Dynamic description method for running state interval of high-proportion photovoltaic power distribution system
By constructing a dynamic adjustment mechanism, the net power time-series data of photovoltaic access nodes is converted into relative angles for data point identification and time-series features are iteratively extracted. This solves the efficiency and accuracy problem of characterizing the uncertainty of distribution network operation status in high-proportion photovoltaic distribution networks and achieves rapid and accurate quantitative assessment of the status.
Patent Information
- Application Number
- CN202510962456.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-14
- Publication Date
- 2025-11-28
AI Technical Summary
Existing technologies struggle to characterize the uncertainties of distribution network operation in real time and dynamically in distribution networks with a high proportion of photovoltaic (PV) integration, especially when PV output is difficult to predict and node net power fluctuates randomly. Existing methods are limited in computational efficiency and accuracy.
By constructing a dynamic adjustment mechanism, the net power time series data of distributed photovoltaic access nodes is converted into relative angles for data point identification, time series features are iteratively extracted, net power trends and interval boundaries are constructed, and distribution network state mapping is established to achieve rapid quantitative assessment.
It enables rapid quantitative assessment of the distribution network operation status in high-proportion photovoltaic distribution systems, improves computational efficiency and accuracy, and can dynamically capture state changes under frequent photovoltaic fluctuation scenarios.
Smart Images
Figure CN121036166A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of power distribution network state interval dynamic characterization, in particular to a high proportion photovoltaic power distribution system operation state interval dynamic characterization method. BACKGROUND
[0002] With the increasing proportion of distributed photovoltaic power in the power distribution network, the difficulty in predicting its output and the random fluctuations of node net power caused thereby significantly enhance the uncertainty of the operation state of the power distribution network. How to effectively characterize the uncertainty of the power distribution state has become a key problem in the current quantitative evaluation of power distribution network uncertainty.
[0003] In order to realize the quantification of the uncertainty of the power distribution network state, the current mainstream methods can be roughly divided into two categories according to the characteristics of uncertain variables and states: probability method and interval method. The probability method maps these uncertain variables into the probability characteristics of the power distribution network state by constructing the probability density function of the uncertain variables and combining Monte Carlo simulation, point estimation method or analytical method. However, this kind of method relies on a large amount of historical statistical data and a long sampling period, and it is difficult to meet the demand of real-time dynamic characterization of the operation state of the power distribution network.
[0004] The interval method is more suitable for scenarios with limited statistical data and period. Common interval methods include iterative method, optimization method and Taylor expansion method. These methods usually express uncertain variables in an interval affine form, and construct an explicit interval mapping from variables to states through a specific algorithm. However, the precision and computational efficiency of the interval result are greatly limited by the mapping error and the algorithm design itself. In addition, since these methods often rely on a unique state mapping model constructed under a certain fluctuation interval, they need to frequently update the baseline operating point and the mapping relationship when facing the scenario of frequent fluctuations of photovoltaic power during the day, which seriously restricts the computational efficiency in dynamic scenarios. SUMMARY
[0005] The purpose of the present application is to overcome the shortcomings of the prior art and provide a high proportion photovoltaic power distribution system operation state interval dynamic characterization method, which can improve the rapid quantitative evaluation level of the operation state of the power distribution network.
[0006] The technical problem of the present application is solved by adopting the following technical solution: The high proportion photovoltaic power distribution system operation state interval dynamic characterization method comprises the following steps: Step 1, according to the selected active power distribution network, input the net power time series data of the distributed photovoltaic access node in the continuous previous Step 2, set the net power of all distributed photovoltaic access nodes as an uncertain variable vector Node voltage, line active power, and line reactive power are used as indicators of the distribution network status. Set mapping parameters and dynamic adjustment mechanism parameters; Step 2: Based on the net power time series data of the distributed photovoltaic access node in Step 1, convert each net power time series data into a relative angle in each time period, identify data points based on the relative angle, and extract time series features through data point iteration to obtain the net power trend and interval boundary of the distributed photovoltaic access node. Step 3, based on the continuous preceding The net power trend and interval boundaries of distributed photovoltaic access nodes in each time period are analyzed to determine volatility, and a dynamic adjustment mechanism is constructed to obtain the next unknown time period. Net power range boundary prediction values; Step 4: Based on the next unknown time period obtained in Step 3 The net power interval boundary prediction value is used to establish a distribution network state mapping, and the distribution network state interval boundary is calculated based on the state mapping to output the next unknown time period. Prediction range of distribution network status.
