New energy multi-station output scene generation method and system

Abstract

The invention discloses a new energy multi-station output scene generation method. The method comprises the steps that historical operation data information of a target new energy multi-station is acquired and preprocessed to construct a data set; fitting a historical power prediction error CDF quantile of the target new energy multi-station based on Tukey distribution; carrying out scene reordering considering the time-space correlation so as to construct an optimization model of multi-station new energy output scene reordering considering the time-space correlation; and solving the constructed model to obtain a new energy output scene of the target new energy multi-station on the to-be-predicted day, and completing generation of the output scene of the target new energy multi-station. The invention also discloses a system for realizing the new energy multi-station output scene generation method. According to the method, the generation of the new energy multi-station output scene can be realized, the reliability is higher, and the accuracy is better.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of electrical automation, and particularly relates to a new energy multi-station output scene generation method and system. BACKGROUND

[0002] With the development of economy and technology and the improvement of people's living standards, electric energy has become an essential secondary energy in people's production and life, bringing endless convenience to people's production and life. Therefore, ensuring the stable and reliable supply of electric energy has become one of the most important tasks of the power system.

[0003] At present, with the increasingly serious environmental problems, more and more new energy power generation systems begin to be integrated into the power system and generate electricity. The randomness and volatility of the output of the new energy power generation system bring great risks to the safe and stable operation of the power system. Therefore, simulating the output scene of the new energy multi-station has important significance for the safe and stable operation and risk assessment of the power system.

[0004] At present, the traditional scene generation method of the new energy multi-station usually only considers the time correlation of errors (application number 202210687578.3 corresponding patent) or relies on the Copula function related parameter assumption to describe the dependence structure between random variables (application number 202110772633.4 corresponding patent). However, the former scheme ignores the spatial correlation between stations, resulting in distortion of the spatial distribution of the generated joint scene, which is difficult to effectively support market equilibrium analysis; and the latter scheme faces the challenges of model selection difficulty, parameter estimation complexity and large amount of calculation when dealing with multi-station (high-dimensional) problems, which makes it difficult to guarantee the consistency of the time and space correlation of the generated scene and historical samples, limiting its application in engineering practice. SUMMARY

[0005] One of the purposes of the present application is to provide a new energy multi-station output scene generation method with high reliability and good accuracy.

[0006] The second purpose of the present application is to provide a system for implementing the new energy multi-station output scene generation method.

[0007] The new energy multi-station output scene generation method provided by the present application comprises the following steps:

[0008] S1. Obtain the historical operation data information of the target new energy multi-station;

[0009] S2. Preprocess the data information obtained in step S1 to construct a data set;

[0010] S3. According to the data information obtained in step S2, based on the Tukey distribution, fitting the historical power prediction error CDF quantile of the target new energy multi-station;

[0011] S4. According to the data information obtained in step S3, scene reordering considering the spatio-temporal correlation is performed to construct an optimization model of multi-station new energy output scene reordering considering the spatio-temporal correlation;

[0012] S5. The model constructed in step S4 is solved to obtain the new energy output scene of the target new energy multi-station on the day to be predicted, and the generation of the output scene of the target new energy multi-station is completed.

[0013] The historical operation data information of the target new energy multi-station obtained in step S1 specifically includes the following steps:

[0014] The historical operation data information of the target new energy multi-station is obtained.

[0015] The historical operation data information includes the historical output actual value and the corresponding historical output prediction value of the target new energy multi-station.

