High-proportion new energy power distribution network voltage stability risk assessment method, equipment and medium

Abstract

The invention relates to a high-proportion new energy power distribution network voltage stability risk assessment method and device and a medium, and the method comprises the following steps: S1, obtaining weather forecast information, and photovoltaic output data and load measurement data at the current moment, and predicting future photovoltaic output and load through an improved BILSTM algorithm; s2, according to the prediction result, the topological structure of the current power distribution network and a standby topological structure formed based on a potential tie line, carrying out load flow calculation by adopting a Newton-Raphson method, and obtaining voltage vectors of nodes of the power distribution network under the topological structures in the future moment; and S3, calculating a voltage stability risk L index of the power distribution network under each topological structure based on the voltage vector in combination with the admittance matrix of each node of the power distribution network, and judging whether a voltage instability risk exists or not. Compared with the prior art, the online voltage stability risk under the high-proportion new energy power distribution network can be accurately evaluated, and early warning and guidance are provided for operation of a power system in real time.

Description

Technical Field

[0001] The present invention belongs to the field of power system operation risk assessment, and in particular relates to a method, device and medium for assessing voltage stability risk of a high-proportion new energy distribution network. Background Art

[0002] With the gradual popularization of renewable energy in the distribution network, the focus of power system operation has gradually shifted from the relatively stable "grid-load" problem to the balance problem under the "source-grid-load" state. The fluctuation of renewable energy will bring greater operation risks to the operation of the distribution network. With the trend of large-scale integration of renewable energy, the frequency and amplitude of voltage vector fluctuations in the operation state of the power system have increased sharply. In the report on the definition and classification of power system stability, IEEE / CIGRE believes that the distinction between power angle stability and voltage stability is not based on the weak coupling relationship between active power / power angle and reactive power / voltage amplitude. Both power angle stability and voltage stability are affected by the active and reactive power flows before the disturbance. Therefore, studying the voltage stability risk caused by the source-grid-load relationship is an important means to provide early warning of operation risks for the distribution network.

[0003] Since voltage stability itself is dynamic, the study of the mechanism often starts from the dynamics and solves the differential equations of the system. However, the differential equation method has the disadvantages of complex calculations and time-consuming iterations, and the difference between the offline dynamic model and the actual distribution network state makes this method difficult to use in practice, especially in real-time online application. Therefore, in the actual offline voltage stability analysis or online operation stability warning, the continuous power flow method is usually used to analyze the voltage stability of the power system. The power flow is continuously calculated by increasing the load until the power flow divergence point. At this moment, the distribution network voltage is unstable and enters the voltage collapse state. This method is simple to calculate and is relatively accurate based on the power flow results. However, in the operation of a large distribution network, it is difficult to enumerate the load growth of each node, and only offline simulation calculations can be performed in advance to form a plan. Due to the uncertainty of the direction of load increase, the number of nodes selected for load growth, the nodes of growth, and the growth rate will affect the calculated risk assessment results to varying degrees. Therefore, it is necessary to design a method to accurately evaluate the online voltage stability risk under a high proportion of new energy distribution network, and timely warn the operation risk of a high proportion of new energy distribution network. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method, device and medium for voltage stability risk assessment of a high-proportion renewable energy distribution network, which can accurately assess the online voltage stability risk under a high-proportion renewable energy distribution network and provide early warning and guidance for the operation of the power system.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] The present invention provides a method for assessing voltage stability risk of a high-proportion new energy distribution network, comprising the following steps:

[0007] S1. Obtaining weather forecast information, photovoltaic output data and load measurement data at the current moment, and predicting future photovoltaic output and load by using an improved BILSTM algorithm, wherein the photovoltaic output data and load measurement data at the current moment are used to correct the prediction result;

[0008] S2, according to the prediction result of step S1, as well as the topological structure of the current distribution network and the backup topological structure formed based on the potential tie line, the Newton-Raphson method is used to perform power flow calculation to obtain the voltage vector of each node of the distribution network under each topological structure at the future moment;

[0009] S3. Based on the voltage vector obtained in step S2 and in combination with the admittance matrix of each node in the distribution network, the voltage stability risk L index of the distribution network under each topology is calculated and compared with the set threshold to determine whether there is a voltage instability risk. If so, a distribution network topology optimization suggestion is generated.

[0010] Furthermore, in the improved BILSTM algorithm, the input data in the time window before and after the prediction moment is standardized and used as an input vector.

