Power distribution network area reliable power supply scene generation method under multi-type fault scene

Abstract

The invention discloses a power distribution network area reliable power supply scene generation method under a multi-type fault scene, which is suitable for the technical field of data processing, and comprises the following steps: S1, generating a sample initial scene by applying a Latin hypercube sampling technology, and cutting down the sample initial scene into a plurality of photovoltaic scenes by applying a scene cutting-down technology, calculating the photovoltaic scene probability of the corresponding photovoltaic scene; s2, constructing a fault data set according to node voltage phasor data in a fault situation corresponding to the photovoltaic scene, performing scene reduction on the fault data set by applying a scene reduction technology to obtain a fault scene, and determining a corresponding fault scene probability; s3, determining a power distribution network region load loss risk index according to the photovoltaic scene probability and the fault scene probability; according to the cloud digital characteristics, the power distribution network regional load loss risk indexes are sorted, and the optimal fault scene is determined by combining the fault scene combination, so that the operation state and the power supply reliability level of the power distribution network in the fault state can be accurately reflected.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network data processing, and in particular to a method for generating reliable power supply scenarios in distribution network areas under multiple types of fault scenarios. Background Art

[0002] With the global warming, extreme climates that occur once every few decades frequently occur. Such low-probability and high-loss events have gradually evolved into normal climate events. The resulting system failures have led to the decoupling of some power plants, causing serious damage to the power system. The actual distribution network has multiple branches and complex nodes. If all fault scenarios are traversed for analysis and calculation, the workload will be very large. Therefore, we need to screen the fault set. Clustering algorithms have been widely used in the electrical field. With the continuous development of power system informatization, the redundancy of measurement data has increased significantly, providing more fault data. Although different types of expected faults have their own characteristics, the similarities or differences between them also conform to the "clustering" characteristics, and clustering can be used to select appropriate fault set data. However, on the one hand, with the diversification of fault scenarios, the fault set samples are becoming larger and larger, and a more adaptable and computationally efficient clustering algorithm is needed; on the other hand, due to the influence of factors such as the volatility of distributed output, there is a certain uncertainty in the various fault scenario sets generated by the fault scenario. In order to accurately screen the fault scenarios in autonomous areas, it is necessary to further study the impact of potential complex uncertainties on the load loss risk index.

[0003] The above information disclosed in the Background section is only for enhancement of understanding of the background of the present application and therefore it may contain information that does not constitute the prior art that is already known to a person of ordinary skill in the art. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the generated fault scenarios are inaccurate due to the influence of complex uncertainty factors on the volatility of the power grid. A method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios is proposed, which fully considers the uncertainty of photovoltaic output, uses the Latin hypercube sampling method to generate initial scenario sets under multi-type faults, and adopts the K-means clustering method to reduce the original scenarios to obtain scenario data and scenario probabilities under the fault state; considering the influence of fault location and fault type, a fault data set is generated, and the K-means clustering method is used to reduce the fault scenarios of the distribution network to obtain various types of fault scenarios and their corresponding scenario probabilities; photovoltaic scenarios and fault scenarios are superimposed, and considering the photovoltaic scenario probability and the fault probability, an autonomous regional load loss risk index under the photovoltaic scenario is established, and a load loss risk cloud model is constructed, and the obtained scenario set is comprehensively sorted and screened to obtain the optimal fault scenario, which is not only representative but also can accurately reflect the operating status and power supply reliability level of the distribution network under the fault state.

[0005] In a first aspect, a technical solution provided in an embodiment of the present invention is a method for generating a reliable power supply scenario in a distribution network area under multiple types of fault scenarios, comprising the following steps: S1. Use Latin hypercube sampling technology to generate sample initial scenes, apply scene reduction technology to reduce sample initial scenes into several photovoltaic scenes, and calculate the photovoltaic scene probability of the corresponding photovoltaic scene based on the photovoltaic output power and probability model; S2. Construct a fault data set based on the node voltage phasor data in the fault scenario corresponding to the photovoltaic scenario, apply the scenario reduction technology to reduce the fault data set to obtain the fault scenario, and determine the corresponding fault scenario probability based on the number of fault lines and the probability model; S3. Determine the load loss risk index of the distribution network area according to the probability of photovoltaic scenarios and fault scenarios; generate fault scenario combinations through photovoltaic scenarios and fault scenarios, build cloud digital feature models under various fault scenario combinations, sort the load loss risk index of the distribution network area according to the cloud digital features, and determine the optimal fault scenario based on the fault scenario combination.