[0016] The data information obtained in step S1 is preprocessed to construct a data set, and the preprocessing specifically includes the following steps:

[0017] The data information obtained in step S1 is supplemented with default values by using a cubic spline interpolation algorithm;

[0018] Unit root test algorithm is used for stationarity test of data;

[0019] The data is standardized by using the following formula:

[0020] In the formula, is the actual power of the mth new energy station at the jth day and the tth hour after standardization; is the actual power of the mth new energy station at the jth day and the tth hour before standardization; is the prediction power of the mth new energy station at the jth day and the tth hour after standardization; is the prediction power of the mth new energy station at the jth day and the tth hour before standardization; is the installed capacity of the mth new energy station;

[0021] The historical power prediction error of each new energy station is calculated by using the following formula:

[0022] In the formula is the power prediction error of the jth historical day and the tth time of the mth new energy field station;

[0023] The following formula is used to normalize the new energy power prediction value of each new energy field station in the day to be predicted:

[0024] In the formula is the predicted power of the mth new energy field station at the tth time of the day to be predicted after normalization; is the predicted power of the mth new energy field station at the tth time of the day to be predicted before normalization.

[0025] The data information obtained according to step S2 is used to fit the historical power prediction error CDF quantile of the target new energy multi-field station based on Tukey distribution, and specifically includes the following steps:

[0026] According to the data information obtained in step S2, an error sample matrix is constructed

[0027] In the formula is the total number of time periods; is the total number of historical days;

[0028] The quantile function of Tukey distribution is set as , which is expressed as:

[0029] In the formula is the first adjustment parameter of the Tukey quantile function; is the second adjustment parameter of the Tukey quantile function; is the quantile of the standard normal distribution CDF; is the probability value;

[0030] C typical probability values are selected to construct a probability set ;

[0031] The empirical distribution quantile of the probability of each row error sample in the statistical calculation ;

[0032] Taking as the benchmark, the parameter optimization objective function of Tukey distribution is defined, which is expressed as:

[0033] In the formula ​​​Let be the weight of the c-th typical probability value, and ; The first adjustment parameter of the Tukey distribution function of the m-th renewable energy power station at time t; The second adjustment parameter is the Tukey distribution function of the m-th renewable energy power station at time t.

[0034] The objective function for optimizing the parameters of the constructed Tukey distribution is solved to obtain the optimal distribution fitting parameters. ;

[0035] according to Calculate any probability CDF quantiles of prediction errors for each new energy power station at each time period :

[0036] In the formula Let be the historical prediction mean of the m-th renewable energy power station at time t; Let m be the historical prediction error standard deviation of the m-th renewable energy power station at time t.

[0037] Based on Latin hypercube sampling, probabilities are set sequentially. for , Given the total number of scenarios, the prediction error CDF matrix for each new energy power station is constructed, as follows:

[0038] In the formula Let be the prediction error CDF matrix of the m-th renewable energy power station, and let be This serves as the initial set of prediction error scenarios.

[0039] Step S4, which involves reordering scenarios based on the data obtained in step S3, considering spatiotemporal correlation, to construct an optimization model for reordering multi-station renewable energy output scenarios based on spatiotemporal correlation, specifically includes the following steps:

[0040] right The scene vector at each time step, after undergoing the corresponding elementary transformation matrix The prediction error scenario vectors for each new energy power station in each time period are calculated and obtained. , is represented as:

[0041] In the formula For the m-th renewable energy power station at time t, the N-th I The prediction error value for each scenario; Let m be the probability value corresponding to the m-th renewable energy power station at time t. The prediction error value;

[0042] The optimization objective is to minimize the spatiotemporal correlation differences between the power output scenarios of each renewable energy power station and the sample data, thereby optimizing... ;

[0043] The following formula is used to calculate the value of each new energy power station at any time. and Time correlation coefficient between historical error samples :

[0044] In the formula This is a collection of historical operating days for new energy power plants.

