Optimized scheduling method and system for photovoltaic source network load storage

Abstract

The invention discloses an optimal scheduling method and system for photovoltaic source network load storage. The method comprises the following steps: acquiring a power distribution network topological structure, photovoltaic output data and a feeder section switch state in each photovoltaic grid-connected scene; calculating the equivalent impedance of a photovoltaic grid-connected point according to the topological structure of the power distribution network and the section switch state of a feeder line to obtain impedance change characteristics; extracting output fluctuation characteristics according to the photovoltaic output data; calculating a correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scene so as to adjust a feeder section switch state and obtain an initial topology adjustment scheme; calculating transient characteristics of power grid nodes under the initial topology adjustment scheme, and performing iterative simulation on the initial topology adjustment scheme to obtain an optimized topology adjustment scheme; and carrying out optimal scheduling of photovoltaic source network load storage by adopting an optimal topology adjustment scheme. By adopting the embodiment of the invention, the consumption capability of the power distribution network on photovoltaic power generation can be effectively improved, and meanwhile, safe and stable operation of the power grid is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power dispatching, and in particular to a photovoltaic source-grid-load-storage optimization dispatching method and system. Background Art

[0002] Under different time periods, different load levels and different photovoltaic output conditions, there are multiple operating scenarios for the distribution network, each of which has its own specific network topology, feeder segment switch status and photovoltaic grid-connected characteristics.

[0003] There is a complex coupling relationship between the dynamic changes of the network topology and the fluctuation of photovoltaic output during the distribution network reconstruction process, and the operation behavior of the feeder segment switches will also affect the stability of the system. At present, the interaction mechanism and coupling relationship between the three have not been clearly explained. The traditional planning of photovoltaic source, grid, load and storage based only on experience is difficult to give full play to the power supply quality benefits of the distribution network. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a photovoltaic source-grid-load-storage optimization scheduling method and system, which can effectively improve the distribution network's ability to absorb photovoltaic power generation while ensuring the safe and stable operation of the power grid.

[0005] The embodiment of the present invention provides a photovoltaic source grid load storage optimization scheduling method, comprising: Obtain the distribution network topology, photovoltaic output data and feeder section switch status under each photovoltaic grid-connected scenario; According to the distribution network topology and the feeder segment switch state, the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario is calculated to obtain the impedance change characteristics of each photovoltaic grid-connected scenario; Extracting output fluctuation characteristics under each photovoltaic grid-connected scenario according to the photovoltaic output data; Calculating the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, adjusting the feeder segment switch state according to the correlation coefficient, and obtaining an initial topology adjustment scheme; Calculating transient characteristics of power grid nodes under the initial topology adjustment scheme, and iteratively simulating the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme; The optimized topology adjustment scheme is used to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario.

[0006] As an improvement of the above solution, the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario is calculated according to the distribution network topology and the feeder segment switch state, and the impedance change characteristics of each photovoltaic grid-connected scenario are obtained, including: Obtain the three-phase voltage and current values ​​of each feeder section switch and the feeder impedance parameters of the distribution line; Calculating the sequence impedance value of the distribution line according to the three-phase voltage and current values ​​and the distribution network topology; Calculating an equivalent impedance value of a photovoltaic grid-connected point according to the sequence impedance value and the feeder impedance parameter; According to the feeder segment switch state, calculating the change data of the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario; An exponential smoothing method is used to perform trend analysis on the change data of the equivalent impedance to obtain impedance change characteristics of each photovoltaic grid-connected scenario.

[0007] As an improvement of the above solution, the output fluctuation characteristics in each photovoltaic grid-connected scenario are extracted according to the photovoltaic output data, including: A sliding window method is used to reduce noise and process outliers on the photovoltaic output data to obtain photovoltaic output time series data; Wavelet transform is used to perform multi-scale decomposition on the photovoltaic output time series data, extract the fluctuation amplitude, frequency and duration, construct a photovoltaic output fluctuation feature matrix, and obtain the output fluctuation characteristics under each photovoltaic grid-connected scenario.

[0008] As an improvement of the above scheme, the calculation of the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics in each photovoltaic grid-connected scenario, and adjusting the feeder segment switch state according to the correlation coefficient to obtain an initial topology adjustment scheme include: The Pearson correlation coefficient method is used to calculate the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario; According to the correlation coefficient and a preset correlation coefficient threshold, the photovoltaic grid-connected scenario is divided into an impedance optimization scenario and a multi-objective optimization scenario; For the impedance optimization scenario, a genetic algorithm is used to find and calculate an impedance optimization target value, and the feeder segment switch state is adjusted according to the impedance optimization target to generate an initial topology adjustment plan; For the multi-objective optimization scenario, according to the preset optimization index, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0009] As an improvement of the above scheme, for the impedance optimization scenario, a genetic algorithm is used to find and calculate the impedance optimization target value, and the feeder segment switch state is adjusted according to the impedance optimization target to generate an initial topology adjustment scheme, including: Constructing an optimization objective function of the impedance optimization scenario; the optimization objective function includes a node impedance term and a power loss term; According to the optimization objective function, a genetic algorithm is used to find and calculate the impedance optimization target value; According to the impedance optimization target value, an algorithm based on graph theory is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0010] As an improvement of the above scheme, the optimization index includes a voltage stability margin index and a power balance index; then for the multi-objective optimization scenario, according to the preset optimization index, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment scheme, including: Obtaining a voltage curve and a power curve of the multi-objective optimization scenario; Calculating a voltage stability margin index according to the voltage curve; Calculating a power balance index according to the power curve; A fuzzy comprehensive evaluation algorithm is used to automatically calculate the weight values ​​of the voltage stability margin index and the power balance index to obtain a multi-objective optimization evaluation matrix; According to the correlation coefficient, the optimization target of the multi-objective optimization evaluation matrix is ​​obtained; According to the optimization target of the multi-objective optimization evaluation matrix, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0011] As an improvement of the above scheme, the calculating of the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulating the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme include: Acquiring voltage data and current data of power grid nodes under the initial topology adjustment scheme; According to the voltage data, calculating the transient power transmission characteristics between each grid node to obtain the transient stability risk; Calculating the line power distribution data of the initial topology adjustment scheme according to the voltage data and the current data to obtain a static safety margin; If the transient stability risk is greater than a preset stability risk threshold, a recursive neural network is used to predict the transient disturbance propagation law, and the transient response process caused by the short-circuit fault and the switching operation is calculated through time domain simulation to obtain transient response optimization data; If the static safety margin is less than the preset safety margin threshold, the power flow distribution of the distribution network is calculated based on the Newton iteration method, and the power transmission constraint and the voltage over-limit constraint are checked to obtain static constraint optimization data; According to the transient response optimization data and the static constraint optimization data, the initial topology adjustment scheme is iteratively simulated to obtain an optimized topology adjustment scheme.