[0006] In this scheme, Latin hypercube sampling technology is used to generate representative and diverse sample initial scenarios, which can fully reflect the operating status of the distribution network under different conditions; by applying scenario reduction technology, the sample initial scenarios are reduced to several photovoltaic scenarios, which can accurately reflect the status of the distribution network under different photovoltaic output powers; at the same time, the photovoltaic scenario probability of the corresponding photovoltaic scenario is calculated based on the photovoltaic output power and probability model, providing basic data for subsequent risk assessment; by measuring the voltage phasor of each node and comparing it with the difference during standard operation, the fault node can be accurately identified, thereby constructing a fault data set containing fault information; through scenario reduction, representative fault scenarios can be extracted from a large amount of fault data, and the number of fault scenarios is not only small, but also can accurately reflect the operating status of the distribution network under different fault conditions; the generation of fault scenario combinations provides more possibilities for subsequent risk assessment, and by constructing a cloud digital feature model, the uncertainty characteristics of the fault scenario combination can be more comprehensively described, thereby providing a more accurate and reliable basis for risk assessment; through sorting and comparison, the optimal fault scenario can be found from a large number of fault scenario combinations, and this optimal fault scenario is not only representative, but also can accurately reflect the operating status and power supply reliability level of the distribution network under fault conditions.

[0007] Preferably, a normal distribution function is used to describe the probability model.

[0008] As a preferred method, a Latin hypercube sampling technique is used to generate a sample initial scene, and a scene reduction technique is used to reduce the sample initial scene into a number of photovoltaic scenes; the steps include: Set the random number interval, divide the random number interval into a parts, generate a random number from a uniform distribution in each part to obtain a samples; Disrupt the order of a random numbers in any way and calculate the value of the random numbers to obtain the probability value of each random sample; Calculate the sample value corresponding to the probability value according to the inverse function of the probability distribution function; The K-means clustering method is used to reduce the original samples and obtain photovoltaic scenarios under various fault conditions.

[0009] As a preferred method, the original samples are reduced by using a K-means clustering method to obtain the number of photovoltaic scenes under various fault conditions, including the following steps: Randomly select K sample points as the initial clustering center points of each cluster; Calculate the Euclidean distance from each sample to the initial cluster center point, and perform the nearest allocation according to the principle of minimum Euclidean distance, and allocate the samples to the sample cluster with the minimum Euclidean distance; The cluster center is updated according to the cluster center update formula, and the sample cluster is reconstructed according to the update result until the update termination condition is reached.

[0010] Preferably, the voltage phasor of each node is measured by a micro phase measurement unit to obtain the difference between the voltage phasor of the entire network nodes after various faults and that during standard operation, and a fault data set is constructed based on the nodes whose differences fall within a set range.

[0011] Preferably, a scenario reduction technology is applied to reduce the fault data set to obtain a fault scenario, and the corresponding fault scenario probability is determined according to the number of fault lines and the probability model, including the following steps: Set the characteristic vector y of the fault scenario s =(a s ,ΔA s ), where ΔA s is the energy loss of scenario s during the fault period, a s is the number of faulty lines in scenario s; The fault data set is divided into K groups, and a set of data is randomly selected from each group as the center of the initial clustering. The interval between each initial clustering center and all data is measured, and each data item is assigned to the category to which the closest clustering center belongs to form a sample cluster. The cluster center is updated according to the cluster center update formula, and the sample cluster is reconstructed according to the update result until the update termination condition is reached; Calculate the failure scenario probability of the failure scenario in each sample cluster.

[0012] As a preference, the energy loss ΔA of scenario s during the fault periods The calculation formula is as follows: Where: T2 is the set of annual average failure time of the distribution network; Ω N Represents a collection of nodes; pc s i,t is the load reduction of node i at time t in scenario s.

[0013] Preferably, the formula for calculating the probability of a fault scenario in each sample cluster is as follows: Where: Pr(s,i) is the scene s in cluster φ i The probability of N s Indicates the number of failure scenarios.