[0007] Furthermore, the mapping parameters in step 1 include: the number of uncertain variables. The highest order of the polynomial used in the state transition function binary conservative parameters The dynamic adjustment mechanism parameters include: the threshold for determining the smooth point. Threshold for determining turning points Feature extraction threshold Volatility determination threshold Time period length adjustment amount Maximum time period length Minimum time period length Maximum interval conservative coefficient Minimum interval conservative coefficient Interval Conservative Coefficient Adjustment .
[0008] Furthermore, in step 1, the input of continuous preceding... The specific implementation method for obtaining the net power time-series data of distributed photovoltaic access nodes for each time period is as follows:
[0009] in, This is a sequence of node net power values. express The Middle Net power measurement value This represents the number of measurement points during that time period.
[0010] Furthermore, step 2 includes the following steps: Step 2.1: Convert the net power time-series data into relative angles within each time period: This is done by transforming the angles of two consecutive points. Equivalent transformation :
[0011] in, As the initial point Relative angle Equivalent time series data, relative angle for:
[0012] in, and They are respectively The front absolute angle and the rear absolute angle; Step 2.2: Identify data points based on relative angles:
[0013] In the formula, For a smooth point set, For the set of turning points, for The Middle One feature point, for In sequence The position in the middle, for The Middle One feature point, for In sequence The position in the middle, and The table is divided into sub-tables. and number of elements Step 2.3: Iteratively extract temporal features: If the threshold is determined based on the flat point Calculated feature extraction rate Less than the given feature extraction threshold Then determine The trend characteristics are not significant at present, so the threshold adjustment amount is used. Gradually increase the threshold for determining smooth points and update the feature extraction rate. Until , obtain Trend characteristics; Step 2.4: Calculate the net power trend and interval boundaries of the distributed photovoltaic access nodes:
[0014] in, The net power trend of distributed photovoltaic access nodes. and These are the upper and lower bounds of the net power range for distributed photovoltaic access nodes, respectively.
[0015] Furthermore, step 3 includes the following steps: Step 3.1: Determine the volatility of the net power trend and interval boundaries of the distributed photovoltaic access nodes:
[0016] in, and Let them represent upward volatility variables and downward volatility variables, respectively. Indicates the time period The vector formed by the net power trends of each distributed photovoltaic access node. and They are respectively in the time period The vector formed by the upper and lower bounds of the net power range of each distributed photovoltaic access node. They are consecutive front A matrix consisting of the net power trends of each distributed photovoltaic access node over a given time period. and They are consecutive front A matrix consisting of the upper and lower bounds of the net power range of each distributed photovoltaic access node in each time period; Step 3.2: Construct a dynamic adjustment mechanism based on the fluctuation determination results: Adjust the next time period length according to volatility.
[0017] in, For the next unknown period The length of the time period For the current time period The length of the time period This is the adjustment amount for the duration of the time period. upward volatile variable and downward volatility variables A unified form; Step 3.3: Obtain the next unknown time period based on the constructed dynamic adjustment mechanism. Net power range boundary prediction values:
[0018] in, and For the next unknown period The upper and lower bounds of the net power prediction interval for distributed photovoltaic (PV) grid-connected nodes. and They are respectively in the time period The upper and lower bounds of the range of net power of distributed photovoltaic access nodes.
[0019] Furthermore, step 4 includes the following steps: Step 4.1: Based on the obtained next unknown time period The net power range boundary prediction value is used to establish the distribution network state mapping;
[0020] in, , , These represent composite functions, linear functions, and coupled functions, respectively. Represents the state transition function. express inverse function, and The first and the An uncertain variable, express coefficient, express coefficient, express The coefficient; Step 4.2: Calculate the boundary of the distribution network state interval based on state mapping: Calculate the upper and lower bounds of the composite function prediction interval for each state; The method for calculating the upper and lower bounds of the prediction interval of the composite function of node voltage and line active power is as follows:
[0021] in, and For the next unknown period Upper and lower bounds of the prediction interval for a composite function. and For the next unknown period No. The upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period No. Correction values for the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period A vector consisting of the correction values of the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node, including subscripts. and The parameters are the mapping parameters for overestimation and underestimation, respectively. and They represent overestimation and underestimation of the mapping, respectively. coefficient, and These are the composite functions in the overestimation and underestimation mappings, respectively; The calculation method for the upper and lower bounds of the composite function of reactive power of the line is as follows:
[0022] in, and For the next unknown period The vector formed by the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. This is the transpose operation for a matrix or vector. For the next unknown period The vector formed by the various uncertain variables, and The coefficient vectors of the coupled function and the linear function respectively Step 4.3: Based on the calculated boundaries of the distribution network state intervals, after... Processing and predicting the distribution network status for the next period. interval :
[0023] in, and For the next unknown period The upper and lower bounds of the prediction interval for the distribution network status. and They represent overestimation and underestimation of the mapping, respectively. The inverse function of .
