A source-load behavior portrait-based active power distribution network voltage distribution modeling method

Abstract

The application specifically provides a source-load behavior portrait-based active power distribution network voltage distribution modeling method. The method of the application comprises the following steps: obtaining multi-source original data of a target active power distribution network, and obtaining a multi-source data set based on the multi-source original data, wherein the multi-source original data comprises original data of photovoltaic, load and micro-grid, and line structure data; constructing a source-load behavior portrait based on the multi-source data set; obtaining a dynamic parameter set, wherein the dynamic parameter set comprises a dynamic coupling coefficient and a line parameter; constructing a voltage mechanism model based on the source-load behavior portrait and the dynamic parameter set; and dynamically optimizing the voltage mechanism model to obtain an optimized voltage mechanism model. The source-load behavior portrait is constructed according to the characteristics of photovoltaic, load and micro-grid, the voltage distribution modeling is established, and the self-checking and dynamic adjustment of the established model are performed, so that the workload of manual calibration is avoided, the accuracy of the established model is ensured, and the accurate prediction through the established model is ensured.

Description

Technical Field

[0001] This invention relates to the field of power system grid modeling technology, specifically providing a method for modeling the voltage distribution of active distribution networks based on source-load behavior profiling. Background Technology

[0002] With the increasing integration of distributed photovoltaic (PV), load, and microgrid renewable energy sources into active distribution networks in recent years, distributed PV, load, and microgrids have become crucial components in building active distribution networks. These three elements work together to drive the energy system towards cleaner and smarter transformation. However, existing voltage modeling methods applied to active distribution networks suffer from several drawbacks: vague source-load characteristics, lack of accurate analysis of voltage distribution influencing factors under different network structures, and poor timeliness.

[0003] Accordingly, there is a need in this field for a new active distribution network voltage distribution modeling scheme based on source-load behavior profiling to solve the above problems. Summary of the Invention

[0004] To overcome the above-mentioned shortcomings, this invention is proposed to provide a method for modeling the voltage distribution of active distribution networks based on source-load behavior profiles. This method addresses, or at least partially addresses, the technical problems in the prior art, such as vague source-load characteristics, lack of accurate analysis of the influencing factors of voltage distribution in active distribution networks under different structural forms, and poor timeliness.

[0005] In a first aspect, the present invention provides a method for modeling voltage distribution in an active distribution network based on source-load behavior profiling, the method comprising: Acquire multi-source raw data of the target active distribution network, and obtain a multi-source dataset based on the multi-source raw data, wherein the multi-source raw data includes at least raw data of photovoltaic, load, microgrid and line structure data; Construct source load behavior profiles based on multi-source datasets; Obtain a dynamic parameter set, wherein the dynamic parameter set includes at least dynamic coupling coefficients and line parameters; Based on the source-load behavior profile and dynamic parameter set, a voltage mechanism model is constructed; The voltage mechanism model is dynamically optimized to obtain the optimized voltage mechanism model.

[0006] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the multi-source dataset obtained based on multi-source raw data includes: Preprocessing of multi-source raw data yields multi-source raw data in a unified format; Outlier handling and missing value repair are performed on the raw data in a uniform format to obtain the processed data and the repaired data, thus obtaining the processed multi-source raw data. The processed multi-source raw data is subjected to time-series block processing to obtain segmented multi-source data, which is used for subsequent sliding window updates and LSTM input. The segmented multi-source data is used as the multi-source dataset.

[0007] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the step of constructing source-load behavior profiling based on multi-source datasets includes: Based on multi-source datasets, a photovoltaic power output model is constructed, and load behavior characteristics and microgrid behavior characteristics are extracted. Based on the construction of a photovoltaic power output model, the extraction of load behavior characteristics and microgrid behavior characteristics, a source-load behavior profile is constructed.

