Bearing capacity assessment method and system based on distributed power supply output high-dimensional combination
By building a high-dimensional joint distribution model and combining time series analysis and deep learning models, the problems of insufficient variable correlation modeling and incomplete dynamic feature capture in distributed power bearing capacity evaluation are solved, and high-precision real-time prediction and evaluation are achieved, which improves the scientific nature of power grid scheduling.
Patent Information
- Application Number
- CN202510143586.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-10
- Publication Date
- 2025-05-06
AI Technical Summary
The existing distributed power bearing capacity evaluation technology has problems such as insufficient variable correlation modeling and incomplete dynamic feature capture.
By constructing a high-dimensional joint distribution model based on kernel density estimation and Gaussian Copula, combining time series analysis and deep learning model, the complex dependence between distributed power output and environmental variables and grid operation state is accurately captured, and high-precision real-time prediction and bearing capacity evaluation of distributed power output are achieved.
It significantly improves the accuracy and dynamic adaptability of bearing capacity assessment, providing a scientific basis for power grid scheduling and optimization.
Smart Images
Figure CN119944663A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of new energy grid connection, and in particular to a method and system for evaluating carrying capacity based on high-dimensional combination of distributed power output. Background Art
[0002] In recent years, with the increasing application of distributed energy resources (DERs) in modern power systems, their advantages in improving energy efficiency, reducing carbon emissions and enhancing grid flexibility have gradually emerged. There are many types of distributed power sources, including photovoltaics, wind power, small hydropower and energy storage equipment, and their output is intermittent, random and regional. In order to adapt to this complex power supply mode, grid dispatching and carrying capacity assessment have become key links to ensure the safe and stable operation of the power grid.
[0003] Traditional distributed generation capacity assessment methods usually rely on linear analysis and static models, which can describe the operating status of distributed generation and the grid carrying capacity to a certain extent. However, with the expansion of the scale of power systems and the increase in the penetration rate of distributed generation, the assessment methods based on single variables and linear assumptions can no longer meet the needs of complex systems. In order to more comprehensively describe the dynamic nonlinear relationship between the output of distributed generation and environmental conditions and the operating status of the grid, researchers have gradually introduced time series analysis, probability statistics models and machine learning methods in recent years. However, these technologies still have certain limitations in practice, such as the lack of comprehensive modeling capabilities for complex correlations between multiple variables and insufficient capture of nonlinear dynamic characteristics, which limits their assessment accuracy and reliability in high-penetration distributed generation environments. Summary of the invention
[0004] Therefore, the technical problem solved by the present invention is that the existing distributed power supply carrying capacity assessment technology has the problems of insufficient variable correlation modeling and incomplete capture of dynamic characteristics.
[0005] In order to solve the above technical problems, the present invention provides a carrying capacity assessment method and system based on high-dimensional joint distributed power output. By constructing a high-dimensional joint distribution model based on kernel density estimation and Gaussian Copula, the complex dependency relationship between distributed power output and environmental variables and power grid operation status is accurately captured; combined with time series analysis and deep learning models, high-precision real-time prediction of distributed power output is achieved, and the distributed power carrying capacity is quantified.
[0006] The technical solution adopted by the present invention is as follows:
[0007] The carrying capacity assessment method based on high-dimensional joint output of distributed power sources includes:
[0008] Collect target data, build a high-dimensional joint distribution model based on the target data, and obtain the characteristic relationship between the target data;
[0009] Use high-dimensional joint distribution models combined with time series to make real-time predictions of the output of distributed power sources;
[0010] Evaluate the carrying capacity of distributed power sources based on real-time prediction results.
[0011] As a preferred solution of the carrying capacity assessment method based on high-dimensional combination of distributed power output described in the present invention, wherein: the target data includes distributed power output data, environmental monitoring data and power grid operation data.
[0012] As a preferred solution of the method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to the present invention, after collecting the target data, it also includes:
[0013] The distributed power output features are extracted through an autoencoder with three hidden layers, and the top three environmental monitoring data with the highest correlation with the distributed power output data are selected as the input features of the high-dimensional joint distribution model;
[0014] The number of neurons in the middle layer of the autoencoder is less than that in the input layer, so feature compression is performed.