[0014] Preferably, the cloud digital features include expectation, entropy and super entropy.

[0015] As a preferred method, a cloud digital feature model is constructed based on expectation, entropy, and super entropy, specifically: expect is the sample mean; entropy x i is the i-th sample value, n is the number of sample points; Super Entropy S 2 is the sample variance.

[0016] Beneficial effects of the present invention: (1) Using Latin hypercube sampling technology to generate sample initial scenarios can ensure the uniform distribution of samples in multidimensional space and improve sampling efficiency. Using scenario reduction technology to reduce sample initial scenarios and fault data sets to a limited number of photovoltaic scenarios and fault scenarios can significantly reduce the number of scenarios that need to be analyzed while maintaining the diversity and representativeness of the scenarios, thereby improving the efficiency of distribution network fault scenario generation. (2) By comprehensively considering the uncertainty of photovoltaic output power, the randomness of the number of faulty lines and their impact on the risk of load loss in the distribution network, the regional load loss risk index of the distribution network can be calculated. By constructing a cloud digital feature model, the risk index can be sorted and compared. The cloud digital feature model can describe the uncertainty characteristics of the fault scenario combination and can achieve an accurate assessment of the regional load loss risk of the distribution network, providing an important reference for the planning, operation and maintenance of the distribution network. By sorting and comparing the risk indicators under different fault scenario combinations, the most dangerous fault scenario can be found, so that corresponding preventive measures can be taken to reduce the loss of the distribution network under the fault state.

[0017] The above invention content is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0019] Figure 1 It is a flow chart of the method for generating reliable power supply scenarios in distribution network areas under multiple types of fault scenarios of the present invention.

[0020] Figure 2 It is a modified IEEE33 node structure diagram of a specific implementation mode of the present invention.

[0021] Figure 3 It is a photovoltaic scene generation result diagram of a specific implementation mode of the present invention.

[0022] Figure 4 They are respectively the photovoltaic scene reduction result diagrams according to the specific implementation modes of the present invention.

[0023] Figure 5 It is a graph showing the probability of occurrence of typical fault scenarios according to a specific implementation of the present invention. DETAILED DESCRIPTION

[0024] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0026] Example: Figure 2 As shown, an improved 33-node radial distribution network is taken as an object, and the distribution network consists of 33 nodes, among which nodes 7, 13, 15, and 27 are equipped with photovoltaics, and the nodes where important loads are located are nodes 6, 7, 13, and 31. At this time, the relevant methods and means in this embodiment are used to generate and reduce the fault scenario set for the distribution network area.

[0027] like Figure 1 As shown, a method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios includes the following steps: S1, using Latin hypercube sampling technology to generate sample initial scenarios, using scenario reduction technology to reduce the sample initial scenarios into several photovoltaic scenarios, and calculating the photovoltaic scenario probability of the corresponding photovoltaic scenario based on the photovoltaic output power and probability model.

[0028] Specifically, for the photovoltaic output prediction error, it is generally considered to satisfy the normal distribution, and the prediction error probability density function is: Among them, μ and σ are the expectation and standard deviation of the prediction error, respectively, and μ is 0.

[0029] As a preferred embodiment, according to the photovoltaic output power and prediction error probability model, Latin hypercube sampling technology is used to generate an initial scene of a sufficiently large sample, and the scene reduction technology is applied to reduce the initial scene to several typical scenes; specifically, Latin hypercube sampling technology is used to generate sample initial scenes, and the scene reduction technology is applied to reduce the sample initial scenes to several photovoltaic scenes; the following steps are included: Set the random number interval and divide the random number interval into a parts. Generate a random number P in each part by uniform distribution. a Get a samples; Disrupt the order of a random numbers in any way and calculate the value of the random numbers to obtain the probability value of each random sample; According to the inverse function f of the probability distribution function -1 (P a )=x a Calculate the sample value corresponding to the probability value a; The K-means clustering method is used to reduce the original samples and obtain photovoltaic scenarios under various fault conditions.