[0045] The following formula is used to calculate the scene vectors of all prediction errors. The time correlation coefficient of the error scenario matrix:

[0046] In the formula For the m-th new energy power station Time and the Correlation coefficient of error scenarios over time; A collection of scenes; For the m-th new energy power station The prediction error of the i-th scene at time i;

[0047] To ensure that the temporal correlation of error scenarios for each new energy power station is as close as possible to the error samples after sorting, a first objective function is constructed:

[0048] In the formula The value of the first objective function; For the collection of new energy power stations; It is a 2-norm function;

[0049] The correlation coefficients between historical power output samples of each renewable energy power station in each time period are calculated using the following formula:

[0050] In the formula For time t, the first... The first new energy power station and the first Spatial correlation coefficient between the historical power of each new energy power station;

[0051] The renewable energy output scenario for the predicted date is calculated using the following formula:

[0052]

[0053] In the formula Let be the power of the m-th renewable energy power station at time t in scenario i;

[0054] The following formula is used to calculate the value of time t on the day to be predicted. The first new energy power station and the first Spatial correlation coefficient between the power output scenarios of individual new energy power plants :

[0055] To ensure and Construct a second objective function that is as close as possible to the target function:

[0056] In the formula This is the value of the second objective function; For the first The first new energy power station and the first The weights of the spatial correlation coefficients among individual new energy power stations;

[0057] Finally, the optimization model for reordering the renewable energy output scenarios of multiple power stations based on the spatiotemporal correlation of construction and design includes:

[0058] Objective functions: Minimize the first objective function and minimize the second objective function;

[0059] Constraints: Prediction error scenario vectors for each renewable energy power station after sorting by time period. The calculation formula for each new energy power station at any time and Time correlation coefficient between historical error samples The calculation formula for the m-th new energy power station Time and the Time-of-flight error correlation coefficient The calculation formula, time t, the first The first new energy power station and the first Spatial correlation coefficient between the historical power of individual new energy power plants The calculation formula for the power of the m-th renewable energy power station at time t in scenario i. The calculation formula and the time of the day to be predicted at time t The first new energy power station and the first Spatial correlation coefficient between the power output scenarios of individual new energy power plants The calculation formula;

[0060] Decision variables: .

[0061] Step S5 involves solving the model constructed in step S4 to obtain the new energy output scenario of the target new energy multi-site on the predicted date. Specifically, this includes the following steps:

[0062] A multi-objective optimization algorithm is used to solve the optimization model for reordering the spatiotemporal correlation of new energy output scenarios of multiple power stations constructed in step S4.

[0063] After solving the problem, a sorting algorithm that approximates the ideal solution is used to obtain the bi-objective Pareto optimal solution, and the results are calculated. The value; will This serves as the final scenario for the new energy output of each new energy power station on the predicted date.

[0064] This invention also provides a system for implementing the method for generating power output scenarios from multiple new energy power plants, comprising a data acquisition module, a data processing module, a quantile fitting module, a model building module, and a scenario generation module; the data acquisition module, data processing module, quantile fitting module, model building module, and scenario generation module are connected in series; the data acquisition module is used to acquire historical operating data information of the target multiple new energy power plants and upload the data information to the data processing module; the data processing module is used to preprocess the acquired data information according to the received data information to construct a dataset and upload the data information to the quantile fitting module; the quantile fitting module is used to preprocess the acquired data information according to the received data information. Based on the received data, the system fits the CDF quantiles of the historical power prediction error of the target renewable energy multi-station based on the Tukey distribution, and uploads the data to the model building module. The model building module is used to perform scene reordering based on the received data, calculating spatiotemporal correlation, to construct an optimized model for reordering renewable energy output scenarios of the multi-station based on spatiotemporal correlation, and uploads the data to the scene generation module. The scene generation module is used to solve the constructed model based on the received data to obtain the renewable energy output scenarios of the target renewable energy multi-station on the forecast date, thus completing the generation of the output scenarios of the target renewable energy multi-station.