[0024] The advantages and positive effects of this invention are: This invention, based on selected active power distribution network input parameters, and using the net power time-series data of distributed photovoltaic (PV) access nodes within these parameters, transforms each net power time-series data into a relative angle within each time period. Data points are identified based on these relative angles, and time-series features are iteratively extracted to obtain the net power trend and interval boundaries of the PV access nodes. Volatility is then assessed, and a dynamic adjustment mechanism is constructed to obtain the predicted net power interval boundary value for the next unknown time period. A distribution network state mapping is established, and the distribution network state interval boundaries are calculated based on the state mapping, outputting the predicted interval of the distribution network state for the next unknown time period. This invention addresses the real-time quantification problem of distribution network state uncertainty, fully utilizing historical measurement data to dynamically model and adjust the net power of nodes connected to distributed PV. Simultaneously, the net power interval boundaries constructed based on this dynamic mechanism are transformed into interval boundaries of state variables (node voltage and line power) through decoupling mapping technology, ultimately achieving rapid quantitative assessment of the distribution network operating state. Attached Figure Description
[0025] Figure 1 This is a flowchart of the present invention; Figure 2 This is a diagram of the improved IEEE 33 architecture used in this invention; Figure 3 This is a schematic diagram of the node net power fluctuation curve and the results of interval dynamic identification and adjustment in this invention; Figure 4 This is a schematic diagram comparing the fluctuation curves and interval results of the voltage at node 17 and the power between line 2 and line 3 during the period from 6:00 to 12:00. Figure (a) is a schematic diagram of the node voltage range; Figure (b) is a schematic diagram of the line active power range; Figure (c) is a schematic diagram of the line reactive power range. Figure 5 This is a schematic diagram comparing the voltage of all nodes and line power across the network and the results of the interval during the time period of 10:30 to 10:35, according to an embodiment of the present invention. Figure (a) is a schematic diagram of the voltage range of nodes in the entire network; Figure (b) is a schematic diagram of the active power range of lines in the entire network; Figure (c) is a schematic diagram of the reactive power range of lines in the entire network. Detailed Implementation
[0026] The present invention will be further described in detail below with reference to the accompanying drawings.
[0027] A method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system, such as... Figure 1 As shown, it includes the following steps: Step 1: Based on the selected active distribution network, input the continuous... The net power time-series data of distributed photovoltaic (PV) access nodes for each time period; the net power of all distributed PV access nodes is set as a vector of uncertain variables. Node voltage, line active power, and line reactive power are used as indicators of the distribution network status. Set the mapping parameters and dynamic adjustment mechanism parameters.
[0028] The mapping parameters set include: the number of uncertain variables. The highest order of the polynomial used in the state transition function binary conservative parameters The parameters set for the dynamic adjustment mechanism include: the threshold for determining the smooth point. Threshold for determining turning points Feature extraction threshold Volatility determination threshold Time period length adjustment amount Maximum time period length Minimum time period length Maximum interval conservative coefficient Minimum interval conservative coefficient Interval Conservative Coefficient Adjustment .
[0029] consecutive previous The method for constructing the net power time series data of distributed photovoltaic access nodes for each time period is as follows: in the consecutive previous The net power time-series data of the distributed photovoltaic access nodes for each time period is as follows:
[0030] in, This is a sequence of node net power values. express The Middle Net power measurement value This represents the number of measurement points during that time period.
[0031] Step 2: Based on the net power time series data of the distributed photovoltaic access nodes in Step 1, convert each net power time series data into a relative angle in each time period, identify data points based on the relative angle, iteratively extract time series features, and obtain the net power trend and interval boundary of the distributed photovoltaic access nodes. Step 2.1: Based on the net power time series data of the distributed photovoltaic access nodes, convert each net power time series data into a relative angle in each time period; By changing the angle of two consecutive points, Equivalent transformation :
[0032] in, As the initial point Relative angle Equivalent time series data, relative angle Represented as:
[0033] in, and They are respectively The front absolute angle and the back absolute angle.