[0015] As a preferred solution of the carrying capacity evaluation method based on high-dimensional joint distribution of distributed power output described in the present invention, wherein: the high-dimensional joint distribution model includes:
[0016] Kernel density estimation is used to estimate the probability density function of distributed power output characteristics and screened environmental monitoring data:
[0017] ,
[0018] in, Indicates the estimated The probability density function value at ; Represents sample data; Indicates bandwidth; Indicates the target point of the current probability density calculation; Indicates the sample data points; represents the kernel function;
[0019] right Integrate to get the cumulative distribution function and get the cumulative distribution function value set for each data point .
[0020] As a preferred solution of the carrying capacity evaluation method based on high-dimensional joint output of distributed power sources described in the present invention, wherein: the high-dimensional joint distribution model also includes:
[0021] The cumulative distribution function is set As the input of the high-dimensional joint distribution model, Gaussian Copula is used to construct the joint distribution between inputs. The joint distribution density function is expressed as:
[0022] ,
[0023] in, Represents the density function of the Gaussian Copula joint distribution; Indicates The cumulative distribution function value of the data points; Represents the correlation matrix of Gaussian Copula; Representation Matrix The determinant of ; A vector of inverse transformed values representing the standard normal distribution; express The transpose of .
[0024] As a preferred solution of the method for evaluating the carrying capacity based on high-dimensional joint output of distributed power sources described in the present invention, wherein: the high-dimensional joint distribution model is used in combination with time series to perform real-time prediction of the output of distributed power sources, including:
[0025] Cumulative distribution function values using marginal distributions As input for time series analysis;
[0026] Use the autoregressive integrated moving average model to independently model each data point and make short-term forecasts on the output of distributed generation;
[0027] After the standardized As input, the long short-term memory network is used to make long-term predictions of the distributed generation output;
[0028] The prediction results of the autoregressive integrated moving average model and the long short-term memory network are combined and the final prediction value of the distributed power output is generated through weighted fusion.
[0029] As a preferred solution of the method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to the present invention, wherein: the method for evaluating the carrying capacity of distributed power according to the real-time prediction result includes:
[0030] Calculate the maximum power supply capacity of distributed power sources in the future based on the joint distribution density function;
[0031] Based on the joint distribution density function, the joint distribution probability of the distributed power output characteristics and environmental monitoring data within the stable range, as well as the joint distribution probability of the distributed power output characteristics exceeding the safety threshold, are quantified to obtain the reliability index of the distributed power output;
[0032] The power supply capacity index and the reliability index are weighted and fused, and output as a distributed power supply carrying capacity evaluation index, wherein the power supply capacity index refers to the maximum power supply capacity of the distributed power supply calculated based on the joint distribution density function.
[0033] A load-bearing capacity evaluation system based on high-dimensional combination of distributed power outputs using any of the methods described in the present invention comprises:
[0034] The acquisition module is used to collect target data, build a high-dimensional joint distribution model based on the target data, and obtain the characteristic relationship between the target data;
[0035] The analysis module is used to use the high-dimensional joint distribution model combined with time series to make real-time predictions on the output of distributed power sources;
[0036] The evaluation module is used to evaluate the carrying capacity of distributed power sources based on real-time prediction results.
[0037] A computer device comprises: a memory and a processor; the memory stores a computer program, comprising: the steps of implementing any one of the methods of the present invention when the processor executes the computer program.
[0038] A computer-readable storage medium stores a computer program, comprising: when the computer program is executed by a processor, the steps of implementing any one of the methods of the present invention are implemented.
[0039] Beneficial effects of the present invention: The method of the present invention accurately captures the complex dependency between the output of distributed power sources and environmental variables and the operating status of the power grid by constructing a high-dimensional joint distribution model based on kernel density estimation and Gaussian Copula; combines time series analysis and deep learning models to achieve high-precision real-time prediction of the output of distributed power sources; integrates the power supply capacity, power supply stability and reliability evaluation results, quantifies the system carrying capacity and outputs optimization indicators. Compared with the existing technology, the present invention significantly improves the accuracy and dynamic adaptability of carrying capacity evaluation, and provides a scientific basis for power grid scheduling and optimization. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:
[0041] Figure 1 An overall flow chart of a method for evaluating carrying capacity based on high-dimensional combination of distributed power outputs provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0042] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.
