Risk early warning method, device and equipment for unbalanced power distribution network line and medium

Abstract

The invention discloses a risk early warning method, device and equipment for an unbalanced power distribution network line and a medium, and the method comprises the steps: constructing an overload risk index of each branch at each moment according to the historical power data of a target power distribution network, obtaining a first Warisstein fuzzy set based on the prediction deviation of the renewable energy power generation output of each node in the target power distribution network, obtaining a second Warisstein fuzzy set based on the prediction deviation of the load demand of each node, and solving the maximum value of the overload risk index in the combined power distribution fuzzy set obtained through fusion. Obtaining a line overload risk value; according to the method, the line overload risk value is converted into the branch power failure probability through the preset mapping relation, and the power distribution network all-day system power failure risk early warning value is obtained through aggregation calculation, so that the dependence on accurate probability distribution in the prior art is overcome, and the reliability of power distribution network risk assessment under the condition that high-proportion renewable energy sources are accessed to the power distribution network is improved.

Description

Technical Field

[0001] This application relates to the field of risk warning technology for distribution networks, and in particular to risk warning methods, devices, equipment and media for unbalanced distribution network lines. Background Technology

[0002] As the penetration rate of distributed photovoltaic, wind power and other renewable energy sources in the distribution network continues to increase, the power flow of the distribution network is changing from the traditional unidirectional and deterministic mode to a bidirectional and highly uncertain mode. Due to the intermittency and volatility of renewable energy output and the randomness of load demand, the imbalance of the three-phase power flow of the lines will be further aggravated, leading to operational risks such as line overload and voltage exceeding limits. In severe cases, it may lead to line power outages or even partial system collapse. Therefore, it is necessary to conduct forward-looking risk assessments of the distribution network in advance and take preventive measures in a timely manner.

[0003] Most existing risk assessment methods for distribution networks are based on stochastic optimization or probabilistic power flow frameworks, relying on accurate probability distribution models of input random variables. However, in actual operation, due to factors such as limited data samples, prediction errors, and environmental changes, the true probability distribution is often difficult to obtain accurately. Traditional empirical distributions may deviate significantly from the actual situation, resulting in unreliable risk assessment results and low accuracy of early warnings. In particular, tail risks are easily underestimated in extreme uncertainty scenarios.

[0004] Therefore, improving the reliability of overload risk assessment for unbalanced distribution network lines when a high proportion of renewable energy is integrated into the distribution network has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This application provides a risk warning method, device, equipment, and medium for unbalanced distribution network lines to solve the technical problem that the existing technology is unreliable in assessing the overload risk of unbalanced distribution network lines when a high proportion of renewable energy is connected to the distribution network.

[0006] To address the aforementioned technical problems, embodiments of this application provide a risk warning method for unbalanced distribution network lines.

[0007] Based on the historical power data of the target distribution network, an overload risk index for each branch at each time is constructed. The overload risk index is the conditional expectation of a quadratic function that characterizes the severity of branch power exceeding the limit under a given power distribution model.

[0008] Based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, the empirical distribution model of renewable energy power generation output of each node is first amplified to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node;

[0009] The empirical distribution model of load demand for each node is augmented based on the prediction deviation of load demand for each node to obtain the second Wasserstein fuzzy set of the load demand reconstruction model for each node.

[0010] The first Wassstein fuzzy set and the second Wassstein fuzzy set are fused to obtain a joint power distribution fuzzy set for each branch; a power prediction distribution model that maximizes the overload risk index is found in the joint power distribution fuzzy set, and the line overload risk value of each branch is calculated based on the power prediction distribution model.

[0011] The overload risk value of each branch at each time is converted into the corresponding branch outage probability using a preset mapping relationship. Based on the branch outage probability of all branches at all times, the all-day system outage risk warning value of the target distribution network is calculated. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

[0012] As one preferred embodiment, the step of constructing overload risk indicators for each branch at each time point based on historical power data of the target distribution network includes:

[0013] The historical power data of each branch in the target distribution network at each time point is compared with the upper power limit threshold and the lower power limit threshold of each branch to obtain the historical power limit exceedance value of each branch at each time point.

[0014] The historical power over-limit values ​​are processed using a nonlinear increasing function to calculate the historical overload risk assessment value of each branch at each time point;

[0015] The expected value of the historical overload risk assessment value is calculated based on the power distribution model of the historical power data of each branch, and is used as the overload risk index of each branch at each time.

[0016] As one preferred embodiment, the first amplification of the empirical distribution model of renewable energy generation output for each node based on the prediction deviation of renewable energy generation output for each node in the target distribution network, to obtain the first Wasserstein fuzzy set of the reconstructed renewable energy generation output model for each node, includes:

[0017] The predicted power generation of each node at each of the stated times and the actual power generation of each node at each of the stated times are obtained respectively.