[0008] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the step of constructing a photovoltaic output model, extracting load behavior features, and microgrid behavior features based on multi-source datasets includes: The photovoltaic parameter set is determined based on multi-source datasets; A photovoltaic power output model is constructed based on the photovoltaic parameter set; Obtain multiple preset time segments; Determine the optimal distribution parameters based on multi-source datasets and multiple time periods; Based on multi-source datasets and optimal distribution parameters, determine the load behavior feature matrix; Based on multi-source datasets, the power direction and operating mode of the microgrid are determined; Based on the power direction and operating mode of the microgrid, the behavioral characteristics of the microgrid are determined.

[0009] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the acquisition of the dynamic parameter set includes: The changes in active power and voltage are obtained through a pre-set power disturbance test. The dynamic coupling coefficient is determined based on the changes in active power and voltage. The updated line parameters are obtained through a pre-built graph neural network; The dynamic parameter set is based on the dynamic coupling coefficient and the updated line parameters.

[0010] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the construction of the voltage mechanism model based on source-load behavior profiling and dynamic parameter set includes: The voltage mechanism model is constructed using the following formula: In the formula, Representing the The voltage value of each node, Representing the Total active power of the line segment This represents the upstream power supply voltage of the target active distribution network. Representing the The total reactive power of the line segment Representing the The voltage of each node, Represents the dynamic coupling coefficient. Representing the The updated resistance parameter values ​​of the line segment. Representing the The updated reactance parameter values ​​of the line segment.

[0011] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the step of dynamically optimizing the voltage mechanism model to obtain the optimized voltage mechanism model includes: A fusion prediction model is generated based on the voltage mechanism model. Based on the fusion prediction model, the fusion voltage prediction value is obtained; Obtain the current measured voltage value; Based on the current measured voltage value and the fused voltage prediction value, the graph neural network, LSTM voltage prediction model and dynamic weights are selectively updated, and the "acquiring dynamic parameter set" and subsequent steps are re-executed to achieve dynamic optimization of the voltage mechanism model.

[0012] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the generation of the fusion prediction model based on the voltage mechanism model includes: Based on source-load behavior profiles and dynamic parameter sets, an LSTM voltage prediction model is constructed. A fusion prediction model is generated by using the voltage mechanism model and the LSTM voltage prediction model as sub-models.

[0013] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the step of obtaining the fused voltage prediction value based on the fused prediction model includes: Based on the voltage mechanism model and the LSTM voltage prediction model, two prediction parameter values ​​are obtained respectively; Update the dynamic weights to obtain the current dynamic weights; The fused voltage prediction value is determined based on the two prediction parameter values ​​and the current dynamic weights.

[0014] In one technical solution of the above-mentioned active distribution network voltage distribution modeling method based on source-load behavior profiling, the step of obtaining updated line parameters through a pre-constructed graph neural network includes: The updated line parameters are obtained using the following formula: In the formula, This represents the updated line resistance value. This represents the original resistance value of the circuit. This represents the adjustment amount of the line resistance. This represents the updated line reactance value. Represents the original reactance value of the line. This represents the adjustment amount of the line reactance.

[0015] The above-described technical solutions of the present invention have at least one or more of the following beneficial effects: By analyzing the electricity consumption data of distributed photovoltaic (PV), load, and microgrid in the target active distribution network, a PV output model is established, and a source-load behavior profile is constructed. This enables comprehensive processing of the behavioral characteristics of PV, load, and microgrid, providing a high-accuracy source-load power profile for subsequent voltage modeling. This reduces voltage prediction errors at the source. A non-parametric kernel density estimation method is used to construct an output model reflecting the influence of sunlight and temperature on distributed PV power generation, achieving accurate descriptions of PV output, load, and microgrid, thereby improving the accuracy of PV output, load power, and voltage prediction. A normal distribution with equal probability density function is used to characterize load behavior, and factors such as weekdays, rest days, and seasonality are comprehensively considered. Time series segment analysis is used to extract the electricity consumption characteristics of PV output, load, and microgrid, thus constructing a source-load behavior profile of the microgrid and the target active distribution network. Furthermore, by combining the structural parameters of the distribution network lines and considering the strong coupling relationship between voltage and active power, a mechanism model for the influence of line state parameters and source-load behavior on the voltage of the target active distribution network is established. This ensures the stability of the voltage model, achieves real-time accuracy of the voltage model, maintains the prediction accuracy of the voltage mechanism model, and improves the universality of the voltage model for different scenarios. Attached Figure Description

