A Method and System for Dynamic Prediction of Distributed Photovoltaic Capacity Based on Monte Carlo Simulation

Abstract

This invention discloses a method and system for dynamic prediction of the openable capacity of distributed photovoltaic (PV) power grids based on Monte Carlo simulation. The method includes: acquiring historical load data of feeders, short-term power prediction data of distributed PV power grids, and feeder topology data of the target area power grid; constructing a real-time remaining capacity calculation model for feeders to dynamically calculate the real-time remaining capacity of feeders; using the Monte Carlo simulation algorithm, forming independent scenarios based on historical load data of feeders and short-term power prediction data of distributed PV power grids, and calculating the voltage over-limit probability for each scenario based on feeder topology data; classifying risk levels and marking restricted areas based on the ratio of real-time remaining capacity of feeders to rated capacity, combined with the voltage over-limit probability; and evaluating the adaptability of existing power grid conditions to PV access through a multi-objective optimization model based on the marked restricted areas, and identifying the maximum accessible capacity threshold of feeders. This invention improves the accuracy and real-time performance of power grid carrying capacity assessment.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and specifically to a method and system for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation. Background Technology

[0002] Photovoltaic energy management technology falls under the field of power system operation and optimization. It primarily utilizes data integration, algorithmic prediction, and intelligent control to achieve efficient operation of distributed photovoltaic (PV) systems and grid coordination. This technology monitors PV power generation, predicts power output, optimizes energy storage scheduling, and combines meteorological data, historical load data, and real-time operating status to construct dynamic evaluation models. Its aim is to improve energy utilization efficiency, ensure grid stability, and support the large-scale integration of renewable energy. Core applications include PV power prediction, grid capacity analysis, voltage stability control, and dynamic assessment of grid connection capacity. Through multi-dimensional data fusion and intelligent algorithms, it addresses grid volatility and uncertainty issues arising from distributed PV integration, promoting the development of energy systems towards low-carbon and intelligent directions.

[0003] Traditional methods for assessing voltage risk rely on single-scenario analysis, such as typical daily load curves or fixed photovoltaic output, and fail to comprehensively evaluate the combined impact of multiple uncertainties on line voltage using probabilistic methods. For example, when photovoltaic installations are located near the end of the line, even with sufficient remaining capacity, reverse power flow can still cause voltage increases, and traditional methods cannot identify such risks through probabilistic analysis. Furthermore, existing technologies lack dynamic correlation analysis between remaining capacity and voltage risk, making it difficult to generate intuitive risk level classifications and visual heat maps, thus limiting the accuracy of practical engineering applications.

[0004] With the increasing penetration rate of distributed photovoltaic (PV) power, grid operating conditions are becoming increasingly complex, and existing technologies are insufficient to meet the dynamic assessment needs across multiple time scales and dimensions. For example, traditional static assessment methods cannot adapt to the real-time data update requirements of sliding time windows, resulting in delayed capacity predictions. Furthermore, without dynamic modeling of node power balance equations in conjunction with feeder topology, it is difficult to accurately locate high-risk areas for voltage exceedances. Therefore, there is an urgent need for a distributed PV grid connection capacity prediction method that integrates dynamic data updates, probabilistic risk analysis, and multi-dimensional assessment to improve the accuracy and real-time performance of grid capacity assessment and support the efficient absorption and safe operation of distributed energy resources. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a dynamic prediction method and system for the open capacity of distributed photovoltaic power based on Monte Carlo simulation, which addresses the above-mentioned problems in the prior art. By integrating historical load data of feeders, short-term power prediction data of photovoltaic power, and feeder topology data, a dynamic evaluation model is constructed, which overcomes the limitation of traditional static evaluation methods in quantifying grid carrying capacity and voltage risk.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for dynamically predicting the available capacity of distributed photovoltaic power based on Monte Carlo simulation includes the following steps: Acquire historical load data of feeders, short-term power forecast data of distributed photovoltaic power, and feeder topology data of the target area power grid; A real-time remaining capacity calculation model for feeders is constructed. A sliding time window is used to dynamically calculate the real-time remaining capacity of feeders based on historical load data and short-term power prediction data of distributed photovoltaics, resulting in a time-series curve of the real-time remaining capacity of feeders. Based on historical load data of the feeder and short-term power prediction data of distributed photovoltaics, the load fluctuation range and photovoltaic prediction fluctuation range are determined respectively. The Monta Carlo simulation algorithm is used to extract samples from the load fluctuation range and photovoltaic prediction range to form independent scenarios. After power flow calculation for each scenario based on feeder topology data, the voltage over-limit probability of the feeder is statistically analyzed. Based on the ratio of the real-time remaining capacity of the feeder to its rated capacity, and combined with the voltage over-limit probability, risk levels are classified and areas where new photovoltaic access is restricted are marked. Based on the annotation results of the restricted area, the performance data of the power grid in the target area is obtained. Based on the feeder topology and power grid performance data, the adaptability of the existing power grid conditions to photovoltaic access is evaluated through a multi-objective optimization model, and the maximum accessible capacity threshold of the feeder is identified.

[0007] Furthermore, the mathematical expression for the feeder real-time remaining capacity calculation model is as follows:

[0008] in, This is the real-time remaining capacity value of the feeder. This is the rated capacity of the feeder. It is the short-term power forecast value of all distributed photovoltaic power. Peak values ​​of historical load data for feeders The maximum value in the sum.

[0009] Furthermore, when using a sliding time window to dynamically calculate the real-time remaining capacity of the feeder based on historical feeder load data and distributed photovoltaic short-term power forecast data, the current time is taken as the end point of the window. The historical feeder load data and distributed photovoltaic short-term power forecast data of the corresponding time period are traced back according to the window size. Within the window, the historical feeder load data and distributed photovoltaic short-term power forecast data are substituted into the feeder real-time remaining capacity calculation model for each time period to calculate the real-time remaining capacity value of the feeder, and the window position is updated on a rolling basis for each time period.

