Multi-level risk identification method, system and equipment for distributed energy access power distribution network

Abstract

The invention discloses a distributed energy access power distribution network multi-level risk identification method, system and equipment. The method comprises the following steps: acquiring voltage data, power data and load data of a transformer area, a line and a main transformer; performing risk assessment based on the voltage data, the power data and the load data in combination with a node voltage out-of-limit risk assessment calculation formula and a line power flow out-of-limit risk assessment calculation formula to obtain a risk assessment result; performing risk identification based on the risk assessment result, the voltage data, the power data and the load data to obtain a risk type identification result; according to the invention, risk assessment is carried out based on transformer area line main transformer multi-stage data, so that risk assessment is more efficient and accurate, and system risks are assessed more comprehensively; according to the invention, the abnormal condition is analyzed based on data, so that the risk identification result is more accurate; the risk identification model obtained through training of the random vector function link classifier is high in identification precision, high in calculation speed and suitable for coping with complex and changeable power grid risk management requirements after distributed energy access.

Description

Technical Field

[0001] The present application relates to the field of distribution network risk identification and assessment, and specifically to a multi-level risk identification method, system and equipment for distributed energy access distribution network. Background Art

[0002] Distributed resources are increasingly widely used, and there are operational risks such as random switching and line short circuits; a large number of various loads are connected to the distribution network, and there are a large number of impact loads among them, which faces greater challenges in the stability and safety of distribution network operation. The safe and stable operation of the distribution network requires timely perception of operational risks, so efficient and accurate risk assessment and identification technology support is required. Distribution network safety risk assessment is to evaluate the distribution network as a whole and conduct quantitative analysis of the system's operational risks, so as to assess the consequences and impacts of power outages; distribution network risk identification is to identify the local distribution network and identify the components or links that will affect the distribution network outage.

[0003] Some developed countries have achieved a lot of results in the risk assessment of distributed power access to distribution networks, mainly focusing on assessment technology and model construction. Assessment technologies include analytical methods such as network method and state space method, as well as simulation methods such as Monte Carlo and Bayesian network. In terms of model construction, the acceptance capacity of distributed power sources is calculated by dynamic simulation methods, and the access power level is indirectly obtained by considering the working mode, disturbance method and stability criteria. However, there is a lack of unified and effective models and algorithms for real-time fault probability prediction. Existing methods are mostly limited to a single level and cannot comprehensively assess the risks of distribution network systems.

[0004] For example, Turner R, Walton S, Duke R. published "Stability and Bandwidth Implications of Digitally Controlled Grid-Connected Parallel Inverters" in IEEE Transactions on Industrial Electronics, which analyzed the impact of distributed photovoltaic power sources of different capacities connected to different locations of the distribution network on the voltage distribution of the distribution network. Studies have shown that the grid connection of distributed power sources will increase the voltage of each node in the system, and the grid connection capacity of distributed power sources should be limited to prevent the node voltage from exceeding the limit. The grid connection of distributed power sources will affect the voltage distribution, and the grid connection capacity needs to be limited to prevent the voltage from exceeding the limit. In addition, since photovoltaic output has a time sequence characteristic, and users also have power loads at different times of the day, with the cooperation of these factors, the operating risks of the distribution network at different times are obviously different, so it is necessary to conduct a time sequence risk analysis of the distribution network. However, the overall risk of the distribution network station-line-transformer multi-level was not evaluated; the risk assessment method has a single time scale, which leads to limited results.

[0005] The "Distribution Network Reliability Assessment Algorithm for Switches" published by Xie Kaigui, Yi Wu, Xia Tian and others in the Journal of Electric Power System Automation and the "Comparative Study of Reliability Assessment Methods for Medium Voltage Distribution Networks" published by Zhao Hua, Wang Zhuding, Xie Kaigui and others in the Journal of Power System Technology are research objects that use components such as lines, transformers or load points to achieve simple early warning of damage to electrical components. However, the impact of the uncertainty of distributed resources and new energy power generation equipment on system operation failures is not considered.

[0006] The above prior art has the following disadvantages:

[0007] (1) Existing risk assessment indicators are mainly concentrated on a single level and do not assess the overall risk of the distribution network at multiple levels: station, line, and transformer;

[0008] (2) The risk assessment method has a single time scale, which leads to limited results;

[0009] (3) The impact of the uncertainty of distributed resources and new energy equipment on system failures is not considered.

[0010] Therefore, how to solve the problem that existing risk assessment indicators are mainly concentrated at a single level, the risk assessment time scale is single, and the impact of the uncertainty of distributed resources and new energy equipment on system failures is not considered is an urgent problem that needs to be solved. Summary of the invention

[0011] In order to solve the problems in the prior art that risk assessment indicators are mainly concentrated at a single level, the risk assessment time scale is single, and the impact of the uncertainty of distributed resources and new energy equipment on system failures is not considered, this application proposes a multi-level risk identification method for distributed energy access distribution network, including:

[0012] Obtain voltage data, power data and load data of substations, lines and main transformers;

[0013] Based on the voltage data, power data and load data combined with the node voltage over-limit risk assessment calculation formula and the line power flow over-limit risk assessment calculation formula, a risk assessment is performed to obtain a risk assessment result;

[0014] Risk identification is performed based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

[0015] Preferably, the risk assessment is performed based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line power flow over-limit risk assessment calculation formula to obtain a risk assessment result, including:

[0016] Performing deterministic power flow calculation based on the voltage data, power data and load data to obtain the values ​​of each node voltage and each line power flow;

[0017] Based on the values ​​of the voltages of each node and the power flows of each line, statistics are performed to obtain the probability distribution function and cumulative distribution function of the node and the probability distribution function and cumulative distribution function of the power flows of the line;

[0018] Based on the cumulative distribution function of the node and the node voltage over-limit risk assessment calculation formula, the node voltage over-limit risk value is obtained;

[0019] The cumulative distribution function of the line flow is combined with the line flow over-limit risk assessment calculation formula to obtain the line flow over-limit risk value.

[0020] Preferably, the node voltage over-limit risk assessment calculation formula is as follows:

[0021]

[0022] In the formula, Risk v is the node voltage over-limit risk value; Prob(V i ) is the voltage over-limit risk probability of node i; Sev(w i ) is the severity of overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i; Ω i is a set of nodes; i is the order of nodes.