[0043] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a carrying capacity evaluation method based on high-dimensional combination of distributed power output, including:
[0044] S1: Collect target data, build a high-dimensional joint distribution model based on the target data, and capture the characteristic relationship between the target data.
[0045] Furthermore, in order to accurately evaluate the output characteristics of distributed power sources and the carrying capacity of the power grid, after collecting the target data, a high-dimensional joint distribution model is constructed to fully capture the characteristic relationship between the data. The target data mainly includes three categories: distributed power output data, environmental monitoring data, and power grid operation data. Among them, the distributed power output data describes the dynamic power supply characteristics of distributed power sources such as photovoltaics and wind power at different times; environmental monitoring data includes key environmental parameters such as wind speed, light intensity and temperature, which directly affect the output fluctuations of distributed power sources; power grid operation data records the changes in power grid load, voltage and frequency, reflecting the dynamic state of power grid operation.
[0046] After the target data is collected, an autoencoder with three hidden layers is used to extract the main features of the output and perform data dimensionality reduction. The autoencoder is trained to map the input data to a low-dimensional latent space, so that the number of neurons in the middle layer is less than that in the input layer, thereby achieving feature compression. The compressed feature representation retains the main information of the original data while reducing the interference of redundancy and noise.
[0047] Specifically, the autoencoder includes an input layer, an output layer, and three hidden layers. The number of neurons in the input layer is the same as the number of original features, and is used to input the collected target data. The first hidden layer has a large number of neurons, and an activation function (such as ReLU) is used to introduce nonlinearity; the number of neurons in the middle layer is reduced to achieve feature compression, and the output of this layer will be used as a low-dimensional representation of the data; the number of neurons in the third hidden layer gradually increases, which is symmetrical with the first hidden layer and also uses an activation function; finally, the extracted features are output through the output layer.
[0048] In the feature extraction process, the autoencoder focuses on environmental factors that are highly correlated with the output of distributed power sources. By calculating the correlation index, the top three variables with the greatest impact on the output are selected from the environmental monitoring data. For example, in the output modeling of a wind farm, wind speed, wind direction, and temperature may be selected as the main input features. This feature screening method not only improves the representativeness of the model input, but also significantly reduces the computational complexity of high-dimensional joint distribution modeling.
[0049] Furthermore, in order to fully describe the dependency of the target data, this embodiment uses kernel density estimation (KDE) and Gaussian Copula to construct a high-dimensional joint distribution model. First, the probability density function (PDF) is estimated for each input feature (such as output features and selected environmental variables). The expression of kernel density estimation is:
[0050] ,
[0051] in, Indicates the estimated The probability density function value at ; Represents sample data; Indicates bandwidth; Indicates the target point of the current probability density calculation; Indicates the sample data points; Represents the kernel function.
[0052] right Integrate to get the cumulative distribution function and get the cumulative distribution function set for each data point , as the input of the high-dimensional joint distribution model. Standardization not only facilitates subsequent modeling, but also eliminates the impact of variable dimension differences on the construction of joint distribution.
[0053] Distributed power data comes from a variety of heterogeneous systems, and their distribution forms vary. Traditional linear modeling methods cannot accurately describe the nonlinear and asymmetric dependencies between variables. The Gaussian Copula method is used to construct the joint distribution between input features to solve the modeling problem of the complex dependencies between distributed power output, environmental monitoring data, and power grid operation data. The joint distribution density function is expressed as:
[0054] ,
[0055] in, Represents the density function of the Gaussian Copula joint distribution; Indicates The cumulative distribution function value of the variables; Represents the correlation matrix of Gaussian Copula; Representation Matrix The determinant of ; A vector of inverse transformed values representing the standard normal distribution; express The transpose of .
[0056] The Gaussian Copula model separates the nonlinear dependencies between variables from the marginal distribution, making the joint distribution modeling flexible to adapt to various marginal distribution forms while ensuring a more accurate description of the dependencies.