[0012] As an improvement of the above solution, the transient power transmission characteristics between the grid nodes are calculated according to the voltage data to obtain the transient stability risk, including: Processing the voltage data by wavelet transform to obtain voltage fluctuation characteristics; According to the voltage data, a voltage phase angle variation trend before and after the impedance mutation under the initial topology adjustment scheme is obtained; Obtaining a node admittance matrix according to the state of the feeder section switch and the distribution network topology; According to the voltage phase angle variation trend and the node admittance matrix, the transient power transmission characteristics between the power grid nodes are calculated to obtain a transient stability margin curve; According to the transient stability margin curve, a recursive neural network is used to predict the transient response of key network nodes under impedance mutation, and the transient disturbance propagation path is calculated through the phase angle stability criterion to obtain the transient stability risk.

[0013] As an improvement of the above solution, the calculating of the line power distribution data of the initial topology adjustment solution according to the voltage data and the current data to obtain the static safety margin includes: Calculating line power distribution data of the initial topology adjustment solution according to the voltage data and the current data; The line overload level and node voltage over-limit degree are calculated through power flow sensitivity to obtain the static safety margin.

[0014] The embodiment of the present invention also provides a photovoltaic source grid load storage optimization scheduling system, including: Multi-scenario data acquisition module, used to obtain the distribution network topology, photovoltaic output data and feeder section switch status under various photovoltaic grid-connected scenarios; An impedance change characteristic calculation module is used to calculate the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario according to the distribution network topology and the feeder segment switch state, and obtain the impedance change characteristics of each photovoltaic grid-connected scenario; An output fluctuation characteristic calculation module, used to extract the output fluctuation characteristics under each photovoltaic grid-connected scenario according to the photovoltaic output data; An initial scheme generating module, used to calculate the correlation coefficient between the impedance variation characteristics and the output fluctuation characteristics in each photovoltaic grid-connected scenario, and adjust the feeder segment switch state according to the correlation coefficient to obtain an initial topology adjustment scheme; A scheme optimization module, used to calculate the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulate the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme; The optimization scheduling module is used to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario by adopting the optimization topology adjustment scheme.

[0015] Compared with the prior art, the present invention discloses a photovoltaic source grid load storage optimization scheduling method and system, by obtaining the distribution network topology structure, photovoltaic output data and feeder segment switch state in each photovoltaic grid-connected scenario; according to the distribution network topology structure and the feeder segment switch state, the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario is calculated to obtain the impedance change characteristics of each photovoltaic grid-connected scenario; according to the photovoltaic output data, the output fluctuation characteristics of each photovoltaic grid-connected scenario are extracted; the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics in each photovoltaic grid-connected scenario is calculated, and the feeder segment switch state is adjusted according to the correlation coefficient to obtain an initial topology adjustment scheme; the transient characteristics of the grid nodes under the initial topology adjustment scheme are calculated, and according to the transient characteristics, the initial topology adjustment scheme is iteratively simulated to obtain an optimized topology adjustment scheme; the optimized topology adjustment scheme is used to optimize the scheduling of photovoltaic source grid load storage in the photovoltaic grid-connected scenario. The embodiment of the present invention can effectively improve the distribution network's ability to absorb photovoltaic power generation, while ensuring the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the steps of the photovoltaic source grid load storage optimization scheduling method provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a photovoltaic source-grid-load-storage optimization scheduling system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In the description of the specification and claims, it is to be understood that the terms first, second, etc. in the specification and claims are only used for the purpose of describing the same technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated, nor necessarily describing the order or time sequence. The terms are interchangeable where appropriate. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features.

[0019] The embodiment of the present invention provides a method for optimizing the scheduling of photovoltaic power generation, grid-load and storage. Figure 1 In this embodiment, the photovoltaic source grid load storage optimization scheduling method is specifically performed through steps S1 to S6: S1. Obtain the distribution network topology, photovoltaic output data and feeder segment switch status under each photovoltaic grid-connected scenario.

[0020] It should be noted that, according to different time periods, load levels and photovoltaic output conditions, several photovoltaic grid-connected scenarios can be obtained, and at least one of the network topology, feeder segment switch status and photovoltaic grid-connected characteristics in each photovoltaic grid-connected scenario is different.

[0021] In some preferred embodiments, the distribution network topology includes the position status of feeder section switches, the distribution information of photovoltaic grid connection points, and the distribution information of distribution network nodes. The feeder section switch status includes two types of status information: closed / open status and normal / fault status.

[0022] It can be understood that, in the embodiment of the present invention, the contents included in the distribution network topology structure and the feeder section switch state are partially the same, because the feeder switch operation will inevitably cause the change of the distribution network topology structure.

[0023] S2. According to the distribution network topology and the feeder segment switch state, the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario is calculated to obtain the impedance change characteristics of each photovoltaic grid-connected scenario.

[0024] It should be noted that the opening and closing states of the feeder segment switches will cause changes in the topology of the distribution network, and will also cause changes in the load of the photovoltaic grid-connected points, thereby affecting the stability and controllability of the photovoltaic output. In the embodiment of the present invention, by calculating the equivalent impedance of the photovoltaic grid-connected points in each photovoltaic grid-connected scenario, the interaction mechanism and coupling relationship between the feeder segment switch operation and the network topology changes and photovoltaic output fluctuations are measured from the perspective of feeder segment switch operation.

[0025] S3. Extracting output fluctuation characteristics in each photovoltaic grid-connected scenario based on the photovoltaic output data.

[0026] It should be noted that there is a complex coupling relationship between the dynamic changes in the topology of the distribution network and the fluctuations in the photovoltaic output during the distribution network reconstruction process. The two influence and restrict each other. Changes in the topology of the distribution network will cause the redistribution of the power flow, resulting in fluctuations in the local node voltage and line power, which in turn affects the maximum power point tracking and reactive power regulation capabilities of the photovoltaic inverter, that is, causes fluctuations in the photovoltaic output. The fluctuations in photovoltaic output will in turn affect the network power flow and voltage level, exacerbating the system instability factors caused by changes in the network topology.

[0027] S4. Calculate the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, and adjust the feeder segment switch state according to the correlation coefficient to obtain an initial topology adjustment solution.

[0028] In an embodiment of the present invention, by calculating the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic, the differentiated impact degree of the distribution network topology adjustment on the photovoltaic absorption capacity can be expressed, and then based on the differentiated impact degree, different feeder segmented switch state operations are performed for different photovoltaic grid-connected scenarios. The obtained initial topology adjustment scheme takes into account the complex coupling between the grid topology structure and the photovoltaic output fluctuation.

[0029] S5. Calculate transient characteristics of power grid nodes under the initial topology adjustment scheme, and perform iterative simulation on the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme.

[0030] It should be noted that in different photovoltaic grid-connected scenarios, the evaluation indicators for the power grid may be different. Therefore, the transient characteristics of the power grid nodes are further analyzed to ensure that the final optimized topology adjustment solution can take into account safety and stability constraints and ensure the reliable operation of the power grid.

[0031] S6. Adopting the optimized topology adjustment scheme to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario.