[0065] The method and system for generating power output scenarios of multiple new energy power plants disclosed in this invention are based on fitting the error quantiles of the Tukey distribution and solving the bi-objective optimization function based on spatiotemporal correlation. Therefore, this invention can not only generate power output scenarios of multiple new energy power plants, but also has higher reliability and better accuracy. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0067] Figure 2 This is a schematic diagram of the functional modules of the system of the present invention. DETAILED DESCRIPTION

[0068] As Figure 1 The method flowchart of the method of the application is shown: the new energy multi-field station output scene generation method disclosed by the application comprises the following steps:

[0069] S1. Obtain historical operation data information of a target new energy multi-field station; specifically comprising the following steps:

[0070] Obtain historical operation data information of the target new energy multi-field station;

[0071] The historical operation data information comprises historical output actual values and corresponding historical output prediction values of the target new energy multi-field station;

[0072] S2. Preprocess the data information obtained in step S1 to construct a data set; specifically comprising the following steps:

[0073] The data information obtained in step S1 is supplemented with default values by using a cubic spline interpolation algorithm (for reference, patent application file: CN120405310A);

[0074] Unit root test algorithm (for reference, patent application file: CN119917985B) is used to test the stationarity of the data;

[0075] The data is standardized by using the following formula:

[0076] In the formula, is the actual power of the mth new energy field station at the jth day and the tth time after standardization; is the actual power of the mth new energy field station at the jth day and the tth time before standardization; is the prediction power of the mth new energy field station at the jth day and the tth time after standardization; is the prediction power of the mth new energy field station at the jth day and the tth time before standardization; is the installed capacity of the mth new energy field station;

[0077] The historical power prediction error of each new energy field station is calculated by using the following formula:

[0078] In the formula, is the power prediction error of the mth new energy field station at the jth day and the tth time;

[0079] The following formula is used to standardize the predicted renewable energy power output of each renewable energy power station on the forecast date:

[0080] In the formula The predicted power of the m-th renewable energy power station at time t on the day to be predicted is standardized. The predicted power of the m-th renewable energy power station at time t on the day to be predicted before standardization;

[0081] The date to be predicted is a future date, and historical data is a historical date. Error scenarios obtained through the statistical analysis of historical data will be superimposed with the predicted power of the date to be predicted to obtain the new energy power scenario of the date to be predicted. The new energy power scenario of the date to be predicted serves as the benchmark for spatial correlation ranking and will be used to construct the second objective function.

[0082] S3. Based on the data obtained in step S2, and using the Tukey distribution, fit the CDF quantiles of the historical power prediction error for the target renewable energy multi-station; specifically, this includes the following steps:

[0083] Based on the data obtained in step S2, construct the error sample matrix. :

[0084] In the formula This represents the total number of time periods; Total number of historical days;

[0085] Define the quantile function of the Tukey distribution , is represented as:

[0086] In the formula This is the first adjustment parameter for the Tukey quantile function; This is the second adjustment parameter for the Tukey quantile function; For the standard normal distribution CDF quantiles; This is a probability value;

[0087] Select C typical probability values ​​to construct a probability set. In specific implementation, 10 typical probability values ​​are selected, namely: 1%, 5%, 10%, 25%, 40%, 50%, 60%, 75%, 95%, and 99%.

[0088] Statistical calculation The probability corresponding to each row error sample empirical distribution quantiles ;

[0089] With as the benchmark, the parameter optimization objective function of the Tukey distribution is defined, denoted as:

[0090] In the formula is the weight of the c-th typical probability value, and ; is the first adjustment parameter of the Tukey distribution function of the m-th new energy station at the t-th time; is the second adjustment parameter of the Tukey distribution function of the m-th new energy station at the t-th time; the subscript t represents that each error sample of each station at each time period corresponds to a group of Tukey quantile parameters to be optimized, that is, the Tukey distribution parameter optimization problem of each station at each time period is independent of each other;

[0091] Solve the constructed parameter optimization objective function of the Tukey distribution (the least square correlation algorithm can be used for solving, such as the trust region method), to obtain the optimal distribution fitting parameter ;

[0092] According to , the prediction error CDF quantile of each new energy station at each time period is calculated :

[0093] In the formula is the historical prediction error mean of the m-th new energy station at the t-th time; is the historical prediction error standard deviation of the m-th new energy station at the t-th time;