[0023] Preferably, the line power flow over-limit risk assessment calculation formula is as follows:

[0024] Pisk s =ΣProb(S j )·Sev(S j )

[0025] In the formula, Pisk s is the risk value of line power flow exceeding the limit; Prob(S j ) is the crossing probability of branch j; Sev(S j ) is the size of the deviation of branch j from the threshold.

[0026] Preferably, the risk identification is performed based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result, including:

[0027] Based on the risk assessment result, voltage over-limit risk data and branch circuit overload risk data are screened out from the voltage data, power data and load data;

[0028] Based on the voltage over-limit risk data and the branch overload risk data, risk identification is performed in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result;

[0029] The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data;

[0030] The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data.

[0031] Preferably, the training process of the voltage risk identification model includes:

[0032] Based on the voltage over-limit risk data, each feature subset is constructed according to each node being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0033] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a voltage risk identification model;

[0034] The voltage risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained voltage risk identification model is obtained; otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0035] Preferably, the step of training a random vector function link classifier using the training set and determining parameters of the random vector function link classifier to obtain a voltage risk identification model includes:

[0036] Constructing an input matrix of a random vector function link classifier based on the training set;

[0037] Constructing a hidden layer matrix based on the input matrix, randomly generated input weights and bias and activation functions of the hidden layer;

[0038] Constructing an output matrix based on the input matrix and the hidden layer matrix;

[0039] Calculate based on the output matrix and the true label matrix to obtain output weights;

[0040] Parameters of the classifier are linked as a random vector function based on the output weights, the randomly generated input weights, and the bias of the hidden layer.

[0041] Preferably, the training process of the tidal risk identification model includes:

[0042] Based on the branch overload risk number, each feature subset is constructed according to each branch being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0043] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a power flow risk identification model;

[0044] The tidal current risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained tidal current risk identification model is obtained. Otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0045] Based on the same application concept, this application also proposes a multi-level risk identification system for distributed energy access to distribution networks, including:

[0046] Data acquisition module, used to obtain voltage data, power data and load data of the substation, line and main transformer;

[0047] A risk assessment module, used to perform risk assessment based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line power flow over-limit risk assessment calculation formula to obtain a risk assessment result;

[0048] The risk identification module is used to perform risk identification based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

[0049] Preferably, the risk assessment module is specifically used for:

[0050] Performing deterministic power flow calculation based on the voltage data, power data and load data to obtain the values ​​of each node voltage and each line power flow;

[0051] Based on the values ​​of the voltages of each node and the power flows of each line, statistics are performed to obtain the probability distribution function and cumulative distribution function of the node and the probability distribution function and cumulative distribution function of the power flows of the line;

[0052] Based on the cumulative distribution function of the node and the node voltage over-limit risk assessment calculation formula, the node voltage over-limit risk value is obtained;

[0053] The cumulative distribution function of the line flow is combined with the line flow over-limit risk assessment calculation formula to obtain the line flow over-limit risk value.

[0054] Preferably, the node voltage over-limit risk assessment calculation formula in the risk assessment module is as follows:

[0055]

[0056] In the formula, Risk v is the node voltage over-limit risk value; Prob(V i ) is the voltage over-limit risk probability of node i; Sev(w i ) is the severity of the overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i; Ω i is a set of nodes; i is the order of nodes.

[0057] Preferably, the line flow over-limit risk assessment calculation formula in the risk assessment module is as follows:

[0058] Pisk s =∑Prob(S j )·Sev(S j )

[0059] In the formula, Pisk s is the risk value of line power flow exceeding the limit; Prob(S j ) is the crossing probability of branch j; Sev(S j ) is the size of the deviation of branch j from the threshold.

[0060] Preferably, the risk identification module is specifically used to:

[0061] A data screening submodule, for screening out voltage over-limit risk data and branch overload risk data from the voltage data, power data and load data based on the risk assessment result;

[0062] A model calculation submodule, used to perform risk identification based on the voltage over-limit risk data and the branch overload risk data in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result;

[0063] The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data;

[0064] The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data.

[0065] Preferably, the risk identification module further includes a voltage risk identification model training submodule; the voltage risk identification model training submodule is specifically used for:

[0066] Based on the voltage over-limit risk data, each feature subset is constructed according to each node being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0067] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a voltage risk identification model;

[0068] The voltage risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained voltage risk identification model is obtained; otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0069] Preferably, the voltage risk identification model training submodule uses the training set to train a random vector function link classifier, and determines the parameters of the random vector function link classifier to obtain a voltage risk identification model, including:

[0070] Constructing an input matrix of a random vector function link classifier based on the training set;

[0071] Constructing a hidden layer matrix based on the input matrix, randomly generated input weights and bias and activation functions of the hidden layer;

[0072] Constructing an output matrix based on the input matrix and the hidden layer matrix;

[0073] Calculate based on the output matrix and the true label matrix to obtain output weights;

[0074] Parameters of the classifier are linked as a random vector function based on the output weights, the randomly generated input weights, and the bias of the hidden layer.

[0075] Preferably, the risk identification module further includes a tidal current risk identification model training submodule; the tidal current risk identification model training submodule is specifically used for:

[0076] Based on the branch overload risk number, each feature subset is constructed according to each branch being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0077] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a power flow risk identification model;

[0078] The tidal current risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained tidal current risk identification model is obtained. Otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0079] On the other hand, the present application also proposes an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0080] The memory is used to store one or more programs;

[0081] When the one or more programs are executed by the at least one processor, a multi-level risk identification method for accessing a distributed energy source to a distribution network as described above is implemented.

[0082] On the other hand, the present application also proposes a readable storage medium having an execution program stored thereon, and when the execution program is executed, a multi-level risk identification method for the access of distributed energy to a distribution network as described above is implemented.