[0032] In the above scheme, by calculating the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics, and then adjusting the feeder segment switch state, the complex coupling between the grid topology and the photovoltaic output fluctuation can be fully considered, and the transient characteristics of the grid nodes can be further analyzed to optimize the topology adjustment scheme, which can verify the feasibility and optimization effect of the scheme. The embodiment of the present invention can effectively improve the distribution network's ability to absorb photovoltaic power generation, while ensuring the safe and stable operation of the power grid.

[0033] In some preferred embodiments, the photovoltaic grid-connected scenario described in step S1 is obtained in advance based on different combinations of time periods, load levels and photovoltaic output conditions, and can be directly obtained. Exemplarily, the method for obtaining the photovoltaic grid-connected scenario includes: calculating a distribution network topology data set according to the feeder segment switch state; normalizing the distribution network topology data set, and establishing an output time series matrix in combination with photovoltaic active power data; extracting features from the output time series matrix using a kernel density estimation method to obtain output features; hierarchically clustering based on the output features to obtain a set of typical photovoltaic output scenarios; based on the typical photovoltaic output scenario set, establishing a mapping relationship between node injection power and node voltage in combination with distribution line impedance parameters to obtain a multi-scenario network flow calculation model; inputting real-time feeder segment switch states and photovoltaic output data into the multi-scenario network flow calculation model to obtain a photovoltaic grid-connected scenario.

[0034] For example, in a regional distribution network, there are 12 outgoing circuit breakers, 8 tie switches, 15 section switches, 14 photovoltaic grid-connected points, and a sampling period of 100 milliseconds. The system records the switch state quantity, including position state and fault state. Each switch contains two types of information: closing / opening state and normal / fault. The switch state association matrix is ​​a 35×35 matrix composed of 0 and 1, where 1 indicates that there is an electrical connection relationship between the two circuit breakers. The current topology of the distribution network is obtained based on the photovoltaic grid-connected capacity, the geographical coordinates of the grid-connected point, and the switch association matrix. During the standardization process, the original data of the three dimensions of voltage level, load level, and voltage margin are normalized. Taking the voltage level as an example, the rated voltage of the substation bus is 10 kV, the measured voltage range is 9.5-10.5 kV, and the normalized value range is 0-1. The load level is standardized based on the rated capacity of the transformer, and abnormal data exceeding 3 times the standard deviation is eliminated. The historical operation data of the photovoltaic power station includes operating parameters such as active power, reactive power, and DC side voltage at a sampling interval of 15 minutes. A 14×96 time series feature matrix was constructed for the power data of 96 time points of 14 photovoltaic grid-connected points. The kernel density estimation method was used to extract the 24-hour output characteristic curve of each grid-connected point. The hierarchical clustering algorithm obtained four typical scenarios by calculating the Euclidean distance, corresponding to cloudy, cloudy, sunny and rainy conditions. In the distribution network flow calculation model, the substation is used as the balancing node, the photovoltaic grid-connected point is used as the injection power node, and the load point is used as the power extraction node. The voltage margin constraint range is plus or minus 7% of the rated voltage, and the line impedance parameters are calculated according to the standard parameters of overhead lines and cables. The influence coefficient of each grid-connected point on the bus voltage is obtained through flow sensitivity analysis, and a 14×35 voltage influence coefficient matrix is ​​constructed to divide the photovoltaic grid-connected scenarios.

[0035] As a preferred implementation, step S2, according to the distribution network topology and the feeder segment switch state, calculates the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario to obtain the impedance change characteristics of each photovoltaic grid-connected scenario, including: Obtain the three-phase voltage and current values ​​of each feeder section switch and the feeder impedance parameters of the distribution line; Calculating the sequence impedance value of the distribution line according to the three-phase voltage and current values ​​and the distribution network topology; Calculating an equivalent impedance value of a photovoltaic grid-connected point according to the sequence impedance value and the feeder impedance parameter; According to the feeder segment switch state, calculating the change data of the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario; An exponential smoothing method is used to perform trend analysis on the change data of the equivalent impedance to obtain impedance change characteristics of each photovoltaic grid-connected scenario.

[0036] It should be noted that the feeder impedance parameters include line length and line cross-section data; the sequence impedance value includes positive sequence impedance value and zero sequence impedance value. When calculating the sequence impedance value, the distribution network line is segmented based on the position of the feeder segmentation switch to obtain the sequence impedance values ​​of several segments of the line, and then the equivalent impedance value of the photovoltaic grid-connected point is calculated.

[0037] In some preferred embodiments, the electrical distance and electrical signal transmission information between each node are obtained from the distribution network topology, and the dynamic change characteristics of the distribution network topology are obtained based on the feeder segment switch state. Then, the voltage stability margin curve is obtained according to the voltage amplitude data of each node, and the change rate of the equivalent impedance of the photovoltaic grid-connected point under different scenarios is calculated to obtain a dynamic impedance characteristic curve. According to the dynamic impedance characteristic curve, the impedance change trends of different scenarios of the photovoltaic grid-connected point are classified by the support vector machine algorithm to obtain the impedance change characteristics of each photovoltaic grid-connected scenario.

[0038] For example, a 10 kV distribution line includes 12 sectionalizers, and the switch state matrix is ​​a 12×12 symmetric matrix. The matrix element value is 0 or 1, where 1 indicates that there is an electrical connection between the two switches, and 0 indicates that there is no electrical connection. The switch online monitoring device collects three-phase voltage and current data every 100 milliseconds, and the sampled data includes phase voltage amplitude, phase current amplitude and phase angle information.

[0039] During the measurement of line impedance parameters, the positive sequence impedance value is obtained by calculating the three-phase voltage and current data. For an overhead line with a cross-section of 35 square millimeters, the positive sequence resistance value of the line is about 0.927 ohms per kilometer, the positive sequence reactance value is about 0.372 ohms, and the zero sequence impedance value is 2.5 to 3 times the positive sequence impedance value. The online measurement device for feeder parameters measures the line length by time domain reflection method. For distribution lines with more branches, a multi-point injection method is used to improve the measurement accuracy. The photovoltaic grid-connected point operation data contains parameters such as active power, reactive power, and grid-connected point voltage, with a sampling interval of 1 minute. For the mutation points and outliers in the original data, the 5-point median filtering method is used for processing, and the filtered data is used to construct a 75×75 electrical connection relationship matrix. Each element in the matrix represents the comprehensive impedance value between two nodes, taking into account the influence of the line impedance itself and the equivalent impedance of the load.

[0040] The electrical distance between nodes is calculated using an improved Euclidean distance method, and a voltage sensitivity weight factor is introduced on the basis of the traditional Euclidean distance. For a photovoltaic power station with an installed capacity of 500 kilowatts, the voltage fluctuation range allowed at the grid-connected point is plus or minus 7% of the rated value. The voltage stability margin curve is drawn through 24 hours of continuous voltage monitoring data. The moment corresponding to the lowest point of the curve often occurs during the period of rapid change in photovoltaic output. The dynamic impedance characteristic curve reflects the change law of the impedance of the photovoltaic grid-connected point under different operating scenarios. The shape of the curve is closely related to the photovoltaic output, load level and network topology. The support vector machine algorithm is used to extract the curve feature quantity for scene classification, and the impedance change characteristics of each photovoltaic grid-connected scenario are obtained. The feature quantity includes information such as the slope of the curve, the fluctuation amplitude, and the periodicity.