[0018] The empirical distribution model of renewable energy power generation output is constructed based on the distribution characteristics of the actual power generation of each node;

[0019] The power generation prediction deviation value of each node at each time moment is calculated based on the difference characteristics between the predicted power generation and the corresponding actual power generation of each node at each time moment;

[0020] The deviation of each of the renewable energy power generation output reconstruction models and the renewable energy power generation output empirical distribution models is adjusted according to the power generation prediction deviation value, the first output Wasserstein metric radius and the second output Wasserstein metric radius to obtain a first Wasserstein fuzzy set, wherein the renewable energy power generation output reconstruction model represents the probability distribution model that the renewable energy power generation output of each node may follow.

[0021] As one preferred embodiment, the second augmentation of the empirical distribution model of load demand for each node based on the prediction deviation of load demand for each node, to obtain a second Wasserstein fuzzy set of the load demand reconstruction model for each node, includes:

[0022] The predicted load demand of each node at each time point and the actual load demand of each node at each time point are obtained respectively.

[0023] The load demand empirical distribution model is constructed based on the distribution characteristics of the actual load demand of each node;

[0024] The load prediction deviation value of each node at each time moment is calculated based on the difference between the predicted load demand and the corresponding actual load demand of each node at each time moment.

[0025] The deviation of each of the load demand reconfiguration models and the load demand empirical distribution models is adjusted based on the load forecast deviation value, the first load Wassstein metric radius, and the second load Wassstein metric radius to obtain a second Wassstein fuzzy set, wherein the load demand reconfiguration model represents a probability distribution model that the load demand of each node may follow.

[0026] As one preferred embodiment, the nonlinear increasing function is a quadratic function.

[0027] As one preferred embodiment, the calculation of the all-day system outage risk warning value of the target distribution network based on the outage probability of all said branches at all said times includes:

[0028] The difference between the value 1 and the probability of power outage of each branch at each time is taken as the probability of safe operation of each branch at each time.

[0029] The product of the safe operation probabilities of all branches in the target distribution network at all times is taken as the safe operation probability of the target distribution network throughout the day.

[0030] The difference between the value 1 and the probability of safe operation throughout the day is used as the warning value for the risk of system outage throughout the day for the target distribution network.

[0031] As one preferred embodiment, the step of converting the line overload risk value of each branch at each time point into the corresponding branch power outage probability using a preset mapping relationship includes:

[0032]

[0033] in, Let be the probability of branch ij being out of power at time t, exp() be an exponential function with base to the natural constant, and γ be the proportionality constant. Let be the line overload risk value of branch ij at time t.

[0034] Another embodiment of this application provides a risk warning device for unbalanced distribution network lines, including:

[0035] The risk assessment model construction module is used to construct overload risk indicators for each branch at each time based on the historical power data of the target distribution network. The overload risk indicator is the conditional expectation of a quadratic function that characterizes the severity of branch power exceeding the limit under a given power distribution model.

[0036] A power generation output fuzzy set construction module is used to perform a first amplification on the empirical distribution model of renewable energy power generation output of each node based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, so as to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node.

[0037] The load demand fuzzy set construction module is used to perform a second augmentation on the load demand empirical distribution model of each node based on the prediction deviation of the load demand of each node, so as to obtain the second Wasserstein fuzzy set of the load demand reconstruction model of each node.

[0038] The fuzzy set risk assessment module is used to fuse the first Wassstein fuzzy set and the second Wassstein fuzzy set to obtain a joint power distribution fuzzy set for each branch; to find the power prediction distribution model that maximizes the overload risk index in the joint power distribution fuzzy set; and to calculate the line overload risk value for each branch based on the power prediction distribution model.

[0039] The all-day power outage probability mapping module is used to convert the line overload risk value of each branch at each time into the corresponding branch power outage probability according to the preset mapping relationship, and to calculate the all-day system power outage risk warning value of the target distribution network based on the branch power outage probability of all branches at all times. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

[0040] Another embodiment of this application provides a risk warning device for unbalanced distribution network lines, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the risk warning method for unbalanced distribution network lines as described above.

[0041] Another embodiment of this application provides a computer-readable storage medium storing a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the risk warning method for unbalanced power distribution lines as described above.