[0010] Furthermore, based on historical load data of the feeder and short-term power forecast data of distributed photovoltaics, the load fluctuation range and photovoltaic forecast fluctuation range are determined respectively. When using the Monta Carlo simulation algorithm to extract samples from the load fluctuation range and photovoltaic forecast fluctuation range to form independent scenarios, the following steps are included: Based on historical load data of the feeder, the standard deviation of load fluctuation is extracted. And adopt a normal distribution Random variables that generate load fluctuations This makes the load fluctuation range of :

[0011] in The historical load benchmark value is the feeder load rate benchmark value generated by obtaining real-time operating data of the feeder and combining it with the load characteristic analysis model to determine the peak load fluctuation range of the feeder on a typical day. Based on distributed photovoltaic short-term power forecast data, the historical standard deviation of photovoltaic forecast error is extracted. And adopt a normal distribution Random variables that generate photovoltaic power output error This makes the photovoltaic prediction fluctuation range of 0. :

[0012] in This represents the short-term power forecast for distributed photovoltaic systems. A specified number of scenarios are generated through independent random sampling, respectively from... Distribution and Distribution Independently draw a random value ,Will and By index Pairing, forming the first The scenarios are grouped so that the combination of load and photovoltaic output in each scenario satisfies the joint probability distribution characteristics.

[0013] Furthermore, when calculating the voltage over-limit probability of the feeder after performing power flow calculations for each scenario based on the feeder topology data, the following steps are included: Based on the line impedance parameters and node location relationships in the feeder topology data, construct the distribution network node-branch correlation matrix; The load power and photovoltaic predicted power in each scenario are used as the node injected power. The node power balance equation is established and the node voltage is solved. Then, the node voltage deviation is calculated based on the node voltage value and the feeder rated voltage value. Finally, the number of scenarios with deviations exceeding the preset threshold is counted, and the number of scenarios is divided by the total number of scenarios to obtain the voltage over-limit probability.

[0014] Furthermore, when classifying risk levels and marking areas where new photovoltaic grid connection is restricted based on the ratio of the feeder's real-time remaining capacity to its rated capacity, combined with the voltage over-limit probability, the following steps are included: If the real-time remaining capacity of the feeder is less than 10% of the rated capacity of the feeder and the voltage over-limit probability is greater than 20%, the corresponding area is a high-risk area. The high-risk area will be marked in red and a warning message prohibiting the addition of new photovoltaic grid connection will be displayed. If the real-time remaining capacity of the feeder is 10%-30% of the feeder's rated capacity and the voltage over-limit probability is greater than 10%, the corresponding area is a medium-risk area. The medium-risk area will be marked in orange and a prompt message indicating that a grid transformation plan is required will be added. If the real-time remaining capacity of the feeder is greater than 30% of the rated capacity and the voltage over-limit probability is less than 10%, the corresponding area is a low-risk area. The low-risk area is marked in green and labeled with recommended information that it can be freely connected.

[0015] Furthermore, the objectives of the multi-objective optimization model are to maximize grid security, minimize grid transformation costs, and maximize grid stability indicators. The objective function is as follows:

[0016] in, These are the objective functions for maximizing power grid security, minimizing power grid upgrade costs, and maximizing power grid stability indicators, respectively. These are the weights of the three objectives; The objective function for maximizing power grid security is as follows:

[0017] in, This represents the total number of nodes in the target area's power grid. Indicates the first Voltage deviation at each node; This is the rated line voltage of the feeder; This refers to the total number of feeder lines in the target area's power grid; It is the first The actual load rate of the feeder; It is the safe threshold for line load rate; The objective function for minimizing the cost of power grid upgrades is as follows:

[0018] in, It is the unit capacity expansion cost of transformers; It refers to the expansion of transformer capacity. ,in This is the original rated capacity of the transformer; It is the unit capacity cost of reactive power compensation equipment; It is the first The reactive power compensation capacity required for each node ,in For the first The reactance of the line where each node is located; It is the total number of nodes; The objective function for maximizing the power grid stability index is as follows:

[0019] in, This represents the total number of feeder lines; They are the first Resistance and reactance of the feeder; These are the safety thresholds for line resistance and reactance, respectively. It is the first Voltage margin of each node, ,in , , This is the rated line voltage; 0.05 It is the maximum allowable value for node voltage margin; It is the total number of nodes; The constraints of the multi-objective optimization model include:

[0020] in, For the safety impedance threshold, For the load safety threshold, Rated line voltage, Based on risk level; The constraints of the multi-objective optimization model also include photovoltaic access constraints:

[0021] in, It is 80% of the rated capacity of the line transformer.

[0022] Furthermore, when evaluating the adaptability of existing grid conditions to photovoltaic (PV) grid integration using a multi-objective optimization model and identifying the maximum connectable capacity threshold for feeders, the multi-objective optimization model is solved using the NSGA-II algorithm to generate a Pareto optimal solution set. Feasible solutions that meet the requirements for safe grid operation are then selected. Based on the selection results of feasible solutions, the maximum connectable capacity is calculated using the following formula:

[0023] in, The rated capacity of the transformer or transmission line; Current load; SF Leave room for safety.

[0024] Furthermore, after identifying the maximum accessible capacity threshold for the feeder, the following steps are also included: Generate a grid access condition constraint matrix, in which key indicators include transformer load factor, line voltage margin, and load fluctuation tolerance. Based on the grid connection constraint matrix, and integrating the dynamic assessment results of remaining capacity and feeder topology parameters, dynamic connection suggestions for the exploitable distributed photovoltaic capacity within the target area are generated, including: Extract real-time index values ​​and safety thresholds from the power grid access condition constraint matrix, conduct preliminary verification of access feasibility, and mark indexes that exceed the limits, which will serve as the core basis for subsequent access recommendation corrections. By integrating the constraint matrix and the dynamic evaluation results of remaining capacity, time-segmented access power thresholds are generated. Specifically, the upper limit of access power for each time period is determined based on the maximum accessible capacity threshold of the feeder and the constraint adaptation parameters in the feasible solution. Based on the high, medium, and low time periods divided by the remaining capacity level, and combined with the constraint matrix index verification results, the access power for each time period is quantitatively calculated and corrected to clarify the time-segmented access power thresholds and access requirements. By integrating the constraint matrix and feeder topology parameters, topology-differentiated access suggestions are generated. Specifically, topology-differentiated access suggestions are generated based on the feeder topology parameters and the optimal access node number in the feasible solution. Integrate time-based access power thresholds and topology-differentiated access suggestions to generate final dynamic access suggestions.

[0025] The present invention also proposes a dynamic prediction system for the available capacity of distributed photovoltaic power based on Monte Carlo simulation, comprising a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, it implements the steps of the dynamic prediction method for the available capacity of distributed photovoltaic power based on Monte Carlo simulation.