[0041] As a preferred implementation, step S3, extracting output fluctuation characteristics in each photovoltaic grid-connected scenario according to the photovoltaic output data, includes: A sliding window method is used to reduce noise and process outliers on the photovoltaic output data to obtain photovoltaic output time series data; Wavelet transform is used to perform multi-scale decomposition on the photovoltaic output time series data, extract the fluctuation amplitude, frequency and duration, construct a photovoltaic output fluctuation feature matrix, and obtain the output fluctuation characteristics under each photovoltaic grid-connected scenario.

[0042] Exemplarily, the data collection period of the photovoltaic output data is 1s, and a 60s sliding window is used to preprocess the photovoltaic output data. For a 500-kilowatt photovoltaic power station, the output fluctuation amplitude is within 5% during normal operation, and the fluctuation amplitude can reach 20% when the cloud layer changes rapidly, and the duration is 3 to 5 minutes. The high-frequency fluctuation components caused by equipment jitter are eliminated through noise reduction processing, and the fluctuation characteristics caused by natural factors such as cloud occlusion are retained. The wavelet transform performs multi-scale decomposition on the photovoltaic output fluctuation curve, selects the db4 wavelet basis function, and decomposes the original signal into 5 layers. After decomposition, the fluctuation characteristics of different frequency bands are obtained, in which the high-frequency component reflects the instantaneous fluctuation characteristics, and the low-frequency component reflects the long-term change trend. The fluctuation amplitude characteristics are characterized by standard deviation, the frequency characteristics are calculated by zero-crossing rate, and the duration is obtained based on the peak detection method.

[0043] As a preferred implementation, step S4, calculating the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics in each photovoltaic grid-connected scenario, adjusting the feeder segment switch state according to the correlation coefficient, and obtaining an initial topology adjustment scheme, is specifically performed through steps S41-S44: S41. Calculate the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario using the Pearson correlation coefficient method.

[0044] S42. According to the correlation coefficient and a preset correlation coefficient threshold, the photovoltaic grid-connected scenario is divided into an impedance optimization scenario and a multi-objective optimization scenario.

[0045] It should be noted that by calculating the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics through the Pearson correlation coefficient method, the coupling relationship between the grid impedance and the output fluctuation in each of the photovoltaic grid-connected scenarios can be obtained.

[0046] S43: For the impedance optimization scenario, a genetic algorithm is used to find and calculate an impedance optimization target value, and the feeder segment switch state is adjusted according to the impedance optimization target to generate an initial topology adjustment plan.

[0047] S44. For the multi-objective optimization scenario, according to preset optimization indicators, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0048] In some preferred embodiments, step S42, according to the correlation coefficient and the preset correlation coefficient threshold, the photovoltaic grid-connected scenario is divided into an impedance optimization scenario and a multi-objective optimization scenario, by grouping the correlation coefficients based on the coupling relationship between the grid impedance and the output fluctuation, using a hierarchical clustering method, and calculating the similarity between each group by Euclidean distance to obtain a scenario classification result. Further, according to the scenario classification result, the photovoltaic grid-connected scenarios are arranged in descending order based on the values ​​of the correlation coefficients to achieve the division of the photovoltaic grid-connected scenarios.

[0049] For example, taking a photovoltaic power station as an example, in the normal operation scenario on a sunny day, the correlation coefficient between output fluctuation and impedance change is 0.15, showing a weak correlation. In the scenario of frequent fluctuations in rainy weather, the correlation coefficient rises to 0.68, showing a significant positive correlation. The hierarchical clustering process uses Euclidean distance to measure the similarity of scenarios, and the clustering results are displayed through a dendrogram. The clustering threshold is set to 0.5, and all operation scenarios are divided into 4 categories. The first category corresponds to the stable operation scenario, characterized by small output fluctuations and slow impedance changes. The second category corresponds to the violent fluctuation scenario, where rapid output fluctuations cause significant impedance changes. The third category corresponds to the topology reconstruction scenario, where impedance changes suddenly and the output is relatively stable. The fourth category corresponds to the fault recovery scenario, where both output and impedance show violent fluctuations. According to the correlation coefficient sorting results, the coupling degree of the violent fluctuation scenario is the highest, with a mean correlation coefficient of 0.72. The second is the fault recovery scenario, with a mean correlation coefficient of 0.65. The coupling degree of the stable operation scenario and the topology reconstruction scenario is relatively low, with mean correlation coefficients of 0.18 and 0.25, respectively.

[0050] Further, based on the above analysis, in an embodiment of the present invention, a photovoltaic grid-connected scenario with a correlation coefficient greater than 0.7 is classified as an impedance optimization scenario, and a photovoltaic grid-connected scenario with a correlation coefficient not greater than 0.7 is classified as a multi-objective optimization scenario.

[0051] It should be noted that the impedance optimization scenario mainly considers improving the equivalent impedance characteristics of the grid connection point and reducing the impedance value by switching the tie switch. The multi-objective optimization scenario focuses on optimizing the overall operation status of the network and balancing the load rate of each feeder under the premise of meeting the voltage stability constraint.

[0052] Further, preferably, step S43, for the impedance optimization scenario, adopting a genetic algorithm to find and calculate an impedance optimization target value, adjusting the feeder segment switch state according to the impedance optimization target, and generating an initial topology adjustment scheme, includes: Constructing an optimization objective function of the impedance optimization scenario; the optimization objective function includes a node impedance term and a power loss term; According to the optimization objective function, a genetic algorithm is used to find and calculate the impedance optimization target value; According to the impedance optimization target value, an algorithm based on graph theory is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0053] In some preferred embodiments, based on the topology of the distribution network, the switch state of each feeder segment is marked by a graph coloring algorithm, the voltage amplitude data of each node is obtained from the voltage measurement terminal, the voltage margin and transmission loss between nodes are calculated, and the basic topology evaluation data set is obtained. According to the basic topology evaluation data set, the path combination from each grid connection point to the substation is calculated by a depth-first search algorithm, and the comprehensive impedance value of each path is calculated by impedance accumulation to obtain a path impedance optimization data set. If the optimization object focuses on the impedance characteristics, the impedance optimal path is calculated based on the ant colony algorithm, and the optimization goal includes two dimensions: minimizing the comprehensive impedance of the path and minimizing the transmission loss.