[0057] It should be noted that the Gaussian Copula model successfully combines variables of various distribution forms into a model, overcoming the limitation of traditional modeling methods that require variables to satisfy a specific distribution. For example, the output of distributed power sources may present a peak distribution, while environmental monitoring variables are mostly normally distributed. The Gaussian Copula model can seamlessly integrate these variables with different distributions.
[0058] S2: Use high-dimensional joint distribution models combined with time series analysis to make real-time predictions on the output of distributed power sources.
[0059] Furthermore, the marginal distribution (CDF) of the high-dimensional joint distribution model is used to standardize the main output characteristics and related environmental variables to generate dimensionless input data as the input for time series analysis.
[0060] ,
[0061] in, Represents dimensionless input data; Representation variables The cumulative distribution function of .
[0062] The autoregressive integrated moving average model is used to independently model each variable to predict short-term dynamic behavior, and the ARIMA model is used to model each variable. The output characteristics and environmental variables of distributed power sources usually show a stable trend in the short term, and their short-term dynamic behavior can be captured by univariate time series modeling.
[0063] The ARIMA model uses the autoregressive characteristics of historical data and the moving average characteristics of residuals to perform short-term dynamic modeling of the time evolution of output characteristics and environmental factors. For example, the output characteristics of distributed power sources are disturbed by changes in environmental factors in the short term. The ARIMA model can capture this short-term dynamic fluctuation and provide high-precision short-term forecasts.
[0064] After the standardized As input, the LSTM model is used to capture its long-term and short-term dynamic characteristics and nonlinear dependencies. The gating mechanism of LSTM (such as forget gate and input gate) can effectively handle the long-term dependency characteristics of variables and provide more comprehensive dynamic prediction capabilities for distributed power output.
[0065] The LSTM model receives the combined input , capturing complex nonlinear dependencies between variables through multi-layer recursive networks. For example, in photovoltaic power generation systems, the nonlinear effects of light intensity and temperature on output characteristics cannot be fully expressed by single-variable modeling, while LSTM can model the long-term and short-term dependencies between these variables and provide comprehensive prediction results for system dynamic characteristics.
[0066] The prediction results of ARIMA and LSTM are fused by weighted method to generate the final prediction value. The weights are adjusted dynamically during the fusion process, and different models are given appropriate contribution ratios according to the prediction stability of the model (such as the size of the residual), ensuring that the final prediction result is balanced between short-term accuracy and long-term trend.
[0067] S3: Evaluate the carrying capacity of distributed power sources based on real-time prediction results.
[0068] Furthermore, the maximum power supply capacity of distributed power sources in the future is calculated through the joint distribution density function:
[0069] ,
[0070] in, Indicates the predicted maximum power supply capacity; Indicates the target value or upper limit of the power supply capacity; Output characteristics The standardized cumulative distribution function value of .
[0071] Calculated by points ,The high-dimensional joint distribution model can quantify the maximum power supply capacity of distributed power sources in the future time period. This process not only takes into account the dynamic changes in output characteristics, but also combines the complex correlation between environmental variables and grid operation status, providing a more comprehensive quantitative basis for power supply capacity evaluation.
[0072] According to the joint distribution probability density function, the probability of the joint behavior of output characteristics and environmental variables being within a stable range is analyzed:
[0073] ,
[0074] in, It represents the joint probability that the output characteristics of distributed generation and environmental variables are within the stable range; and Indicates the upper and lower bounds of the stable output range.
[0075] Traditional analysis mostly assumes independent variables and is difficult to capture the dependencies between multiple variables. However, this embodiment uses Gaussian Copula to describe the joint behavior between variables, which can comprehensively consider the joint impact of output characteristics and environmental variables on power supply stability.
[0076] Furthermore, based on the joint distribution probability density, the reliability of the system under critical conditions is evaluated and the probability of system variables exceeding the safety range is quantified. The joint distribution probability under certain specific conditions (such as environmental variables exceeding critical values) is calculated:
[0077] ,
[0078] in, It represents the probability that the output characteristic exceeds the safety threshold, that is, the joint probability of failure under certain specific conditions; The standardized CDF value corresponding to the safety threshold of the output characteristic is shown. The failure probability is quantified through the high-dimensional joint distribution model, and the reliability of distributed power generation can be analyzed under specific conditions and the reliability index can be calculated.