[0042] Compared with the prior art, the beneficial effects of the embodiments of this application are as follows:

[0043] First, overload risk indicators for each branch at each time point are constructed based on historical power data of the target distribution network. Simultaneously, the nonlinear growth of the probability and severity of exceeding limits is considered, accurately characterizing the characteristic of overload risk deteriorating rapidly with increasing overload severity, providing a high-resolution metric benchmark for subsequent analysis. Then, based on the prediction deviations of renewable energy generation output and load demand at each node in the target distribution network, the uncertainties of renewable energy generation output and load demand at each node are modeled. Using the empirical distribution formed by historical data as the center, a first Wassstein fuzzy set for the renewable energy generation output reconstruction model and a second Wassstein fuzzy set for the load demand reconstruction model are constructed, thus simultaneously integrating the uncertainty of the data distribution itself and the inaccuracy of prediction, improving the robustness of subsequent risk assessment of the target distribution network system. Then, the first Wassstein fuzzy set is used to model the first Wassstein fuzzy set for the load demand reconstruction model. The fuzzy sets of Wassstein and the second Wassstein are fused to obtain the joint power distribution fuzzy set for each node. A power prediction distribution model that maximizes the overload risk index is found within this joint power distribution fuzzy set. The line overload risk value for each branch is calculated based on this model. The upper bound of the risk faced by the target distribution network system is assessed across all reasonably possible uncertainty scenarios. Finally, the line overload risk value of each branch at each time moment is converted into the corresponding branch outage probability using a pre-defined mapping relationship. Based on the branch outage probabilities at all times, the all-day system outage risk warning value for the target distribution network is calculated. Based on the probability independence assumption, system-level risk aggregation is performed, and the all-day system outage risk warning value is output. This approach overcomes the dependence of existing technologies on precise probability distributions and improves the reliability of distribution network risk assessment under conditions of high-proportion renewable energy integration. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart illustrating a risk warning method for unbalanced distribution network lines in one embodiment of this application.

[0046] Figure 2 This is a schematic diagram of an unbalanced IEEE 33-node system in one embodiment of this application;

[0047] Figure 3 This is a structural block diagram of a risk warning device for an unbalanced power distribution network line in one embodiment of this application;

[0048] Figure label:

[0049] 11. Risk assessment model construction module; 12. Power generation output fuzzy set construction module; 13. Load demand fuzzy set construction module; 14. Fusion fuzzy set risk assessment module; 15. All-day power outage probability mapping module. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of this application more thorough and comprehensive. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0051] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0052] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting this application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0053] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this application is for the purpose of describing specific embodiments only and is not intended to limit the application. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0054] It should be noted that, in the case of a high proportion of renewable energy access to the distribution network, in order to conduct risk assessment of the distribution network in advance, avoid causing large-scale damage to the distribution network system, solve the problem of inaccurate construction of the probability distribution model of the distribution network due to the uncertainty of distributed renewable energy, and thus improve the reliability of the risk assessment of the distribution network, an embodiment of this application provides a risk warning method for unbalanced distribution network lines, including steps S1 to S5.

[0055] Step S1: Construct overload risk indicators for each branch at each time point based on the historical power data of the target distribution network. The overload risk indicator is the conditional expectation of a quadratic function representing the severity of branch power over-limit when an over-limit occurs, given the power distribution model.

[0056] It should be noted that traditional risk assessments often only focus on the possibility of exceeding limits or apply linear weighting, failing to accurately depict the nonlinear characteristics of risk increasing rapidly with the degree of exceeding limits. The risk warning method for unbalanced distribution network lines provided in this embodiment uses a quadratic function to significantly distinguish between the risk values ​​of minor and severe exceeding limits, accurately depicting the nonlinear increasing characteristics of risk, that is, the closer the line load is to the limit, the faster the risk increases, thus improving the accuracy of risk assessment.

[0057] Step S2: Based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, perform the first amplification of the empirical distribution model of renewable energy power generation output of each node to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node.

[0058] It should be noted that in an unbalanced distribution network environment with a high proportion of renewable energy access, the probability distribution model that the actual renewable energy power generation output and power load follow is often uncertain. Traditional empirical distributions often deviate significantly from the actual situation. In order to fully consider and estimate the possible errors in the power probability distribution, this application constructs fuzzy sets for renewable energy power generation output and power load respectively.

[0059] Specifically, the Wasserstein distance is used to perform a first augmentation on the empirical distribution model of renewable energy power generation output of each node based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, so as to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node.

[0060] It should be noted that the Wasserstein fuzzy set is a set of probability distributions that may approximate the true distribution, centered on the empirical distribution. The Wasserstein distance measures the difference between two distributions by calculating the minimum cost required to "transfer" one probability distribution to another. The calculation of the Wasserstein distance is a well-known technique and will not be elaborated here in this embodiment. The first Wasserstein fuzzy set of the renewable energy power generation output reconfiguration model is the set of probability distributions that the renewable energy power generation output of each node in the target distribution network may follow.

