacing L is strongly correlated with the comprehensive resilience index R, the switch placement parameters are adjusted in the direction of reducing the segment spacing L within the optimal threshold range matching the corresponding distribution network type until the comprehensive resilience index R reaches its maximum.

[0033] When the correlation coefficient satisfy At that time, the interconnection density of the connection switch is determined. The overall resilience index R is strongly correlated, and the interconnection density of the tie switches is increased within the optimal threshold range matching the corresponding distribution network type. Adjust the switch placement parameters in the direction until the comprehensive resilience index R reaches its maximum.

[0034] The adjusted switch placement parameters are output as a switch layout scheme that meets the system resilience requirements.

[0035] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0036] This invention introduces a timing coordination constraint between automatic switches and tie switches in the Nk fault reconfiguration model (the automatic switch upstream of the fault point operates first, the tie switch operates with a delay, and the time difference between the actions meets a preset threshold), which enables the switches to operate in an orderly manner during the fault reconfiguration process, avoids transient overvoltages and inrush currents caused by the simultaneous operation of multiple switches, and improves the success rate of fault reconfiguration and load restoration efficiency.

[0037] This invention employs a deep convolutional generative adversarial network (DCGAN) to perform time-series prediction of distributed power generation output, and combines it with Kalman filtering to correct the prediction results. The corrected renewable energy output prediction values ​​are embedded into the active power balance constraints and renewable energy output constraints of the fault reconstruction model, enabling the fault reconstruction model to adapt to the randomness and volatility of renewable energy output. This avoids the problem of underestimating the scale of power outages in scenarios with high renewable energy penetration using traditional deterministic modeling methods.

[0038] This invention constructs the segment spacing L and connection density through Monte Carlo simulation and regression analysis. The quantitative correlation model between switch placement and the scale of power loss elevates the relationship between switch placement and the scale of system power loss from traditional empirical qualitative analysis to mathematical quantitative analysis, providing a scientific basis for the formulation of switch placement schemes.

[0039] This invention constructs a resilience assessment index system from four dimensions: power outage scale, recovery efficiency, voltage safety, and disturbance immunity resilience. It also uses the entropy weight method to objectively determine the weight of each index, overcoming the shortcomings of traditional resilience assessment that rely on a single index or subjective weighting, and making the resilience assessment results more comprehensive and objective.

[0040] Based on the entropy weight method to determine the initial weights, this invention introduces a dynamic correction factor to adaptively adjust the index weights according to the real-time load level of the distribution network, the penetration rate of new energy sources, and meteorological conditions. This enables the resilience assessment system to adapt to the differentiated assessment needs under different operating scenarios and avoids the problem of assessment distortion in multiple scenarios with fixed weights.

[0041] This invention quantifies the segment spacing L and connection density using the Pearson correlation coefficient. The correlation with the comprehensive resilience index R is used to identify key parameters affecting resilience. Combined with the optimal threshold range determined by the quantitative correlation model, a univariate optimization algorithm is used to search for switch placement parameters that maximize the comprehensive resilience index, providing a precise decision-making basis for the formulation of switch layout schemes. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0043] Figure 1 This is a flowchart of a power distribution network switch layout optimization method according to the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0045] Example 1 discloses a method for optimizing the layout of distribution network switches, which includes 6 steps.

[0046] S1. Obtain basic data of the distribution network, including network topology and operating status data, line fault rate data, load data, and historical output data of distributed power sources.

[0047] The data includes distribution network topology data such as node distribution, line connection relationships, number of feeders and line impedance parameters; line fault rate data such as historical fault count, fault location, fault duration, fault recovery time and system stable operation time after fault; load data such as node load size, load distribution and load level; and historical output data of distributed power sources such as photovoltaic and wind power, including historical output data, fluctuation characteristics and grid connection point location.

[0048] After data collection, the following methods are used: The criteria include removing outlier data, supplementing missing data using linear interpolation, and standardizing the data using the following formula:

[0049] ;in, For the first Class parameter number One set of raw data, For standardized data, , The first Minimum and maximum values ​​of class parameters.

[0050] A dedicated database, using MySQL, was built to store the preprocessed data. A data access interface was written in Python to ensure the timeliness and accuracy of data support for subsequent scenario simulations and model calculations. The database employs a master-slave architecture (the master database handles data writing, and the slave database handles data retrieval) to guarantee data security. The Python access interface was developed using the Flask framework, with a data caching mechanism (10-minute cache time) to improve call efficiency.

[0051] S2. Based on the line fault rate data, distribution network topology data, and load data, fault scenario simulations are performed for different switch placement parameters to obtain fault index data reflecting the scale of power outages under each switch placement scheme. The fault scenarios include single-line fault scenarios and dual-line fault scenarios. The switch placement parameters include the automatic switch segment spacing L and the interconnection density of tie switches. .

[0052] Specifically, the Monte Carlo simulation method is used to generate Nk fault scenarios, where k=1,2 (i.e., single-line fault N-1, dual-line fault N-2). 10,000 fault scenario simulations are performed for each switch placement scheme. In each simulation, the historical fault rate of each line is considered. Non-uniform random sampling is used, with lines having a higher failure rate having a greater probability of being selected as faulty lines: in the N-1 fault scenario, one line is randomly selected as the faulty line, and in the N-2 fault scenario, two lines are randomly selected as the faulty lines. The power loss load rate and power loss node rate of each simulation are recorded, and the average of 10,000 results is taken as the average power loss level of this switch placement scheme in a probabilistic sense.