[0026] Compared with the prior art, the advantages of the present invention are as follows: This invention integrates the entire process of distributed photovoltaic (PV) grid connection, dividing the prediction method into five stages: data acquisition, remaining capacity calculation, voltage over-limit probability assessment, grid adaptability analysis, and dynamic access suggestion generation. For the grid adaptability analysis stage, a multi-objective optimization model based on the NSGA-II algorithm is constructed, comprehensively considering grid performance constraints such as feeder topology, node voltage levels, and load fluctuation characteristics, dynamically correcting the PV access capacity threshold. Monte Carlo simulation is used to quantify the joint probability distribution of load fluctuations and short-term PV prediction errors. The number of voltage over-limit scenarios is statistically analyzed through power flow calculations, generating a voltage over-limit probability as a basis for access risk assessment. Finally, risk levels are classified based on remaining capacity and voltage over-limit probability, generating an open capacity heatmap, and access power thresholds are determined through topology parameter differentiation. This is a distributed PV access prediction method that integrates dynamic capacity assessment and grid topology optimization, enabling accurate dynamic prediction of PV access capacity. It provides time-based and region-based access suggestions for grid dispatch, effectively improving the economy and security of distributed PV grid connection, and is suitable for distribution network scenarios with large load fluctuations and complex topologies. Attached Figure Description

[0027] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0028] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0029] Example 1 To overcome the shortcomings of existing technologies, such as the inability to dynamically quantify grid capacity and voltage limit exceedance risks, and the difficulty in accurately assessing distributed photovoltaic (PV) grid connection capacity, this embodiment proposes a Monte Carlo simulation-based dynamic prediction method for distributed PV open capacity for grid capacity assessment and voltage risk quantification analysis. Figure 1 As shown, it includes the following steps: S1: Collect power grid operation data for the target area. Specifically, based on the power grid operation data for the target area, acquire historical load data of feeders, short-term power forecast data of distributed photovoltaic power, and feeder topology data of the power grid in the target area. S2: Based on historical load data and photovoltaic power forecast, analyze the matching relationship between the rated capacity of the feeder and the historical load peak. Combined with the spatiotemporal distribution characteristics of photovoltaic power forecast, construct a real-time remaining capacity calculation model for the feeder. Use a sliding time window to dynamically calculate the real-time remaining capacity of the feeder based on historical load data and short-term power forecast data of distributed photovoltaics, and obtain the time series curve of the real-time remaining capacity of the feeder. S3: Quantify the voltage exceedance probability under different installed capacity combinations based on Monte Carlo simulation. Specifically, the load fluctuation range and photovoltaic prediction fluctuation range are determined based on historical load data of feeders and short-term power prediction data of distributed photovoltaics, respectively. Using the Monte Carlo simulation algorithm, samples are extracted from the load fluctuation range and photovoltaic prediction fluctuation range to form independent scenarios. After power flow calculation for each scenario based on feeder topology data, the voltage exceedance probability of the feeder is statistically analyzed. S4: Based on the ratio of the real-time remaining capacity of the feeder to the rated capacity, and in conjunction with the voltage over-limit probability, classify the risk level and mark the areas where new photovoltaic access is restricted; S5: Based on the annotation results of the restricted area, obtain the performance data of the power grid in the target area (including line impedance, node voltage level, and load fluctuation characteristics). Based on the feeder topology and power grid performance data, evaluate the adaptability of the existing power grid conditions to photovoltaic access through a multi-objective optimization model, identify the maximum accessible capacity threshold of the feeder, and generate a power grid access condition constraint matrix. The key indicators in the constraint matrix include transformer load factor, line voltage margin, and load fluctuation tolerance. S6: Based on the grid access condition constraint matrix, and by integrating the dynamic assessment results of remaining capacity and feeder topology parameters, dynamic access suggestions are generated for the exploitable capacity of distributed photovoltaic power in the target area.

[0030] The above steps quantify the risks of photovoltaic grid connection through Monte Carlo simulation, and combine dynamic assessment of remaining capacity with grid topology optimization to generate grid connection suggestions based on time periods and regions. This can effectively improve the economy and security of distributed photovoltaic grid connection and is suitable for distribution network scenarios with large load fluctuations and complex topology structures.

[0031] In step S1 of this embodiment, the target area power grid operation data includes historical load data of feeders (voltage, current, power factor, time resolution of 15 minutes), short-term power prediction data of distributed photovoltaics (power prediction value and spatiotemporal distribution characteristics, time resolution of 15 minutes), and feeder topology data (line impedance parameters). The feeder historical load dataset includes time-series data of voltage, current, and power factor, collected through a SCADA system or smart meters, with data stored in timestamp-value pairs format. The photovoltaic short-term forecast dataset is generated by a meteorological data-driven forecasting model and includes predicted power values, prediction timestamps, and prediction confidence intervals. The feeder topology dataset is obtained through a Geographic Information System (GIS) and a distribution network automation system, and includes line impedance parameters, node connection relationships, and geographical location information. When collecting power grid operation data for the target area, the specific steps include: Based on the power grid operation data of the target area, historical load data of feeders are collected, including operating parameters such as voltage, current, and power factor, with a time resolution of 15 minutes, to form a historical load dataset of feeders. Based on distributed photovoltaic short-term power prediction data, predicted power values ​​and their spatiotemporal distribution characteristics are collected with a time resolution of 15 minutes to construct a photovoltaic short-term prediction dataset, in which the prediction error range is controlled within ±10%. Based on feeder topology data, line impedance parameters are collected. Based on the relationship between nodes and their locations, a distribution network node-branch correlation matrix is ​​constructed to form a feeder topology dataset.

[0032] Step S2 of this embodiment preprocesses historical load data and photovoltaic forecast data to construct a real-time remaining capacity calculation model for the feeder. Specifically, it includes: 1) Collect feeder rated capacity data through the power grid operation monitoring system (Unit: kW) and historical peak load (Unit: kW), where the historical load peak data comes from the historical load curve of the feeder, the extraction period is the load peak data of the past 12 months, and the 95th percentile of the load peak of the past 12 months is taken; real-time operation data of the feeder is obtained through SCADA system or smart meter, combined with load characteristic analysis model, to determine the load peak fluctuation range of the feeder on a typical day, and generate the feeder load rate benchmark value.