[0083] Compared with the prior art, the beneficial effects of this application are:

[0084] A multi-level risk identification method, system and device for distributed energy access to a distribution network, including: obtaining voltage data, power data and load data of a substation, a line and a main transformer; performing risk assessment based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line flow over-limit risk assessment calculation formula to obtain a risk assessment result; performing risk identification based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result; the application performs risk assessment based on multi-level data of substation lines and main transformers, making risk assessment more efficient and accurate, and more comprehensive in assessing system risks; the application analyzes abnormal conditions based on data to make risk identification results more accurate;

[0085] This application uses a risk identification model trained by a random vector function link classifier, which has high recognition accuracy and fast calculation speed, and is suitable for coping with the complex and changeable grid risk management needs after distributed energy access. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 This is a flow chart of a multi-level risk identification method for distributed energy access to a distribution network in this application;

[0087] Figure 2 A flow chart of the method of this application;

[0088] Figure 3 This is a structural diagram of the RVFL classifier model of this application;

[0089] Figure 4 This is the online simulation diagram of the carrying capacity of the 10kV 111 line in a certain place - the photovoltaic access distribution network of the whole county in this application;

[0090] Figure 5 A schematic diagram of the voltage over-limit risk value of each node in the substation area of ​​this application;

[0091] Figure 6 The voltage probability density curve and cumulative probability distribution curve of node 70 of the present application;

[0092] Figure 7 The voltage probability density curve and cumulative probability distribution curve of node 40 of the present application;

[0093] Figure 8 A schematic diagram of the risk value of the flow crossing limit of each branch in the substation area of ​​this application;

[0094] Fig. 9 The active power probability density curve and cumulative probability distribution curve of branch 10 of this application;

[0095] Fig.10 The active power probability density curve and cumulative probability distribution curve of branch 40 of this application are shown in FIG.

[0096] Fig.11 A schematic diagram of voltage over-limit risk values ​​for each node on the main line of this application;

[0097] Fig.12 The voltage probability density curve and cumulative probability distribution curve of node 20 of the present application;

[0098] Fig.13 The voltage probability density curve and cumulative probability distribution curve of node 40 of the present application;

[0099] Fig.14 This is a schematic diagram of the risk value of the flow crossing limit of each branch on the main line of this application;

[0100] Fig.15 The active power probability density curve and cumulative probability distribution curve of branch 5 of this application;

[0101] Fig.16 The active power probability density curve and cumulative probability distribution curve of branch 40 of this application;

[0102] Fig.17 A schematic diagram of voltage over-limit risk values ​​of each node on the main transformer of this application;

[0103] Fig.18 The voltage probability density curve and cumulative probability distribution curve of node 70 of the present application;

[0104] Fig.19 The voltage probability density curve and cumulative probability distribution curve of node 12 of the present application;

[0105] Fig. 20 A schematic diagram of the risk value of power flow exceeding the limit for each branch of the main transformer of this application;

[0106] Fig.21 The active power probability density curve and cumulative probability distribution curve of branch 35 of this application;

[0107] Fig. 22 The active power probability density curve and cumulative probability distribution curve of branch No. 12 of this application;

[0108] Fig.23 A schematic diagram of the voltage over-limit risk value of each node on the transformer in this application;

[0109] Fig.24 The voltage probability density curve and cumulative probability distribution curve of node 70 of the present application;

[0110] Fig.25 The voltage probability density curve and cumulative probability distribution curve of node 40 of the present application;

[0111] Fig.26 A schematic diagram of the risk value of the power flow exceeding the limit for each branch of the platform transformer in this application;

[0112] Fig. 27 The active power probability density curve and cumulative probability distribution curve of branch 10 of this application;

[0113] Fig.28 The active power probability density curve and cumulative probability distribution curve of branch No. 40 of this application;

[0114] Fig.29 This is a structural diagram of a multi-level risk identification system for distributed energy access to a distribution network in this application;

[0115] Fig.30 A diagram of an electronic device for the present application. DETAILED DESCRIPTION

[0116] In view of the above-mentioned deficiencies in the prior art, the present invention proposes a data-driven multi-level risk assessment and identification method for distributed energy access to distribution networks. The present invention establishes a multi-level risk assessment model by considering the uncertainty of distributed photovoltaics and the differences in access areas, lines and main transformers, and uses data-driven analysis of abnormal situations to perform risk assessment and identification. This technology solves the problem of a multi-level risk evaluation index system for stations, lines and transformers at multiple time scales within a day in the distribution network. In order to better understand the present application, the contents of the present application are further explained below in conjunction with the drawings and examples of the specification.

[0117] Embodiment 1:

[0118] A multi-level risk identification method for distributed energy access to distribution networks, the specific process is as follows Figure 1 As shown, including:

[0119] Step 1, obtaining voltage data, power data and load data of the substation, line and main transformer;

[0120] Step 2, performing risk assessment based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line power flow over-limit risk assessment calculation formula to obtain a risk assessment result;

[0121] Step 3: Perform risk identification based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

[0122] Combine the following Figure 2 The process shown is used to describe this embodiment in detail.

[0123] Before step 1, it also includes collecting multi-level data related to the substation-line-main transformer from multiple data sources such as photovoltaic power and meteorology to construct sample data, and establish node voltage over-limit risk assessment calculation formulas and line power flow over-limit risk assessment calculation formulas, specifically including:

[0124] Based on the online simulation model of the carrying capacity of the 10kV 111 line in a certain place - photovoltaic access to the whole county, the Monte Carlo simulation method is used to simulate and generate distribution network load scenarios and photovoltaic output scenarios. Each generated scenario contains a 24-hour distribution network load sequence and photovoltaic output sequence to obtain initial sample data.

[0125] The output of solar cells is affected by many factors, such as lighting conditions, geographical conditions, panel surface temperature, panel structure, load conditions, etc. Therefore, the output of photovoltaic power generation has great uncertainty, and the grid connection of photovoltaic power generation will have a great impact on the stable operation of the power grid. The photoelectric conversion efficiency η is generally used to characterize the ability of solar energy to convert into electrical energy.

[0126]

[0127] Generally speaking, for photovoltaic power generation, the most important factor affecting the output power of solar cells is the lighting conditions. A large amount of data research shows that the intensity of sunlight conforms to the Beta distribution over a period of time.