[0011] Furthermore, when predicting photovoltaic output, the input data includes the solar irradiance, temperature, humidity at the prediction time and the photovoltaic output data at the current time; when predicting load, the input data includes the temperature, humidity, rainfall, date, specific time at the prediction time and the load measurement data at the current time.

[0012] Furthermore, the calculation process of the voltage stability risk L index of the distribution network under a certain topology is as follows:

[0013] S301, calculate the local voltage stability risk index of each load node in the distribution network under the topological structure, where the local voltage stability risk index L of load node i is i It is the modulus of the ratio of the superposition of the injected current of all load nodes and the product of the corresponding self-impedance or mutual impedance and the complex voltage of load node i;

[0014] S302: Taking the maximum value of the local voltage stability risk indexes of all load nodes as the voltage stability risk L index under the topology structure.

[0015] Furthermore, the local voltage stability risk index L of load node i is i The calculation formula is as follows:

[0016]

[0017] Among them, N L is the number of load nodes, Z ik is the self-impedance or mutual impedance of the load node, I k is the injected current of the load node, U L,i is the complex voltage at load node i.

[0018] Furthermore, the complex voltage U at load node i L,i The specific expression is as follows:

[0019]

[0020] Among them, N G is the number of generators, F ij is the load participation factor, is the load participation factor matrix F LG The element in row i and column j, U G,j is the complex voltage of the jth load node.

[0021] Furthermore, the load participation factor matrix F LG The specific expression is as follows:

[0022] F LL =-Z LL Y LG

[0023] Among them, Z LL is the load node impedance matrix, Y LG is the block submatrix of the admittance matrix.

[0024] Furthermore, the load node impedance matrix Z LL The specific expression is as follows:

[0025] Z LL =Y LL -1

[0026] Among them, Y LL is the load node admittance matrix.

[0027] The present invention also provides an electronic device, comprising a memory, a processor, and a program stored in the memory, wherein the processor implements the above method when executing the program.

[0028] The present invention also provides a computer-readable storage medium on which a computer program is stored, and the program implements the above method when executed by a processor.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The present invention proposes a voltage stability risk assessment method for a high-proportion new energy distribution network. First, meteorological forecast information, photovoltaic output data and load measurement data at the current moment are obtained, and the future photovoltaic output and load are predicted by an improved BILSTM algorithm. In the improved BILSTM algorithm, the present invention uses the photovoltaic output data and load measurement data at the current moment to correct the "source-load" prediction result, thereby improving the accuracy of the prediction result; then, according to the prediction result, as well as the topological structure of the current distribution network and the backup topological structure formed based on the potential tie line, the Newton-Raphson method is used to perform power flow calculation to obtain the distribution network under each topological structure at the future moment. The voltage vector of the node; finally, based on the obtained voltage vector and combined with the admittance matrix of each node in the distribution network, the voltage stability risk L index of the distribution network under each topological structure is calculated and compared with the set threshold value to determine whether there is a risk of voltage instability. The above method can avoid the problem that the risk assessment results of the existing voltage stability analysis method, the continuous power flow method, have different reliability due to different load growth modes. The voltage stability risk L index can ensure that for any load node, its local voltage stability index increases monotonically with the load, which greatly improves the reliability of the assessment result. At the same time, if there is a risk, the present invention can generate a distribution network topology optimization suggestion to provide guidance for the stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a flow chart of the method of the present invention;

[0032] Figure 2 This is the BILSTM framework diagram.

[0033] Among them, X t-1 ,X t ,X t+1 Represents the input data at each moment. LSTM F and LSTM B Represent the hidden layer states of the forward LSTM layer and the backward LSTM network layer, O t-1 ,O t ,O t+1 Represents the output data at each moment, ω1, ω2, …, ω6 represent the weight transfer between different layers;

[0034] Figure 3 This is a schematic diagram of the BILSTM time window;

[0035] Figure 4 This is the flow chart of voltage stability risk assessment under “source-grid-load” estimation;

[0036] Figure 5 It is the IEEE33 node diagram;

[0037] Figure 6This is a comparison chart of photovoltaic output prediction;

[0038] Figure 7 This is a load forecast comparison chart;

[0039] Figure 8 It is the estimated voltage amplitude diagram of 22 nodes in the distribution network;

[0040] Fig. 9 is the estimated voltage phase angle diagram of 22 nodes in the distribution network;

[0041] Fig.10 It is the analysis diagram of the estimated distribution network voltage vector R2;

[0042] Fig.11 It is the estimated L index diagram of 29-node distribution network;

[0043] Fig.12 It is the L index contour and profile of the distribution network;

[0044] Fig.13 L index heat map of different topological structures at 120 moments. DETAILED DESCRIPTION

[0045] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0046] Example 1

[0047] This embodiment provides a method for assessing voltage stability risk in a high-proportion renewable energy distribution network. Figure 1 As shown, the following steps are included:

[0048] S1. Obtain weather forecast information, real-time photovoltaic output data and load measurement data, and predict future photovoltaic output and load through the improved BILSTM algorithm.