[0030] As a preferred embodiment, the original samples are reduced by using a K-means clustering method to obtain the number of photovoltaic scenes under various fault conditions, including the following steps: For the data sample X that needs to be clustered M×N (M is the sample size, N is the sample dimension), randomly select K sample points as the initial clustering center points of each cluster; Calculate the Euclidean distance from each sample to the initial cluster center point, implement the nearest allocation according to the principle of minimum Euclidean distance, allocate the samples to the sample cluster with the minimum Euclidean distance, and calculate the number of samples C contained in the new cluster i ; Update the cluster center according to the cluster center update formula and recalculate the cluster center point o in each new cluster i , where the cluster center update expression is: Where x is the new cluster C i Sample data in ; The sample cluster is reconstructed according to the update results until the update termination condition is reached and the update is stopped.

[0031] Furthermore, this embodiment extracts 3000 sample data of photovoltaic based on historical data, such as Figure 3 As shown, five typical scenarios are selected according to the method in S1, such as Figure 4 As shown, these scenarios are equivalent to obtain the equivalent distributed output.

[0032] S2. Construct a fault data set based on the node voltage phasor data in the fault situation corresponding to the photovoltaic scenario, apply scenario reduction technology to reduce the fault data set to obtain fault scenarios, and determine the corresponding fault scenario probability based on the number of fault lines and the probability model.

[0033] It can be understood that when various faults occur in the system, the voltage phasor data of each node measured by μPMU (micro phase measurement unit) constitutes a fault data set. The formed fault data set is used for K-means clustering to reduce scenarios. The clustering algorithm extracts multiple fault types from the fault data set, thereby reducing similar fault scenarios and generating severity indicators for various types of faults, thereby achieving preliminary screening of the expected fault set. By using μPMU for measurement, the difference between the voltage phasor of the entire network node after various faults and the standard operation can be obtained, and a fault data set is constructed based on the nodes whose differences fall within the set range.

[0034] As a preferred embodiment, a scenario reduction technology is applied to reduce the fault data set to obtain a fault scenario, and the corresponding fault scenario probability is determined according to the number of fault lines and the probability model, including the following steps: Set the characteristic vector y of the fault scenario s =(a s ,ΔA s ), where ΔA s is the energy loss of scenario s during the fault period, a s is the number of faulty lines under scenario s; where the energy loss ΔA of scenario s during the fault period s The calculation formula is as follows: Where: T2 is the set of annual average failure time of the distribution network; Ω N Represents a collection of nodes; pc s i,t is the load reduction of node i at time t under scenario s; The fault data set is divided into K groups, and a set of data is randomly selected from each group as the center of the initial clustering. The interval between each initial clustering center and all the data is measured, and each data item is assigned to the category to which the closest clustering center belongs to form a sample cluster.

[0035] The cluster center is updated according to the cluster center update formula, and the sample cluster is reconstructed according to the update result until the update termination condition is reached. Specifically, in the K-mean clustering algorithm, N s dimensional vector {y1,y2,...,y Ns} is divided into k clusters {φ1,φ2,...,φ k}, as follows: In the formula, dis(k) represents the sum of distances of the kth class, μ i is the cluster center.

[0036] Choose φ arbitrarily i A scene in a cluster is taken as a representative. By accumulating the probabilities of all scenes in the cluster, the probability expression of the representative scene can be obtained as follows: Where: Pr(s,i) is the scene s in cluster φ i The probability of N s Indicates the number of failure scenarios.

[0037] The aforementioned clustering algorithm is used to reduce the scenarios of a large number of faults generated by μPMU measurements, thereby preliminarily screening out the fault types; after collecting and analyzing the historical scenario data of different fault types at each node in a certain area's distribution network, a fault data set containing 1,000 sets of data is generated by subtracting the voltage phasor of each node after the fault measured by μPMU from the standard voltage phasor under normal operating conditions. The K-means mean clustering algorithm is used to reduce the fault scenarios 995 times, and 5 typical scenarios that can describe the uncertainty of the fault are extracted, and their corresponding probability of occurrence is as follows: Figure 5 shown.

[0038] S3. Determine the load loss risk index of the distribution network area according to the probability of photovoltaic scenarios and fault scenarios; generate fault scenario combinations through photovoltaic scenarios and fault scenarios, build cloud digital feature models under various fault scenario combinations, sort the load loss risk index of the distribution network area according to the cloud digital features, and determine the optimal fault scenario based on the fault scenario combination.