[0033] 2) Obtain short-term power forecasts for distributed photovoltaic power plants. Historical load peak With photovoltaic predicted power Normalization is performed to remove outlier data (such as data points with load fluctuations exceeding ±15% or photovoltaic prediction errors exceeding ±10%). 3) By dynamically comparing the rated capacity of the feeder The combined value of historical load peaks and predicted photovoltaic power is used to generate the real-time remaining capacity of the feeder. The calculation formula is as follows:

[0034] in, This is the real-time remaining capacity value of the feeder. This is the rated capacity of the feeder. It is the short-term power forecast value of all distributed photovoltaic power. Peak values ​​of historical load data for feeders The maximum value in the sum; 4) The calculation results are dynamically updated using a sliding time window (set to 1 hour) to generate a feeder capacity time-series curve. Using the current moment as the window endpoint, historical load and photovoltaic (PV) forecast data are traced back 1 hour. Within the window, historical load data and short-term power forecast data for distributed PV are substituted into the feeder real-time remaining capacity calculation model at 15-minute intervals to calculate the feeder's real-time remaining capacity value. The window position is updated continuously for each time interval, generating a continuous remaining capacity time-series curve that reflects the dynamic changes in feeder capacity over time. The time-series curve is smoothed using time-series analysis tools (such as Python's Pandas library) to extract the fluctuation characteristics and trend information of the remaining capacity. Negative values ​​indicate insufficient feeder capacity, requiring the triggering of grid expansion or power dispatch strategies.

[0035] Step S3 in this embodiment includes: S301: Generate 1000 sets of random scenarios, considering a load fluctuation range of ±15% and a photovoltaic prediction error range of ±10%, including the following sub-steps: 1): Extract the standard deviation of load fluctuations based on historical load fluctuation data of the feeder. And adopt a normal distribution Random variables that generate load fluctuations This makes the load fluctuation range of :

[0036] in Based on historical load baseline values, ; 2) Based on the short-term forecast curve of distributed photovoltaic power, extract the historical standard deviation of photovoltaic forecast error. And adopt a normal distribution Random variables that generate photovoltaic power output error This makes the photovoltaic prediction fluctuation range of 0. :

[0037] in To predict power, ; 3) Generate 1000 scene groups through independent random sampling, respectively from... Distribution and Distribution Independently draw a random value ),Will and By index Pairing, forming the first The load value after considering errors for the group scenario is:

[0038] The photovoltaic prediction value after considering the error is:

[0039] Repeat the above steps 1000 times to generate 1000 independent scenarios. In each scenario, the combination of load and photovoltaic output satisfies the joint probability distribution characteristics. S302: Perform power flow calculations for each scenario group and count the number of scenarios where the line voltage deviation exceeds a preset threshold (±5%). This includes the following sub-steps: 1) Based on the line impedance parameters in the feeder topology data ( Based on the relationship between nodes and their locations, construct a node-branch correlation matrix for the distribution network; 2) The first generated in step S301 Load power in group scenarios and photovoltaic power forecast As node injected power, combined with the feeder rated voltage Establish the node power balance equations:

[0040] in, For nodes Complex power, For nodes voltage, For the elements of the node-branch admittance matrix, This represents the total number of nodes. 3) Solve for node voltages using the Newton-Raphson method. Calculate the voltage deviation at the end node of the line:

[0041] in, This is the rated voltage of the feeder. For nodes The percentage of voltage deviation; 4) Count the number of scenarios in which the line voltage deviation exceeds the preset threshold (set to ±5%). ; S303: Calculate the voltage over-limit probability based on the number of scenarios. :

[0042] This represents the probability that the line voltage deviation exceeds a preset threshold. The total number of scenes generated in step S301 (set to 1000 sets).

[0043] Step S4 of this implementation includes the following steps: S401) Risk Level Classification Standard: By dynamically comparing the ratio of the real-time remaining capacity to the rated capacity of the feeder, and combining the voltage over-limit probability generated by Monte Carlo simulation, a threshold cross-matching algorithm is used to classify the risk level. The risk level classification standards are as follows: High risk: the remaining capacity is less than 10% of the rated capacity of the feeder and the voltage over-limit probability is greater than 20%; Medium risk: the remaining capacity is 10%-30% of the rated capacity of the feeder and the voltage over-limit probability is greater than 10%; Low risk: the remaining capacity is greater than 30% of the rated capacity of the feeder and the voltage over-limit probability is less than 10%.

[0044] S402) Restricted Area Labelling: Using a Geographic Information System (GIS), the risk level is correlated with the feeder topology data to generate a visual heatmap, in which: High-risk areas are marked in red, with a warning message indicating that no new photovoltaic grid connection is allowed. Medium-risk areas are marked in orange and include a message indicating that a power grid upgrade plan is required. Low-risk areas are marked in green, along with recommended information that can be freely accessed.

[0045] In this embodiment, high-risk areas are prohibited from adding new photovoltaic (PV) power, marked with a red border on the heatmap, and accompanied by the text "Current feeder capacity is insufficient; priority should be given to capacity expansion or grid structure optimization." Low-risk areas are allowed to connect freely. Medium-risk areas require a supporting grid upgrade plan, including: 1) Transformer capacity expansion, the formula is:

[0046] in This refers to the expanded transformer capacity. This represents the current remaining capacity.

[0047] 2) Installation of reactive power compensation equipment, based on the voltage deviation percentage ( Calculate the required reactive power compensation amount. The formula is:

[0048] in This is the rated voltage of the feeder. This refers to the line reactance.

[0049] Step S5 of this embodiment evaluates grid adaptability based on feeder topology and grid performance data using a multi-objective optimization model, including the following steps: S501) Based on the access impact assessment information, the following data were collected and integrated: 1) Basic topology and performance data: i.e., the feeder topology data (line resistance) obtained in S1. Reactance Node-branch correlation matrix), real-time performance data of the target area power grid (line load rate, real-time node voltage values, reused with power flow calculation data of S3). 2) Dynamic capacity results: namely, the real-time remaining capacity time-series curve of the feeder output by S2 (including the fluctuation range of remaining capacity in different time periods, peak / valley values). 3) Risk quantification results: namely, the voltage over-limit probability output by S3 (statistical results by region and time series), the risk level classification results and restricted area labeling information output by S4 (clearly defining the boundaries and access constraints of high / medium / low risk sub-regions). 4) Load and photovoltaic characteristic data: namely, the historical load fluctuation data (σ1) of feeder S3 and the photovoltaic predicted fluctuation range (σ2) of S3, which are used to quantify the uncertainty constraints in the optimization model.

[0050] This yields performance data of the target area's power grid, including line impedance, node voltage levels, and load fluctuation characteristics. Combined with feeder topology and historical operating data, a power grid performance dataset is generated.