[0128] Beta distribution, also known as B distribution, is a set of continuous probability density distributions defined in the interval (0,1). Its mathematical expression is:

[0129]

[0130] Where α is the first shape parameter of the Beta distribution; β is the second shape parameter of the Beta distribution, and s is an arbitrary variable. This variable is used in the following practical application. Γ() is the gamma function, that is Since the expectation μ and variance σ of the Beta distribution 2 The expression is:

[0131]

[0132] Combining the expectation and variance expressions, we can find the two parameters of the Beta distribution based on the expectation and variance of the historical observation data of the random variable:

[0133]

[0134] In practical applications, the probability properties of sunlight intensity can be described by the probability density function of the Beta distribution. Therefore, the probability model of photovoltaic power generation at a certain moment can be expressed as:

[0135]

[0136] in, is the probability of photovoltaic power generation when the current light intensity is γ; γ is the light intensity; γ max For maximum illumination.

[0137] For a photovoltaic field containing N solar arrays, the photoelectric conversion efficiency of each array is the same as η O , then the total output power and total maximum output power of the optical field can be expressed as:

[0138] P=γAη O

[0139] R = γ max Aη O

[0140] Where P is the actual output power of the photovoltaic power plant; γ is the light intensity; A is the total area of ​​the photovoltaic array; η O is the photoelectric conversion efficiency; max is the maximum light intensity; R is the maximum output power of the photovoltaic power plant.

[0141] Therefore, the probability model of the photovoltaic power plant is:

[0142]

[0143] Where f(P) is the probability corresponding to the actual output power of the photovoltaic power plant being P.

[0144] Its cumulative probability distribution function is:

[0145]

[0146] Where F(P) is the cumulative probability distribution within a certain period of time; f(t) is the probability at a certain moment.

[0147] The load is generally obtained through prediction, which has certain uncertainties. The load size in a certain period of time is not a constant value. It is affected by many factors and presents certain rules. For example, climate factors, social factors, holidays, etc. are all factors that affect load changes. A large number of studies have shown that the load of the power system approximately obeys the normal distribution, and the probability density of its active load and reactive load are:

[0148]

[0149] Where g(P) is the probability density of active load; σ P is the variance of active power; μ P is the expected value of active power; g(Q) is the probability density of reactive load; σ Qis the variance of reactive power; μ Q is the expected value of reactive power; Q is the reactive load.

[0150] According to the load and photovoltaic probability distribution, n samples are randomly selected to obtain a large number of random number samples that conform to this distribution;

[0151] The initial sample data under the same simulation is normalized to obtain the processed sample data;

[0152] The node voltage over-limit risk assessment calculation formula is as follows:

[0153]

[0154] In the formula, Risk v is the node voltage over-limit risk; i is the node order; Ω i is a node set; Prob(V i ) is the probability of node i voltage exceeding the limit; Sev(w i ) is the severity of overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i.

[0155] It is usually stipulated that when the voltage per unit value is between 0.95-1.05, it can be considered that the voltage is within the limit. The greater the voltage fluctuation, the greater the severity of the voltage risk. Therefore, the severity of voltage exceeding the limit is expressed by a risk preference utility function:

[0156]

[0157] In the formula, w i is the voltage loss value of node i; V i is the voltage per unit value of node i.

[0158] The line flow over-limit risk assessment calculation formula is as follows:

[0159] Risk s =∑Prob(S j )·Sev(S j )

[0160] In the formula, Risk s is the risk of line power flow exceeding the limit; Prob(S j ) is the probability of branch j crossing the limit; Sev(S j ) is the size of the deviation of branch j from the threshold.

[0161]

[0162] Where Sev(S j) is the deviation of branch j from the threshold; θ is the ratio of the actual transmission power to the maximum transmission power allowed by the line, and the maximum transmission power allowed by the line depends on the model and size of the line.

[0163] In step 1, the voltage data, power data and load data of the substation, line and main transformer are obtained, including:

[0164] Collect a set of voltage data, power data and load data of the substation-line-main transformer.

[0165] In step 2, risk assessment is performed based on the voltage data, power data and load data in combination with the node voltage over-limit risk assessment calculation formula and the line power flow over-limit risk assessment calculation formula to obtain a risk assessment result, which specifically includes:

[0166] Based on voltage data, power data and load data, the deterministic power flow calculation is used to obtain the values ​​of node voltage and line power flow;

[0167] The obtained node voltage and line power flow values ​​are statistically analyzed to obtain their probability distribution function and cumulative distribution function;

[0168] According to each cumulative distribution function combined with the node voltage over-limit risk assessment calculation formula and the line power flow over-limit risk assessment calculation formula, the node voltage over-limit risk value and the line power flow over-limit risk value are calculated to carry out risk identification.

[0169] In step 3, risk identification is performed based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result, which specifically includes:

[0170] Based on the risk assessment result, voltage over-limit risk data and branch circuit overload risk data are screened out from the voltage data, power data and load data;

[0171] Based on the voltage over-limit risk data and the branch overload risk data, risk identification is performed in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result;

[0172] The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data;

[0173] The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data. Figure 3 As shown, Hidden layer is a hidden layer; Input layer is an input layer; Output layer is an output layer.

[0174] The training process of the voltage risk identification model includes:

[0175] Input raw data, including voltage and power risk data, including constructed sample data;

[0176] The voltage over-limit risk data and branch overload risk data can be obtained from the sample data;

[0177] According to each risk data, 80% of the sample data are randomly selected to form key features, which constitute the training set to train the RVFL classifier, and the trained RVFL classifier is obtained, which specifically includes:

[0178] The processed voltage sample data is recorded as different feature vectors according to each node, the power sample data is recorded as different feature vectors according to each branch, and the load sample data is recorded as different feature vectors according to each load; each feature vector is set to a different dimension to obtain n different feature subsets;

[0179] Use the existing classification algorithms to obtain the feature subset with the highest test accuracy under each risk mode, and randomly select 80% to obtain the training set. The existing classification algorithms include support vector machine classification model, decision tree classification model, random forest classification model, random nearest neighbor classification model and neural network classification model;

[0180] Construct the input matrix H1 of the RVFL classifier according to the formula:

[0181]

[0182] In the formula, x 11 is the first feature vector of the first sample in the training set; x 1d is the dth feature vector of the first sample in the training set; x N1 is the first feature vector of the Nth sample in the training set; x nd is the d-th feature vector of the N-th sample in the training set.

[0183] Construct the hidden layer matrix H2 of the RVFL classifier according to the formula:

[0184]

[0185] Where ω1 is the first input weight; x1 is the first sample in the test set; b1 is the bias of the first hidden layer; g(·) is the activation function; ω j is the jth input weight; x N is the Nth sample in the test set; b j is the bias of the jth hidden layer.