[0054] Preferably, the optimization index includes a voltage stability margin index and a power balance index. Then, step S44, for the multi-objective optimization scenario, according to the preset optimization index, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment scheme, including: Obtaining a voltage curve and a power curve of the multi-objective optimization scenario; Calculating a voltage stability margin index according to the voltage curve; Calculating a power balance index according to the power curve; A fuzzy comprehensive evaluation algorithm is used to automatically calculate the weight values ​​of the voltage stability margin index and the power balance index to obtain a multi-objective optimization evaluation matrix; According to the correlation coefficient, the optimization target of the multi-objective optimization evaluation matrix is ​​obtained; According to the optimization target of the multi-objective optimization evaluation matrix, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0055] In some preferred embodiments, the fuzzy comprehensive evaluation adopts a triangular membership function to obtain the weights of voltage stability and power balance according to a preset evaluation matrix.

[0056] It should be noted that, for the distribution network, even if the distribution network topology structure and feeder segment switch status are completely consistent, it may be divided into different photovoltaic grid-connected scenarios due to different weather conditions. In the embodiment of the present invention, the initial topology adjustment scheme and the optimized adjustment scheme under each photovoltaic grid-connected scenario are recorded through the scenario-based topology adjustment instruction library to realize the scenario-adaptive adjustment operation.

[0057] In a preferred embodiment, a 10 kV distribution line is taken as an example. The line includes 15 segment switches and 12 contact switches to form a switch state association matrix. Each element in the matrix takes a value of 0 or 1, and 1 indicates that the two switches are in a connected state. A network topology diagram is constructed based on the minimum spanning tree algorithm, and the weight value of each edge is determined by the line impedance and the transmission loss. For an overhead line with a cross-section of 300 square millimeters, the impedance value per kilometer is about 0.12 ohms. The graph coloring algorithm uses 4 colors to mark the switch state, including normally open, normally closed, switchable and faulty states. The node voltage measurement data shows that the voltage of each node fluctuates between 9.7 and 10.3 kV during normal operation, and the voltage margin is maintained within the range of plus or minus 3%. The transmission loss of each path is calculated based on the voltage and impedance data, and the average loss rate is 2.5%. Depth-first search calculates all feasible paths from the grid connection point to the substation. There are 8 feasible paths from a 500-kilowatt photovoltaic grid connection point to the substation, and the comprehensive impedance value of each path is calculated by accumulation. The comprehensive impedance of the shortest path is 0.85 ohms, and the comprehensive impedance of the longest path is 1.42 ohms. Taking into account the influence of transmission loss, the path with smaller comprehensive impedance is preferred in actual operation. During the optimization process of the ant colony algorithm, the initial pheromone concentration is set to 1 and the pheromone volatility coefficient is 0.3. After 200 iterations, the optimal path is obtained. The comprehensive impedance of the path is 0.78 ohms and the transmission loss is reduced to 1.8%. The particle swarm algorithm considers voltage stability and power balance at the same time. The population size is set to 100, and the inertia weight decreases linearly from 0.9 to 0.4.

[0058] The multi-objective optimization results show that the voltage deviation is controlled within plus or minus 2%, and the load rate of each feeder tends to be balanced. The topology reconstruction sequence includes a combination of switch operations, and its feasibility is verified by power flow calculation. The verification indicators include that the node voltage allowable deviation does not exceed plus or minus 7%, the line current carrying capacity does not exceed 85% of the rated capacity, and the power factor is not less than 0.95. The neural network prediction model adopts a three-layer structure. The input layer contains characteristic quantities such as voltage, power, and impedance. The number of hidden layer nodes is 20, and the output layer predicts the dynamic characteristic indicators of each scheme.

[0059] In actual applications, in scenarios where PV output fluctuates violently (i.e., PV grid-connected scenarios with correlation coefficients greater than 0.7), the solution with the best impedance characteristics is preferred. In scenarios where load distribution is uneven (i.e., PV grid-connected scenarios with correlation coefficients less than 0.7), the solution with the best overall performance is selected. Actual measured data shows that the distribution network operation indicators have been significantly improved through scenario adaptive adjustment.

[0060] As a preferred implementation, step S5, calculating the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulating the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme, is specifically performed through steps S51-S56: S51. Obtain voltage data and current data of power grid nodes under the initial topology adjustment scheme.

[0061] S52. Calculate transient power transmission characteristics between grid nodes based on the voltage data to obtain transient stability risk.

[0062] S53. Calculate the line power distribution data of the initial topology adjustment solution according to the voltage data and the current data to obtain a static safety margin.

[0063] It should be noted that the transient stability risk reflects the voltage phase angle change trend before and after the impedance mutation, while the static safety margin reflects the line overload level and the node voltage over-limit degree. In some preferred embodiments, for the impedance optimization scenario, the transient stability risk is used to evaluate whether the transient characteristics of the initial topology adjustment scheme meet the requirements; for the multi-objective optimization scenario, the static safety margin is used to evaluate whether the transient characteristics of the initial topology adjustment scheme meet the requirements.

[0064] S54. If the transient stability risk is greater than a preset stability risk threshold, a recursive neural network is used to predict the transient disturbance propagation law, and the transient response process caused by the short-circuit fault and the switching operation is calculated through time domain simulation to obtain transient response optimization data.

[0065] S55. If the static safety margin is less than a preset safety margin threshold, the power flow distribution of the distribution network is calculated based on the Newton iteration method, and the power transmission constraint and the voltage over-limit constraint are verified to obtain static constraint optimization data.

[0066] It should be noted that in the embodiment of the present invention, if the transient stability risk is prominent, a transient simulation analysis is carried out; if the static safety margin is insufficient, a static safety verification is carried out; and the initial topology adjustment plan is further iteratively optimized until the requirements for safe and stable operation of the power grid in multiple scenarios are met.

[0067] S56. Perform iterative simulation on the initial topology adjustment scheme according to the transient response optimization data and the static constraint optimization data to obtain an optimized topology adjustment scheme.

[0068] Further, preferably, step S52, calculating the transient power transmission characteristics between the grid nodes according to the voltage data to obtain the transient stability risk, includes: Processing the voltage data by wavelet transform to obtain voltage fluctuation characteristics; According to the voltage data, a voltage phase angle variation trend before and after the impedance mutation under the initial topology adjustment scheme is obtained; Obtaining a node admittance matrix according to the state of the feeder section switch and the distribution network topology; According to the voltage phase angle variation trend and the node admittance matrix, the transient power transmission characteristics between the power grid nodes are calculated to obtain a transient stability margin curve; According to the transient stability margin curve, a recursive neural network is used to predict the transient response of key network nodes under impedance mutation, and the transient disturbance propagation path is calculated through the phase angle stability criterion to obtain the transient stability risk.

[0069] For example, a 10 kV distribution line is used as an example. The line contains 10 section switches and 6 tie switches. A high-speed synchronous phasor measurement device is used to collect data with a sampling frequency of 100 Hz. The wavelet transform uses the db4 wavelet basis function to decompose the voltage fluctuation characteristics and extract the fluctuation components of different frequency bands.

