Method and device for evaluating credible capacity of wind power plant
Through Copula variational inference and I-Spline technology, the problem of incompleteness and nonlinear relationships of wind farm observation data is solved, the accuracy and reliability of wind farm output prediction are improved, and the performance differences of wind turbines are comprehensively considered. , realizes accurate assessment of wind farm power supply capacity.
Patent Information
- Application Number
- CN202411112202.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-08
AI Technical Summary
The incomplete and uncertainty of the observation data of the wind farm affects the accuracy of output prediction. Traditional linear models are difficult to capture the complex nonlinear relationship between wind speed, wind direction and wind turbine power output, and ignore the performance differences of different units.
Copula's variational inference model is used to obtain the posterior distribution of unit reliability in wind farms, and I-Spline is used to construct a wind farm output model with monotonous power output when the wind speed increases, and combines the sequential Monte Carlo method to calculate the transmission system The reliability index of the wind farm is searched through dichotomy.
The accuracy and reliability of wind farm output prediction is improved, and the performance differences and uncertainties of different wind turbines are comprehensively considered, ensuring reliable prediction results when the observation data is incomplete.
Smart Images

Figure CN118917694A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power system reliability, and in particular to a method and device for evaluating the reliable capacity of a wind farm. Background Art
[0002] As the global demand for renewable energy continues to increase, wind power generation, as a clean energy source, occupies an increasingly important position in the power system. The use of wind energy resources not only helps to reduce carbon emissions and mitigate climate change, but also improves energy security and reduces dependence on fossil fuels. However, due to the intermittent and unstable nature of wind power generation, accurately assessing the output capacity and power supply reliability of wind farms has become an important technical challenge. This issue has a profound impact on grid dispatching, the stability of the power system, and the operating efficiency of the power market. Therefore, it is particularly important to develop a method that can accurately assess the output and reliable capacity of wind farms.
[0003] Observation data of wind farms are often incomplete and uncertain. Due to sensor failure, data transmission errors and other reasons, there may be missing values or outliers in the observation data. These incomplete data will affect the accuracy of traditional statistical models, and thus affect the prediction of wind farm output. There is a complex nonlinear relationship between wind speed, wind direction and wind turbine power output. Traditional linear models are difficult to accurately capture these complex relationships, resulting in large deviations in prediction results. In addition, environmental conditions (such as air temperature, humidity, etc.) will also affect the power output of wind turbines, which further increases the difficulty of modeling. Wind farms are composed of multiple wind turbines, and there are differences in performance and reliability between different units. Traditional evaluation methods usually ignore this point and fail to fully consider the performance differences and uncertainties between units, thus affecting the accuracy of the overall output forecast. Summary of the invention
[0004] In view of this, the present invention provides a method and device for evaluating the reliable capacity of a wind farm to solve at least one of the above-mentioned problems.
[0005] In order to achieve the above object, the present invention adopts the following scheme:
[0006] According to a first aspect of the present invention, a method for evaluating credible capacity of a wind farm is provided, the method comprising: obtaining a posterior distribution of reliability of units in the wind farm using a variational inference model of Copula; constructing a wind farm output model whose power output monotonically does not decrease when wind speed increases using I-Spline based on the posterior distribution; generating a joint probability distribution of wind farm output power using the wind farm output model; performing random sampling on the joint probability distribution using a sequential Monte Carlo method to calculate a reliability index of a power transmission system; and performing a credible capacity search using a dichotomy method to ensure that the reliability index of the wind farm after access is equal to that before access.
[0007] As an embodiment of the present invention, the method of using the variational inference model of the Copula to obtain the posterior distribution of the reliability of the wind farm unit includes:
[0008] Preprocessing historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, wherein the historical data includes wind direction D, wind speed W, and unit output power P;
[0009] The preprocessed historical data is used as the observed data X(t), and the unit reliability R(t) is used as the latent variable Z(t);
[0010] Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t):
[0011] q(Z)=C(F(R));
[0012] In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable;
[0013] Combined with the observed data X(t), construct the joint distribution p(X,Z):
[0014] p(X,Z)=p(X|Z)p(Z);
[0015] In the above formula, p(X|Z) is the distribution of observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t);
[0016] The objective function ELBO is constructed as follows:
[0017] ELBO=E q(Z) [logp(D,W,P,Z)-logq(Z)]
[0018] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable.
[0019] By maximizing ELBO using the gradient descent method, optimizing the Copula parameters and marginal distribution parameters, the posterior distribution q'(Z) of the optimal unit reliability R is obtained.
[0020] As an embodiment of the present invention, in the above method, based on the posterior distribution, using I-Spline to construct a wind farm output model whose power output does not decrease monotonically when the wind speed increases includes:
[0021] Using the posterior distribution q(Z), a latent variable sample set {Z i};
[0022] Combining the wind speed W, wind direction D, and unit output power P in the historical data, the I-Spline technology is used to fit the wind farm output model f(W, D, Z) to describe the relationship between wind speed W, wind direction D, potential variable Z and output power P:
[0023] P = f(W, D, Z);
[0024] In the process of fitting the wind farm output model f(W, D, Z) using I-Spline technology, the fitting error is minimized by the least squares method to optimize the coefficient β of I-Spline. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows:
[0025]
[0026] In the above formula, S j represents the node-based basis function;
[0027] The following constraints are added during the optimization process to ensure that the derivative is non-negative:
[0028]
[0029] As an embodiment of the present invention, the reliability index in the above method includes power shortage probability, power shortage expected value and unpowered energy.