[0049] Weather forecast information is obtained from the weather forecast department, and real-time photovoltaic output data and load measurement data are obtained from the SCADA system, which are updated synchronously with the data provided by the data provider. During training and application, the improved BILSTM algorithm not only considers the weather and time factors in other existing photovoltaic output prediction algorithms and load prediction algorithms, but also considers the actual photovoltaic output and load values ​​at this moment, so as to correct the prediction based only on weather information and further improve the accuracy of the prediction.

[0050] BILSTM is an optimization and improvement of unidirectional LSTM. BILSM combines a forward LSTM layer and a backward LSTM layer, and uses LSTM layers in two directions to form output. BILSTM can better utilize historical data of the distribution network, fully consider forward and backward information, and can better predict and improve prediction accuracy. The typical BILSTM structure diagram is as follows: Figure 2 As shown, where X t-1 ,X t ,X t+1 Represents the data input at each moment. LSTM F and LSTM B Represent the hidden layer states of the corresponding forward LSTM layer and backward LSTM network layer respectively. t-1 ,O t ,O t+1 Represents the corresponding output data. ω1, ω2, …, ω6 represent the weight transfer between different layers. The update state of the forward LSTM, the backward hidden layer and the output state of the BILSTM are shown in equations (1) to (3):

[0051] LSTM F,i =f1(ω1X i +ω4LSTM F,i-1 ) (1)

[0052] LSTM B,i =f1(ω2X i +ω3LSTM B,i-1 ) (2)

[0053] O i =f1(ω5LSTM F,i +ω6LSTM B,i ) (3)

[0054] Where f1, f2, f3 are activation functions between different layers, and i is the time corresponding to the output response.

[0055] In the present invention, the input for training and application of BILTM takes the form of a time window, such as Figure 3 As shown in the figure, the standardized data in the time window before and after the prediction moment is taken as the input vector, and the time window will gradually move forward as time goes by, so as to give full play to the characteristics of BILISM.

[0056] Table 1 is a Pearson correlation analysis table of the input vector for photovoltaic output prediction. When the present invention predicts photovoltaic output, the input is the solar irradiance G and the temperature T at the prediction time. a , humidity RH and the current photovoltaic output data M PV , the output is the output P of the corresponding photovoltaic power station at this momentPV,i ,Right now:

[0057] Input PV = {G,T a ,RH,M PV} (4)

[0058] Output PV,i = {P PV,i} (5)

[0059] Table 1 Pearson correlation analysis of PV output prediction input vector

[0060]

[0061]

[0062] Load forecasting is similar to photovoltaic output forecasting and also has a strong correlation with weather factors. However, since power load is closely related to people's living habits, it has a strong correlation in time. While load forecasting relies on meteorological information, it has obvious periodicity and characteristics, the influence of seasonal changes, temperature, factory work and the changing laws of residents' lives. By utilizing the mechanism of BILSTM to connect the upper and lower data in the time series, it can strengthen the impact of special dates and seasonal changes in periodic load forecasting and improve the accuracy of load forecasting. Table 2 is a Pearson coefficient correlation analysis table corresponding to the input of load forecasting. When the present invention predicts the load, the input is the temperature T at the time of prediction. a , humidity RH, rainfall W S , date Day, specific time Hour and current time load measurement data M load , the output is the output of the corresponding photovoltaic power station at this moment i ,Right now:

[0063] Input load ={T a ,RH,W S ,Day,Hour,M load} (6)

[0064] Output load,i = {Load i} (7)

[0065] Table 2 Pearson coefficient correlation analysis table corresponding to load forecast input

[0066]

[0067] S2. According to the prediction result of step S1, as well as the topological structure of the current distribution network and the backup topological structure formed based on the potential tie line, the Newton-Raphson method is used to perform power flow calculation to obtain the voltage vector of each node of the distribution network under each topological structure at the future moment.

[0068] S3. Based on the voltage vector obtained in step S2 and in combination with the admittance matrix of each node in the distribution network, the voltage stability risk L index of the distribution network under each topology is calculated and compared with the set threshold to determine whether there is a voltage instability risk. If so, a distribution network topology optimization suggestion is generated.