[0186] Figure 3 x in IRepresented as the i-th sample among N samples.

[0187] Construct the output matrix H according to the formula:

[0188] H=H1H2

[0189] Where H1 is the input matrix; H2 is the hidden layer matrix.

[0190] Randomly initialize the input weights and the bias of the hidden layer, and calculate the output weights according to the formula:

[0191]

[0192] Where Ω is the output weight; is the Moore-Penrose generalized inverse of the output matrix H; H T is the transpose of matrix H; Y is the true label matrix.

[0193] The current input weight, the bias of the hidden layer and the output weight are used as the trained RVFL classifier parameters to obtain the trained RVFL classifier.

[0194] The remaining 20% ​​of relevant data is taken as test data to construct a test set, the test data is input into the trained RVFL classifier, and the accuracy of the test set test results is calculated;

[0195] The accuracy of each identification model using the test set is determined to see if the accuracy is greater than 98%. If so, a trained voltage risk identification model is obtained; otherwise, the key features are re-formed to construct the training set, and the RVFL classifier is continuously trained until the accuracy is greater than 98%.

[0196] The training process of the tidal risk identification model includes:

[0197] Input raw data, including voltage and power risk data, including constructed sample data;

[0198] The voltage over-limit risk data and branch overload risk data can be obtained from the sample data;

[0199] According to each risk data, 80% of the sample data are randomly selected to form key features, which constitute the training set to train the RVFL classifier, and the trained RVFL classifier is obtained, which specifically includes:

[0200] The processed voltage sample data is recorded as different feature vectors according to each node, the power sample data is recorded as different feature vectors according to each branch, and the load sample data is recorded as different feature vectors according to each load; each feature vector is set to a different dimension to obtain n different feature subsets;

[0201] Use the existing classification algorithms to obtain the feature subset with the highest test accuracy under each risk mode, and randomly select 80% to obtain the training set. The existing classification algorithms include support vector machine classification model, decision tree classification model, random forest classification model, random nearest neighbor classification model and neural network classification model;

[0202] Construct the input matrix H1 of the RVFL classifier according to the formula:

[0203]

[0204] In the formula, x 11 is the first feature vector of the first sample in the training set; x 1d is the dth feature vector of the first sample in the training set; x N1 is the first feature vector of the Nth sample in the training set; x Nd is the d-th feature vector of the N-th sample in the training set.

[0205] Construct the hidden layer matrix H2 of the RVFL classifier according to the formula:

[0206]

[0207] Where ω1 is the first input weight; x1 is the first sample in the test set; b1 is the bias of the first hidden layer; g(·) is the activation function; ω j is the jth input weight; x N is the Nth sample in the test set; b j is the bias of the jth hidden layer.

[0208] Construct the output matrix H according to the formula:

[0209] H=H1H2

[0210] Where H1 is the input matrix; H2 is the hidden layer matrix.

[0211] Randomly initialize the input weights and the bias of the hidden layer, and calculate the output weights according to the formula:

[0212]

[0213] Where Ω is the output weight; is the Moore-Penrose generalized inverse of the output matrix H; H T is the transpose of matrix H; Y is the true label matrix.

[0214] The current input weight, the bias of the hidden layer and the output weight are used as the trained RVFL classifier parameters to obtain the trained RVFL classifier.

[0215] The remaining 20% ​​of relevant data is taken as test data to construct a test set, the test data is input into the trained RVFL classifier, and the accuracy of the test set test results is calculated;

[0216] The accuracy of each identification model using the test set is determined to see if the accuracy is greater than 98%. If so, a trained tidal risk identification model is obtained; otherwise, the key features are re-formed to construct the training set, and the RVFL classifier is continuously trained until the accuracy is greater than 98%.

[0217] Construct a multi-level risk assessment index system for distribution network substations, lines, and transformers at multiple time scales within a day, and give the system risk assessment results based on the weight factors determined by theoretical research; identify abnormal situations and change trends based on data analysis, use the RVFL classifier to train a risk classifier, form key feature quantities, and use the trained model to classify and identify the acquired data samples, and give the risk type identification results including branch overload, voltage over-limit, etc.; This method has high recognition accuracy and can be applied to a variety of risk assessments and identifications. Take the online simulation diagram of the carrying capacity of the 10kV 111 line-whole county photovoltaic access distribution network as an example.

[0218] In order to facilitate calculation and analysis, the original 10 / 0.4kV transformer in the system is simplified into one line, and the entire system can be simplified to the same voltage level. The system has 83 nodes and 80 lines. Figure 4 shown.

[0219] (1) An example of risk assessment of the distribution network under multiple time scales within a day: 10kV 111 line in a certain place - the entire county distribution network (substation);

[0220] A total of 45 nodes in the substation area are connected to photovoltaic power sources. The voltage over-limit risk value of each node is obtained according to the node voltage over-limit risk assessment calculation formula, such as Figure 5 As shown in the figure. Since multiple nodes on the substation are connected to photovoltaic power, not only the voltage over-limit risk of the substation nodes is increased, but also the voltage over-limit risk of the nodes on the main line is increased. In addition, due to the access of photovoltaic power sources, active power is directly provided to nearby loads, which will reduce the power transmitted on the line and reduce voltage loss, thereby increasing the node voltage and increasing the voltage over-limit risk.

[0221] In order to verify the correctness of the Monte Carlo algorithm, Figure 6 The voltage per unit value of the node 70 is shown to be between 0.93 and 1.15, which exceeds the standard between 0.95 and 1.05, and a voltage over-limit occurs. Figure 7The voltage per unit value of node 40 is between 0.98 and 1.04, and its cumulative probability distribution curve reaches 1 after 1.05, and the voltage does not exceed the limit. The above reason is that the nodes near the substation are connected to the distributed power supply, which increases the voltage. However, the voltage of node 40 far away from the substation connected to the distributed power supply does not exceed the limit.

[0222] In the power saving of the substation, a total of 45 nodes are connected to the photovoltaic power source. The power flow over-limit risk value of each branch is obtained according to the line power flow over-limit risk assessment calculation formula, such as Figure 8 As shown in the figure, since multiple nodes on the platform are connected to photovoltaic power, the risk of over-limit power flow on the main line is increased, and since they are associated with or relatively close to the nodes connected to distributed power sources, the reverse transmission of distributed power output power will cause the risk of over-limit power flow.