[0079] Through the correlation matrix in the high-dimensional joint distribution, the variables that have the greatest impact on power supply reliability are identified, and the sensitivity index is output. Traditional methods can only evaluate safety based on a single variable, while the high-dimensional joint distribution model comprehensively considers the interactive effects of multiple variables, and the resulting failure probability is more accurate. Sensitivity analysis identifies key variables through the correlation matrix. For example, if wind speed has the greatest impact on power supply reliability, the wind power generation system can be optimized first.
[0080] The power supply capacity and power supply reliability results are integrated to quantify the carrying capacity of the distributed power system and output the evaluation index. The carrying capacity evaluation index can be expressed as a weighted fusion of the power supply capacity index and the reliability index. Output the carrying capacity evaluation index, and combine charts and visualization to show the comprehensive performance of the distributed power supply. The power supply capacity index refers to the maximum power supply capacity calculated above.
[0081] Embodiment 2, based on the same inventive concept, the second embodiment of the present invention further provides a load-bearing capacity evaluation system based on high-dimensional combination of distributed power output, including:
[0082] The acquisition module collects target data, builds a high-dimensional joint distribution model based on the target data, and captures the characteristic relationship between the target data;
[0083] The analysis module uses a high-dimensional joint distribution model combined with time series analysis to make real-time predictions on the output of distributed power sources;
[0084] The evaluation module evaluates the carrying capacity of distributed power sources based on real-time prediction results.
[0085] It is worth pointing out that the system embodiment corresponds to the above-mentioned method embodiment, and the implementation methods of the above-mentioned method embodiment are all applicable to the system embodiment and can achieve the same or similar technical effects, so they will not be repeated here.
[0086] Example 3, the following is an embodiment of the present invention, which provides a carrying capacity assessment method based on high-dimensional combination of distributed power output. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.
[0087] In order to verify the effectiveness of the carrying capacity assessment method based on high-dimensional joint distributed power output, the experiment took a certain wind farm as the test object and collected the distributed power output data, environmental monitoring data (wind speed, temperature) and power grid operation data (voltage, load) of the wind farm. The experimental goal is to evaluate the power supply capacity, stability and reliability of the wind farm in the next hour through high-dimensional joint distribution model modeling and time series prediction, and verify the innovation and advantages of this method.
[0088] The experimental data are recorded as follows. The data fields include time, wind power output, wind speed, temperature, voltage and load. The recorded data reflect the dynamic characteristics of the experimental object.
[0089] Table 1 Experimental data table
[0090]
[0091] Through the high-dimensional joint distribution model, the maximum power supply capacity of the wind farm in the next hour is calculated and predicted to be 520kW. Compared with the actual data, it can be seen that the model has a high prediction accuracy for dynamic power supply capacity, which improves the reliability and sensitivity of the prediction compared with the traditional linear model.
[0092] According to the joint distribution probability density function, the joint probability of wind power output, wind speed and temperature being within the stable range (450 kW to 520 kW) is calculated to be 85%, which indicates that under the experimental conditions, the output characteristics of the wind farm have a high stability.
[0093] Traditional methods usually assume the independence of variables, while the present invention captures the nonlinear effects of wind speed and temperature on wind power output through a Gaussian Copula model, significantly improving the accuracy of stability assessment.
[0094] According to the joint distribution probability density function, the failure probability is calculated to be 15% when the wind power output exceeds 500 kW (safety threshold). Sensitivity analysis shows that the change of wind speed has the greatest impact on power supply reliability, and its correlation index value reaches 0.92.
[0095] Through model analysis, the wind farm dispatch strategy can be optimized first for periods with large wind speed fluctuations to reduce the risk of failure. Through the above simulation experiments, it can be seen that the high-dimensional joint distribution model of the present invention effectively integrates the data of wind power output, environmental variables and grid operation status, and its results have significant advantages in prediction accuracy and comprehensive evaluation capabilities compared with traditional linear modeling methods.
[0096] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.
[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.
[0098] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.