[0094] Based on Latin hypercube sampling, the probability is set in turn , is the total number of scenarios, so as to construct the prediction error CDF matrix of each new energy station, denoted as:

[0095] In the formula is the prediction error CDF matrix of the m-th new energy station, and is taken as the initial prediction error scenario set;

[0096] S4. According to the data information obtained in step S3, the scenario reordering considering the space-time correlation is performed to construct an optimization model of multi-station new energy output scenario reordering considering the space-time correlation; specifically including the following steps:

[0097] to ​The scene vector at each time step, after undergoing the corresponding elementary transformation matrix The prediction error scenario vectors for each new energy power station in each time period are calculated and obtained. , represented as:

[0098] For the m-th renewable energy power station at time t, the N-th I The prediction error value for each scenario; Let m be the probability value corresponding to the m-th renewable energy power station at time t. The prediction error value;

[0099] The optimization objective is to minimize the spatiotemporal correlation differences between the power output scenarios of each renewable energy power station and the sample data, thereby optimizing... ;

[0100] The following formula is used to calculate the value of each new energy power station at any time. and Time correlation coefficient between historical error samples :

[0101] In the formula This is a collection of historical operating days for new energy power plants;

[0102] The following formula is used to calculate the scene vectors of all prediction errors. The time correlation coefficient of the error scenario matrix:

[0103] In the formula For the m-th new energy power station Time and the Correlation coefficient of error scenarios over time; A collection of scenes; For the m-th new energy power station The prediction error of the i-th scene at time i;

[0104] To ensure that the temporal correlation of error scenarios for each new energy power station is as close as possible to the error samples after sorting, a first objective function is constructed:

[0105] In the formula The value of the first objective function; For the collection of new energy power stations; It is a L2 norm function; given that the correlation between time periods that are far apart is often weak and has little impact on the evolution trend of prediction error, it is set to... The time period range is calculated according to the correlation coefficient, so as to simplify the calculation and improve the reliability, wherein The time period set participating in the time correlation coefficient calculation;

[0106] Different from the time correlation calculation based on the prediction error, since the spatial correlation between the stations is essentially derived from the common geographical-meteorological environment affecting the absolute output level, rather than only affecting the prediction deviation, the spatial correlation coefficient calculation needs to be based on the station output (i.e. the power after the prediction value is superimposed with the error); therefore, the following formula is used to calculate the correlation coefficient between the historical output samples of each new energy station in each time period:

[0107] In the formula, is the spatial correlation coefficient between the historical power of the th new energy station and the th new energy station at the t th time;

[0108] The new energy output scenario of the day to be predicted is calculated by the following formula:

[0109] In the formula, is the power of the i th scenario of the m th new energy station at the t th time;

[0110] The spatial correlation coefficient between the output scenarios of the th new energy station and the th new energy station at the t th time of the day to be predicted is calculated by the following formula:

[0111] In order to ensure that and are as close as possible, a second objective function is constructed:

[0112] In the formula, is the value of the second objective function; is the weight of the spatial correlation coefficient between the th new energy station and the th new energy station;

[0113] Finally, the optimization model of the multi-station new energy output scenario reordering considering the time and space correlation is constructed, including:

[0114] Objective function: minimize the first objective function and minimize the second objective function;

[0115] ​Constraint: the sorted predicted error scenario vector of each new energy station in each time period Calculation formula of the time correlation coefficient of historical error samples between each new energy station at any time and Calculation formula of the error scenario correlation coefficient between the mth new energy station at the tth time and the tth time Calculation formula of the spatial correlation coefficient between the tth time the ith scenario of the mth new energy station and the tth time the ith scenario of the mth new energy station Calculation formula of the power of the tth time the ith scenario of the mth new energy station Calculation formula of the spatial correlation coefficient between the tth time the ith scenario of the mth new energy station and the tth time the ith scenario of the mth new energy station Calculation formula of the power of the tth time the ith scenario of the mth new energy station

[0116] Decision variable: ;