[0016] The disclosure of this invention will become more readily understood with reference to the accompanying drawings. It will be readily understood by those skilled in the art that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Furthermore, similar numbers in the drawings are used to denote similar components, wherein: Figure 1 This is a schematic diagram of the main steps of an active distribution network voltage distribution modeling method based on source-load behavior profiling according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating an active distribution network voltage distribution modeling method based on source-load behavior profiling according to an embodiment of the present invention. Figure 3 This is a schematic diagram illustrating the process of constructing a source-load behavior profile in an active distribution network voltage distribution modeling method based on a source-load behavior profile according to an embodiment of the present invention. Figure 4 This is a schematic diagram of the main structure of an active distribution network voltage distribution modeling method based on source-load behavior profiling according to an embodiment of the present invention. Detailed Implementation

[0017] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0018] In the description of this invention, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and may also include software components, such as program code, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Non-transitory computer-readable storage media includes any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.

[0019] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of an active distribution network voltage distribution modeling method based on source-load behavior profiling according to an embodiment of the present invention. Figures 1 to 4 As shown, the active distribution network voltage distribution modeling method based on source-load behavior profile in this embodiment of the invention mainly includes the following steps S101-S105.

[0020] Step S101: Obtain multi-source raw data of the target active distribution network, and obtain a multi-source dataset based on the multi-source raw data, wherein the multi-source raw data includes at least the raw data of photovoltaic, load, microgrid and line structure data; Specifically, the photovoltaic raw data in the multi-source raw data of the target active distribution network includes at least the light intensity, ambient temperature, and photovoltaic inverter output power; The raw load data should include at least the real-time power consumption of residential, commercial, or industrial loads and the location of the load access node. The raw data of a microgrid includes at least the microgrid access nodes, real-time power (where power supply is positive and power consumption is negative), and operating mode (specifically, grid-connected or off-grid).

[0021] Specifically, obtaining the multi-source dataset based on multi-source raw data includes: Preprocessing of multi-source raw data yields multi-source raw data in a unified format; Outlier handling and missing value repair are performed on the raw data in a uniform format to obtain the processed data and the repaired data, thus obtaining the processed multi-source raw data. The processed multi-source raw data is subjected to time-series block processing to obtain segmented multi-source data, which is used for subsequent sliding window updates and LSTM input. The segmented multi-source data is used as the multi-source dataset.

[0022] Specifically, preprocessing the multi-source raw data to obtain multi-source raw data in a unified format includes: The format of the verified raw data is converted to obtain multi-source raw data in a unified format.

[0023] Specifically, outlier handling and missing value repair are performed on the raw data in a uniform format to obtain processed data and repaired data, resulting in processed multi-source raw data including: The outlier threshold is determined using the following formula: In the formula, Represents outliers. Represents the sample mean. Represents the sample standard deviation; If the time range of missing values ​​in the original data of the uniform format is less than the preset missing time threshold, then the missing value part is filled by linear interpolation. If the time range of missing values ​​in the original data of the uniform format exceeds the preset missing time threshold, the missing value part is filled by LSTM prediction. To obtain the processed multi-source raw data.

[0024] Specifically, the step of performing time-series block processing on the processed multi-source raw data to obtain segmented multi-source data includes: The processed multi-source raw data is divided into blocks of 2 hours each with a 15-minute step size to accommodate subsequent sliding window updates and LSTM inputs.

[0025] Step S102: Construct source load behavior profiles based on multi-source datasets; Specifically, constructing source payload behavior profiles based on multi-source datasets includes: Based on multi-source datasets, a photovoltaic power output model is constructed, and load behavior characteristics and microgrid behavior characteristics are extracted. Based on the construction of a photovoltaic power output model, the extraction of load behavior characteristics and microgrid behavior characteristics, a source-load behavior profile is constructed.