[0061] In one embodiment, the empirical distribution model of renewable energy generation output for each node is first amplified based on the prediction deviation of renewable energy generation output for each node in the target distribution network, resulting in the first Wasserstein fuzzy set of the renewable energy generation output reconstruction model for each node, specifically including:

[0062] First, the predicted power generation and actual power generation of each node at each time point are obtained. Then, an empirical distribution model of renewable energy power generation is constructed based on the distribution characteristics of the actual power generation of each node. Next, the power generation prediction deviation value of each node at each time point is calculated according to the difference characteristics between the predicted power generation and the corresponding actual power generation of each node. Then, the deviation of each renewable energy power generation reconstruction model and the empirical distribution model of renewable energy power generation is adjusted according to the power generation prediction deviation value, the preset first output Wassstein metric radius, and the preset second output Wassstein metric radius to obtain the first Wassstein fuzzy set. Here, the renewable energy power generation reconstruction model represents the probability distribution model that the renewable energy power generation of each node may follow.

[0063] In one embodiment, the uncertainty of renewable energy power generation output at each node is modeled as follows:

[0064]

[0065] in, The uncertainty of the renewable energy power output of node i at time t is represented by phase φ. The predicted power generation of phase φ of renewable energy at node i at time t. Let be the random variable of renewable energy power generation output at time t.

[0066] Furthermore, the deviations of each renewable energy power generation output reconfiguration model and the renewable energy power generation output empirical distribution model are constrained by the following formula, and a first Wasserstein fuzzy set is constructed:

[0067]

[0068] Where EP() is the function for calculating mathematical expectation. The uncertainty of the renewable energy power output of node i at time t is represented by phase φ. N represents the predicted power generation of phase φ of renewable energy at node i at time t. REG The number of historical sample data for renewable energy power generation output. For the first The actual power generation at time t is based on historical sample data of renewable energy power generation output. For the first The power generation prediction deviation value of a historical sample of renewable energy power generation output at time t. To preset the first output Wasserstein metric radius, For the first The 1-norm of the random variable representing the power generation prediction bias of a historical sample of renewable energy power generation output data. To preset the second output Wasserstein metric radius, For the first The infinite norm of the random variable representing the power generation prediction deviation of a historical sample of renewable energy power generation output data.

[0069] It should be noted that the first output Wassstein metric radius is preset to be the Wassstein metric radius based on the 1 norm of the historical sample data of renewable energy power generation output, and the second output Wassstein metric radius is preset to be the Wassstein metric radius based on the infinite norm of the historical sample data of renewable energy power generation output. The 1 norm controls the sum of the absolute values ​​of all sample deviations to reflect the cumulative effect of the overall uncertainty of the system, and the infinite norm controls the maximum single-point deviation in all samples to reflect the uncertainty intensity under extreme conditions.

[0070] The risk warning method for unbalanced distribution network lines provided in this embodiment does not assume precise distribution. Instead, it first constructs an empirical distribution model of renewable energy power generation output based on the distribution characteristics of the actual power generation of each node. Then, taking the empirical distribution model of renewable energy power generation output of each node as the center, it incorporates all distributions that are similar to the empirical distribution model of renewable energy power generation output into a fuzzy set based on Wasserstein distance.

[0071] Step S3: Based on the prediction deviation of the load demand of each node, perform a second expansion on the empirical distribution model of the load demand of each node to obtain the second Wasserstein fuzzy set of the load demand reconstruction model of each node.

[0072] It should be noted that the second Wasserstein fuzzy set, which includes various load demand reconfiguration models, is the set of probability distributions that the load demand of each node in the target distribution network may follow.

[0073] Specifically, the predicted load demand and actual load demand of each node at each time point are first obtained, and an empirical load demand distribution model is constructed based on the distribution characteristics of the actual load demand of each node. Then, the load prediction deviation value of each node at each time point is calculated according to the difference characteristics between the predicted load demand and the corresponding actual load demand of each node at each time point. Finally, the deviation of each load demand reconstruction model and the empirical load demand distribution model is adjusted according to the load prediction deviation value, the preset first load Wassstein metric radius, and the preset second load Wassstein metric radius to obtain the second Wassstein fuzzy set. Here, the load demand reconstruction model represents the probability distribution model that the load demand of each node may follow.

[0074] In one embodiment, the uncertainty of the load demand of each node is modeled as follows:

[0075]

[0076] in, The uncertainty of the φ-phase load demand at node i at time t. Predict the load demand of node i at time t for phase φ. Let be the load demand random variable at time t.