[0053] The fault indicator data includes the power outage rate. and power outage node rate .

[0054] For each fault scenario, the switching action logic is simulated (automatic switch action time ≤ 0.5s, tie switch action time ≤ 5s). Based on the distribution network topology and power flow distribution, the core indicators of the system power loss scale are calculated as follows:

[0055] (1) Power loss rate: ;

[0056] (2) Power failure rate: ;

[0057] in, The power supply status of node i (1 for power supply, 0 for power failure). For the load of node i, This refers to the total active power of the system under normal operating conditions (excluding power injected by distributed generation sources). This represents the number of power outage nodes. This represents the total number of nodes.

[0058] Different switch layout schemes were set up (changing the segment spacing of automated switches and the interconnection density of tie switches). For each scheme, 10,000 fault scenario simulations were performed. The power outage load rate and power outage node rate of each simulation were recorded and averaged to form a control dataset. Specifically, with the distribution network topology, load level, and switch layout scheme remaining unchanged, one (N-1 fault) or two (N-2 fault) lines were randomly selected from the set of lines each time as the faulted lines. The operation of automated switches and tie switches was simulated, and the power outage load rate and power outage node rate under this fault were calculated. After repeating 10,000 times, the average of the 10,000 results was taken as the average power outage level of the switch layout scheme in a probabilistic sense.

[0059] In some embodiments, comparison schemes are set for three scenarios to cover different power distribution network types:

[0060] (1) Urban distribution network (dense load, 400-500 nodes): The spacing of automatic switch segments is set at three gradients of 1km, 2km and 3km, and the interconnection density of tie switches is set at three gradients of 0.3, 0.4 and 0.5, for a total of 9 control schemes; each scheme is simulated 10,000 times, and each time a faulty line is randomly selected, the power loss load rate and power loss node rate are recorded, and the average value is taken as the index value of the scheme.

[0061] (2) Industrial park power distribution network (high new energy penetration rate, number of nodes 300-400): The automatic switch segment spacing is set at three gradients of 1.5km, 2.5km and 3km, and the interconnection density of tie switches is set at three gradients of 0.4, 0.5 and 0.6, for a total of 9 sets of comparison schemes; each set of schemes is simulated 10,000 times, and each time a faulty line is randomly selected, the power loss load rate and power loss node rate are recorded, and the average value is taken as the index value of the scheme.

[0062] (3) Rural power distribution network (dispersed load, number of nodes 200-300): The spacing of the automatic switch segments is set at three gradients of 2km, 3km and 4km, and the interconnection density of the tie switch is set at three gradients of 0.2, 0.3 and 0.4, for a total of 9 control schemes; each scheme is simulated 10,000 times, and each time a faulty line is randomly selected, the power loss load rate and power loss node rate are recorded, and the average value is taken as the index value of the scheme.

[0063] A total of 27 control schemes were implemented across three types of distribution networks. Each scheme underwent 10,000 Monte Carlo fault simulations to obtain an average power outage rate. and an average power outage node rate This resulted in 27 sets of data samples, which were used for subsequent regression equation fitting and to establish switch placement parameters. Scale of power outage A quantitative correlation model between them.

[0064] S3. Perform statistical analysis on the fault index data to determine the nonlinear relationship between the segment spacing L and the power loss scale, as well as the relationship between the segment spacing L and the interconnection density of the tie switch. The coupling and interaction relationships between (hereinafter referred to as "connection density").

[0065] Specifically, the nonlinear relationship is manifested as follows: as the segment spacing L increases, the power outage rate... The corresponding increase occurs; when L exceeds a certain first threshold, The growth rate increases rapidly with increasing L. The coupling interaction relationship is manifested as: the interconnection density of the interconnecting switches... Changes in power outage rate The marginal effect increases with the increase of the segment spacing L; when L exceeds the first threshold, the effect further increases. To reduce The effect is even more significant.

[0066] The propagation of faults in distribution networks exhibits the following objective characteristics:

[0067] Feature 1: Nonlinear relationship between segment spacing and power loss scale

[0068] When the segment spacing is below the second threshold, further reducing the segment spacing has little marginal contribution to reducing the power outage rate; when the segment spacing exceeds the first threshold, the power outage rate increases rapidly with the increase of the segment spacing. Therefore, there is a reasonable range for the segment spacing; below the second threshold, the investment efficiency is low, and above the first threshold, the power outage scale deteriorates sharply.

[0069] Feature 2: Coupling relationship between segment spacing and connection density

[0070] The effects of segment spacing and interconnection density on the scale of power outages are coupled, and their optimization is mutually restrictive. When the segment spacing exceeds the first threshold, increasing the interconnection density has a more significant effect on reducing the power outage load rate; when the segment spacing is below the second threshold, the effect of increasing the interconnection density is relatively limited.