[0026] Specifically, the construction of the photovoltaic output model, extraction of load behavior features, and microgrid behavior features based on multi-source datasets include: The photovoltaic parameter set is determined based on multi-source datasets; A photovoltaic power output model is constructed based on the photovoltaic parameter set; Obtain multiple preset time segments; Determine the optimal distribution parameters based on multi-source datasets and multiple time periods; Based on multi-source datasets and optimal distribution parameters, determine the load behavior feature matrix; Based on multi-source datasets, the power direction and operating mode of the microgrid are determined; Based on the power direction and operating mode of the microgrid, the behavioral characteristics of the microgrid are determined.

[0027] Specifically, based on multi-source datasets, the photovoltaic parameter set is determined as follows: Kernel density estimation is performed on the illumination, temperature, and photovoltaic output data in the multi-source dataset to obtain their respective marginal distribution functions. The marginal distribution functions are then used as their respective cumulative probability parameters to obtain the photovoltaic parameter set.

[0028] Specifically, based on the photovoltaic parameter set, the photovoltaic output model is constructed as follows: The fusion bandwidth parameter is determined using the following formula: In the formula, Represents the converged bandwidth parameter. The standard deviation of the representative data Represents the sample size. Represents the global bandwidth baseline value. Represents the local density estimate. Represents the geometric mean density, where, , Represents the local bandwidth parameter value. Represents the global bandwidth parameter value; The photovoltaic output model is constructed using the following formula: In the formula, These represent the cumulative edge distribution probabilities of illumination, temperature, and power, respectively. Represents Copula parameters; Furthermore, the photovoltaic output model is updated every 15 minutes by calling the sliding window data.

[0029] Specifically, obtaining the preset multiple time segments includes: The time zones are divided into 7 categories based on "weekdays or rest days + spring, summer, autumn or winter".

[0030] Specifically, determining the optimal distribution parameters based on multi-source datasets and multiple time periods includes: Based on multi-source datasets and multiple time intervals, determine the time intervals of load data in the multi-source datasets; The normal distribution parameter value and the log-normal distribution parameter value of the load data are determined by the distribution selection function of AIC / BIC, and the normal distribution parameter value or the log-normal distribution parameter value that minimizes both AIC and BIC is taken as the optimal distribution parameter.

[0031] Specifically, determining the load behavior feature matrix based on multi-source datasets and optimal distribution parameters includes: To construct an LSTM load forecasting model, load data from the same time period of the previous 7 days in the multi-source dataset are imported into the LSTM load forecasting model to obtain the mean load data and standard deviation data for the day. Dynamic volatility can be obtained using the following formula: In the formula, Representing the Dynamic volatility over a period of time Representing the Standard deviation data values ​​for the time period Representing the The average daily load data for the time period; Based on the load data in the multi-source dataset, the statistical characteristic parameters are determined, namely peak-to-valley values ​​and peak-to-valley differences. A load behavior characteristic matrix is ​​constructed based on statistical characteristic parameters, dynamic volatility, and optimal distribution parameters.

[0032] In the above embodiments, by constructing photovoltaic power output models, load characteristic matrices, and microgrid behavior parameters respectively, a source-load behavior profile is constructed, which provides high-fidelity source-load power input for subsequent voltage distribution modeling, reduces voltage prediction deviation from the source, and thus reduces the error rate of the constructed model, thereby making the method of the present invention applicable to various distribution network systems.

[0033] Step S103: Obtain a dynamic parameter set, wherein the dynamic parameter set includes at least dynamic coupling coefficients and line parameters; Specifically, obtaining the dynamic parameter set includes: The changes in active power and voltage are obtained through a pre-set power disturbance test. The dynamic coupling coefficient is determined based on the changes in active power and voltage. The updated line parameters are obtained through a pre-built graph neural network; The dynamic parameter set is based on the dynamic coupling coefficient and the updated line parameters.

[0034] Specifically, determining the dynamic coupling coefficient based on the changing active power and voltage includes: The dynamic coupling coefficient is obtained by the following formula: In the formula, Represents the dynamic coupling coefficient. Represents the change in active power. This represents the voltage change value.