[0070] In normal operation, the voltage fluctuation amplitude is controlled within 0.02 per unit value, and the phase angle deviation does not exceed 2 degrees. The node admittance matrix reflects the electrical connection between nodes in the network. For an overhead line with a cross-section of 300 square millimeters, the positive sequence admittance value per kilometer is about 8.33 Siemens. Through the analysis of transient power transmission characteristics, it is found that the power fluctuation caused by impedance mutation can be propagated to three adjacent nodes within 0.1 seconds, and the fluctuation amplitude decays exponentially with the propagation distance.

[0071] The transient stability margin curve shows that under the condition of impedance mutation, the transient power margin of the key node drops to 65% of the nominal value. The recursive neural network prediction model contains 10 input nodes, corresponding to characteristic quantities such as voltage amplitude, phase angle, and active power. The accuracy of the model obtained through historical data training reached 92%, and the prediction results showed that the duration of the transient process caused by impedance mutation was about 0.5 seconds. The phase angle stability criterion is based on the calculation of the power angle difference between adjacent nodes. When the angle difference exceeds 30 degrees, it is determined that there is a risk of instability.

[0072] The dynamic impedance scan was performed in the frequency range of 0.1 to 10 Hz with a scan step of 0.1 Hz. The scan results showed that in the scenario where the photovoltaic output fluctuates frequently, the equivalent impedance of the grid-connected point has a resonant peak in the frequency range of 1 to 2 Hz, and the peak impedance reaches 3 times the steady-state value.

[0073] Preferably, step S53, calculating the line power distribution data of the initial topology adjustment scheme according to the voltage data and the current data to obtain a static safety margin, includes: Calculating line power distribution data of the initial topology adjustment solution according to the voltage data and the current data; The line overload level and node voltage over-limit degree are calculated through power flow sensitivity to obtain the static safety margin.

[0074] The static safety margin takes into account the voltage over-limit depth and duration, and obtains a comprehensive evaluation value through weighted summation. The power flow sensitivity analysis shows that the line overload level is positively correlated with the node voltage over-limit degree.

[0075] For example, when the load rate of a certain line reaches 90%, the voltage at the node at the end of the line drops to 9.4 kilovolts. The support vector machine uses a Gaussian kernel function to classify the operating status, and the feature vector contains parameters such as line load rate, node voltage, and power factor. All operating scenarios are divided into three categories: safety, warning, and danger. The warning state mainly occurs during the period of rapid daily load growth. The scenario-based safety assessment result library records the transient and static characteristics of various operating scenarios. In the scenario of sudden load increase, the transient stability risk is high but the static constraint satisfaction is good. In the scenario of network topology reconstruction, the opposite characteristics are presented. Comprehensive evaluation data show that the overall operation level of the distribution network is within a controllable range.

[0076] Furthermore, in some preferred embodiments, in step S56, the initial topology adjustment scheme is iteratively simulated according to the transient response optimization data and the static constraint optimization data to obtain an optimized topology adjustment scheme, and a genetic algorithm is used to optimize the parameters of the topology adjustment scheme according to the transient response optimization data and the static constraint optimization data, and the comprehensive evaluation value of each scheme is calculated by the fitness function. According to the comprehensive evaluation value, the optimization scheme is classified based on the support vector machine, and the optimization experience of similar scenarios is extracted from the historical topology scheme library to obtain a scheme optimization iterative data set. According to the scheme optimization iterative data set, each iterative scheme is repeatedly verified by power flow calculation until the voltage stability and power transmission capacity constraints are met.

[0077] For example, a 10 kV distribution line is connected to three photovoltaic power stations with a total installed capacity of 2,000 kW. The sampling period for photovoltaic output fluctuation data is 1 second, and the sampling period for load change data is 1 minute. The digital twin network contains 35 nodes and 42 lines, and a π-type equivalent circuit is used to simulate line parameters. For an overhead line with a cross-section of 300 square millimeters, the positive sequence impedance is 0.12+j0.095 ohms / kilometer. The node admittance matrix reflects the electrical connection between the nodes in the network, and the transient response characteristics are obtained by calculating the sensitivity of the node voltage to the injected power.

[0078] Transient stability risk includes two dimensions: phase angle stability and voltage stability. The phase angle stability threshold is set to 30 degrees, and the voltage stability threshold is 0.1 per unit. When the photovoltaic output drops by 50%, the voltage at the grid point drops to 0.92 per unit, and the phase angle deviation reaches 25 degrees. The recursive neural network prediction model adopts a three-layer structure. The input layer contains characteristic quantities such as voltage amplitude, phase angle, and frequency, and the number of hidden layer nodes is 20. The model obtained by training 1,000 sets of historical data has an accuracy rate of 93% in transient disturbance prediction. Time domain simulation analysis shows that the voltage drop recovery time caused by a three-phase short circuit fault is about 0.3 seconds, and the transient process caused by switch switching lasts about 0.2 seconds.

[0079] The Newton iteration method is used for power flow calculation in static safety verification, and the iterative convergence accuracy is set to 0.0001. The power transmission constraint requires that the line load rate does not exceed 85%, and the voltage over-limit constraint requires that the node voltage deviation does not exceed plus or minus 7%. The calculation results show that under heavy load conditions, the load rates of the three feeders reach 82%, 78% and 75% respectively, and the terminal node voltage drops to 9.5 kV. During the genetic algorithm optimization process, the population size is set to 100 and the evolutionary generation is 200 generations. The fitness function comprehensively considers the transient stability margin and the static safety margin, and the weight coefficients are 0.6 and 0.4 respectively.

[0080] The support vector machine uses Gaussian kernel function to classify the optimization schemes, and divides all schemes into three categories: safe, warning and dangerous. A total of 15 groups of optimization experiences with a similarity of more than 85% were selected from the historical scheme library. During the iterative optimization process, power flow verification is performed in each round of iteration. The voltage stability constraint requires that the voltage deviation of each node is less than 5%, and the power transmission constraint requires that the line overload duration does not exceed 5 minutes. After an average of 6 rounds of iterative optimization, all indicators meet the constraint requirements. The final topology adjustment scheme achieves the optimal configuration of network static parameters while ensuring transient characteristics.

[0081] By adopting a photovoltaic source grid load storage optimization scheduling method provided by an embodiment of the present invention, by calculating the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics, and then adjusting the feeder segment switch state, it is possible to fully consider the complex coupling between the grid topology and the photovoltaic output fluctuation, and further analyze the transient characteristics of the grid nodes to optimize the topology adjustment scheme, and verify the feasibility and optimization effect of the scheme. The embodiment of the present invention can effectively improve the distribution network's ability to absorb photovoltaic power generation, while ensuring the safe and stable operation of the power grid.