[0117] S5. Solve the model constructed in step S4 to obtain the new energy output scenario of the target new energy multi-station on the day to be predicted, and complete the generation of the output scenario of the target new energy multi-station; specifically including the following steps:

[0118] Solve the multi-station new energy output scenario reordering optimization model constructed in step S4 by using a multi-objective optimization algorithm (such as NSGA-III algorithm, which can be referred to in patent application file CN116360355B);

[0119] After solving, the double-objective Pareto optimal solution is obtained by using an ideal solution approximation sorting algorithm (such as TOPSIS method, which can be referred to in patent application file CN120672199A), and the value of is calculated; and the value of is taken as the final new energy output scenario of each new energy station on the day to be predicted.

[0120] The method of the application realizes the joint generation of the output scenarios of multiple new energy stations through the technical route of “error quantile fitting-time and space correlation double-objective optimization”; compared with the traditional method of considering only time correlation or relying on complex parameter assumption for spatial correlation fitting, the application can synchronously and efficiently generate joint output scenarios that maintain historical statistical characteristics in time and space dimensions, and provides a more reliable data basis for the market environment of multi-type power supply competition.

[0121] ​​​​​​​The method of the application innovatively adopts a double optimization target of "time correlation based on prediction error and space correlation based on output" in scene reordering; this design makes the generated scene accurately depict the time sequence fluctuation rule of the prediction deviation of each station itself, and truly reflect the actual power coordination distribution mode of multiple stations as a whole in the power grid, significantly improving the rationality and practicability of the scene.

[0122] The method of the application directly realizes the mapping of historical error samples to CDF quantiles by using Tukey distribution, and can flexibly adapt to the non-Gaussian and thick-tailed characteristics of new energy prediction error by optimizing the tail parameters; compared with traditional high-dimensional distribution modeling methods such as Copula, the application avoids complex model selection and parameter estimation, has better calculation efficiency and numerical stability while ensuring accurate description of extreme probability points (such as 1% / 99%), and effectively solves the dimension disaster problem in high-dimensional scene generation.

[0123] As Figure 2 The system for implementing the new energy multi-station output scene generation method disclosed by the application comprises a data acquisition module, a data processing module, a quantile fitting module, a model construction module and a scene generation module; the data acquisition module, the data processing module, the quantile fitting module, the model construction module and the scene generation module are sequentially connected; the data acquisition module is used for acquiring historical operation data information of a target new energy multi-station, and uploading the data information to the data processing module; the data processing module is used for pre-processing the acquired data information according to the received data information, constructing a data set, and uploading the data information to the quantile fitting module; the quantile fitting module is used for fitting the historical power prediction error CDF quantiles of the target new energy multi-station based on Tukey distribution according to the received data information and the obtained data information, and uploading the data information to the model construction module; the model construction module is used for performing scene reordering considering time and space correlation according to the received data information and the obtained data information, constructing an optimization model of multi-station new energy output scene reordering considering time and space correlation, and uploading the data information to the scene generation module; and the scene generation module is used for solving the constructed model according to the received data information, obtaining the new energy output scene of the target new energy multi-station on a to-be-predicted day, and completing the generation of the output scene of the target new energy multi-station.

Claims

1. A method for generating power output scenarios across multiple new energy power plants, comprising the following steps: S1. Obtain historical operational data information of multiple target new energy power stations; S2. Preprocess the data obtained in step S1 to construct a dataset; S3. Based on the data obtained in step S2, fit the CDF quantile of the historical power prediction error of the target new energy multi-power station based on the Tukey distribution; S4. Based on the data information obtained in step S3, the scenarios with time and space correlation are reordered to construct an optimization model for reordering the multi-station new energy output scenarios with time and space correlation. S5. Solve the model constructed in step S4 to obtain the new energy output scenario of the target new energy multi-site on the day to be predicted, thus completing the generation of the output scenario of the target new energy multi-site.