[0051] The data acquisition includes line impedance data: the resistance of the feeder lines is collected through the power grid monitoring system. and reactance The unit is Ω; Feeder topology and historical operating data: The feeder node-branch relationship matrix was obtained through a GIS system, and historical operating data (such as peak load and power factor) was integrated. Node voltage level data: including real-time voltage data of key feeder nodes (such as transformer outlets and load concentration points), in kV units, with a time resolution of 15 minutes; Load fluctuation characteristics data: Analysis of the volatility of historical load curves ( The calculation formula is:

[0052] in, Let be the load value for time period t. Where is the average load value, and T is the total number of time periods.

[0053] S502) Based on the collected power grid performance data, a multi-objective optimization model is constructed. The objectives are to maximize power grid security (voltage deviation and line load are within normal ranges) and minimize power grid transformation costs. ), Maximizing power grid stability index ( ).

[0054] 1) Objective 1: Maximize power grid security The core is to control node voltage deviation and line load rate within safe thresholds, combined with the power flow calculation results mentioned above (node ​​voltage deviation ΔV). i ) and grid performance data (line load factor) Feeder rated voltage The objective function is constructed as follows:

[0055] Parameter meaning explanation: : Power grid security evaluation index, with a value range of [0,1]. The closer it is to 1, the higher the power grid security level. : The total number of nodes in the target area power grid (from the "feeder topology data" of the power grid performance dataset, i.e., the total number of nodes in the node-branch correlation matrix); Indicates the first The voltage deviation of each node (unit: kV) is calculated as shown in S302. ),in For the first The actual voltage of each node (from the "Node Voltage Level Data" dataset of the power grid performance dataset); : Rated line voltage of feeder (unit: kV) (from "Feeder topology data" in the power grid performance dataset, which is a core performance parameter of feeder); 0.05 : Safety threshold for node voltage deviation (corresponding to a preset ±5% deviation, associated with the "node voltage level data" in the power grid performance dataset, and serving as the benchmark for voltage safety control); : The total number of feeder lines in the target area power grid (from the "feeder topology data" in the power grid performance dataset, i.e., the total number of branches in the node-branch association matrix). : No. Actual load rate of each feeder (unitless) (from "Line Load Rate Data" in the power grid performance dataset, collected in real time from the power grid operation monitoring system); The safe threshold for line load rate (unitless) is 80% (the industry standard safe threshold, which is associated with the "line load rate data" in the power grid performance dataset and is used to constrain line load from exceeding the limit).

[0056] 2) Minimize the cost of power grid transformation The cost of power grid upgrades mainly includes the cost of transformer capacity expansion and the cost of installing reactive power compensation equipment, both of which are related to the core parameters of the power grid performance dataset (transformer rated capacity). Line reactance Node voltage deviation Directly related to this, the objective function is constructed as follows:

[0057] Parameter meaning explanation: Total cost of power grid renovation (unit: 10,000 yuan). The smaller the value, the better the economic efficiency of the renovation. Transformer unit capacity expansion cost (unit: RMB 10,000 / kVA) (Industry standard parameter, not a power grid performance dataset parameter; the value can be adjusted according to the actual project, such as RMB 0.08 million / kVA). Transformer expansion capacity (unit: kVA) (compared to "Transformer rated capacity" in the power grid performance dataset) "Related, ,in (This refers to the original transformer's rated capacity, derived from "feeder topology data" in the power grid performance dataset). Unit capacity cost of reactive power compensation equipment (unit: RMB 10,000) (These are standard industry parameters, not parameters from the power grid performance dataset. Values ​​can be adjusted based on actual project conditions, e.g., 0.02 million yuan / ...) ); : No. Reactive power compensation capacity required for each node (unit: (and the "line reactance" data in the power grid performance dataset) "Node voltage level data ( The calculation method for the ")" association is shown in S402: ,in For the first The reactance of the line where each node is located comes from the "line impedance data" in the power grid performance dataset. Total number of nodes (from the "Feeder Topology Data" in the power grid performance dataset, same as Target 1).

[0058] 3) Maximize power grid stability indicators The core of power grid stability is related to line impedance and node voltage margin, combined with the "line impedance data" in the power grid performance dataset. , "Node voltage level data ( , The objective function is constructed as follows (reflecting line transmission capacity and voltage stability):

[0059] Parameter meaning explanation: : Power grid stability index, with a value range of [0,1]. The closer it is to 1, the higher the power grid stability. Total number of feeder lines (from "Feeder Topology Data" in the power grid performance dataset, same as Target 1); Resistance and reactance of the l-th feeder (unit: Ω) (from "Line Impedance Data" in the power grid performance dataset, i.e., line impedance parameters) , ); Safety thresholds for line resistance and reactance (unit: Ω) (from the "Line Impedance Data" in the power grid performance dataset, representing the maximum allowable impedance during feeder design, which is a core performance parameter of the power grid). : No. Voltage margin of each node (unit: kV) (correlated with "node voltage level data" in the power grid performance dataset). ,in , , (Rated line voltage, from a power grid performance dataset). 0.05 : The maximum allowable value for node voltage margin (corresponding to ±5% voltage deviation, associated with the "node voltage level data" in the power grid performance dataset); Total number of nodes (from the "Feeder Topology Data" in the power grid performance dataset, same as Target 1).

[0060] 4) Comprehensive form of multi-objective optimization model Combining the three objectives, the comprehensive expression of the multi-objective optimization model is as follows (using a linear weighting method, the weights can be adjusted according to the actual engineering situation, and the sum of the weights is 1):

[0061] in: The weights of the three objectives are respectively ( ), can be adjusted according to the power grid dispatch priority (e.g., when power grid safety takes priority, , , ); These are the objective functions for the three objectives mentioned above.

[0062] The constraints include grid performance constraints (upper limit of line impedance, allowable deviation range of node voltage, load fluctuation rate) and photovoltaic access constraints (upper limit of photovoltaic installed capacity, location of access node).

[0063] Considering the upper limit of line impedance, the allowable deviation range of node voltage, and the load fluctuation rate limit, the following constraints are set:

[0064] in For the safety impedance threshold, For the load safety threshold, Rated line voltage, According to the S4 risk level setting, low-risk areas medium-risk area Access is prohibited in high-risk areas.

[0065] Simultaneously, photovoltaic access constraints are added, as shown in the following formula:

[0066] in, It is 80% of the rated capacity of the line transformer.

[0067] S503) The Pareto optimal solution set is solved by the NSGA-II algorithm, and feasible solutions that meet the requirements of safe operation of the power grid are selected.