[0035] Specifically, obtaining the updated line parameters through a pre-constructed graph neural network includes: The updated line parameters are obtained using the following formula: In the formula, This represents the updated line resistance value. This represents the original resistance value of the circuit. This represents the adjustment amount of the line resistance. This represents the updated line reactance value. Represents the original reactance value of the line. This represents the adjustment amount of the line reactance.

[0036] Step S104: Construct a voltage mechanism model based on the source-load behavior profile and dynamic parameter set; Specifically, the construction of the voltage mechanism model based on the source-load behavior profile and dynamic parameter set includes: The voltage mechanism model is constructed using the following formula: In the formula, Representing the The voltage value of each node, Representing the Total active power of the line segment This represents the upstream power supply voltage of the target active distribution network. Representing the The total reactive power of the line segment Representing the The voltage of each node, Represents the dynamic coupling coefficient. Representing the The updated resistance parameter values ​​of the line segment. Representing the The updated reactance parameter values ​​of the line segment.

[0037] Specifically, the first Total active power of the line segment It can be obtained through the following formula: In the formula, Representing the Section of the line Real-time photovoltaic power output at any given moment Representing the Section of the line Real-time load power at any given moment Representing the Section of the line The real-time power of the microgrid at any given time, where, The symbol is determined by the operating mode. If the power supply is connected to the grid, the symbol is "+"; if the power supply is off-grid, the symbol is "-".

[0038] Step S105: Dynamically optimize the voltage mechanism model to obtain the optimized voltage mechanism model.

[0039] In the above embodiments, by sequentially executing a complete closed-loop operation of data acquisition and preprocessing, source-load behavior feature extraction, and voltage distribution modeling and verification, the characteristics of photovoltaics, loads, and microgrids are extracted and modeled based on voltage distribution. The established model is then self-checked and dynamically adjusted based on the self-check results, avoiding the workload of manual calibration, ensuring the accuracy of the established model, and ensuring that the established model can accurately predict. This enables the modeling of the complex distribution network structure of photovoltaics, loads, and microgrids, so as to facilitate timely knowledge of the voltage distribution of active distribution networks under different structural forms.

[0040] Specifically, the dynamic optimization of the voltage mechanism model to obtain the optimized voltage mechanism model includes: A fusion prediction model is generated based on the voltage mechanism model. Based on the fusion prediction model, the fusion voltage prediction value is obtained; Obtain the current measured voltage value; Based on the current measured voltage value and the fused voltage prediction value, the graph neural network, LSTM voltage prediction model and dynamic weights are selectively updated, and the "acquiring dynamic parameter set" and subsequent steps are re-executed to achieve dynamic optimization of the voltage mechanism model.

[0041] Specifically, the generation of the fusion prediction model based on the voltage mechanism model includes: Based on source-load behavior profiles and dynamic parameter sets, an LSTM voltage prediction model is constructed. A fusion prediction model is generated by using the voltage mechanism model and the LSTM voltage prediction model as sub-models.

[0042] Specifically, obtaining the fused voltage prediction value based on the fused prediction model includes: Based on the voltage mechanism model and the LSTM voltage prediction model, two prediction parameter values ​​are obtained respectively; Update the dynamic weights to obtain the current dynamic weights; The fused voltage prediction value is determined based on the two prediction parameter values ​​and the current dynamic weights.

[0043] Specifically, the dynamic weights are obtained using the following formula: In the formula, Representing the Dynamic weights of each sub-model Representing the The prior probabilities of each sub-model, specifically, the initial values ​​of the prior probabilities of each sub-model are as follows: , , Representing the The likelihood function of each sub-model This represents the measured voltage data from the previous day. Representing the The parameter set of each sub-model Representing the Prior probabilities of each sub-model Representing the The likelihood function of each sub-model Representing the The parameter set of each sub-model.