[0039] As a preferred embodiment, the load loss risk index describes the loss of important loads after a power system failure, which is the product of the failure probability and the degree of power loss of important loads in the distribution network area. The load loss risk index of the distribution network area in the photovoltaic scenario is: Where N is the number of scenes, P s is the photovoltaic field probability, P f is the probability of the fault scenario, and S is the degree of load power loss under the fault scenario.

[0040] As a preferred embodiment, due to various influencing factors such as changes in distributed output, the various scenario sets generated by the fault scenario have certain unpredictability. This application uses cloud digital features to rank the load loss risk index under various fault scenarios of the distribution network by deeply studying the uncertainty factors that may cause fluctuations in the load loss risk index.

[0041] Construction of cloud digital feature model for load loss risk: Cloud digital features include three index values: expectation, entropy, and super entropy. Cloud heuristics can extract cloud digital features. To extract cloud digital features, we first need to calculate the sample point x i The average value of (i=1,2,3,…,n) Sample first-order absolute central moment variance Then, calculate the expectation entropy Super Entropy

[0042] Furthermore, the fault scenario combinations are screened: the scenarios are sorted according to the output values ​​of the expectation, entropy, and super entropy cloud digital characteristic indicators. The larger the value, the more serious the fault. The optimal fault scenario combination is screened out according to the severity.

[0043] As an example of the implementation methods in the above embodiments, the five expected fault sets obtained by the preliminary screening of the clustering algorithm and the five distributed output scenarios are combined to obtain 25 different fault scenario combinations. Taking the autonomous area load loss risk index as the expectation, a cloud digital feature model for various scenarios is constructed. The larger the expectation, entropy, and super entropy cloud digital feature indicators, the more serious the fault. The 25 results are screened and sorted according to the expectation, entropy, and super entropy cloud digital feature indicators according to the severity of their fault impact, and the cloud digital features for the final five scenarios are shown in Table 1. Table 1 Digital characteristics of scene cloud <![CDATA[Expected E x > <![CDATA[Entropy E n > <![CDATA[Hyperentropy H e > Scenario 1 12.04 0.67 0.038 Scenario 2 8.07 0.30 0.025 Scene 3 6.93 0.29 0.029 Scene 4 4.40 0.17 0.044 Scene 5 2.68 0.28 0.018

[0044] By comparing and analyzing the expected value, entropy and super entropy cloud digital features under the above-mentioned fault conditions, it can be seen that the expected value, entropy and super entropy cloud digital features in these five scenarios are all high. For scenario 1, its expected value and entropy are the highest, and the super entropy is also relatively high, indicating that in this case, the possibility of overload in the autonomous area is the highest, the load loss risk index is the largest, the potential risk is the most serious and the impact is widespread, so it is necessary to adopt methods such as autonomous area division and network reconstruction to strengthen preventive measures to reduce the impact of the fault.

[0045] This embodiment has at least the following technical effects: by using Latin hypercube sampling technology, representative and diverse sample initial scenarios can be generated, which can fully reflect the operating status of the distribution network under different conditions; by applying scenario reduction technology, the sample initial scenarios are reduced to several photovoltaic scenarios, which can accurately reflect the distribution network status under different photovoltaic output powers; at the same time, the photovoltaic scenario probability of the corresponding photovoltaic scenario is calculated according to the photovoltaic output power and probability model, providing basic data for subsequent risk assessment; by measuring the voltage phasor of each node and comparing it with the difference during standard operation, the fault node can be accurately identified, thereby constructing a fault data set containing fault information; through scenario reduction, representative fault scenarios can be extracted from a large amount of fault data, and the number of fault scenarios is not only small, but also can accurately reflect the operating status of the distribution network under different fault conditions; the generation of fault scenario combinations provides more possibilities for subsequent risk assessment, and by constructing a cloud digital feature model, the uncertainty characteristics of the fault scenario combination can be more comprehensively described, thereby providing a more accurate and reliable basis for risk assessment; through sorting and comparison, the optimal fault scenario can be found from a large number of fault scenario combinations, and this optimal fault scenario is not only representative, but also can accurately reflect the operating status and power supply reliability level of the distribution network under a fault state.