[0068] Specifically, after setting the constraints, an optimization model is built using Python's Pyomo library, and multi-objective optimization is performed by calling the NSGA-II algorithm solver. The NSGA-II algorithm solves the multi-objective optimization model, generating a Pareto optimal solution set, and then selecting feasible solutions that meet the requirements for safe power grid operation. The NSGA-II algorithm is configured as follows: population size: 100; number of iterations: 200; crossover probability: 0.9; mutation probability: 0.1. Set the Pareto optimal solution set selection rules: 1) Power grid safety first: Screening based on voltage deviation ≤±5% and line load rate ≤80% of the solutions; 2) Economic optimization: Selecting the modification cost from the safe solution set. The lowest 20% of solutions; 3) Stability Verification: Verify the power grid stability indices of the solution set through power flow calculations. ≥0.8).

[0069] The feasible solution in this embodiment is a complete combination of grid operation and photovoltaic access parameters, which can be directly used for the calculation of maximum grid-accessible capacity and the generation of subsequent dynamic access suggestions. Specifically, it includes the following core parameters: 1) Core parameters for photovoltaic grid connection: Candidate value of the maximum installed capacity of the photovoltaic system to be connected. (Unit: kW), Optimal Access Node Number (The node number corresponds to the node number in the feeder topology, derived from the node-branch correlation matrix). 2) Power grid operating status parameters: steady-state voltage values ​​at each node (Unit: kV, satisfying node voltage constraints) ), load rate of each feeder line (No unit, meets the requirements) ), voltage over-limit probability (No unit, meets the constraint requirements of the corresponding risk area); 3) Parameters related to power grid upgrade: Required transformer expansion capacity (Unit: kVA, 0 if no expansion is needed) Required reactive power compensation capacity for each node (unit: (If the voltage margin meets the requirements, then it is 0). 4) Constraint adaptation parameters: Real-time remaining capacity of the feeder during the current time period. (Unit: kW, from the time-series curve in step S2), rated capacity of transformer or transmission line (Unit: kVA, from power grid performance dataset) Current actual load (Unit: kW, from real-time power grid operation data), candidate values ​​for safety margin factor (range of values) (Dynamically determined based on power grid security priority).

[0070] The selection process for feasible solutions must strictly adhere to preset rules, and the final selected feasible solutions must simultaneously satisfy: grid security constraints (voltage deviation, load factor, voltage over-limit probability), economic constraints (lowest retrofit cost), and stability constraints (grid stability indicators). This ensures that the selected feasible solutions have practical engineering application value, and that all parameters can be directly obtained through the data acquisition and model calculation process described above.

[0071] Finally, the maximum accessible capacity is calculated using the following formula:

[0072] in, Rated capacity of transformers or transmission lines (unit: kVA); The current load (unit: kW) is taken from the grid operating state parameters in the feasible solution, corresponding to the steady-state value (unit: kW) of the "current actual load" in the filtered feasible solution. Its value must satisfy the load factor constraint in the feasible solution. ,Right now This ensures that the current load is within the safe operating range of the power grid and reserves space for the capacity that can be connected. A safety margin factor (ranging from 0.1 to 0.2) is included in the feasible solution. The optimal value (0.1~0.2) is dynamically determined by the feasible solution selection process—combining the voltage over-limit probability in the feasible solutions. Node voltage margin These parameters ensure that the power grid can still meet the voltage exceedance probability constraint (low-risk area) after photovoltaic grid connection. medium-risk area To avoid problems such as grid voltage fluctuations and excessive load rates caused by photovoltaic grid connection; maximum connectable capacity The calculation process essentially involves analyzing the "core parameters for photovoltaic grid connection" in the feasible solution. Verification and confirmation. Among the filtered feasible solutions, The maximum value needs to be calculated by the formula. Exact match, i.e. Ensure that the photovoltaic power connected to the grid does not exceed the grid's carrying capacity limit. If there are multiple candidate values ​​that satisfy the constraints in the feasible solution (such as different safety margin factors corresponding to...), Then, the formula is used to calculate... The optimal value is determined by selecting the candidate value with the lowest transformation cost and the highest grid stability among the feasible solutions, substituting it into the formula for verification, and finally determining the unique maximum access capacity threshold.

[0073] Generate the power grid connection condition constraint matrix, and calculate the matrix as follows:

[0074] In step S6 of this embodiment, the dynamic access suggestion generation rules are as follows: In this step, the grid access constraint matrix serves as the quantitative verification benchmark for photovoltaic access. Its key indicators include transformer load factor, line voltage margin, and load fluctuation tolerance. Each indicator corresponds to a preset safety threshold, and the real-time values ​​of these indicators are updated synchronously from multi-objective optimization feasible solutions, power flow calculation results, and remaining capacity time-series curves. The generation of dynamic access suggestions must be based on this constraint matrix. The steps include matrix indicator threshold judgment, joint screening of remaining capacity and topology parameters, and quantitative output of differentiated suggestions, specifically including the following sub-steps: S601: Extract the real-time index values ​​and safety thresholds of the power grid access condition constraint matrix to complete the preliminary verification of access feasibility; Extracting transformer load rate from constraint matrix Line voltage margin Tolerance for load fluctuations The real-time values ​​are compared with the preset safety thresholds. , , The comparison identifies and marks indicators that exceed the limits, serving as the core basis for subsequent adjustments to the grid connection recommendations; among these, transformer load rate is correlated with remaining capacity. The probability of voltage exceedance is related to the load distribution of topology nodes and the line voltage margin. Related to topology line impedance, load fluctuation tolerance, remaining capacity time-series fluctuation, and photovoltaic forecasting error .

[0075] S602: Combine the constraint matrix and the dynamic evaluation results of remaining capacity to generate time-based access power thresholds; "Constraint fitting parameters" in feasible solutions Used to generate time-segmented access power thresholds—for different time periods in the feasible solution. The time series value corresponds to the division of high, medium, and low remaining capacity periods, and is calculated using a formula. To determine the upper limit of access power for each time period, based on the high, medium, and low time periods defined by the remaining capacity level, and combined with the constraint matrix indicator verification results, the access power for each time period is quantitatively calculated and corrected to clarify the access power threshold and access requirements for each time period, specifically: 1) High remaining capacity period ( >30%): If the constraint matrix satisfies , , The period is determined to be a feasible time for rapid access; the formula for calculating the upper limit of access power is as follows: Finally, the calculated value is taken as the transformer's rated capacity. The minimum value ensures that the transformer load rate does not exceed [a certain value] after connection. A safety threshold allows for rapid grid connection of solar power.