[0223] In order to verify the correctness of the Monte Carlo algorithm, the probability distribution of branch power is obtained: Fig. 9 The power fluctuation range of branch 10 is shown as -2.5-0.98, which exceeds the limit value and will bring the risk of overcurrent exceeding the limit to the branch; Fig.10 The power fluctuation range of branch 40 is shown as -0.3-0.11, which is within the limit because it is far away from the node where the distributed power source is connected.

[0224] (2) An example of risk assessment of multiple levels of distribution network, including substation, line and transformer, at multiple time scales within a day - 10kV line 111 in a certain place - the entire county distribution network (main line);

[0225] There are 24 nodes on the main line, 10 of which are connected to photovoltaic power sources. According to the established risk assessment index system, the voltage over-limit risk value of each node is obtained by the node voltage over-limit risk assessment calculation formula, as shown in Fig.11 As shown in the figure. Since multiple nodes on the main line are connected to photovoltaic power, not only the voltage over-limit risk of the main line nodes is increased, but also the voltage over-limit risk of nodes in some substations is increased. In addition, due to the access of photovoltaic power sources, active power is directly provided to nearby loads, which will reduce the power transmitted on the line and the voltage loss, thereby increasing the node voltage and increasing the voltage over-limit risk.

[0226] In order to verify the correctness of the Monte Carlo algorithm, Fig.12 The voltage per unit value of the node 20 is shown to be between 0.93 and 1.09, which exceeds the standard between 0.95 and 1.05, and a voltage over-limit occurs. Fig.13 The voltage per unit value of node 40 is between 0.98-0.994, and its cumulative probability distribution curve reaches 1 after 1.05, and the voltage does not exceed the limit. The above reason is that the distributed power supply is connected to the node near the main line, which increases the voltage. However, the voltage of node 40 far away from the main line connected to the distributed power supply does not exceed the limit.

[0227] There are 24 nodes on the main line, 10 of which are connected to photovoltaic power sources. According to the established risk assessment index system, the risk value of the power flow over-limit of each branch is obtained by the line power flow over-limit risk assessment calculation formula, as follows: Fig.14 As shown in the figure, since multiple nodes on the main line are connected to photovoltaic power, the risk of power flow crossing the line on the main line is increased, and since they are associated with or relatively close to the nodes connected to the distributed power generation, the reverse transmission of the output power of the distributed power generation will cause the risk of power flow crossing the limit.

[0228] In order to verify the correctness of the Monte Carlo algorithm, the probability distribution of branch power is obtained: Fig.15 The power fluctuation range of branch 5 is shown as -2.4-0.97, which exceeds the limit and will bring the risk of overcurrent exceeding the limit to the branch; Fig.16 The power fluctuation range of branch No. 40 is shown as 0.01-0.05, which is within the limit. The reason is that it is far away from the node where the distributed power source is connected.

[0229] (3) Evaluation example of multi-level risk of distribution network downtime at multiple time scales within a day - 10kV 111 line in a certain place - whole county distribution network (main transformer)

[0230] A total of 25 nodes near the main transformer are connected to the photovoltaic power source. According to the established risk assessment index system, the voltage over-limit risk value of each node is obtained by the node voltage over-limit risk assessment calculation formula, such as Fig.17 shown.

[0231] As multiple nodes are connected to photovoltaic power, the risk of node voltage exceeding the limit increases. The access of photovoltaic power directly provides active power to nearby loads, which will reduce the power transmitted on the line and reduce voltage loss, thereby increasing the node voltage and increasing the risk of voltage exceeding the limit.

[0232] In order to verify the correctness of the Monte Carlo algorithm, Fig.18 The voltage per unit value of the node 70 is shown to be between 0.935 and 0.975, which exceeds the standard between 0.95 and 1.05, and a voltage over-limit occurs. Fig.19 The voltage per unit value of node 12 is between 0.96 and 0.99, and its cumulative probability distribution curve reaches 1 after 1.05, and the voltage does not exceed the limit. The above reason is that the node near the main transformer is connected to the distributed power supply, which increases the voltage. However, the voltage of node 12, which is far away from the main transformer connected to the distributed power supply, does not exceed the limit.

[0233] According to the established risk assessment index system, the risk value of the power flow exceeding the limit of each branch is obtained by the line power flow exceeding the limit risk assessment calculation formula, such as Fig. 20As shown in the figure, the risk of branch overload is greatest in branches 2 and 3. This is because these branches are associated with or relatively close to the nodes where distributed generation is connected, and the risk of branch power flow exceeding the limit is generated due to the reverse transmission of distributed generation output power. Branches farther from the distributed generation access node have a lower risk of power flow exceeding the limit.

[0234] In order to verify the correctness of the Monte Carlo algorithm, the probability distribution of branch power is obtained: Fig.21 The power fluctuation range of branch 35 is shown as -1.8-0.4, which exceeds the limit value and will bring the risk of overcurrent exceeding the limit to the branch; Fig. 22 The power fluctuation range of branch 12 is shown as 0.32-0.63, which is within the limit. The reason is that it is far away from the node where the distributed power source is connected.

[0235] (4) An example of risk assessment of the distribution network under multiple time scales within a day: a 10kV distribution network with 111 lines in a certain place and a whole county (multi-level distribution network with station, line and transformer);

[0236] A total of 30 nodes on the substation are connected to photovoltaic power sources; a total of 10 nodes on the main line are connected to photovoltaic power sources; a total of 15 nodes near the main transformer are connected to photovoltaic power sources. According to the established risk assessment index system, the voltage over-limit risk value of each node is obtained by the node voltage over-limit risk assessment calculation formula, such as Fig.23 shown.

[0237] In order to verify the correctness of the Monte Carlo algorithm, Fig.24 The voltage per unit value of the node 70 is shown to be between 0.9-1.18, which exceeds the standard between 0.95-1.05, and a voltage over-limit occurs. Fig.25 The voltage per unit value of the node 40 shown is between 0.97 and 1.04, and its cumulative probability distribution curve reaches 1 after 1.05, and the voltage does not exceed the limit.