[0046] The specific implementation described above is a preferred implementation of the method for generating reliable power supply scenarios in distribution network areas under multiple types of fault scenarios of the present invention. It is not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for generating reliable power supply scenarios in distribution network areas under multiple types of fault scenarios, characterized in that: The steps include: S1. Use Latin hypercube sampling technology to generate sample initial scenes, apply scene reduction technology to reduce sample initial scenes into several photovoltaic scenes, and calculate the photovoltaic scene probability of the corresponding photovoltaic scene based on the photovoltaic output power and probability model; S2. Construct a fault data set based on the node voltage phasor data in the fault scenario corresponding to the photovoltaic scenario, apply the scenario reduction technology to reduce the fault data set to obtain the fault scenario, and determine the corresponding fault scenario probability based on the number of fault lines and the probability model; S3. Determine the load loss risk index of the distribution network area according to the probability of photovoltaic scenarios and fault scenarios; generate fault scenario combinations through photovoltaic scenarios and fault scenarios, build cloud digital feature models under various fault scenario combinations, sort the load loss risk index of the distribution network area according to the cloud digital features, and determine the optimal fault scenario based on the fault scenario combination.

2. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 1 is characterized in that: The normal distribution function is used to describe the probability model.

3. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 1 is characterized in that: The Latin hypercube sampling technique is used to generate the sample initial scene, and the scene reduction technique is used to reduce the sample initial scene into several photovoltaic scenes; the steps include: Set the random number interval, divide the random number interval into a parts, generate a random number from a uniform distribution in each part to obtain a samples; Disrupt the order of a random numbers in any way and calculate the value of the random numbers to obtain the probability value of each random sample; Calculate the sample value corresponding to the probability value according to the inverse function of the probability distribution function; The K-means clustering method is used to reduce the original samples and obtain photovoltaic scenarios under various fault conditions.

4. The method for generating a reliable power supply scenario in a distribution network area under multiple types of fault scenarios according to claim 3 is characterized in that: The original samples are reduced by K-means clustering method to obtain the number of photovoltaic scenes under various fault conditions, including the following steps: Randomly select K sample points as the initial clustering center points of each cluster; Calculate the Euclidean distance from each sample to the initial cluster center point, and perform the nearest allocation according to the principle of minimum Euclidean distance, and allocate the samples to the sample cluster with the minimum Euclidean distance; The cluster center is updated according to the cluster center update formula, and the sample cluster is reconstructed according to the update result until the update termination condition is reached.

5. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 1, characterized in that: The voltage phasor of each node is measured by a micro-phase measurement unit to obtain the difference between the voltage phasor of the entire network node after various faults and the voltage phasor during standard operation, and a fault data set is constructed based on the nodes whose differences fall within the set range.

6. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 1, characterized in that: The scenario reduction technology is used to reduce the fault data set to obtain the fault scenario, and the corresponding fault scenario probability is determined according to the number of fault lines and the probability model, including the following steps: Set the characteristic vector y of the fault scenario s =(a s ,ΔA s ), where ΔA s is the energy loss of scenario s during the fault period, a s is the number of faulty lines in scenario s; The fault data set is divided into K groups, and a set of data is randomly selected from each group as the center of the initial clustering. The interval between each initial clustering center and all data is measured, and each data item is assigned to the category to which the closest clustering center belongs to form a sample cluster. The cluster center is updated according to the cluster center update formula, and the sample cluster is reconstructed according to the update result until the update termination condition is reached; Calculate the failure scenario probability of the failure scenario in each sample cluster.

7. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 6, characterized in that: Energy loss ΔA of scenario s during fault s The calculation formula is as follows: Where: T2 is the set of annual average failure time of the distribution network; Ω N Represents a collection of nodes; pc s i,t is the load reduction of node i at time t in scenario s.

8. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 6, characterized in that: The formula for calculating the failure scenario probability in each sample cluster is as follows: Where: Pr(s,i) is the scene s in cluster φ i The probability of N s Indicates the number of failure scenarios.

9. The method for generating reliable power supply scenarios in distribution network areas under multi-type fault scenarios according to claim 1, characterized in that: The cloud digital features include expectation, entropy, and super entropy.

10. The method for generating a reliable power supply scenario in a distribution network area under multiple types of fault scenarios according to claim 9, characterized in that: The cloud digital feature model is constructed based on expectation, entropy and super entropy, specifically: expect is the sample mean; entropy x i is the i-th sample value, n is the number of sample points; Super Entropy S 2 is the sample variance.

Images