[0077] Furthermore, the deviations between each load demand reconfiguration model and the load demand empirical distribution model are constrained by the following formula, and a second Wasserstein fuzzy set is constructed:

[0078]

[0079] Where EP() is the function for calculating mathematical expectation. The uncertainty of the φ-phase load demand at node i at time t. Predict the load demand of node i at time t for phase φ. The number of historical sample data for load demand. For the first The actual load demand at time t is based on historical sample data of load demand. For the first The load forecast deviation value of a historical sample of load demand at time t. To preset the Wasserstein metric radius for the first load, For the first The 1-norm of the load forecast deviation variable based on historical load demand sample data. To preset the second load Wasserstein metric radius, For the first The infinite norm of the load forecast deviation variable of a historical sample of load demand.

[0080] It should be noted that the preset first load Wassstein metric radius is the Wassstein metric radius based on the 1 norm of the historical load demand sample data, and the preset second load Wassstein metric radius is the Wassstein metric radius based on the infinite norm of the historical load demand sample data.

[0081] The risk warning method for unbalanced distribution network lines provided in this embodiment does not assume precise distribution. Instead, it first constructs an empirical distribution model of load demand based on the distribution characteristics of load demand at each node. Then, taking the empirical distribution model of load demand at each node as the center, it incorporates all distributions that are similar to the empirical distribution model of load demand into a fuzzy set based on Wasserstein distance.

[0082] Step S4: Merge the first Wassstein fuzzy set and the second Wassstein fuzzy set to obtain the joint power distribution fuzzy set for each branch; find the power prediction distribution model that maximizes the overload risk index in the joint power distribution fuzzy set, and calculate the line overload risk value for each branch based on the power prediction distribution model.

[0083] It should be noted that by constructing the first Wassstein fuzzy set in step S2 and the second Wassstein fuzzy set in step S3, the risk of distribution network overload under the worst-case distribution is transformed into: finding the random variable of generation prediction deviation. Deviation from load forecast variable The worst-case scenario is achieved by maximizing the value of the overload risk index, rather than directly searching for the worst-case result of the original uncertainty. Through this mathematical reconstruction, the calculation of the distribution network line overload risk index becomes easier to handle and can be effectively solved using conventional optimization techniques.

[0084] Specifically, the first Wassstein fuzzy set of the renewable energy generation output reconfiguration model and the second Wassstein fuzzy set of the load demand reconfiguration model are fused to generate a joint power distribution fuzzy set describing the power uncertainty of each branch. The candidate power distribution model that maximizes the overload risk index in the joint power distribution fuzzy set of each branch is taken as the power prediction distribution model. The overload risk index of the power prediction distribution model is taken as the line overload risk value of each branch at each time.

[0085] It should be noted that the joint power distribution fuzzy set of each node is a set including various power prediction distribution models. The risk warning method for unbalanced distribution network lines provided in this embodiment fuses the first Wassstein fuzzy set of the renewable energy generation output reconfiguration model and the second Wassstein fuzzy set of the load demand reconfiguration model to form a joint probability distribution fuzzy set describing the branch power distribution model. On this set, the probability distribution model that maximizes the overload risk index, i.e., the power prediction distribution model, is solved, thereby obtaining the line overload risk value, which represents the most unfavorable risk level that the target distribution network system may encounter under the premise of acknowledging all possible distributions. It does not depend on any specific preset distribution, but gives an upper bound on the risk. Based on this worst-case risk value, early warning and planning can ensure that even when the uncertainty moves in the most unfavorable direction, the system still has a high safety margin, thereby greatly improving the robustness and reliability of decision-making and overcoming the problem of inaccurate risk assessment caused by inaccurate distribution assumptions in traditional stochastic optimization.

[0086] Step S5: Convert the line overload risk value of each branch at each time into the corresponding branch outage probability using a preset mapping relationship, and calculate the all-day system outage risk warning value of the target distribution network based on the branch outage probability of all branches at all times. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

[0087] For details, please see Figure 1 , Figure 1 The diagram shows a flowchart illustrating a risk warning method for unbalanced distribution network lines in one embodiment of this application.

[0088] In one embodiment, step S1 involves constructing overload risk indicators for each branch at each time point based on historical power data of the target distribution network, including:

[0089] The historical power data of each branch in the target distribution network at each time point is compared with the upper and lower power thresholds of each branch to obtain the historical power over-limit value of each branch at each time point. The historical power over-limit value is then processed by a nonlinear increasing function to calculate the historical overload risk assessment value of each branch at each time point. The mathematical expectation of the historical overload risk assessment value is calculated based on the power distribution model of the historical power data of each branch, which serves as the overload risk index of each branch at each time point.