[0238] A total of 30 nodes on the substation are connected to photovoltaic power sources; a total of 10 nodes on the main line are connected to photovoltaic power sources; a total of 15 nodes near the main transformer are connected to photovoltaic power sources. According to the established risk assessment index system, the voltage over-limit risk value of each node is obtained by the line flow over-limit risk assessment calculation formula, such as Fig.26 shown.

[0239] In order to verify the correctness of the Monte Carlo algorithm, the probability distribution of branch power is obtained: Fig. 27 The power fluctuation range of branch 10 is shown as -2.98-1, which exceeds the limit value and will bring the risk of overcurrent exceeding the limit to the branch; Fig.28 The power fluctuation range of branch 40 is shown as -0.3-0.07, which is within the limit.

[0240] Embodiment 2:

[0241] A multi-level risk identification system for distributed energy access to distribution networks, with a structure such as Fig.29 As shown, including:

[0242] Data acquisition module, used to obtain voltage data, power data and load data of the substation, line and main transformer;

[0243] A risk assessment module, used to perform risk assessment based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line power flow over-limit risk assessment calculation formula to obtain a risk assessment result;

[0244] The risk identification module is used to perform risk identification based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

[0245] Risk assessment module, specifically for:

[0246] Performing deterministic power flow calculation based on the voltage data, power data and load data to obtain the values ​​of each node voltage and each line power flow;

[0247] Based on the values ​​of the voltages of each node and the power flows of each line, statistics are performed to obtain the probability distribution function and cumulative distribution function of the node and the probability distribution function and cumulative distribution function of the power flows of the line;

[0248] Based on the cumulative distribution function of the node and the node voltage over-limit risk assessment calculation formula, the node voltage over-limit risk value is obtained;

[0249] The cumulative distribution function of the line flow is combined with the line flow over-limit risk assessment calculation formula to obtain the line flow over-limit risk value.

[0250] The node voltage over-limit risk assessment calculation formula in the risk assessment module is as follows:

[0251]

[0252] In the formula, Risk v is the node voltage over-limit risk; i is the node order; Ω i is a node set; Prob(V i ) is the probability of node i voltage exceeding the limit; Sev(w i ) is the severity of overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i.

[0253] The line flow over-limit risk assessment calculation formula in the risk assessment module is as follows:

[0254] Pisk s =∑Prob(S j )·Sev(S j )

[0255] In the formula, Pisk s is the risk value of line power flow exceeding the limit; Prob(S j ) is the crossing probability of branch j; Sev(S j ) is the size of the deviation of branch j from the threshold.

[0256] The risk identification module is specifically used for:

[0257] A data screening submodule, for screening out voltage over-limit risk data and branch overload risk data from the voltage data, power data and load data based on the risk assessment result;

[0258] A model calculation submodule, used to perform risk identification based on the voltage over-limit risk data and the branch overload risk data in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result;

[0259] The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data;

[0260] The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data.

[0261] The risk identification module also includes a voltage risk identification model training submodule; the voltage risk identification model training submodule is specifically used for:

[0262] Based on the voltage over-limit risk data, each feature subset is constructed according to each node being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0263] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a voltage risk identification model;

[0264] The voltage risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained voltage risk identification model is obtained; otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0265] The voltage risk identification model training submodule uses the training set to train the random vector function link classifier, determines the parameters of the random vector function link classifier to obtain the voltage risk identification model, including:

[0266] Constructing an input matrix of a random vector function link classifier based on the training set;

[0267] Constructing a hidden layer matrix based on the input matrix, randomly generated input weights and bias and activation functions of the hidden layer;

[0268] Constructing an output matrix based on the input matrix and the hidden layer matrix;

[0269] Calculate based on the output matrix and the true label matrix to obtain output weights;

[0270] Parameters of the classifier are linked as a random vector function based on the output weights, the randomly generated input weights, and the bias of the hidden layer.

[0271] The risk identification module also includes a tidal current risk identification model training submodule; the tidal current risk identification model training submodule is specifically used for:

[0272] Based on the branch overload risk number, each feature subset is constructed according to each branch being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm;

[0273] Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a power flow risk identification model;

[0274] The tidal current risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained tidal current risk identification model is obtained. Otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

[0275] Embodiment 3:

[0276] like Fig.30 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0277] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a multi-level risk identification method for distributed energy access to a distribution network in the above-mentioned embodiment.

[0278] Example 4

[0279] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a multi-level risk identification method for distributed energy access to a distribution network in the above embodiment.

[0280] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0281] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0282] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0283] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0284] The above are merely embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application are included in the scope of the claims of the present application to be approved.

Claims

1. A multi-level risk identification method for distributed energy access to distribution network, characterized in that: include: Obtain voltage data, power data and load data of substations, lines and main transformers; Based on the voltage data, power data and load data combined with the node voltage over-limit risk assessment calculation formula and the line power flow over-limit risk assessment calculation formula, a risk assessment is performed to obtain a risk assessment result; Risk identification is performed based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

2. The method according to claim 1, characterized in that The risk assessment is performed based on the voltage data, power data and load data in combination with the node voltage over-limit risk assessment calculation formula and the line power flow over-limit risk assessment calculation formula to obtain a risk assessment result, including: Performing deterministic power flow calculation based on the voltage data, power data and load data to obtain the values ​​of each node voltage and each line power flow; Based on the values ​​of the voltages of each node and the power flows of each line, statistics are performed to obtain the probability distribution function and cumulative distribution function of the node and the probability distribution function and cumulative distribution function of the power flows of the line; Based on the cumulative distribution function of the node and the node voltage over-limit risk assessment calculation formula, the node voltage over-limit risk value is obtained; The cumulative distribution function of the line flow is combined with the line flow over-limit risk assessment calculation formula to obtain the line flow over-limit risk value.

3. The method according to claim 2, characterized in that The node voltage over-limit risk assessment calculation formula is as follows: In the formula, Risk v is the node voltage over-limit risk value; Prob(V i ) is the voltage over-limit risk probability of node i; Sev(w i ) is the severity of overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i; Ω i is a set of nodes; i is the order of nodes.