[0044] Specifically, determining the fused voltage prediction value based on two prediction parameter values ​​and the current dynamic weights includes: The fused voltage prediction value is obtained using the following formula: In the formula, Represents the predicted value of the fused voltage. , These represent the current dynamic weights corresponding to the voltage mechanism model and the LSTM voltage prediction model, respectively. , These represent the prediction parameter values ​​corresponding to the voltage mechanism model and the LSTM voltage prediction model, respectively.

[0045] Specifically, the process of selectively updating the graph neural network, LSTM voltage prediction model, and dynamic weights based on the current measured voltage value and the fused voltage prediction value, and re-executing the "acquiring dynamic parameter set" and subsequent steps to achieve dynamic optimization of the voltage mechanism model includes: Based on the current measured voltage value and the fused voltage prediction value, the current prediction error value is obtained; If the current prediction error value exceeds the preset prediction error threshold, the model accuracy is deemed unqualified and optimization is required. This involves updating the graph neural network, LSTM voltage prediction model, and dynamic weights, and then re-executing the "obtain dynamic parameter set" and subsequent steps based on the updated graph neural network, LSTM voltage prediction model, and dynamic weights. Otherwise, the model is deemed to have sufficient accuracy and no optimization is required.

[0046] Specifically, the preset prediction error threshold is 1.1 * the smallest predicted or measured voltage value within the past 30 days.

[0047] Specifically, the functions not explicitly stated above all use conventional formulas found in existing technologies.

[0048] Based on steps S101-S105 above, by analyzing the distributed photovoltaic, load, and microgrid electricity consumption data of the target active distribution network, a photovoltaic output model is established, and a source-load behavior profile is constructed. This achieves comprehensive processing of the behavioral characteristics of photovoltaic, load, and microgrid, and constructs a source-load behavior profile, thereby providing a high-accuracy source-load power for subsequent voltage modeling. This reduces voltage prediction errors from the source. A non-parametric kernel density estimation method is used to construct an output model that reflects the influence of sunlight and temperature on distributed photovoltaic power generation, achieving accurate description of photovoltaic output, load, and microgrid, and thus improving the accuracy of photovoltaic output, load power, and voltage prediction. A normal distribution with equal probability density function is used to characterize load behavior, and factors such as weekdays, rest days, and seasonality are comprehensively considered. Using time series segment analysis, the characteristics of photovoltaic output, load, and microgrid electricity consumption are extracted, thereby constructing a source-load behavior profile of the microgrid and the target active distribution network. Furthermore, considering the strong coupling relationship between voltage and active power, and combining the structural parameters of the distribution network lines, a mechanism model of the influence of line state parameters and source-load behavior on the voltage of the target active distribution network is established. This ensures the stability of the voltage model, achieves real-time accuracy, maintains the predictive precision of the voltage mechanism model, and improves the universality of the voltage model for different scenarios. This avoids the technical problems of existing technologies, such as vague source-load characteristics, lack of accurate analysis of the influencing factors of voltage distribution under different structural forms of active distribution networks, and poor timeliness.

[0049] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of the present invention.

[0050] Those skilled in the art will understand that all or part of the processes in the method of the above embodiment of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0051] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of software and hardware. Therefore, the number of modules shown in the figures is merely illustrative.

[0052] Those skilled in the art will understand that the various modules in the device can be adaptively split or combined. Such splitting or combining of specific modules will not cause the technical solution to deviate from the principles of the present invention; therefore, the technical solutions after splitting or combining will fall within the protection scope of the present invention.

[0053] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for modeling voltage distribution in active distribution networks based on source-load behavior profiling, characterized in that, The method includes: Acquire multi-source raw data of the target active distribution network, and obtain a multi-source dataset based on the multi-source raw data, wherein the multi-source raw data includes at least raw data of photovoltaic, load, microgrid and line structure data; Construct source load behavior profiles based on multi-source datasets; Obtain a dynamic parameter set, wherein the dynamic parameter set includes at least dynamic coupling coefficients and line parameters; Based on the source-load behavior profile and dynamic parameter set, a voltage mechanism model is constructed; The voltage mechanism model is dynamically optimized to obtain the optimized voltage mechanism model.

2. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 1, characterized in that, The multi-source dataset obtained based on multi-source raw data includes: Preprocessing of multi-source raw data yields multi-source raw data in a unified format; Outlier handling and missing value repair are performed on the raw data in a uniform format to obtain the processed data and the repaired data, thus obtaining the processed multi-source raw data. The processed multi-source raw data is subjected to time-series block processing to obtain segmented multi-source data, which is used for subsequent sliding window updates and LSTM input. The segmented multi-source data is used as the multi-source dataset.

3. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 2, characterized in that, The construction of source payload behavior profiles based on multi-source datasets includes: Based on multi-source datasets, a photovoltaic power output model is constructed, and load behavior characteristics and microgrid behavior characteristics are extracted. Based on the construction of a photovoltaic power output model, the extraction of load behavior characteristics and microgrid behavior characteristics, a source-load behavior profile is constructed.

4. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 3, characterized in that, The process of constructing a photovoltaic power output model and extracting load behavior features and microgrid behavior features based on multi-source datasets includes: The photovoltaic parameter set is determined based on multi-source datasets; A photovoltaic power output model is constructed based on the photovoltaic parameter set; Obtain multiple preset time segments; Determine the optimal distribution parameters based on multi-source datasets and multiple time periods; Based on multi-source datasets and optimal distribution parameters, determine the load behavior feature matrix; Based on multi-source datasets, the power direction and operating mode of the microgrid are determined; Based on the power direction and operating mode of the microgrid, the behavioral characteristics of the microgrid are determined.

5. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 4, characterized in that, The acquisition of the dynamic parameter set includes: The changes in active power and voltage are obtained through a pre-set power disturbance test. The dynamic coupling coefficient is determined based on the changes in active power and voltage. The updated line parameters are obtained through a pre-built graph neural network; The dynamic parameter set is based on the dynamic coupling coefficient and the updated line parameters.

6. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 5, characterized in that, The voltage mechanism model constructed based on source-load behavior profiles and dynamic parameter sets includes: The voltage mechanism model is constructed using the following formula: In the formula, Representing the The voltage value of each node, Representing the Total active power of the line segment This represents the upstream power supply voltage of the target active distribution network. Representing the The total reactive power of the line segment Representing the The voltage of each node, Represents the dynamic coupling coefficient. Representing the The updated resistance parameter values ​​of the line segment. Representing the The updated reactance parameter values ​​of the line segment.

7. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling as described in claim 6, characterized in that, The dynamic optimization of the voltage mechanism model to obtain the optimized voltage mechanism model includes: A fusion prediction model is generated based on the voltage mechanism model. Based on the fusion prediction model, the fusion voltage prediction value is obtained; Obtain the current measured voltage value; Based on the current measured voltage value and the fused voltage prediction value, the graph neural network, LSTM voltage prediction model and dynamic weights are selectively updated, and the "acquiring dynamic parameter set" and subsequent steps are re-executed to achieve dynamic optimization of the voltage mechanism model.

8. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 7, characterized in that, The generation of the fusion prediction model based on the voltage mechanism model includes: Based on source-load behavior profiles and dynamic parameter sets, an LSTM voltage prediction model is constructed. A fusion prediction model is generated by using the voltage mechanism model and the LSTM voltage prediction model as sub-models.

9. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 8, characterized in that, The fusion voltage prediction value obtained based on the fusion prediction model includes: Based on the voltage mechanism model and the LSTM voltage prediction model, two prediction parameter values ​​are obtained respectively; Update the dynamic weights to obtain the current dynamic weights; The fused voltage prediction value is determined based on the two prediction parameter values ​​and the current dynamic weights.

10. The method for modeling voltage distribution in active distribution networks based on source-load behavior profiling according to claim 9, characterized in that, The process of obtaining updated line parameters through a pre-constructed graph neural network includes: The updated line parameters are obtained using the following formula: In the formula, This represents the updated line resistance value. This represents the original resistance value of the circuit. This represents the adjustment amount of the line resistance. This represents the updated line reactance value. Represents the original reactance value of the line. This represents the adjustment amount of the line reactance.

Images