[0082] The embodiment of the present invention provides a photovoltaic source grid load storage optimization scheduling system. Figure 2The photovoltaic source grid load storage optimization scheduling system includes a multi-scenario data acquisition module 11, an impedance change characteristic calculation module 12, an output fluctuation characteristic calculation module 13, an initial scheme generation module 14, a scheme optimization module 15 and an optimization scheduling module 16, wherein: The multi-scenario data acquisition module 11 is used to obtain the distribution network topology, photovoltaic output data and feeder segment switch status in each photovoltaic grid-connected scenario; The impedance change characteristic calculation module 12 is used to calculate the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario according to the distribution network topology and the feeder segment switch state, and obtain the impedance change characteristic of each photovoltaic grid-connected scenario; An output fluctuation characteristic calculation module 13 is used to extract output fluctuation characteristics in each photovoltaic grid-connected scenario according to the photovoltaic output data; An initial scheme generating module 14 is used to calculate the correlation coefficient between the impedance variation characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, and adjust the feeder segment switch state according to the correlation coefficient to obtain an initial topology adjustment scheme; A scheme optimization module 15 is used to calculate the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulate the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme; The optimization scheduling module 16 is used to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario by adopting the optimized topology adjustment scheme.

[0083] As a preferred implementation, the impedance change characteristic calculation module 12 is used to: Obtain the three-phase voltage and current values ​​of each feeder section switch and the feeder impedance parameters of the distribution line; Calculating the sequence impedance value of the distribution line according to the three-phase voltage and current values ​​and the distribution network topology; Calculating an equivalent impedance value of a photovoltaic grid-connected point according to the sequence impedance value and the feeder impedance parameter; According to the feeder segment switch state, calculating the change data of the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario; An exponential smoothing method is used to perform trend analysis on the change data of the equivalent impedance to obtain impedance change characteristics of each photovoltaic grid-connected scenario.

[0084] As a preferred implementation, the output fluctuation characteristic calculation module 13 is used to: A sliding window method is used to reduce noise and process outliers on the photovoltaic output data to obtain photovoltaic output time series data; Wavelet transform is used to perform multi-scale decomposition on the photovoltaic output time series data, extract the fluctuation amplitude, frequency and duration, construct a photovoltaic output fluctuation feature matrix, and obtain the output fluctuation characteristics under each photovoltaic grid-connected scenario.

[0085] As a preferred implementation, the initial solution generation module 14 includes: A correlation coefficient calculation unit, used for calculating the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario by using a Pearson correlation coefficient method; A scenario division unit, configured to divide the photovoltaic grid-connected scenario into an impedance optimization scenario and a multi-objective optimization scenario according to the correlation coefficient and a preset correlation coefficient threshold; A first initial solution generating unit is used to use a genetic algorithm to find and calculate an impedance optimization target value for the impedance optimization scenario, adjust the feeder segment switch state according to the impedance optimization target, and generate an initial topology adjustment solution; The second initial solution generating unit is used to adjust the feeder segment switch state according to preset optimization indicators and generate an initial topology adjustment solution for the multi-objective optimization scenario by using a heuristic search algorithm.

[0086] Further, preferably, the first initial solution generating unit is specifically used to: Constructing an optimization objective function of the impedance optimization scenario; the optimization objective function includes a node impedance term and a power loss term; According to the optimization objective function, a genetic algorithm is used to find and calculate the impedance optimization target value; According to the impedance optimization target value, an algorithm based on graph theory is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0087] Preferably, the second initial solution generating unit is specifically used for: Obtaining a voltage curve and a power curve of the multi-objective optimization scenario; Calculating a voltage stability margin index according to the voltage curve; Calculating a power balance index according to the power curve; A fuzzy comprehensive evaluation algorithm is used to automatically calculate the weight values ​​of the voltage stability margin index and the power balance index to obtain a multi-objective optimization evaluation matrix; According to the correlation coefficient, the optimization target of the multi-objective optimization evaluation matrix is ​​obtained; According to the optimization target of the multi-objective optimization evaluation matrix, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

[0088] As a preferred implementation, the solution optimization module 15 includes: Acquiring voltage data and current data of power grid nodes under the initial topology adjustment scheme; A transient stability risk calculation unit, used to calculate the transient power transmission characteristics between each grid node according to the voltage data to obtain the transient stability risk; A static safety margin calculation unit, configured to calculate the line power distribution data of the initial topology adjustment scheme according to the voltage data and the current data, so as to obtain a static safety margin; A transient response optimization data calculation unit is used to predict the transient disturbance propagation law by using a recursive neural network if the transient stability risk is greater than a preset stability risk threshold, and calculate the transient response process caused by the short-circuit fault and the switching operation by time domain simulation to obtain transient response optimization data; A static constraint optimization data calculation unit is used to calculate the power flow distribution of the distribution network based on the Newton iteration method if the static safety margin is less than a preset safety margin threshold, and to obtain static constraint optimization data by checking the power transmission constraint and the voltage over-limit constraint; A scheme optimization unit is used to iteratively simulate the initial topology adjustment scheme according to the transient response optimization data and the static constraint optimization data to obtain an optimized topology adjustment scheme.

[0089] Further, preferably, the transient stability risk calculation unit is specifically used to: Processing the voltage data by wavelet transform to obtain voltage fluctuation characteristics; According to the voltage data, a voltage phase angle variation trend before and after the impedance mutation under the initial topology adjustment scheme is obtained; Obtaining a node admittance matrix according to the state of the feeder section switch and the distribution network topology; According to the voltage phase angle variation trend and the node admittance matrix, the transient power transmission characteristics between the power grid nodes are calculated to obtain a transient stability margin curve; According to the transient stability margin curve, a recursive neural network is used to predict the transient response of key network nodes under impedance mutation, and the transient disturbance propagation path is calculated through the phase angle stability criterion to obtain the transient stability risk.

[0090] Preferably, the static safety margin calculation unit is specifically used to: Calculating line power distribution data of the initial topology adjustment solution according to the voltage data and the current data; The line overload level and node voltage over-limit degree are calculated through power flow sensitivity to obtain the static safety margin.

[0091] The photovoltaic source grid load storage optimization dispatching system provided by the embodiment of the present invention can fully consider the complex coupling between the grid topology and photovoltaic output fluctuations by calculating the correlation coefficient between the impedance change characteristics and the output fluctuation characteristics, and then adjust the feeder segment switch state, and further analyze the transient characteristics of the grid nodes to optimize the topology adjustment scheme, and verify the feasibility and optimization effect of the scheme. The embodiment of the present invention can effectively improve the distribution network's ability to absorb photovoltaic power generation, while ensuring the safe and stable operation of the grid.

[0092] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0093] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A photovoltaic source grid load storage optimization scheduling method, characterized in that: include: Obtain the distribution network topology, photovoltaic output data and feeder section switch status under each photovoltaic grid-connected scenario; According to the distribution network topology and the feeder segment switch state, the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario is calculated to obtain the impedance change characteristics of each photovoltaic grid-connected scenario; Extracting output fluctuation characteristics under each photovoltaic grid-connected scenario according to the photovoltaic output data; Calculating the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, adjusting the feeder segment switch state according to the correlation coefficient, and obtaining an initial topology adjustment scheme; Calculating transient characteristics of power grid nodes under the initial topology adjustment scheme, and iteratively simulating the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme; The optimized topology adjustment scheme is used to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario.