[0076] 2) Remaining capacity period ( The key focus is on verifying the line voltage margin and load fluctuation tolerance. or The time period is determined to require dynamic adjustment of the access time; the recommended access power value is the initial upper limit value multiplied by the voltage margin correction factor. ), load fluctuation correction factor ( ( To ensure the probability of voltage exceeding limits after connection. Furthermore, priority will be given to adjusting access during periods of low load fluctuation; 3) Periods with low remaining capacity ( At this point, if at least one indicator in the constraint matrix exceeds the limit, it triggers the access power limit, and it is recommended to delay access until a period with high remaining capacity. Simultaneously, based on the type of indicator exceeding the limit in the matrix, the criteria for determining delayed access should be clearly defined, such as the transformer load rate. Then it will be delayed until the load drops to At that time; if the line voltage margin The power generation will be delayed until the off-peak period of photovoltaic power output.

[0077] S603: Integrate constraint matrix and feeder topology parameters to generate topology-differentiated access suggestions; Based on feeder topology parameters (line impedance, node connection path redundancy, node load distribution), the "optimal access node number" in the feasible solution... This is directly used to generate topology-differentiated access suggestions—for regions with simple topology structures, feasible solutions are available. The maximum access power for the corresponding load center node is: (80% of the transformer's rated capacity); for regions with complex topologies, feasible solutions include... For nodes with higher voltage margins, the maximum connected power is: (50% of the transformer's rated capacity), specifically: 1) Regions with simple topology (line impedance) High redundancy in node connection paths: The line voltage loss in this area is low, so load center nodes are preferred for connection (load density is confirmed through the node-branch correlation matrix). (Area average density 1.2 times); After connection, the node line voltage margin must be met. The load rate of the transformer Priority access to photovoltaic power is permitted, with the maximum access power capped at 80% of the transformer's rated capacity. 2) Regions with complex topology (line impedance) (Single node connection path): The voltage margin of the lines in this area is prone to insufficient, and the risk of voltage exceeding the limit needs to be strictly controlled. The upper limit of the connected power is set at 50% of the rated capacity of the transformer; at the same time, the required reactive power compensation is calculated based on the line voltage margin index of the constraint matrix. It is required to provide reactive power compensation equipment of appropriate capacity to ensure that the node voltage deviation is correct after connection. Voltage over-limit probability ; S604: Integrate time-based access power thresholds and topology-differentiated access suggestions to generate final dynamic access suggestions. By integrating the quantitative results of S602 and S603, a complete dynamic access recommendation for the exploitable capacity of distributed photovoltaic power in the target area is formed, which clarifies the time-of-use access power limit, access time requirements, topology-adaptive access nodes, power limiting conditions, and supporting equipment requirements. At the same time, the grid access condition constraint matrix index verification results corresponding to each recommendation are marked, forming a traceable correlation between the recommendations and quantitative indicators, providing a clear basis for grid dispatch and photovoltaic access implementation.

[0078] Furthermore, this embodiment also proposes a distributed photovoltaic (PV) capacity dynamic prediction system based on Monte Carlo simulation, including a processor and a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by the processor, implements the steps of the distributed PV capacity dynamic prediction method based on Monte Carlo simulation described in this embodiment.

[0079] Example 2: This embodiment provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores a computer program executable by the processor, which, when executed by the at least one processor, implements the previously described method for dynamic prediction of the available capacity of distributed photovoltaic power generation based on Monte Carlo simulation. This electronic device can be deployed in a power grid dispatch center to process photovoltaic access data in real time and generate dynamic access suggestions, supporting power grid operation decisions.

[0080] Example 3: This embodiment provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the previously described method for dynamically predicting the available capacity of distributed photovoltaic (PV) power grids based on Monte Carlo simulation. This storage medium can be a USB flash drive, hard drive, cloud storage, etc., and is used to store and distribute the PV access prediction model for use and deployment by the power grid management department.

[0081] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0082] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation, characterized in that, Includes the following steps: Acquire historical load data of feeders, short-term power forecast data of distributed photovoltaic power, and feeder topology data of the target area power grid; A real-time remaining capacity calculation model for feeders is constructed. A sliding time window is used to dynamically calculate the real-time remaining capacity of feeders based on historical load data and short-term power prediction data of distributed photovoltaics, resulting in a time-series curve of the real-time remaining capacity of feeders. Based on historical load data of the feeder and short-term power prediction data of distributed photovoltaics, the load fluctuation range and photovoltaic prediction fluctuation range are determined respectively. The Monta Carlo simulation algorithm is used to extract samples from the load fluctuation range and photovoltaic prediction range to form independent scenarios. After power flow calculation for each scenario based on feeder topology data, the voltage over-limit probability of the feeder is statistically analyzed. Based on the ratio of the real-time remaining capacity of the feeder to its rated capacity, and combined with the voltage over-limit probability, risk levels are classified and areas where new photovoltaic access is restricted are marked. Based on the annotation results of the restricted area, the performance data of the power grid in the target area is obtained. Based on the feeder topology and power grid performance data, the adaptability of the existing power grid conditions to photovoltaic access is evaluated through a multi-objective optimization model, and the maximum accessible capacity threshold of the feeder is identified.

2. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, The mathematical expression for the feeder real-time remaining capacity calculation model is as follows: in, This is the real-time remaining capacity value of the feeder. This is the rated capacity of the feeder. It is the short-term power forecast value of all distributed photovoltaic power. Peak values ​​of historical load data for feeders The maximum value in the sum.

3. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, When using a sliding time window to dynamically calculate the real-time remaining capacity of a feeder based on historical feeder load data and short-term power forecast data of distributed photovoltaics, the process involves using the current moment as the end point of the window, backtracking the historical feeder load data and short-term power forecast data of distributed photovoltaics for the corresponding duration according to the window size, and substituting the historical feeder load data and short-term power forecast data of distributed photovoltaics into the feeder real-time remaining capacity calculation model for each time period within the window to calculate the real-time remaining capacity value of the feeder, and updating the window position on a rolling basis for each time period.

4. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, Based on historical load data of the feeder and short-term power forecast data of distributed photovoltaics, the load fluctuation range and photovoltaic forecast fluctuation range are determined respectively. When using the Monta Carlo simulation algorithm to extract samples from the load fluctuation range and photovoltaic forecast range to form independent scenarios, the following steps are included: Based on historical load data of the feeder, the standard deviation of load fluctuation is extracted. And adopt a normal distribution Random variables that generate load fluctuations This makes the load fluctuation range of : in The historical load benchmark value is the feeder load rate benchmark value generated by obtaining real-time operating data of the feeder and combining it with the load characteristic analysis model to determine the peak load fluctuation range of the feeder on a typical day. Based on distributed photovoltaic short-term power forecast data, the historical standard deviation of photovoltaic forecast error is extracted. And adopt a normal distribution Random variables that generate photovoltaic power output error This makes the photovoltaic prediction fluctuation range of 0. : in This represents the short-term power forecast for distributed photovoltaic systems. A specified number of scenarios are generated through independent random sampling, respectively from... Distribution and Distribution Independently draw a random value ,Will and By index Pairing, forming the first The scenarios are grouped so that the combination of load and photovoltaic output in each scenario satisfies the joint probability distribution characteristics.

5. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, When calculating the voltage over-limit probability of the feeder after performing power flow calculations for each scenario based on feeder topology data, the following steps are included: Based on the line impedance parameters and node location relationships in the feeder topology data, construct the distribution network node-branch correlation matrix; The load power and photovoltaic predicted power in each scenario are used as the node injected power. The node power balance equation is established and the node voltage is solved. Then, the node voltage deviation is calculated based on the node voltage value and the feeder rated voltage value. Finally, the number of scenarios with deviations exceeding the preset threshold is counted, and the number of scenarios is divided by the total number of scenarios to obtain the voltage over-limit probability.

6. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, When classifying risk levels and marking areas where new photovoltaic grid connection is restricted based on the ratio of the feeder's real-time remaining capacity to its rated capacity, combined with the voltage over-limit probability, the following steps are included: If the real-time remaining capacity of the feeder is less than 10% of the rated capacity of the feeder and the voltage over-limit probability is greater than 20%, the corresponding area is a high-risk area. The high-risk area will be marked in red and a warning message prohibiting the addition of new photovoltaic grid connection will be displayed. If the real-time remaining capacity of the feeder is 10%-30% of the feeder's rated capacity and the voltage over-limit probability is greater than 10%, the corresponding area is a medium-risk area. The medium-risk area will be marked in orange and a prompt message indicating that a grid transformation plan is required will be added. If the real-time remaining capacity of the feeder is greater than 30% of the rated capacity and the voltage over-limit probability is less than 10%, the corresponding area is a low-risk area. The low-risk area is marked in green and labeled with recommended information that it can be freely connected.

7. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, The objectives of the multi-objective optimization model are to maximize power grid security, minimize power grid transformation costs, and maximize power grid stability indicators. The objective function is as follows: in, These are the objective functions for maximizing power grid security, minimizing power grid upgrade costs, and maximizing power grid stability indicators, respectively. These are the weights of the three objectives; The objective function for maximizing power grid security is as follows: in, This represents the total number of nodes in the target area's power grid. This represents the voltage deviation at the i-th node; This is the rated line voltage of the feeder; This refers to the total number of feeder lines in the target area's power grid; It is the first The actual load rate of the feeder; It is the safe threshold for line load rate; The objective function for minimizing the cost of power grid upgrades is as follows: in, It is the unit capacity expansion cost of transformers; It refers to the expansion of transformer capacity. ,in This is the original rated capacity of the transformer; It is the unit capacity cost of reactive power compensation equipment; It is the first The reactive power compensation capacity required for each node ,in For the first The reactance of the line where each node is located; It is the total number of nodes; The objective function for maximizing the power grid stability index is as follows: in, This represents the total number of feeder lines; These are the resistance and reactance of the l-th feeder, respectively. These are the safety thresholds for line resistance and reactance, respectively. It is the first Voltage margin of each node, ,in , , This is the rated line voltage; 0.05 It is the maximum allowable value for node voltage margin; It is the total number of nodes; The constraints of the multi-objective optimization model include: in, For the safety impedance threshold, For the load safety threshold, Rated line voltage, Based on risk level; The constraints of the multi-objective optimization model also include photovoltaic access constraints: in, It is 80% of the rated capacity of the line transformer.

8. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 1, characterized in that, When evaluating the adaptability of existing power grid conditions to photovoltaic (PV) grid integration using a multi-objective optimization model and identifying the maximum connectable capacity threshold for feeders, the multi-objective optimization model is solved using the NSGA-II algorithm to generate a Pareto optimal solution set. Feasible solutions that meet the requirements for safe grid operation are then selected. Based on the selection results of feasible solutions, the maximum connectable capacity is calculated using the following formula: in, The rated capacity of the transformer or transmission line; Current load; SF Leave room for safety.

9. The method for dynamic prediction of the available capacity of distributed photovoltaic power based on Monte Carlo simulation according to claim 8, characterized in that, After identifying the maximum accessible capacity threshold for the feeder, the following steps are also included: Generate a grid access condition constraint matrix, in which key indicators include transformer load factor, line voltage margin, and load fluctuation tolerance. Based on the grid connection constraint matrix, and integrating the dynamic assessment results of remaining capacity and feeder topology parameters, dynamic connection suggestions for the exploitable distributed photovoltaic capacity within the target area are generated, including: Extract real-time index values ​​and safety thresholds from the power grid access condition constraint matrix, conduct preliminary verification of access feasibility, and mark indexes that exceed the limits, which will serve as the core basis for subsequent access recommendation corrections. By integrating the constraint matrix and the dynamic evaluation results of remaining capacity, time-segmented access power thresholds are generated. Specifically, the upper limit of access power for each time period is determined based on the maximum accessible capacity threshold of the feeder and the constraint adaptation parameters in the feasible solution. Based on the high, medium, and low time periods divided by the remaining capacity level, and combined with the constraint matrix index verification results, the access power for each time period is quantitatively calculated and corrected to clarify the time-segmented access power thresholds and access requirements. By integrating the constraint matrix and feeder topology parameters, topology-differentiated access suggestions are generated. Specifically, topology-differentiated access suggestions are generated based on the feeder topology parameters and the optimal access node number in the feasible solution. Integrate time-based access power thresholds and topology-differentiated access suggestions to generate final dynamic access suggestions.

10. A dynamic prediction system for the open capacity of distributed photovoltaic power based on Monte Carlo simulation, characterized in that, The device includes a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program that, when executed by the processor, implements the steps of the method for dynamic prediction of the available capacity of distributed photovoltaic systems based on Monte Carlo simulation as described in any one of claims 1 to 9.

Images