[0069] The voltage equation of the power system network node established by KVL is:

[0070]

[0071] Among them, I G and I L are the injected currents of the generator and load nodes respectively, U G , U L and U N are the voltages of the generator, load node and interconnection node respectively, Y GG , Y GL , Y GN are block sub-matrices of the admittance matrix, and the G, L, and N in their subscripts represent the generator, load node, and connection node, respectively.

[0072] After eliminating the contact nodes in the network, the nodes in the system can be divided into two groups, one group includes all the generator nodes and the part of the PV nodes, and the other group includes all the other load nodes in the system. Equation (8) can be changed to:

[0073]

[0074] Define the load node impedance matrix Z LL =Y LL -1 And the load participation factor matrix F LG =-Z LL Y LG , we can get the equation of load node voltage:

[0075]

[0076] It can be seen from the above formula that the load node voltage can be regarded as the product superposition of the generator node voltage and the corresponding participation factor and the product superposition of the injected current of all load nodes and the corresponding self-impedance and mutual impedance. The local voltage stability index L of the load node u is defined as the modulus of the complex ratio of the latter, that is:

[0077]

[0078] Where, L i is the local index of the i-th load node, i∈α L , U G,j is the complex voltage of the jth load node, j∈α G ; U L,i is the complex voltage of the ith load node, F ij is the load participation factor, is the load participation factor matrix F LG The ijth element of .

[0079] The local voltage stability index L of all load nodes in the system i Can constitute the system voltage stability index L sys =[L1L2 … L n ], define the system L index as L = max(L sys ). For the system L index, if L<1, the system voltage is stable; if L>1, the system voltage is unstable; if L=1, the system is critically stable. Usually, the threshold of the L index warning system is designed to be 0.4-0.6.

[0080] After the early warning is issued, the topology with lower voltage instability risk is recommended to the distribution network dispatching center. Figure 4 As shown, since the distribution network is usually designed in a closed-loop design and open-loop operation, the state of the tie switch can be adjusted to open and close the tie line to transfer power to the isolated area after the fault is isolated. By calculating the L index of the backup tie line under the N-1 fault of the distribution network line, on the one hand, it can provide an early warning when the backup topology cannot achieve voltage stability when the line N-1 fails in a certain operating state, and on the other hand, it can optimize the operation of the distribution network topology for the current distribution network or after the distribution network fault occurs. When the L index analysis of each topological structure calculated after the N-1 fault of a certain line is greater than or equal to 1, it means that the distribution network enters an emergency voltage instability state. When a fault occurs, the dispatch center cannot restore power to the faulty area by switching the topology of the switch tie line.

[0081] The above method first uses the improved BISLTM prediction method as a framework to predict the "source-load". This method not only considers factors such as meteorology, but also considers the measured value at the current moment for correction. On the basis of the predicted "source-load", different distribution network topologies are considered to estimate the distribution network voltage vector at the corresponding prediction moment. Then, based on the amplitude and phase angle of the estimated voltage vector combined with the corresponding distribution network line parameters, the voltage stability risk of the distribution network at the prediction moment is analyzed through the L index, which can realize the deterministic distribution network voltage stability risk assessment under the access of new energy such as photovoltaics.

[0082] Since the existing methods usually calculate the distribution network voltage stability risk assessment based on the continuous power flow method. This method has different nodes and different load increase modes, which will form more possible power flow modes. Usually, the continuous power flow method will adopt a single parameterized load growth method, that is, only the active and reactive power of a single node will be increased at the same ratio. Different PV curves and power limit points will be obtained under different load growth methods. If the load growth methods of all possible nodes are enumerated for analysis, it will seriously affect the timeliness of the voltage stability risk calculation, and may cause the evaluation results to be too extensive and lose validity. In the present invention, the load prediction value and the photovoltaic output prediction value are regarded as the load "true value" or photovoltaic output "true value" of each node in the distribution network at the corresponding prediction time, and the distribution network power flow calculation is performed on the basis of the "true value", the distribution network power flow mode at the corresponding prediction time is determined, and the voltage phasor estimation value of each node is calculated for voltage stability risk assessment. Therefore, the present invention uses the predicted load value and photovoltaic output value as the node active true value at the prediction time, which can effectively reduce the possibility of power flow mode and improve the credibility of voltage stability assessment.