[0071] Feature 3: The amplifying effect of renewable energy fluctuations on the scale of power outages

[0072] The higher the penetration rate of distributed power generation, the more significant the amplification effect of fluctuations in renewable energy output on the scale of power outages. When the penetration rate is below the third threshold, the impact of fluctuations is negligible; when the penetration rate exceeds the third threshold, the impact of fluctuations increases significantly; when the penetration rate exceeds the fourth threshold, the impact of fluctuations increases dramatically.

[0073] The above nonlinear relationship is determined using the controlled variable method: the interconnection density of the tie switch is fixed. Change the segment spacing L, and fit the segment spacing L with the power outage rate. The nonlinear curve between the segments increases exponentially after the segment spacing L exceeds a threshold. The above coupling interaction relationship was determined through a full-factor experiment: the interconnection density of the interconnecting switches was calculated for different values ​​of segment spacing L. Changes in power outage rate The marginal effect, which increases with increasing segment spacing L. In this invention, the connection density... Changes in power outage rate The marginal impact refers to For every increase of 0.1, How many percentage points did it decrease?

[0074] The first to fourth thresholds mentioned above are critical parameters in the distribution network fault propagation process. Specifically: the first threshold represents the critical value at which the segment spacing changes from a reasonable range to an excessively wide range; the second threshold represents the critical value at which the segment spacing changes from an excessively dense range to a reasonable range, and the first threshold is greater than the second threshold; the third threshold represents the critical value at which the new energy penetration rate changes from a negligible range to a significantly amplified range; and the fourth threshold represents the critical value at which the new energy penetration rate changes from a significantly amplified range to a rapidly amplified range, and the fourth threshold is greater than the third threshold.

[0075] S4. Based on the nonlinear relationship and the coupling interaction relationship, construct a quadratic term that includes the piecewise spacing. and the interaction term between the segment spacing and the interconnection density of the tie switch. The regression equation serves as a quantitative correlation model between the switch placement parameters and the scale of system power loss.

[0076] Specifically, the regression equation is:

[0077] ;

[0078] ;

[0079] in, This is used to characterize the coupling effect of segment spacing and interconnection density on the scale of power outage. Power outage rate, The power outage node rate, , , , , , , , , , For regression coefficients, , The error term is solved using the least squares method to obtain the regression coefficients.

[0080] In some embodiments, in the regression equation The term is used to characterize the relationship between the segment spacing L and the power failure rate. The non-linear relationship between them (i.e., after L exceeds the threshold) (Exponential growth) Interaction terms are used to characterize the segment spacing L and the interconnection density of the interconnection switch. The coupling and interaction relationship between them (i.e.) right The marginal effect increases with increasing L. This is achieved by simultaneously introducing [the marginal effect] into the regression equation. Item and This method can accurately reflect the comprehensive impact mechanism of switch layout parameters on the power outage scale, avoiding optimization deviations caused by ignoring nonlinear or coupling effects.

[0081] The regression coefficients are solved using the least squares method, ensuring that the regression equation accurately fits the aforementioned 27 sets of data samples. The solution formula is as follows:

[0082] ;

[0083] in, Let X be the column vector formed by the regression coefficients mentioned above, and let X be the matrix of independent variables (including L, ...). , , Y is the dependent variable vector ( or ).

[0084] In some embodiments, based on the regression equation and the requirements for safe operation of the distribution network, the optimal threshold range of the segment spacing L and the interconnection density of the tie switches are determined. The optimal threshold range; wherein, the safe upper limit of the power failure load rate is 5%, and the safe upper limit of the power failure node rate is 3%.

[0085] Specifically, based on the safe upper limit of 5% for the power outage load rate and the safe upper limit of 3% for the power outage node rate, and in conjunction with the aforementioned regression equation, the segment spacing L and interconnection density that meet the requirements for safe operation are deduced. The boundary values ​​of L. For example, when L... When a certain fixed value is taken, Substituting into the regression equation, the upper limit of L can be solved; when L is below a certain threshold, the effect of L on reducing... The marginal contribution of L is no longer significant; this value is the lower limit of L. Therefore, L and L are determined. Each has its own optimal threshold range, within which the system's power failure load rate and power failure node rate do not exceed the safe upper limit.

[0086] Furthermore, the optimal threshold for the segment spacing L and the interconnection density of the interconnection switches... The optimal threshold is set differently based on the type of distribution network:

[0087] For urban power distribution networks (with dense loads and high power supply reliability requirements), the optimal threshold range for the segment spacing L is 1-2 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5;

[0088] For industrial park power distribution networks (where renewable energy penetration is high and renewable energy fluctuations significantly amplify the scale of power outages), the optimal threshold range for the segment spacing L is 1.5-2.5 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5;

[0089] For rural power distribution networks (with dispersed loads and limited investment budgets), the optimal threshold range for the segment spacing L is 2-3 km, and the interconnection density of the tie switches... The optimal threshold range is 0.3-0.4.

[0090] Differentiated settings for different types of power distribution networks can balance power supply reliability and investment economy, avoiding the problem of excessive investment in urban areas or insufficient investment in rural areas caused by a "one-size-fits-all" uniform standard.

[0091] The above-mentioned optimal threshold range was obtained through simulation regression based on a typical 10kV distribution network structure. In practical applications, it can be adaptively adjusted according to the specific network structure and voltage level.