2. The method for generating power output scenarios from multiple new energy power plants according to claim 1, characterized in that... Step S1, which involves obtaining historical operational data information of the target new energy multi-station, specifically includes the following steps: Obtain historical operational data information of multiple target new energy power stations; The historical operational data includes the actual historical power output of the target new energy power plants and the corresponding predicted historical power output.

3. The method for generating power output scenarios from multiple new energy power plants according to claim 2, characterized in that... Step S2 involves preprocessing the data obtained in step S1 to construct a dataset, specifically including the following steps: The data obtained in step S1 is supplemented with default values ​​using a cubic spline interpolation algorithm; The unit root test algorithm is used to test the stationarity of the data. The following formula is used for data standardization: In the formula This represents the standardized actual power output of the m-th renewable energy power station at time t on day j in history. This represents the actual power output of the m-th renewable energy power station on day j, at time t, before standardization. The predicted power of the m-th renewable energy power station at time t on day j of the historical time standardization; The predicted power of the m-th renewable energy power station on day j, time t, before standardization; Let m be the installed capacity of the m-th renewable energy power station; The historical power prediction error of each new energy power station is calculated using the following formula: In the formula The power prediction error of the m-th renewable energy power station at time t on day j in history; The following formula is used to standardize the predicted renewable energy power output of each renewable energy power station on the forecast date: In the formula The predicted power of the m-th renewable energy power station at time t on the day to be predicted is standardized. The predicted power of the m-th renewable energy power station at time t on the day to be predicted, before standardization.

4. The method for generating power output scenarios from multiple new energy power plants according to claim 3, characterized in that... Step S3, which involves fitting the CDF quantiles of the historical power prediction error of the target renewable energy multi-power station based on the data information obtained in step S2 and the Tukey distribution, specifically includes the following steps: Based on the data obtained in step S2, construct the error sample matrix. : In the formula This represents the total number of time periods; Total number of historical days; Define the quantile function of the Tukey distribution , represented as: In the formula This is the first adjustment parameter for the Tukey quantile function; This is the second adjustment parameter for the Tukey quantile function; For the standard normal distribution CDF quantiles; This is a probability value; Select C typical probability values ​​to construct a probability set. ; Statistical calculation The probability corresponding to each row error sample empirical distribution quantiles ; by Based on this, we define the objective function for optimizing the parameters of the Tukey distribution as follows: In the formula Let be the weight of the c-th typical probability value, and ; The first adjustment parameter of the Tukey distribution function at time t for the m-th renewable energy power station; The second adjustment parameter is the Tukey distribution function of the m-th renewable energy power station at time t. The objective function for optimizing the parameters of the constructed Tukey distribution is solved to obtain the optimal distribution fitting parameters. ; according to Calculate any probability CDF quantiles of prediction errors for each new energy power station at each time period : In the formula Let be the historical prediction mean of the m-th renewable energy power station at time t; Let m be the historical prediction error standard deviation of the m-th renewable energy power station at time t. Based on Latin hypercube sampling, probabilities are set sequentially. for , Given the total number of scenarios, the prediction error CDF matrix for each new energy power station is constructed, as follows: In the formula Let be the prediction error CDF matrix of the m-th renewable energy power station, and let... This serves as the initial set of prediction error scenarios.