[0034] Step 2.2: Identify data points based on relative angles:
[0035] in, For a smooth point set, For the set of turning points, for The Middle One feature point, for In sequence The position in the middle, for The Middle One feature point, for In sequence The position in the middle, and The table is divided into sub-tables. and The number of elements.
[0036] Step 2.3: Iteratively extract temporal features; The rule for iteratively extracting temporal features is as follows: if the threshold is determined based on the flat point... Calculated feature extraction rate Less than the given feature extraction threshold Then determine The trend characteristics are not significant at present, so the threshold adjustment amount is used. Gradually increase the threshold for determining smooth points and update the feature extraction rate. Until to obtain The trend characteristics. For those satisfying the feature extraction rate... flat point set Using the mean as trend Take the maximum and minimum values as interval .
[0037] Step 2.4: Based on the extracted time-series features, obtain the net power trend and interval boundaries of the distributed photovoltaic access nodes:
[0038] in, The net power trend of distributed photovoltaic access nodes. and These are the upper and lower bounds of the net power range for distributed photovoltaic access nodes, respectively.
[0039] Step 3, based on the continuous preceding The net power trend and interval boundaries of distributed photovoltaic access nodes in each time period are analyzed to determine volatility, and a dynamic adjustment mechanism is constructed to obtain the next unknown time period. Net power range boundary prediction values; Step 3.1, based on the continuous preceding... The volatility is determined by analyzing the net power trend and interval boundaries of distributed photovoltaic access nodes over a given time period. During the period When determining volatility, volatility can be categorized into upward volatility and downward volatility:
[0040] in, and Let them represent upward volatility variables and downward volatility variables, respectively. Indicates the time period The vector formed by the net power trends of each distributed photovoltaic access node. and They are respectively in the time period The vector formed by the upper and lower bounds of the net power range of each distributed photovoltaic access node. They are consecutive front A matrix consisting of the net power trends of each distributed photovoltaic access node over a given time period. and They are consecutive front A matrix consisting of the upper and lower bounds of the net power range of each distributed photovoltaic access node in each time period.
[0041] Step 3.2: Based on the volatility assessment results, construct a dynamic adjustment mechanism: The length of the next time period is adjusted according to volatility as follows:
[0042] in, For the next unknown period The length of the time period For the current time period The length of the time period This is the adjustment amount for the duration of the time period. upward volatile variable and downward volatility variables A unified form.
[0043] The conservative coefficient for adjusting the volatility variable range is obtained as follows: like ,but It should remain unchanged.
[0044]
[0045] in, and The current time period The conservative coefficients for the upward and downward intervals.
[0046] like ,but Should be maintained according to Adjustments will be made.
[0047]
[0048] in, and They are respectively in the time period The vector formed by the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and They are consecutive front A matrix consisting of the upper and lower bounds of the net power prediction intervals for each distributed photovoltaic access node in each time period.
[0049] Step 3.3: Based on the constructed dynamic adjustment mechanism, predict the next unknown period. Net power range boundary prediction values:
[0050] in ,and For the next unknown period The upper and lower bounds of the net power prediction interval for distributed photovoltaic (PV) grid-connected nodes. and They are respectively in the time period The upper and lower bounds of the range of net power of distributed photovoltaic access nodes.
[0051] Step 4: Based on the next unknown time period obtained in Step 3 Based on the predicted net power interval boundary values, a distribution network state mapping is established. The distribution network state interval boundary is calculated based on the state mapping, and the next unknown time period is output. Prediction range of distribution network status.
[0052] Step 4.1, based on the next unknown time period Based on the predicted net power interval boundaries, construct a state mapping:
[0053] in, , , These represent composite functions, linear functions, and coupled functions, respectively. Represents the state transition function. express inverse function, and The first and the An uncertain variable, express coefficient, express coefficient, express The coefficient.
[0054] Step 4.2: Calculate the boundary of the distribution network state interval based on the state mapping: First, calculate the upper and lower bounds of the composite function prediction interval for each state. The calculation method for the upper and lower bounds of the composite function prediction interval for node voltage and line active power is as follows:
[0055] in, and For the next unknown period Upper and lower bounds of the prediction interval for a composite function. and For the next unknown period No. The upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period No. Correction values for the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period A vector consisting of the correction values of the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node, with subscripts... and These represent the mapping parameters for overestimation and underestimation, respectively. and They represent overestimation and underestimation of the mapping, respectively. coefficient, and These are the composite functions in the overestimation and underestimation mappings, respectively.