[0083] To verify the effectiveness of the above method, this embodiment conducted the following simulation experiment: the load and photovoltaic output forecast data are from the Power Consumption of Tetouan City dataset. Since the sampling interval of the above two datasets is 10 minutes, the simulation interval selected in this embodiment is 10 minutes. The simulation verification uses the IEEE33 distribution network system for testing. The nodes and tie lines set in the photovoltaic station are Figure 5 shown.

[0084] Scenario 1: “Source-load” prediction. This scenario is used to test the prediction accuracy of the improved BILSTM algorithm proposed in the present invention for photovoltaic output and load prediction.

[0085] Figure 6 and Figure 7 They are comparison diagrams of the photovoltaic output prediction and load prediction by the method of the present invention. The BILSTM prediction value refers to the prediction value that uses only BILSTM as the prediction framework, and only meteorological factors are used in the photovoltaic output prediction; and only meteorological factors and time factors are used in the load prediction. The predictions at time 1 and time 24 adopt the improved BILSTM prediction method proposed by the present invention that adds the measured value at the current time to the prediction. Time 1 and time 24 respectively represent the prediction of the value at the current time and 24 moments after the value at the moment before the current time. As the time interval between the actual measured photovoltaic output value and the actual load value at the previous moment and the predicted moment increases, the correction effect of the actual measured value at the previous moment on the predicted value is further attenuated, but its prediction accuracy is still higher than the accuracy of the conventional photovoltaic output and load time series prediction that only uses meteorological factors.

[0086] Scenario 2: Estimation of distribution network voltage vector under “source-load” prediction. This scenario verifies the accuracy of the estimation of distribution network voltage vector under the “source-grid-load” prediction method proposed in the present invention.

[0087] Figure 8 and Fig. 9 This is a comparison diagram of the voltage amplitude and voltage phase angle of the estimated 22-node distribution network obtained through the "source-load" prediction calculation. Fig.10 It is the voltage amplitude and voltage phase angle estimation statistical R2 diagram of all nodes. The distribution network nodes are divided into four sections according to the branches by dotted lines: 1-18 nodes, 19-22 nodes, 23-25 ​​nodes, and 26-33 nodes. The R2 estimation accuracy of the voltage vector of all nodes in the distribution network is higher than 0.9. Especially the estimation in the branch section of nodes 19-22, because this branch is closer to the main generator node and there are photovoltaic equipment in it, it is relatively stable relative to other nodes, so that the estimated voltage vector estimated amplitude can be closer to the actual value. In general, the method of calculating the power flow through "source-load" prediction proposed in the present invention can better estimate the voltage vector state of the distribution network at future moments.

[0088] Scenario 3: Accuracy of the estimated voltage vector L index calculation. This scenario verifies the error between the L index calculated by the estimated voltage vector and the actual L index. Fig.11 This is a comparison chart of the actual and predicted L index of 29 nodes. For the simulation test system of this embodiment, the R2 value between the predicted L index value and the actual L index is 0.9051. The calculation result of the L index in the real distribution network is well restored, and the estimated voltage risk status of the distribution network can be well reflected.

[0089] Scenario 4: This simulation scenario verifies the accuracy of voltage stability risk assessment under the "source-grid-load" model. The voltage risk assessment method in the proposed method not only considers the operation mode of the existing distribution network, but also needs to comprehensively consider the voltage stability risk caused by the topological structure change when the distribution network has N-1 faults.

[0090] Fig.12 In the simulation system after adding photovoltaic nodes, all loads increase at a rate of 0.01% per simulation time. The L index contour map of the voltage collapse point is determined by calculating the singular point of the Jacobian matrix of the distribution network through the continuous power flow algorithm. The collapse point is located at the simulation time 412. Through the L index calculated by the estimated voltage vector under the "source-grid-load" prediction, the method proposed in the present invention can give an early warning before the L index threshold is reached in the future, and the dispatcher can make plans in advance to transfer loads, change topology, and adjust the power generation plan.

[0091] In the simulation test of voltage risk assessment of different topologies under N-1 fault, time 120 is selected. At this time, the load value of each load node increases to 1.2 times of the initial value without increasing the reactive load of the distribution network. Fig.13 The L index is obtained by analyzing the voltage risk under different topologies after N-1 faults occur. The value represents the maximum value of the L index of each node after calculation. The color block represents the voltage stability. The "load loss" color block means that in this case, switching the tie line after a branch fault will still cause load loss, so it is not calculated. The "unstable" color block means that in this case, switching to this topology after a branch fault will cause voltage collapse.