2. The photovoltaic source grid load storage optimization scheduling method according to claim 1, characterized in that: The calculating the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario according to the distribution network topology and the feeder segment switch state to obtain the impedance change characteristics of each photovoltaic grid-connected scenario includes: Obtain the three-phase voltage and current values ​​of each feeder section switch and the feeder impedance parameters of the distribution line; Calculating the sequence impedance value of the distribution line according to the three-phase voltage and current values ​​and the distribution network topology; Calculating an equivalent impedance value of a photovoltaic grid-connected point according to the sequence impedance value and the feeder impedance parameter; According to the feeder segment switch state, calculating the change data of the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario; An exponential smoothing method is used to perform trend analysis on the change data of the equivalent impedance to obtain impedance change characteristics of each photovoltaic grid-connected scenario.

3. The photovoltaic source grid load storage optimization scheduling method according to claim 1, characterized in that: The step of extracting output fluctuation characteristics in each photovoltaic grid-connected scenario according to the photovoltaic output data includes: A sliding window method is used to reduce noise and process outliers on the photovoltaic output data to obtain photovoltaic output time series data; Wavelet transform is used to perform multi-scale decomposition on the photovoltaic output time series data, extract the fluctuation amplitude, frequency and duration, construct a photovoltaic output fluctuation feature matrix, and obtain the output fluctuation characteristics under each photovoltaic grid-connected scenario.

4. The photovoltaic source grid load storage optimization scheduling method according to claim 1, characterized in that: The calculating the correlation coefficient between the impedance variation characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, and adjusting the feeder segment switch state according to the correlation coefficient to obtain an initial topology adjustment scheme includes: The Pearson correlation coefficient method is used to calculate the correlation coefficient between the impedance change characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario; According to the correlation coefficient and a preset correlation coefficient threshold, the photovoltaic grid-connected scenario is divided into an impedance optimization scenario and a multi-objective optimization scenario; For the impedance optimization scenario, a genetic algorithm is used to find and calculate an impedance optimization target value, and the feeder segment switch state is adjusted according to the impedance optimization target to generate an initial topology adjustment plan; For the multi-objective optimization scenario, according to the preset optimization index, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

5. The photovoltaic source grid load storage optimization scheduling method according to claim 4, characterized in that: For the impedance optimization scenario, a genetic algorithm is used to find and calculate an impedance optimization target value, and the feeder segment switch state is adjusted according to the impedance optimization target to generate an initial topology adjustment plan, including: Constructing an optimization objective function of the impedance optimization scenario; the optimization objective function includes a node impedance term and a power loss term; According to the optimization objective function, a genetic algorithm is used to find and calculate the impedance optimization target value; According to the impedance optimization target value, an algorithm based on graph theory is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

6. The photovoltaic source grid load storage optimization scheduling method according to claim 4, characterized in that: The optimization index includes a voltage stability margin index and a power balance index; then for the multi-objective optimization scenario, according to the preset optimization index, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment scheme, including: Obtaining a voltage curve and a power curve of the multi-objective optimization scenario; Calculating a voltage stability margin index according to the voltage curve; Calculating a power balance index according to the power curve; A fuzzy comprehensive evaluation algorithm is used to automatically calculate the weight values ​​of the voltage stability margin index and the power balance index to obtain a multi-objective optimization evaluation matrix; According to the correlation coefficient, the optimization target of the multi-objective optimization evaluation matrix is ​​obtained; According to the optimization target of the multi-objective optimization evaluation matrix, a heuristic search algorithm is used to adjust the feeder segment switch state to generate an initial topology adjustment plan.

7. The photovoltaic source grid load storage optimization scheduling method according to claim 1, characterized in that: The calculating the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulating the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme, comprises: Acquiring voltage data and current data of power grid nodes under the initial topology adjustment scheme; According to the voltage data, calculating the transient power transmission characteristics between each grid node to obtain the transient stability risk; Calculating the line power distribution data of the initial topology adjustment scheme according to the voltage data and the current data to obtain a static safety margin; If the transient stability risk is greater than a preset stability risk threshold, a recursive neural network is used to predict the transient disturbance propagation law, and the transient response process caused by the short-circuit fault and the switching operation is calculated through time domain simulation to obtain transient response optimization data; If the static safety margin is less than the preset safety margin threshold, the power flow distribution of the distribution network is calculated based on the Newton iteration method, and the power transmission constraint and the voltage over-limit constraint are checked to obtain static constraint optimization data; According to the transient response optimization data and the static constraint optimization data, the initial topology adjustment scheme is iteratively simulated to obtain an optimized topology adjustment scheme.

8. The photovoltaic source grid load storage optimization scheduling method according to claim 7, characterized in that: The step of calculating the transient power transmission characteristics between the power grid nodes according to the voltage data to obtain the transient stability risk includes: Processing the voltage data by wavelet transform to obtain voltage fluctuation characteristics; According to the voltage data, a voltage phase angle variation trend before and after the impedance mutation under the initial topology adjustment scheme is obtained; Obtaining a node admittance matrix according to the state of the feeder section switch and the distribution network topology; According to the voltage phase angle variation trend and the node admittance matrix, the transient power transmission characteristics between the power grid nodes are calculated to obtain a transient stability margin curve; According to the transient stability margin curve, a recursive neural network is used to predict the transient response of key network nodes under impedance mutation, and the transient disturbance propagation path is calculated through the phase angle stability criterion to obtain the transient stability risk.

9. The photovoltaic source grid load storage optimization scheduling method according to claim 7, characterized in that: The calculating, according to the voltage data and the current data, the line power distribution data of the initial topology adjustment scheme to obtain a static safety margin comprises: Calculating line power distribution data of the initial topology adjustment solution according to the voltage data and the current data; The line overload level and node voltage over-limit degree are calculated through power flow sensitivity to obtain the static safety margin.

10. An optimized dispatching system for photovoltaic source, grid, load and storage, characterized in that: include: Multi-scenario data acquisition module, used to obtain the distribution network topology, photovoltaic output data and feeder segment switch status under various photovoltaic grid-connected scenarios; An impedance change characteristic calculation module is used to calculate the equivalent impedance of the photovoltaic grid-connected point in each photovoltaic grid-connected scenario according to the distribution network topology and the feeder segment switch state, and obtain the impedance change characteristics of each photovoltaic grid-connected scenario; An output fluctuation characteristic calculation module, used to extract the output fluctuation characteristics under each photovoltaic grid-connected scenario according to the photovoltaic output data; An initial scheme generating module, used for calculating the correlation coefficient between the impedance variation characteristic and the output fluctuation characteristic in each photovoltaic grid-connected scenario, adjusting the feeder segment switch state according to the correlation coefficient, and obtaining an initial topology adjustment scheme; A scheme optimization module, used to calculate the transient characteristics of the power grid nodes under the initial topology adjustment scheme, and iteratively simulate the initial topology adjustment scheme according to the transient characteristics to obtain an optimized topology adjustment scheme; The optimization scheduling module is used to optimize the scheduling of photovoltaic source, grid, load and storage in the photovoltaic grid-connected scenario by adopting the optimization topology adjustment scheme.

Images