[0099] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0100] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A method for evaluating the carrying capacity based on high-dimensional integration of distributed power output, characterized in that: include: Collect target data, build a high-dimensional joint distribution model based on the target data, and obtain the characteristic relationship between the target data; Use high-dimensional joint distribution models combined with time series to make real-time predictions of the output of distributed power sources; Evaluate the carrying capacity of distributed power sources based on real-time prediction results.
2. According to claim 1, a method for evaluating the carrying capacity based on high-dimensional combination of distributed power output is characterized in that: The target data includes distributed power output data, environmental monitoring data and power grid operation data.
3. According to claim 2, a method for evaluating the carrying capacity based on high-dimensional combination of distributed power output is characterized in that: After collecting the target data, the method further includes: The distributed power output features are extracted through an autoencoder with three hidden layers, and the top three environmental monitoring data with the highest correlation with the distributed power output data are selected as the input features of the high-dimensional joint distribution model; The number of neurons in the middle layer of the autoencoder is less than that in the input layer, so feature compression is performed.
4. The method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to claim 3 is characterized in that: The high-dimensional joint distribution model includes: Kernel density estimation is used to estimate the probability density function of distributed power output characteristics and screened environmental monitoring data: ， in, Indicates the estimated The probability density function value at ; Represents sample data; Indicates bandwidth; Indicates the target point of the current probability density calculation; Indicates the first data points; represents the kernel function; right Integrate to get the cumulative distribution function and get the cumulative distribution function value set for each data point .
5. The method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to claim 4 is characterized in that: The high-dimensional joint distribution model also includes: The cumulative distribution function is set As the input of the high-dimensional joint distribution model, Gaussian Copula is used to construct the joint distribution between inputs. The joint distribution density function is expressed as: ， in, Represents the density function of the Gaussian Copula joint distribution; Indicates The cumulative distribution function value of the data points; Represents the correlation matrix of Gaussian Copula; Representation Matrix The determinant of ; A vector of inverse transformed values representing the standard normal distribution; express The transpose of .
6. The method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to claim 5 is characterized in that: The high-dimensional joint distribution model is used in combination with time series to make real-time predictions on the output of distributed power sources, including: Cumulative distribution function values using marginal distributions As input for time series analysis; Use the autoregressive integrated moving average model to independently model each data point and make short-term forecasts on the output of distributed generation; After the standardized As input, the long short-term memory network is used to make long-term predictions of the distributed generation output; The prediction results of the autoregressive integrated moving average model and the long short-term memory network are combined and the final prediction value of the distributed power output is generated through weighted fusion.
7. The method for evaluating the carrying capacity based on high-dimensional combination of distributed power output according to claim 6 is characterized by: The evaluation of the carrying capacity of the distributed power source according to the real-time prediction results includes: Calculate the maximum power supply capacity of distributed power sources in the future based on the joint distribution density function; Based on the joint distribution density function, the joint distribution probability of the distributed power output characteristics and environmental monitoring data within the stable range, as well as the joint distribution probability of the distributed power output characteristics exceeding the safety threshold, are quantified to obtain the reliability index of the distributed power output; The power supply capacity index and the reliability index are weighted and fused, and output as a distributed power supply carrying capacity evaluation index, wherein the power supply capacity index refers to the maximum power supply capacity of the distributed power supply calculated based on the joint distribution density function.
8. A load-bearing capacity assessment system based on high-dimensional integration of distributed power output, characterized in that: The system is used to implement the carrying capacity assessment method based on high-dimensional combination of distributed power output as described in any one of claims 1 to 7, wherein the system comprises: The acquisition module is used to collect target data, build a high-dimensional joint distribution model based on the target data, and obtain the characteristic relationship between the target data; The analysis module is used to use the high-dimensional joint distribution model combined with time series to make real-time predictions on the output of distributed power sources; The evaluation module is used to evaluate the carrying capacity of distributed power sources based on real-time prediction results.
9. A computer device comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the carrying capacity assessment method based on high-dimensional combination of distributed power output as described in any one of claims 1-7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for carrying capacity assessment based on high-dimensional combination of distributed power outputs as described in any one of claims 1-7 are implemented.