[0092] S5. Based on the historical output data of the distributed power source, a deep convolutional generative adversarial network is used to generate time-series output data of new energy sources, and Kalman filtering is used to correct the generated data. The corrected data is then embedded into the fault reconstruction model to establish an Nk fault reconstruction optimization model that takes into account the volatility of new energy sources.

[0093] S5.1 Handling the volatility of new energy sources.

[0094] The output of distributed power sources (represented by photovoltaics and wind power) exhibits significant randomness and volatility, with output characteristics varying considerably across different seasons and time periods. Using deterministic output modeling would lead to an underestimation of the power outage scale during fault scenarios, resulting in a decrease in the effectiveness of optimized switch layout schemes in actual operation. Therefore, this invention employs a deep convolutional generative adversarial network (DCGAN) to generate and predict the active power output time-series data of distributed power sources.

[0095] DCGAN consists of a generator and a discriminator. The generator is used to generate simulated power output data consistent with the actual power output distribution from random noise. The input is random noise z (100 dimensions), and the output is the time series length. The output data (i.e., the output value at 24 times a day). The discriminator is used to distinguish whether the input output data is real historical data or simulated data generated by the generator. The input is real output data or generated data, and the output is the probability that the data is real data.

[0096] The generator and discriminator networks are constructed using convolutional layers. The generator employs four convolutional layers (3×3 kernel size, stride of 2) to progressively map low-dimensional random noise into high-dimensional temporal output data. The discriminator uses four convolutional layers plus one fully connected layer to progressively extract features from the temporal data and output the discrimination result. The LeakyReLU activation function (slope 0.2) is chosen to avoid the vanishing gradient problem during training.

[0097] The model was trained using the Adam optimizer with a learning rate of 0.0002, a momentum parameter of 0.5, and at least 50,000 iterations. During training, the generator and discriminator engaged in a game-like interaction: the generator attempted to generate simulated data that could fool the discriminator, while the discriminator attempted to accurately distinguish between real and simulated data. The loss function was as follows:

[0098] ;

[0099] Where G is the generator, D is the discriminator, x is the real data, z is random noise, and E is the mathematical expectation.

[0100] Through the aforementioned adversarial training, the generator can learn the true distribution characteristics of distributed power source output data and generate high-precision simulated output data.

[0101] The Kalman filter algorithm is used to correct the output data generated by DCGAN to further improve prediction accuracy. The Kalman filter recursively combines the predicted value at the current moment with the observed value using a weighted fusion method to obtain the corrected optimal estimate. The correction formula is as follows:

[0102] ;

[0103] ;

[0104] ;

[0105] ;

[0106] in, Let k be the predicted value at time k. Let A be the predicted value at time k-1, A be the state transition matrix, and B be the control input matrix. This is the control input vector at time k (representing meteorological data such as light intensity, wind speed, and temperature; if real-time meteorological data cannot be obtained, this parameter is set to 0). Let be the covariance matrix at time k. Let be the covariance matrix at time k-1, and Q be the process noise covariance matrix. represents the inherent uncertainty of the new energy power output prediction model. Since there is a deviation between the power output prediction value generated by DCGAN and the actual power output, Q is used to quantify the statistical characteristics of this prediction deviation. Its value reflects the degree of confidence in the prediction model; the larger Q is, the less reliable the prediction, and the more the filtering algorithm tends to trust the observed values. Its value is pre-set based on historical data statistics and does not require real-time adjustment. Here, H is the Kalman gain, H is the observation matrix, and R is the observation noise covariance. This is the corrected optimal estimate. Let be the observation value at time k.

[0107] In some embodiments, historical power output data of distributed power sources (photovoltaics and wind power) over the past three years are collected, and the sample set is divided by season (spring, summer, autumn, and winter), with no less than 1000 samples per season. The DCGAN model described above is trained separately for each season to generate typical daily power output time-series data for each season. This data is then corrected using Kalman filtering, with the prediction error controlled within 8%. The latest power output data is collected daily to supplement the sample set, and the DCGAN model is retrained every 7 days to ensure that the model can adapt to the dynamic changes in renewable energy power output. The predicted active power output of renewable energy after Kalman filtering correction is used as the input parameter for the injected active power of distributed power source nodes in the Nk fault reconfiguration optimization model.

[0108] S5.2 Construct the Nk fault reconstruction optimization model.

[0109] ;

[0110] in, This represents the power supply status of node i, with a value of 1 indicating power supply and 0 indicating power failure. Let i be the load of node i. Let be the resistance of line ij. Let be the current in line ij. is the weighting coefficient (taken as 0.3, used to balance the two objectives of minimizing power outage load and minimizing network loss), and n is the number of nodes.

[0111] The first term of the objective function is the total power loss load after the fault, and the second term is the power flow loss of the network after reconstruction. By adjusting... The value of can be adjusted to accommodate the different reliability and economic requirements of various power distribution networks. For scenarios with high power supply reliability requirements (such as urban areas and industrial parks), the value can be appropriately increased. The value of can be appropriately reduced for scenarios with high economic requirements (such as rural areas). The value of .

[0112] The constraints of the Nk fault reconstruction optimization model include:

[0113] (1) Active power balance constraint:

[0114] ;

[0115] in, Let be the active power flowing from node i to node j. The active power output of the distributed power source at node i is determined by the predicted value after Kalman filtering correction in S5.1. Let be the active power loss at node i.