[0056] The calculation method for the upper and lower bounds of the composite function of reactive power of the line is as follows:
[0057] in, and For the next unknown period The vector formed by the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. This is the transpose operation for a matrix or vector. For the next unknown period The vector formed by the various uncertain variables, and The coefficient matrix of the coupling function and the coefficient vector of the linear function are respectively coupled.
[0058] Step 4.3: Based on the calculated boundaries of the distribution network state intervals, after... Processing and predicting the distribution network status for the next period. interval :
[0059] in, and For the next unknown period The upper and lower bounds of the prediction interval for the distribution network status. and They represent overestimation and underestimation of the mapping, respectively. The inverse function of .
[0060] Based on the above-mentioned dynamic characterization method for the operating state range of high-proportion photovoltaic power distribution systems, the following is adopted: Figure 2 The improved IEEE 33 architecture diagram is shown, and the advancement of the proposed method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system is verified through comparative analysis of the following six methods: Option I: Adopt the dynamic characterization method of state intervals proposed in this invention; Scheme II: The iterative affine method represents uncertain variables as affine operators, and combined with the forward-backward substitution method for distribution networks, the affine mapping and interval results of the state are obtained; Option III: The extended affine method uses first-order and second-order sensitivity coefficients to construct the second-order affine expansion form of the distribution network state, without the need for optimization to obtain the second-order affine mapping and interval results of the state. Scheme IV: Optimized Affine Method. First, construct the affine forms of node voltage magnitude and phase angle at the base point; then reconstruct the active and reactive power balance equations; finally, construct a linear optimization model to solve the state affine mapping and interval results. Option V: The traditional Monte Carlo method uses a sample size of 60,000; Option VI: Because the Monte Carlo method ignores extreme scenarios, the interval result is often smaller than the true interval. Enhanced Monte Carlo method, to improve accuracy, additionally considers extreme scenarios that are difficult to sample using the Monte Carlo method. See the node net power fluctuation curve and the results of interval dynamic identification and adjustment. Figure 3 For a comparison of the voltage fluctuation curves and interval results of node 17 and the power fluctuation between line 2 and line 3 during the period from 6:00 to 12:00 for schemes I-VI, see [link to relevant documentation]. Figure 4 A comparison of network node voltage, line power, and interval results during the period of 10:30-10:35 is shown below. Figure 5 The computation time comparison of each scheme is shown in Table 1.
[0061] Table 1 Comparison of computation time for each scheme
[0062] The computer hardware environment for performing the optimized calculations was an Intel(R) Core(TM) i7-12700 with a clock speed of 2.10GHz and 16GB of memory; the software environment was a Windows 11 operating system.
[0063] like Figure 3 As shown, during the period of severe photovoltaic fluctuations from 10:00 to 13:00, the net power at each node fluctuates sharply downwards. In this case, increasing the conservatism coefficient and actively widening the interval for the next time period provides more prediction margin and helps the prediction interval cover the actual fluctuation curve as much as possible. Maintaining a shorter time interval for a period helps quickly capture significant changes in the net power at each node, improving the sensitivity of interval identification. During the period of smoother photovoltaic fluctuations from 06:00 to 10:00, the net power at each node fluctuates less and tends to stabilize, and the conservatism coefficient and interval width gradually return to their default values.
[0064] like Figure 4As shown, during periods of intense fluctuation, the width of each state interval increases significantly. For node voltage, the interval results of Method II expand significantly, while those of Method III contract slightly. For line active power, the interval results of Method IV are excessively contracted. For line reactive power, the interval of Method III deviates from the benchmark interval, while the interval of Method IV is excessively expanded. During periods of gentle fluctuation, the interval width is correspondingly smaller, reflecting the weak fluctuations in photovoltaics. For voltage, the overlap of the interval results is high, indicating that each method has good accuracy in characterizing the voltage interval. However, for line power, the interval results of the affine methods deviate significantly from those of Method VI, which is likely due to the large mapping errors of these affine methods. In terms of interval accuracy comparison, Method VI is obtained by traversing all extreme scenarios and is therefore used as the interval benchmark. Method V, due to its difficulty in capturing the most extreme scenarios and its small interval range, is not suitable as a reference for interval accuracy analysis. The interval results of Method IV can cover and are close to the intervals of Method VI for all state variables (node voltage, line active power and line reactive power), indicating that the method in this paper can accurately characterize the state interval including extreme scenarios. This is because the decoupling mapping error is small and the positive and negative values are consistent.