5. The method for generating power output scenarios from multiple new energy power plants according to claim 4, characterized in that... Step S4, which involves reordering scenarios based on the data obtained in step S3, considering spatiotemporal correlation, to construct an optimization model for reordering multi-station renewable energy output scenarios based on spatiotemporal correlation, specifically includes the following steps: right The scene vector at each time step, after undergoing the corresponding elementary transformation matrix The prediction error scenario vectors for each new energy power station in each time period are calculated and obtained. , represented as: In the formula For the m-th renewable energy power station at time t, the N-th I The prediction error value for each scenario; Let m be the probability value corresponding to the m-th renewable energy power station at time t. The prediction error value; The optimization objective is to minimize the spatiotemporal correlation differences between the power output scenarios of each renewable energy power station and the sample data, thereby optimizing... ; The following formula is used to calculate the value of each new energy power station at any time. and Time correlation coefficient between historical error samples : In the formula This is a collection of historical operating days for new energy power plants; The following formula is used to calculate the scene vectors of all prediction errors. The time correlation coefficient of the error scenario matrix: In the formula For the m-th new energy power station Time and the Correlation coefficient of error scenarios over time; A collection of scenes; For the m-th new energy power station The prediction error of the i-th scene at time i; To ensure that the temporal correlation of error scenarios for each new energy power station is as close as possible to the error samples after sorting, a first objective function is constructed: In the formula The value of the first objective function; For the collection of new energy power stations; It is a 2-norm function; The correlation coefficients between historical power output samples of each renewable energy power station in each time period are calculated using the following formula: In the formula For time t, the first... The first new energy power station and the first Spatial correlation coefficient between the historical power of each new energy power station; The renewable energy output scenario for the predicted date is calculated using the following formula: In the formula Let be the power of the m-th renewable energy power station at time t in scenario i; The following formula is used to calculate the value of time t on the day to be predicted. The first new energy power station and the first Spatial correlation coefficient between the power output scenarios of individual new energy power plants : To ensure and Construct a second objective function that is as close as possible to the target function: In the formula This is the value of the second objective function; For the first The first new energy power station and the first The weights of the spatial correlation coefficients among individual new energy power stations; Finally, the optimization model for reordering the renewable energy output scenarios of multiple power stations based on the spatiotemporal correlation of construction and design includes: Objective functions: Minimize the first objective function and minimize the second objective function; Constraints: Prediction error scenario vectors for each renewable energy power station after sorting by time period. The calculation formula for each new energy power station at any time and Time correlation coefficient between historical error samples The calculation formula for the m-th new energy power station Time and the Time-of-time error scenario correlation coefficient The calculation formula, time t, the first The first new energy power station and the first Spatial correlation coefficient between the historical power of individual new energy power plants The calculation formula for the power of the m-th renewable energy power station at time t in scenario i. The calculation formula and the time of the day to be predicted at time t The first new energy power station and the first Spatial correlation coefficient between the power output scenarios of individual new energy power plants The calculation formula; Decision variables: .

6. The method for generating power output scenarios from multiple new energy power plants according to claim 5, characterized in that... Step S5 involves solving the model constructed in step S4 to obtain the new energy output scenario of the target new energy multi-site on the predicted date. Specifically, this includes the following steps: A multi-objective optimization algorithm is used to solve the optimization model for reordering the spatiotemporal correlation of new energy output scenarios of multiple power stations constructed in step S4. After solving the problem, a sorting algorithm that approximates the ideal solution is used to obtain the bi-objective Pareto optimal solution, and the results are calculated. The value; will This serves as the final scenario for the new energy output of each new energy power station on the predicted date.

7. A system for implementing the method for generating output scenarios of multiple new energy power plants as described in any one of claims 1 to 6, comprising a data acquisition module, a data processing module, a quantile fitting module, a model building module, and a scenario generation module; the data acquisition module, data processing module, quantile fitting module, model building module, and scenario generation module are connected in series; the data acquisition module is used to acquire historical operating data information of the target multiple new energy power plants and upload the data information to the data processing module; The data processing module is used to preprocess the acquired data information based on the received data information to construct a dataset and upload the data information to the quantile fitting module; the quantile fitting module is used to fit the historical power prediction error CDF quantile of the target new energy multi-station based on the received data information and the acquired data information, and upload the data information to the model building module. The model building module is used to reorder the scenarios based on the received data information and the obtained data information, in order to build an optimized model for reordering the multi-station new energy output scenarios based on the spatiotemporal correlation, and upload the data information to the scenario generation module. The scenario generation module is used to solve the constructed model based on the received data information to obtain the new energy output scenario of the target new energy multi-site on the forecast date, thus completing the generation of the output scenario of the target new energy multi-site.

Images