[0090] It should be noted that a nonlinear increasing function refers to a function whose value increases continuously as the independent variable increases within a certain interval, but its growth rate is not linear.

[0091] In one embodiment, the nonlinear increasing function is a quadratic function, and the formula for calculating the historical overload risk assessment value of each branch at each time step is:

[0092]

[0093] in, branch road At any moment of Historical overload risk assessment value, branch road At any moment of Historical power data branch road The upper limit threshold of power, branch road The lower limit threshold of power.

[0094] It should be noted that this embodiment calculates the expected severity only when the limit is exceeded, rather than simply calculating the expected severity of all historical data. This avoids the influence of non-limit-exceeding data, measures the average severity when the limit is exceeded, and organically integrates the limit-exceeding probability and severity into a single historical overload risk assessment value, thereby improving the comprehensiveness and accuracy of overload risk assessment.

[0095] In one embodiment, the overload risk index for each branch at each time step is calculated using the following formula:

[0096]

[0097] in, branch road At any moment of Phase overload risk indicators Assuming a branch When the historical power data follows a P-distribution, at time... of Under the condition that the historical power data of the phase is greater than or equal to the upper limit threshold, the branch At any moment of The mathematical expectation of the historical overload risk assessment value, Assuming a branch When the historical power data follows a P-distribution, at time... of Under the condition that the historical power data of the phase is less than or equal to the lower power threshold, the branch At any moment of Mathematical expectation of historical overload risk assessment value.

[0098] It should be noted that traditional risk assessments often only focus on the probability of exceeding limits or use linear weighting, failing to characterize the nonlinear characteristics of risk increasing rapidly with the degree of exceeding limits. This embodiment, by adding the expected values ​​of historical overload risk assessments under the condition that the historical power data of the branch at each time point is greater than or equal to the upper power limit threshold and the expected values ​​of historical overload risk assessments under the condition that the historical power data is less than or equal to the lower power limit threshold, covers the probability of various line overloads. Furthermore, it captures the severity of the nonlinear increase in risk caused by line overload through a quadratic function, simultaneously capturing both the probability of line overload and the severity of the nonlinear increase in risk, thus improving the reliability of characterizing the severity of branch power exceeding limits under a given power distribution model.

[0099] In one embodiment, the preset mapping relationship in step S5 is a linear mapping relationship. Step S5 converts the line overload risk value of each branch at each time into the corresponding branch power outage probability using the preset mapping relationship, including:

[0100]

[0101] in, branch road At any moment The probability of a branch circuit power outage. It is an exponential function with the natural constant as its base. This is the proportionality coefficient. branch road At any moment The line overload risk value.

[0102] It should be noted that, in order to generate intuitive early warning information, this embodiment uses the records of "overload risk index" and "whether a power outage has occurred" in historical operating data to calibrate a linear mapping relationship through statistical analysis. This mapping relationship is made to fit the actual situation of the specific power distribution network, such as the equipment tolerance and protection configuration characteristics. The linear mapping relationship links the line overload risk value with the branch power outage probability, thereby improving the accuracy of inferring the possibility of a power outage from the risk value.

[0103] In one embodiment, step S5, which calculates the all-day system outage risk warning value of the target distribution network based on the branch outage probability of all branches at all times, includes:

[0104]

[0105] in, This is the 24-hour system outage risk warning value for the target distribution network. It is a multiplication function. branch road At any moment The probability of a branch circuit power outage.

[0106] It should be noted that, under the assumption that the power outage events of each branch are independent, this embodiment aggregates the power outage probabilities of all branches and all time periods in the entire system into a single all-day system power outage risk warning value through probability calculation, forming system-level warning information that can be directly used for decision-making.

[0107] In one embodiment, the risk warning method for unbalanced distribution network lines provided in this application further includes:

[0108] The risk warning value for system power outage throughout the day is compared and analyzed with the preset risk probability threshold to determine the risk level of the target distribution network. Based on the risk level, early warning information and preventive control strategies are generated. The preventive control strategies include optimizing power generation plans and deploying emergency resources in advance.

[0109] In one embodiment, the proposed risk warning method for unbalanced distribution network lines is verified using an unbalanced IEEE 33-bus system. The system is configured as follows to highlight the risk: the load is set to 1.2 times the original value, the maximum allowable line power flow is 7MW, and four distributed photovoltaic power stations, each with an installed capacity of 800 kW, are connected to the system.

[0110] To address the uncertainties in photovoltaic (PV) output and load demand, the prediction errors for both were simulated using sampling. Using the day-ahead forecast as the mean, and 10%, 15%, and 20% of the forecast as standard deviations, 50 samples were generated for each. These samples were then used to construct fuzzy sets based on Wasserstein distance, with a confidence level of 90%. All modeling and calculations were performed in MATLAB, with optimization modeling using the YALMIP toolbox and solutions obtained using an optimization solver.