[0116] (2) Reactive power balance constraint:

[0117] ;

[0118] in, The reactive power flowing from node i to node j is determined by the power factor of that node or the inverter's reactive power control strategy. Let i be the reactive power output of the distributed power source at node i. Let i be the reactive load. Let i be the reactive power loss at node i.

[0119] (3) Node voltage safety constraints:

[0120] ;

[0121] in, The actual voltage at node i. This is the rated voltage. This constraint ensures that the voltage of each node remains within ±5% of the rated voltage after reconfiguration, preventing equipment damage or power quality degradation due to voltage exceeding limits.

[0122] (4) Line current carrying constraints:

[0123] ;

[0124] in, Let be the rated current-carrying capacity of line ij. This constraint ensures that the transmission current of each line after reconfiguration does not exceed its rated capacity, thus avoiding line overload.

[0125] (5) Operating time constraints of automatic switches and interconnecting switches:

[0126] The automatic switch actuation time is ≤0.5s, and the tie switch actuation time is ≤5s. The actuation time constraints include timing coordination constraints: the nearest upstream automatic switch to the fault point actuates first, disconnecting the faulty section from the upstream power supply; the tie switch actuates after a delay (≥0.1s) following the automatic switch's actuation, closing the tie path to transfer load to the non-faulty section. The actuation time difference between the automatic switch and the tie switch meets a preset threshold, specifically, this time difference is not less than 0.05 seconds, to avoid simultaneous actuation of both switches causing transient overvoltages or inrush currents, ensuring the safety and stability of the fault reconstruction process.

[0127] (6) Constraints on new energy output:

[0128] ;

[0129] in, and These are the minimum and maximum output values ​​of the distributed generation at node i, determined based on the new energy prediction results after Kalman filtering correction. This constraint ensures that the output of the distributed generation is within its actual adjustable range.

[0130] The renewable energy output constraints include islanding support constraints: when the output of a distributed power source is not less than its downstream load, the distributed power source is given priority to start operation as a black-start power source to provide islanding power supply to the faulty downstream section. Specifically, when If the distributed power source is deemed capable of islanding, it will be prioritized for activation during fault reconfiguration to minimize the scope of power loss. It is the sum of the active power of all load nodes within the power supply area covered by the distribution line extending downstream (away from the main power source) from the grid connection point of the distributed power source.

[0131] The model solution employs a Nested Column Constraint Generation (NCCG) algorithm, decomposing the model into a main problem and sub-problems. The main problem is switch placement optimization, and the sub-problem is fault reconstruction verification. The solution process is as follows: Initialize the switch placement scheme → Verify the reconstruction feasibility of the sub-problem → If infeasible, generate a cutting plane → Add the cutting plane to the constraints of the main problem → Update the main problem and solve it again → Iterative convergence → Output the optimal reconstruction scheme.

[0132] Solution accuracy The solution time is ≤60s (for large-scale distribution networks with ≤500 nodes).

[0133] S6. Construct a resilience assessment system. Calculate the resilience assessment system based on the quantitative correlation model and the fault reconstruction optimization model. Based on the calculation results, comprehensively evaluate the switch placement parameters and determine a switch layout scheme that meets the system resilience requirements.

[0134] The quantitative correlation model outputs the segment spacing L and interconnection density for each distribution network type. The optimal threshold range provides an adjustment boundary for switch placement parameters for resilience assessment; the fault reconfiguration optimization model outputs the power supply status of each node by solving the objective function and constraints. The network power flow distribution (including the voltage of each node) is obtained through power flow calculation. and power of each branch , Based on this, operational indicators such as power outage load rate, node voltage safety level, and recovery time are further calculated, providing calculation data for various dimensions of indicators for resilience assessment.

[0135] S6.1 Construct a resilience assessment index system.

[0136] In some embodiments, the resilience assessment system includes four dimensions of assessment indicators: power outage scale indicators. Recovery efficiency indicators Voltage safety indicators and disturbance resistance index .

[0137] The calculation methods for each evaluation indicator are as follows:

[0138] (1) Power loss scale indicators ( (Negative indicator): Reflects the degree of power loss of the system after a fault.

[0139] ;

[0140] This indicator ranges from 0 to 1. A smaller value indicates a smaller power loss after a fault and better system resilience. Let n represent the power supply status of node i, and n be the total number of nodes in the distribution network. Let i be the load of node i. The total system load capacity, Indicates the power loss indicator factor.

[0141] (2) Recovery efficiency indicators ( Positive indicator: reflects the speed of load recovery after a fault.

[0142] ;

[0143] in, The power supply status of node i at time t, where T is the total evaluation duration. This indicator ranges from 0 to 1; a larger value indicates faster load recovery after a fault and better system resilience.

[0144] (3) Voltage safety index ( Positive indicators: Reflect the voltage stability of each node in the system after reconstruction.

[0145] ;

[0146] The value of this indicator ranges from 0 to 1. The larger the value, the closer the node voltage is to the rated voltage after reconstruction, and the better the power supply quality of the system.

[0147] (4) Disturbance resistance toughness index ( Positive indicators: These reflect the system's overall ability to withstand and recover from fault disturbances.