[0092] The method proposed in the present invention can calculate the voltage risk of the distribution network under the existing distribution network topology and N-1 fault in the future through the L index. While warning of the voltage collapse risk, it can also analyze the margin of the distribution network after the current potential fault occurs for the dispatching personnel, and assist in optimizing the topology selection and switching of the distribution network voltage stability after the fault occurs. Taking the distribution network status at 120 as an example, in the "source-load" case, when a line fault occurs between 3-4 nodes, all alternative topological solutions will cause the distribution network to enter a voltage collapse state, and when a fault occurs, it is necessary to enter the distribution network ring network operation state. For the line fault between 6-7 nodes at the same time, it is recommended to enable C or D connecting lines.

[0093] Example 2

[0094] This embodiment provides an electronic device, including a memory and a processor, wherein the processor is used to execute a program stored in the memory, wherein the program includes a number of instructions and can execute all or part of the steps of the method described in Example 1. The memory includes a computer-readable storage medium, which can be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0095] The above description of the embodiments is to facilitate the understanding and use of the invention by those skilled in the art. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative work. Therefore, the present invention is not limited to the above embodiments, and improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the present invention should be within the scope of protection of the present invention.

Claims

1. A method for assessing voltage stability risk in a high-proportion renewable energy distribution network, characterized in that: The following steps are involved: S1. Obtaining weather forecast information, photovoltaic output data and load measurement data at the current moment, and predicting future photovoltaic output and load by using an improved BILSTM algorithm, wherein the photovoltaic output data and load measurement data at the current moment are used to correct the prediction result; S2, according to the prediction result of step S1, as well as the topological structure of the current distribution network and the backup topological structure formed based on the potential tie line, the Newton-Raphson method is used to perform power flow calculation to obtain the voltage vector of each node of the distribution network under each topological structure at the future moment; S3. Based on the voltage vector obtained in step S2 and in combination with the admittance matrix of each node in the distribution network, the voltage stability risk L index of the distribution network under each topology is calculated and compared with the set threshold to determine whether there is a voltage instability risk. If so, a distribution network topology optimization suggestion is generated.

2. The method for assessing voltage stability risk of a high-proportion renewable energy distribution network according to claim 1, characterized in that: In the improved BILSTM algorithm, the input data in the time window before and after the prediction moment is standardized and used as an input vector.

3. The method for assessing voltage stability risk of a high-proportion renewable energy distribution network according to claim 2 is characterized in that: When predicting photovoltaic output, the input data includes the solar irradiance, temperature, humidity at the prediction time and the photovoltaic output data at the current time; when predicting load, the input data includes the temperature, humidity, rainfall, date, specific time at the prediction time and the load measurement data at the current time.

4. The method for assessing voltage stability risk of a high-proportion renewable energy distribution network according to claim 1 is characterized in that: The calculation process of the voltage stability risk L index of the distribution network under a certain topology is as follows: S301, calculate the local voltage stability risk index of each load node in the distribution network under the topological structure, where the local voltage stability risk index L of load node i is i It is the modulus of the ratio of the superposition of the injected current of all load nodes and the product of the corresponding self-impedance or mutual impedance and the complex voltage of load node i; S302: Taking the maximum value of the local voltage stability risk indexes of all load nodes as the voltage stability risk L index under the topology structure.

5. The method for assessing voltage stability risk of a high-proportion renewable energy distribution network according to claim 4 is characterized in that: The local voltage stability risk index L of load node i i The calculation formula is as follows: Among them, N L is the number of load nodes, Z ik is the self-impedance or mutual impedance of the load node, I k is the injected current of the load node, U L,i is the complex voltage at load node i.

6. A method for assessing voltage stability risk in a high-proportion renewable energy distribution network according to claim 5, characterized in that: Complex voltage U at load node i L,i The specific expression is as follows: Among them, N G is the number of generators, F ij is the load participation factor, is the load participation factor matrix F LG The element in row i and column j, U G,j is the complex voltage of the jth load node.

7. A method for assessing voltage stability risk in a high-proportion renewable energy distribution network according to claim 6, characterized in that: The load participation factor matrix F LG The specific expression is as follows: F LG =-Z LL Y LG Among them, Z LL is the load node impedance matrix, Y LG is the block submatrix of the admittance matrix.

8. The method for assessing voltage stability risk of a high-proportion renewable energy distribution network according to claim 7, characterized in that: The load node impedance matrix Z LL The specific expression is as follows: Z LL =Y LL -1 Among them, Y LL is the load node admittance matrix.

9. An electronic device comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images