[0065] like Figure 5 As shown, the node voltage increases with the increase of photovoltaic output, but the determination of the interval boundary is significantly affected by uncertainty. The line active power is significantly affected by the fluctuation of net active power from photovoltaics, and the interval fluctuation range of these two state variables is relatively large. The line reactive power is mainly affected by the coupling effect of net active power from photovoltaics, and this effect is weaker than that of line active power, therefore its fluctuation range is smaller.
[0066] Table I lists the maximum computation time for different methods characterizing a certain state over a certain time interval. It can be seen that the method presented in this paper has a significant advantage in computational efficiency, requiring only 0.2892 s, far lower than other methods. The iterative affine method and the optimized affine method use affine arithmetic, reducing the computation time to 0.3509 s and 0.4283 s, respectively. The extended affine method simplifies the computation process through second-order affine transformation, but still retains some complexity. The Monte Carlo method and the enhanced Monte Carlo method are based on large-scale sampling, resulting in relatively long computation times.
[0067] A comparison of the interval boundaries between Scheme I and Schemes II-VI reveals that Scheme I, with its dynamic adjustment mechanism, exhibits a significant advantage in interval characterization. Scheme VI, considering extreme scenarios, provides more complete interval results and is more suitable as a benchmark than Scheme V. While Scheme VI offers good interval coverage, its computation time is significantly longer than other methods due to its large sample size, making it suitable for static quantization scenarios but less suitable for real-time requirements. Among Schemes III-IV, Scheme I provides the closest interval envelope for node voltage and line power to Scheme VI. In contrast, Schemes II, III, and IV exhibit interval expansion and contraction in certain time periods, resulting in insufficient accuracy in characterizing intervals under different states; Scheme I, however, maintains high accuracy across all states. Considering the computation time comparison, Scheme I effectively captures the state boundaries of the system under time-varying operating conditions, balancing accuracy and dynamic response capability.
[0068] It should be emphasized that the embodiments described in this invention are illustrative rather than limiting. Therefore, this invention includes, but is not limited to, the embodiments described in the specific implementation. Any other implementations derived by those skilled in the art based on the technical solutions of this invention are also within the scope of protection of this invention.
Claims
1. A method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system, characterized in that: Includes the following steps: Step 1: Based on the selected active distribution network, input the continuous... The net power time-series data of distributed photovoltaic (PV) access nodes for each time period; the net power of all distributed PV access nodes is set as a vector of uncertain variables. Node voltage, line active power, and line reactive power are used as indicators of the distribution network status. Set mapping parameters and dynamic adjustment mechanism parameters; Step 2: Based on the net power time series data of the distributed photovoltaic access node in Step 1, convert each net power time series data into a relative angle in each time period, identify data points based on the relative angle, and extract time series features through data point iteration to obtain the net power trend and interval boundary of the distributed photovoltaic access node. Step 3, based on the continuous preceding The net power trend and interval boundaries of distributed photovoltaic access nodes in each time period are analyzed to determine volatility, and a dynamic adjustment mechanism is constructed to obtain the next unknown time period. Net power range boundary prediction; Step 4: Based on the next unknown time period obtained in Step 3 The net power interval boundary prediction value is used to establish a distribution network state mapping, and the distribution network state interval boundary is calculated based on the state mapping to output the next unknown time period. Prediction range of distribution network status.
2. The method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system according to claim 1, characterized in that: The mapping parameters in step 1 include: the number of uncertain variables. The highest order of the polynomial used in the state transition function binary conservative parameters The dynamic adjustment mechanism parameters include: the threshold for determining the smooth point. Threshold for determining turning points Feature extraction threshold Volatility determination threshold Time period length adjustment amount Maximum time period length Minimum time period length Maximum interval conservative coefficient Minimum interval conservative coefficient Interval Conservative Coefficient Adjustment .
3. The method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system according to claim 1, characterized in that: In step 1, input continuous previous... The specific implementation method for obtaining the net power time-series data of distributed photovoltaic access nodes for each time period is as follows: ; in, This is a sequence of node net power values. express The Middle Net power measurement value This represents the number of measurement points during that time period.