[0148] ;

[0149] This refers to the system's stable operating time after a failure. Duration of the fault Recovery time. This indicator ranges from 0 to 1; a higher value indicates a stronger ability of the system to withstand fault disturbances and better resilience.

[0150] S6.2 Determination and comprehensive evaluation of indicator weights.

[0151] The weights of each evaluation indicator are determined using a combination of entropy weighting and dynamic correction factors. Entropy weighting is an objective weighting method that determines the basic weight based on the degree of variation of each indicator in various evaluation schemes—the greater the degree of variation of an indicator, the more information it contains, and the higher its weight. The dynamic correction factor adaptively adjusts the basic weights according to the real-time operating status of the distribution network to overcome the problem of evaluation distortion under different operating scenarios with fixed weights.

[0152] The specific steps for determining the weights are as follows:

[0153] S6.2.1 Calculate the initial weights based on the entropy weight method:

[0154] First, the evaluation indicators are standardized:

[0155] For positive indicators (the larger the better): ;

[0156] For negative indicators (the smaller the better): ;

[0157] in, Let j be the standardized value of the j-th indicator in the i-th evaluation scheme. Let be the original value of the j-th indicator in the i-th evaluation scheme. and These are the maximum and minimum values ​​of the j-th indicator across all evaluation schemes, respectively.

[0158] Calculate the entropy value of each indicator:

[0159] ;

[0160] in, Let the entropy value be the j-th index. To evaluate the total number of options.

[0161] Calculate the initial weights for each indicator:

[0162] ;

[0163] in, Let be the initial weight of the j-th indicator, and .

[0164] S6.2.2 Introduction of dynamic correction factors: Based on the real-time operating status of the distribution network, determine the dynamic correction factors for each indicator. (j=1,2,3,4):

[0165] (1) Power loss scale indicators Correction factor The weight of the power outage scale indicator is determined based on the ratio of the current load level to the historical maximum load. The higher the current load level, the greater the social impact and economic losses caused by the power outage; therefore, the weight of this indicator should be increased accordingly.

[0166] ;

[0167] in, This represents the current total load level of the system. The baseline load level (such as the annual average load). This represents the highest load level in history. The preset maximum correction range ( ).

[0168] (2) Recovery efficiency indicators Correction factor The recovery efficiency index is related to the level of automation in the distribution network. In distribution networks with a high level of automation, its weight can be appropriately increased. However, in the dynamic weighting framework of this invention, The weight of the recovery efficiency index is kept at 1, based on the initial weight, and is not dynamically adjusted with changes in operating status to maintain the stability of the evaluation system. In other embodiments, k2 can also be set according to the actual automation level of the distribution network; the higher the automation level, the higher the efficiency. The larger the value, the better.

[0169] (3) Voltage safety indicators Correction factor The weight of voltage safety indicators should be determined based on the current penetration rate of new energy sources. The higher the penetration rate of new energy sources, the greater the risk of voltage limitations and power fluctuations; therefore, the weight of voltage safety indicators should be increased accordingly.

[0170] ;

[0171] in, Given the current penetration rate of new energy sources, This is the baseline penetration rate (such as the annual average penetration rate of the distribution network). When hour, No correction; when When it increases, Gradually increase but do not exceed 1.5.

[0172] (4) Disturbance resistance index Correction factor The severity of weather conditions (such as thunderstorms, strong winds, and snow) is determined based on the current weather warning level. The more severe the weather conditions (e.g., thunderstorms, strong winds, snow), the greater the probability of equipment failure and the risk of disturbance; therefore, the weight of the disturbance resilience index should be increased accordingly.

[0173] ;

[0174] in, The current weather warning level is as follows: The highest level of weather warning. The preset maximum correction range ( (When there is no weather warning) , The higher the weather warning level, The larger.

[0175] The values ​​of each correction factor range from 0.5 to 1.5.

[0176] S6.2.3 Multiply the correction factor by the above initial weights to obtain the unnormalized correction weights:

[0177] ;

[0178] The unnormalized adjusted weights are normalized to obtain the corrected comprehensive weights:

[0179] ;

[0180] After normalization, the sum of the weights of all indicators is always equal to 1, that is... This ensures that the value of the comprehensive resilience index R is always within the range of [0,1].

[0181] S6.2.4 The comprehensive resilience index is calculated using the following formula:

[0182] ;

[0183] R is the comprehensive resilience index, which ranges from 0 to 1. The larger the value, the stronger the overall resilience of the distribution network.

[0184] S6.3 Quantify the correlation and determine the layout scheme. The resilience assessment is based on the comprehensive statistical results of N-1 and N-2 fault scenarios, without performing specific optimization for a single fault type, so that the optimized switch layout scheme has stable resilience performance under various fault scenarios.

[0185] To quantify the relationship between switch placement parameters and distribution network resilience, the Pearson correlation coefficient is used to quantify the segment spacing L and the interconnection density of tie switches. The correlation with the comprehensive resilience index R. The Pearson correlation coefficient is calculated as follows:

[0186] ;

[0187] in, For switch layout parameters (L or ), The corresponding comprehensive resilience index R, and , respectively, are the mean values ​​of the corresponding parameters and the comprehensive resilience index, where m is the sample size.