4. The method according to claim 2, characterized in that: The line flow over-limit risk assessment calculation formula is as follows: Squeak s =∑Prob(S j )·Sev(S j ) In the formula, Pisk s is the risk value of line power flow exceeding the limit; Prob(S j ) is the crossing probability of branch j; Sev(S j ) is the size of the deviation of branch j from the threshold.

5. The method according to claim 1, characterized in that The risk identification is performed based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result, including: Based on the risk assessment result, voltage over-limit risk data and branch overload risk data are screened out from the voltage data, power data and load data; Based on the voltage over-limit risk data and the branch overload risk data, risk identification is performed in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result; The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data; The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data.

6. The method according to claim 5, characterized in that The training process of the voltage risk identification model includes: Based on the voltage over-limit risk data, each feature subset is constructed according to each node being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm; Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a voltage risk identification model; The voltage risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained voltage risk identification model is obtained; otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

7. The method according to claim 6, characterized in that The step of training the random vector function link classifier using the training set and determining the parameters of the random vector function link classifier to obtain a voltage risk identification model includes: Constructing an input matrix of a random vector function link classifier based on the training set; Constructing a hidden layer matrix based on the input matrix, randomly generated input weights and bias and activation functions of the hidden layer; Constructing an output matrix based on the input matrix and the hidden layer matrix; Calculate based on the output matrix and the true label matrix to obtain output weights; Parameters of the classifier are linked as a random vector function based on the output weights, the randomly generated input weights, and the bias of the hidden layer.

8. The method according to claim 5, characterized in that The training process of the tidal current risk identification model includes: Based on the branch overload risk number, each feature subset is constructed according to each branch being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm; Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a power flow risk identification model; The tidal current risk identification model is tested using the test set. If the accuracy is greater than a set threshold, a trained tidal current risk identification model is obtained. Otherwise, the training set and the test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

9. A multi-level risk identification system for distributed energy access to distribution networks, characterized in that: include: Data acquisition module, used to obtain voltage data, power data and load data of the substation, line and main transformer; A risk assessment module, used to perform risk assessment based on the voltage data, power data and load data in combination with a node voltage over-limit risk assessment calculation formula and a line power flow over-limit risk assessment calculation formula to obtain a risk assessment result; The risk identification module is used to perform risk identification based on the risk assessment result and the voltage data, power data and load data to obtain a risk type identification result.

10. The system according to claim 9, characterized in that The risk assessment module is specifically used for: Performing deterministic power flow calculation based on the voltage data, power data and load data to obtain the values ​​of each node voltage and each line power flow; Based on the values ​​of the voltages of each node and the power flows of each line, statistics are performed to obtain the probability distribution function and cumulative distribution function of the node and the probability distribution function and cumulative distribution function of the power flows of the line; Based on the cumulative distribution function of the node and the node voltage over-limit risk assessment calculation formula, the node voltage over-limit risk value is obtained; The cumulative distribution function of the line flow is combined with the line flow over-limit risk assessment calculation formula to obtain the line flow over-limit risk value.

11. The system according to claim 10, characterized in that The node voltage over-limit risk assessment calculation formula in the risk assessment module is as follows: In the formula, Risk v is the node voltage over-limit risk value; Prob(V i ) is the voltage over-limit risk probability of node i; Sev(w i ) is the severity of overvoltage at node i; V i is the voltage per unit value of node i; w i is the voltage over-limit loss value of node i; Ω i is a set of nodes; i is the order of nodes.

12. The method according to claim 10, characterized in that The line flow over-limit risk assessment calculation formula in the risk assessment module is as follows: Squeak s =∑Prob(S j )·Sev(S j ) In the formula, Pisk s is the risk value of line power flow exceeding the limit; Prob(S j ) is the crossing probability of branch j; Sev(S j ) is the size of the deviation of branch j from the threshold.

13. The system according to claim 9, characterized in that The risk identification module is specifically used for: A data screening submodule, for screening out voltage over-limit risk data and branch overload risk data from the voltage data, power data and load data based on the risk assessment result; A model calculation submodule, used to perform risk identification based on the voltage over-limit risk data and the branch overload risk data in combination with a pre-trained voltage risk identification model and a power flow risk identification model to obtain a risk type identification result; The voltage risk identification model is obtained by training a random vector function link classifier using voltage over-limit risk data; The power flow risk identification model is obtained by training a random vector function link classifier using branch overload risk data.

14. The system according to claim 13, characterized in that The risk identification module further includes a voltage risk identification model training submodule; the voltage risk identification model training submodule is specifically used for: Based on the voltage over-limit risk data, each feature subset is constructed according to each node being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm; Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a voltage risk identification model; The voltage risk identification model is tested using the test set, and if the accuracy is greater than a set threshold, a trained voltage risk identification model is obtained; Otherwise, the training set and test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

15. The system according to claim 14, characterized in that The voltage risk identification model training submodule uses the training set to train the random vector function link classifier, and determines the parameters of the random vector function link classifier to obtain the voltage risk identification model, including: Constructing an input matrix of a random vector function link classifier based on the training set; Constructing a hidden layer matrix based on the input matrix, randomly generated input weights and bias and activation functions of the hidden layer; Constructing an output matrix based on the input matrix and the hidden layer matrix; Calculate based on the output matrix and the true label matrix to obtain output weights; Parameters of the classifier are linked as a random vector function based on the output weights, the randomly generated input weights, and the bias of the hidden layer.

16. The system according to claim 13, characterized in that The risk identification module further includes a tidal current risk identification model training submodule; the tidal current risk identification model training submodule is specifically used for: Based on the branch overload risk number, each feature subset is constructed according to each branch being recorded as a different feature vector, and each feature subset is divided into a training set and a test set by a classification algorithm; Using the training set to train a random vector function link classifier, determining parameters of the random vector function link classifier to obtain a power flow risk identification model; The tidal current risk identification model is tested using the test set, and if the accuracy is greater than a set threshold, a trained tidal current risk identification model is obtained; Otherwise, the training set and test set are reconstructed by selecting feature vectors, and the random vector function link classifier is continuously trained until the accuracy is greater than the set threshold.

17. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a multi-level risk identification method for accessing a distributed energy source to a distribution network as described in any one of claims 1 to 8 is implemented.

18. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a multi-level risk identification method for the access of distributed energy to a distribution network as described in any one of claims 1 to 8 is implemented.

Images