4. The method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system according to claim 1, characterized in that: Step 2 includes the following steps: Step 2.1: Convert the net power time-series data into relative angles within each time period: This is done by transforming the angles of two consecutive points. Equivalent transformation : ; in, As the initial point Relative angle Equivalent time series data, relative angle for: ; in, and They are respectively The front absolute angle and the rear absolute angle; Step 2.2: Identify data points based on relative angles: ; In the formula, For a smooth point set, For the set of turning points, for The Middle One feature point, for In sequence The position in the middle, for The Middle One feature point, for In sequence The position in the middle, and The table is divided into sub-tables. and number of elements Step 2.3: Iteratively extract temporal features: If the threshold is determined based on the flat point Calculated feature extraction rate Less than the given feature extraction threshold Then determine The trend characteristics are not significant at present, so the threshold adjustment amount is used. Gradually increase the threshold for determining smooth points and update the feature extraction rate. Until , obtain Trend characteristics; Step 2.4: Calculate the net power trend and interval boundaries of the distributed photovoltaic access nodes: ; in, The net power trend of distributed photovoltaic access nodes. and These are the upper and lower bounds of the net power range for distributed photovoltaic access nodes, respectively.
5. The method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system according to claim 1, characterized in that: Step 3 includes the following steps: Step 3.1: Determine the volatility of the net power trend and interval boundaries of the distributed photovoltaic access nodes: ; in, and Let them represent upward volatility variables and downward volatility variables, respectively. Indicates the time period The vector formed by the net power trends of each distributed photovoltaic access node. and They are respectively in the time period The vector formed by the upper and lower bounds of the net power range of each distributed photovoltaic access node. They are consecutive front A matrix consisting of the net power trends of each distributed photovoltaic access node over a given time period. and They are consecutive front A matrix consisting of the upper and lower bounds of the net power range of each distributed photovoltaic access node in each time period; Step 3.2: Construct a dynamic adjustment mechanism based on the fluctuation determination results: Adjust the next time period length according to volatility. ; in, For the next unknown period The length of the time period For the current time period The length of the time period This is the adjustment amount for the duration of the time period. upward volatile variable and downward volatility variables A unified form; Step 3.3: Obtain the next unknown time period based on the constructed dynamic adjustment mechanism. Net power range boundary prediction values: ; in, and For the next unknown period The upper and lower bounds of the net power prediction interval for distributed photovoltaic (PV) grid-connected nodes. and They are respectively in the time period The upper and lower bounds of the range of net power of distributed photovoltaic access nodes.
6. The method for dynamically characterizing the operating state range of a high-proportion photovoltaic power distribution system according to claim 1, characterized in that: Step 4 includes the following steps: Step 4.1: Based on the obtained next unknown time period The net power range boundary prediction value is used to establish the distribution network state mapping; ; in, , , These represent composite functions, linear functions, and coupled functions, respectively. Represents the state transition function. express inverse function, and The first and the An uncertain variable, express coefficient, express coefficient, express The coefficient; Step 4.2: Calculate the boundary of the distribution network state interval based on state mapping: Calculate the upper and lower bounds of the composite function prediction interval for each state; The method for calculating the upper and lower bounds of the prediction interval of the composite function of node voltage and line active power is as follows: ; in, and For the next unknown period Upper and lower bounds of the prediction interval for composite functions. and For the next unknown period No. The upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period No. Correction values for the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. and For the next unknown period A vector consisting of the correction values of the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node, including subscripts. and The parameters are the mapping parameters for overestimation and underestimation, respectively. and They represent overestimation and underestimation of the mapping, respectively. coefficient, and These are the composite functions in the overestimation and underestimation mappings, respectively; The calculation method for the upper and lower bounds of the composite function of reactive power of the line is as follows: ; in, and For the next unknown period The vector formed by the upper and lower bounds of the net power prediction interval for each distributed photovoltaic access node. This is the transpose operation for a matrix or vector. For the next unknown period The vector formed by the various uncertain variables, and The coefficient vectors of the coupled function and the linear function respectively Step 4.3: Based on the calculated boundaries of the distribution network state intervals, after... Processing and predicting the distribution network status for the next period. interval : ; in, and For the next unknown period The upper and lower bounds of the prediction interval for the distribution network status. and They represent overestimation and underestimation of the mapping, respectively. The inverse function of .