[0188] Correlation coefficient The range of values ​​for is [-1, 1]: A positive value indicates a positive correlation (the toughness increases as the parameter increases). A negative value indicates a negative correlation (the toughness decreases as the parameter increases). The closer to 1, the stronger the correlation.

[0189] In some embodiments, when When it is determined to be strongly correlated, When it is determined to be of moderate correlation, The correlation was determined to be weak.

[0190] The specific process for determining a switch layout scheme that meets the system resilience requirements is as follows:

[0191] Calculate the correlation coefficient between the segment spacing L and the comprehensive toughness index R respectively. and the interconnection density of the communication switch Correlation coefficient with the comprehensive toughness index R .

[0192] A negative value indicates that L and R are negatively correlated, meaning that the larger the segment spacing, the lower the toughness. Therefore, L needs to be reduced to improve toughness. A positive value indicates It is positively correlated with R, meaning that the greater the connection density, the higher the resilience, therefore it needs to be increased. To enhance resilience.

[0193] when and When the absolute values ​​of the parameters are different, priority should be given to optimizing the parameters that have a greater impact on resilience (i.e., the parameters with larger absolute values ​​of the correlation coefficients) in order to achieve the greatest resilience improvement within a limited investment budget.

[0194] When the correlation coefficient satisfy If the segment spacing L is determined to be strongly correlated with the comprehensive resilience index R, the switch placement parameters are adjusted in the direction of reducing the segment spacing L within the optimal threshold range matching the corresponding distribution network type.

[0195] When the correlation coefficient satisfy At that time, the interconnection density of the connection switch is determined. The overall resilience index R is strongly correlated with the optimal threshold range matching the corresponding distribution network type, thereby increasing the interconnection density of the tie switches. The direction of the switch placement parameters is adjusted.

[0196] like and All meet That is, L and When both are strongly correlated, compare and The size of the correlation coefficient should be prioritized for adjustment, as the parameter with the larger absolute value has a more significant impact on resilience. Optimizing this parameter first can achieve a greater improvement in resilience when investment is limited.

[0197] The specific value within the optimal threshold range is determined as follows:

[0198] Using the switch layout parameters before adjustment as the initial solution, a single-variable optimization algorithm (such as the golden section method, quadratic interpolation method, etc.) is used to search for the switch layout parameters that maximize the comprehensive resilience index R within the optimal threshold range, which is then used as the optimal switch layout scheme for the current distribution network.

[0199] In some embodiments, the univariate optimization algorithm adopts a two-level optimization method combining coarse scanning and fine scanning: a coarse scan is performed on the optimal threshold interval with a first step length to initially locate the sub-interval with the highest comprehensive resilience index R; then a fine scan is performed in the sub-interval with a second step length smaller than the first step length to select the switch layout parameters corresponding to the maximum comprehensive resilience index R as the optimal switch layout scheme for the current distribution network.

[0200] If the segment spacing L and the connection density If both are strongly correlated, then for L and Perform a joint traversal to search for the parameter combination that maximizes the overall resilience index R within the optimal threshold range of both.

[0201] If the correlation coefficients of the two are both moderate or weak, it indicates that the current distribution network switch layout is already reasonable and does not require significant adjustments. Only minor adjustments within the existing threshold range are needed.

[0202] If one is strongly correlated and the other is weakly correlated, then only the parameters of the strong correlation are iterated and optimized, while the parameters of the weak correlation are kept at the middle value of their threshold range.

[0203] The adjusted switch placement parameters are output as a switch layout scheme that meets the system resilience requirements.

[0204] In some embodiments, the above evaluation and adjustment are performed on different types of distribution networks to output differentiated switch layout schemes:

[0205] For urban power distribution networks (with dense loads and high requirements for power supply reliability), L and They are usually strongly correlated, requiring simultaneous optimization of L and Furthermore, when investment is limited, parameters with larger absolute values ​​of correlation coefficients should be prioritized for optimization.

[0206] For industrial park power distribution networks (where the penetration rate of new energy sources is high, and fluctuations in new energy sources have a significant amplifying effect on the scale of power outages). Typically, a strong correlation is observed (because the interconnection switch is crucial for load transfer under conditions of renewable energy fluctuations), and a larger value should be selected within the optimal threshold range of ρ.

[0207] For rural power distribution networks (distributed loads, limited investment budgets), L and The correlation coefficient is usually moderate or weak, indicating that the existing switch layout is already reasonable, and the status quo can be maintained or only minor adjustments can be made while meeting the requirements for safe operation.

[0208] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing the layout of switches in a power distribution network, characterized in that, Includes the following steps: Acquire basic data of the distribution network, including distribution network topology and operating status data, line fault rate data, load data, and historical output data of distributed power sources; Based on the aforementioned fundamental data, fault scenario simulations were conducted for different switch placement parameters to obtain fault index data reflecting the scale of system power loss under each switch placement scheme. The switch placement parameters include the automatic switch segment spacing L and the interconnection density of tie switches. ; Statistical analysis was performed on the fault index data to determine the nonlinear relationship between the segment spacing L and the power outage scale, as well as the relationship between the segment spacing L and the interconnection density of the tie switch. The coupling and interaction relationships between them; Based on the nonlinear relationship and the coupled interaction relationship, a quadratic term containing the piecewise spacing is constructed. and the interaction term between the segment spacing and the interconnection density of the tie switch. The regression equation serves as a quantitative correlation model between the switch placement parameters and the system power loss scale. Based on the historical output data of the distributed power source, a deep convolutional generative adversarial network is used to generate time-series output data of new energy sources, and Kalman filtering is used to correct the generated data. The corrected data is then embedded into the fault reconstruction model to establish an Nk fault reconstruction optimization model that takes into account the volatility of new energy sources. A resilience assessment system is constructed. The resilience assessment system is calculated based on the quantitative correlation model and the fault reconstruction optimization model. The switch placement parameters are comprehensively evaluated based on the calculation results to determine the switch layout scheme that meets the system resilience requirements.

2. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, The fault scenario simulation uses the Monte Carlo simulation method to generate Nk fault scenarios, where k=1,2, and the number of simulations is no less than 10,000; the fault index data includes the power outage rate. and power outage node rate .

3. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, The nonlinear relationship is determined using the controlled variable method: the interconnection density of the tie switches is fixed. Change the segment spacing L, and fit the segment spacing L with the power outage rate. The nonlinear curve between the segments, wherein the nonlinear curve exhibits an exponential growth pattern after the segment spacing L exceeds a threshold; The coupling interaction relationship was determined through a full-factor experiment: the interconnection density of the tie switches was calculated under different values ​​of segment spacing L. Changes in power outage rate The marginal impact, which increases with the increase of the segment spacing L.

4. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, The regression equation is: ； ； in, Power outage rate, The power outage node rate, , , , , , , , , , For regression coefficients, , The error term is solved using the least squares method to obtain the regression coefficients.

5. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, Based on the regression equation and the requirements for safe operation of the distribution network, the optimal threshold range for the segment spacing L and the interconnection density of the tie switches are determined. The optimal threshold range; wherein, the safe upper limit of the power failure load rate is 5%, and the safe upper limit of the power failure node rate is 3%.

6. The method for optimizing the layout of distribution network switches according to claim 5, characterized in that, The optimal threshold range of the segment spacing L and the interconnection density of the interconnection switch The optimal threshold range is set differently based on the type of distribution network: For urban power distribution networks, the optimal threshold range for the segment spacing L is 1-2 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5; For industrial park power distribution networks, the optimal threshold range for the segment spacing L is 1.5-2.5 km, and the interconnection density of the tie switches... The optimal threshold range is 0.4-0.5; For rural power distribution networks, the optimal threshold range for the segment spacing L is 2-3 km, and the interconnection density of the tie switches... The optimal threshold range is 0.3-0.

4.

7. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, The objective function of the Nk fault reconstruction optimization model is: ； in, This represents the power supply status of node i, with a value of 1 indicating power supply and 0 indicating power failure. Let i be the load of node i. Let be the resistance of line ij. Let be the current in line ij. is the weighting coefficient, and n is the number of nodes; the constraints of the Nk fault reconfiguration optimization model include: active power balance constraints, reactive power balance constraints, node voltage safety constraints, line current carrying constraints, operating time constraints of automatic switches and tie switches, and new energy output constraints.

8. The method for optimizing the layout of distribution network switches according to claim 7, characterized in that, The action time constraints include timing coordination constraints: the nearest automatic switch upstream of the fault point takes priority in action, the tie switch takes priority in action after the automatic switch has taken action, and the time difference between the automatic switch and the tie switch meets a preset threshold; the renewable energy output constraints include island support constraints: when the output of the distributed power source meets the downstream load demand, the distributed power source is put into operation as a black start power source.

9. The method for optimizing the layout of distribution network switches according to claim 1, characterized in that, The resilience assessment system includes four dimensions of assessment indicators: power outage scale indicators. Recovery efficiency indicators Voltage safety indicators and disturbance resistance index The weights of each evaluation indicator were determined using the entropy weight method, and the comprehensive resilience index was calculated using... Calculation, where The weight of the j-th indicator is given; the Pearson correlation coefficient is used to quantify the segment spacing L and the interconnection density of the tie switches. Based on the correlation with the comprehensive resilience index R, a switch layout scheme that meets the system resilience requirements is determined.

10. The method for optimizing the layout of distribution network switches according to claim 9, characterized in that, The specific process for determining the switch layout scheme that meets the system resilience requirements based on the correlation calculation results is as follows: Calculate the correlation coefficient between the segment spacing L and the comprehensive toughness index R respectively. and the interconnection density of the communication switch Correlation coefficient with the comprehensive toughness index R : When the correlation coefficient satisfy When it is determined that the segment spacing L is strongly correlated with the comprehensive resilience index R, the switch placement parameters are adjusted in the direction of reducing the segment spacing L within the optimal threshold range matching the corresponding distribution network type until the comprehensive resilience index R reaches its maximum. When the correlation coefficient satisfy At that time, the interconnection density of the connection switch is determined. The overall resilience index R is strongly correlated, and the interconnection density of the tie switches is increased within the optimal threshold range matching the corresponding distribution network type. Adjust the switch placement parameters in the direction until the comprehensive resilience index R reaches its maximum. The adjusted switch placement parameters are output as a switch layout scheme that meets the system resilience requirements.

Images