[0111] The overload risk value of the line and the corresponding all-day system power outage risk warning value obtained by this method were calculated under different prediction error levels (taking standard deviation as an example), assuming a proportionality coefficient γ=0.962 obtained through historical data calibration. The results are compared in the table below:

[0112] Prediction error level Line overload risk value All-day system power outage risk warning value 10% 0.081 7.8% 20% 0.134 12.5% 30% 0.291 19.7%

[0113] As can be seen from the table, as the prediction uncertainty, i.e. the standard deviation, increases, the worst-case overload risk index of the system increases non-linearly and significantly. This indicates that the method can keenly capture the amplification effect of increased uncertainty on system risk.

[0114] For details, please see Figure 2 , Figure 2 The diagram shown is a schematic representation of an unbalanced IEEE 33-node system in one embodiment of this application.

[0115] Another embodiment of this application provides a risk warning device for unbalanced distribution network lines. For details, please refer to [link to relevant documentation]. Figure 3 , Figure 3 The diagram shown is a structural block diagram of a risk warning device for an unbalanced distribution network line according to one embodiment of this application, comprising:

[0116] The risk assessment model construction module 11 is used to construct the overload risk index of each branch at each time based on the historical power data of the target distribution network. The overload risk index is the conditional expectation of a quadratic function that characterizes the severity of the branch power over-limit when the given power distribution model exceeds the limit.

[0117] The power generation output fuzzy set construction module 12 is used to perform a first amplification on the empirical distribution model of renewable energy power generation output of each node based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, so as to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node.

[0118] The load demand fuzzy set construction module 13 is used to perform a second expansion on the load demand empirical distribution model of each node based on the prediction deviation of the load demand of each node, so as to obtain the second Wasserstein fuzzy set of the load demand reconstruction model of each node.

[0119] The fuzzy set risk assessment module 14 is used to fuse the first Wassstein fuzzy set and the second Wassstein fuzzy set to obtain the joint power distribution fuzzy set of each node; it searches for the power prediction distribution model that maximizes the overload risk index in the joint power distribution fuzzy set, and calculates the line overload risk value of each branch based on the power prediction distribution model.

[0120] The all-day power outage probability mapping module 15 is used to convert the line overload risk value of each branch at each time into the corresponding branch power outage probability according to the preset mapping relationship, and to calculate the all-day system power outage risk warning value of the target distribution network based on the branch power outage probability of all branches at all times. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

[0121] Another embodiment of this application provides a risk warning device for unbalanced distribution network lines, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the risk warning method for unbalanced distribution network lines as described above.

[0122] Another embodiment of this application provides a computer-readable storage medium storing a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the risk warning method for unbalanced power distribution lines as described above.

[0123] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A risk early warning method for unbalanced distribution network lines, characterized in that, The method includes the following steps: Based on the historical power data of the target distribution network, an overload risk index for each branch at each time is constructed. The overload risk index is the conditional expectation of a quadratic function that characterizes the severity of branch power exceeding the limit under a given power distribution model. Based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, the empirical distribution model of renewable energy power generation output of each node is first amplified to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node; The empirical distribution model of load demand for each node is augmented based on the prediction deviation of load demand for each node to obtain the second Wasserstein fuzzy set of the load demand reconstruction model for each node. The first Wassstein fuzzy set and the second Wassstein fuzzy set are fused to obtain a joint power distribution fuzzy set for each branch; a power prediction distribution model that maximizes the overload risk index is found in the joint power distribution fuzzy set, and the line overload risk value of each branch is calculated based on the power prediction distribution model. The overload risk value of each branch at each time is converted into the corresponding branch outage probability using a preset mapping relationship. Based on the branch outage probability of all branches at all times, the all-day system outage risk warning value of the target distribution network is calculated. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

2. The risk early warning method for unbalanced distribution network lines according to claim 1, characterized in that, The process of constructing overload risk indicators for each branch at each time point based on historical power data of the target distribution network includes: The historical power data of each branch in the target distribution network at each time point is compared with the upper power limit threshold and the lower power limit threshold of each branch to obtain the historical power limit exceedance value of each branch at each time point. The historical power over-limit values ​​are processed using a nonlinear increasing function to calculate the historical overload risk assessment value of each branch at each time point; The expected value of the historical overload risk assessment value is calculated based on the power distribution model of the historical power data of each branch, and is used as the overload risk index of each branch at each time.

3. The risk early warning method for unbalanced distribution network lines according to claim 1, characterized in that, The first amplification of the empirical distribution model of renewable energy generation output for each node based on the prediction deviation of renewable energy generation output for each node in the target distribution network is performed to obtain the first Wasserstein fuzzy set of the reconstructed renewable energy generation output model for each node, including: The predicted power generation of each node at each of the stated times and the actual power generation of each node at each of the stated times are obtained respectively. The empirical distribution model of renewable energy power generation output is constructed based on the distribution characteristics of the actual power generation of each node; The power generation prediction deviation value of each node at each time moment is calculated based on the difference characteristics between the predicted power generation and the corresponding actual power generation of each node at each time moment; The deviation of each of the renewable energy power generation output reconstruction models and the renewable energy power generation output empirical distribution models is adjusted according to the power generation prediction deviation value, the first output Wasserstein metric radius and the second output Wasserstein metric radius to obtain a first Wasserstein fuzzy set, wherein the renewable energy power generation output reconstruction model represents the probability distribution model that the renewable energy power generation output of each node may follow.

4. The risk early warning method for unbalanced distribution network lines according to claim 1, characterized in that, The second augmentation of the empirical distribution model of load demand for each node based on the prediction deviation of load demand for each node yields a second Wasserstein fuzzy set of the load demand reconstruction model for each node, including: The predicted load demand of each node at each time point and the actual load demand of each node at each time point are obtained respectively. The load demand empirical distribution model is constructed based on the distribution characteristics of the actual load demand of each node; The load prediction deviation value of each node at each time moment is calculated based on the difference between the predicted load demand and the corresponding actual load demand of each node at each time moment. The deviation of each of the load demand reconfiguration models and the load demand empirical distribution models is adjusted based on the load forecast deviation value, the first load Wassstein metric radius, and the second load Wassstein metric radius to obtain a second Wassstein fuzzy set, wherein the load demand reconfiguration model represents a probability distribution model that the load demand of each node may follow.

5. The risk early warning method for unbalanced distribution network lines according to claim 2, characterized in that, The nonlinear increasing function is a quadratic function.

6. The risk early warning method for unbalanced distribution network lines according to claim 1, characterized in that, The calculation of the all-day system outage risk warning value of the target distribution network based on the outage probability of all branches at all times includes: The difference between the value 1 and the probability of power outage of each branch at each time is taken as the probability of safe operation of each branch at each time. The product of the safe operation probabilities of all branches in the target distribution network at all times is taken as the safe operation probability of the target distribution network throughout the day. The difference between the value 1 and the probability of safe operation throughout the day is used as the warning value for the risk of system outage throughout the day for the target distribution network.

7. The risk early warning method for unbalanced distribution network lines according to claim 1, characterized in that, The step of converting the line overload risk value of each branch at each time point into the corresponding branch power outage probability using a preset mapping relationship includes: in, Let be the probability of branch ij being out of power at time t, exp() be an exponential function with base to the natural constant, and γ be the proportionality constant. Let be the line overload risk value of branch ij at time t.

8. A risk warning device for unbalanced distribution network lines, characterized in that, include: The risk assessment model construction module is used to construct overload risk indicators for each branch at each time based on the historical power data of the target distribution network. The overload risk indicator is the conditional expectation of a quadratic function that characterizes the severity of branch power exceeding the limit under a given power distribution model. A power generation output fuzzy set construction module is used to perform a first amplification on the empirical distribution model of renewable energy power generation output of each node based on the prediction deviation of renewable energy power generation output of each node in the target distribution network, so as to obtain the first Wasserstein fuzzy set of the renewable energy power generation output reconstruction model of each node. The load demand fuzzy set construction module is used to perform a second augmentation on the load demand empirical distribution model of each node based on the prediction deviation of the load demand of each node, so as to obtain the second Wasserstein fuzzy set of the load demand reconstruction model of each node. The fuzzy set risk assessment module is used to fuse the first Wassstein fuzzy set and the second Wassstein fuzzy set to obtain a joint power distribution fuzzy set for each branch; to find the power prediction distribution model that maximizes the overload risk index in the joint power distribution fuzzy set; and to calculate the line overload risk value for each branch based on the power prediction distribution model. The all-day power outage probability mapping module is used to convert the line overload risk value of each branch at each time into the corresponding branch power outage probability according to the preset mapping relationship, and to calculate the all-day system power outage risk warning value of the target distribution network based on the branch power outage probability of all branches at all times. The preset mapping relationship is a linear relationship and the proportional coefficient is obtained by calibration through historical data.

9. A risk early warning device for unbalanced distribution network lines, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the risk warning method for unbalanced distribution network lines as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the risk warning method for unbalanced distribution network lines as described in any one of claims 1 to 7.

Images