[0030] As an embodiment of the present invention, the above method uses the binary search method to search for credible capacity to ensure that the reliability index of the wind farm after access is equal to that before access, including:
[0031] Calculate the power shortage when the wind farm is connected to the grid, set the gradient value of load increase on the basis of ensuring the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value;
[0032] Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold value. If the EENS value is lower than the first preset threshold value, continue to increase the load; if the EENS value is higher than the first preset threshold value, stop increasing the load;
[0033] Based on the above process, two boundary load values of the trusted capacity search range are determined;
[0034] Taking the average of the two boundary load values and calculating the EENS at this time, if the obtained EENS value is less than the second preset threshold, updating the average as the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, updating the average as the right endpoint of the search range;
[0035] When the difference between the left and right endpoints is less than a set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
[0036] According to a second aspect of the present invention, there is provided a wind farm credible capacity assessment device, the device comprising: a posterior distribution acquisition unit, for obtaining a posterior distribution of unit reliability in the wind farm using a variational inference model of Copula; an output model construction unit, for constructing a wind farm output model whose power output monotonically does not decrease when the wind speed increases using I-Spline based on the posterior distribution; a probability distribution generation unit, for generating a joint probability distribution of wind farm output power using the wind farm output model; a reliability calculation unit, for randomly sampling the joint probability distribution using a sequential Monte Carlo method to calculate a reliability index of a power transmission system; and a credible capacity determination unit, for searching for credible capacity using a dichotomy method to ensure that the reliability index after the wind farm is connected is equal to that before the wind farm is connected.
[0037] As an embodiment of the present invention, the posterior distribution acquisition unit is specifically used for:
[0038] Preprocessing historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, wherein the historical data includes wind direction D, wind speed W, and unit output power P;
[0039] The preprocessed historical data is used as the observed data X(t), and the unit reliability R(t) is used as the latent variable Z(t);
[0040] Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t):
[0041] q(Z)=C(F(R));
[0042] In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable;
[0043] Combined with the observed data X(t), construct the joint distribution p(X,Z):
[0044] p(X,Z)=p(X|Z)p(Z);
[0045] In the above formula, p(X|Z) is the distribution of observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t);
[0046] The objective function ELBO is constructed as follows:
[0047] ELBO=E q(Z)[logp(D,W,P,Z)-logq(Z)]
[0048] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable.
[0049] By maximizing ELBO using the gradient descent method, optimizing the Copula parameters and marginal distribution parameters, the posterior distribution q'(Z) of the optimal unit reliability R is obtained.
[0050] As an embodiment of the present invention, the output model building unit is specifically used for:
[0051] Using the posterior distribution q(Z), a latent variable sample set {Z i};
[0052] Combining the wind speed W, wind direction D, and unit output power P in the historical data, the I-Spline technology is used to fit the wind farm output model f(W, D, Z) to describe the relationship between wind speed W, wind direction D, potential variable Z and output power P:
[0053] P = f(W, D, Z);
[0054] In the process of fitting the wind farm output model f(W, D, Z) using I-Spline technology, the fitting error is minimized by the least squares method to optimize the coefficient β of I-Spline. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows:
[0055]
[0056] In the above formula, S j represents the node-based basis function;
[0057] The following constraints are added during the optimization process to ensure that the derivative is non-negative:
[0058]
[0059] As an embodiment of the present invention, the above reliability index includes power shortage probability, power shortage expected value and unpowered energy.
[0060] As an embodiment of the present invention, the above-mentioned trusted capacity determination unit is specifically used for:
[0061] Calculate the power shortage when the wind farm is connected to the grid, set the gradient value of load increase on the basis of ensuring the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value;
[0062] Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold value. If the EENS value is lower than the first preset threshold value, continue to increase the load; if the EENS value is higher than the first preset threshold value, stop increasing the load;
[0063] Based on the above process, two boundary load values of the trusted capacity search range are determined;
[0064] Taking the average of the two boundary load values and calculating the EENS at this time, if the obtained EENS value is less than the second preset threshold, updating the average as the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, updating the average as the right endpoint of the search range;
[0065] When the difference between the left and right endpoints is less than a set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
[0066] According to a third aspect of the present invention, there is provided an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.
[0067] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.
[0068] The wind farm reliable capacity assessment method and device proposed in the present invention can accurately capture the complex nonlinear relationship between wind speed, wind direction, unit reliability and power output by introducing Copula variational inference technology. Compared with traditional linear models, Copula variational inference can more effectively handle data uncertainty and dependencies, and improve the accuracy of output prediction. In addition, the I-Spline technology ensures the smoothness and monotonicity of the power curve, and can provide reliable prediction results even when the observation data is incomplete. By constructing the posterior distribution of wind turbine reliability, the performance differences and uncertainties between different wind turbines are fully considered. This method not only improves the accuracy of single unit output prediction, but also improves the reliability of overall wind farm output assessment. Finally, by selecting appropriate reliability indicators and combining the posterior distribution, while considering wind resources and unit reliability, the power supply capacity of the wind farm is accurately assessed at different confidence levels, which helps to improve the accuracy of power supply reliability assessment. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. 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. In the drawings:
[0070] Figure 1 It is a flow chart of a wind farm reliable capacity assessment method provided in an embodiment of the present application;
[0071] Figure 2 It is a flow chart of obtaining the posterior distribution of unit reliability by using the variational inference model of Copula provided in an embodiment of the present invention;
[0072] Figure 3 It is a flow chart of a wind farm output model constructed by using I-Spline provided in an embodiment of the present invention;
[0073] Figure 4 It is a schematic diagram of a process of searching for a trusted capacity using a binary search method provided in an embodiment of the present application;
[0074] Figure 5 It is a structural schematic diagram of a wind farm reliable capacity assessment device provided in an embodiment of the present application;
[0075] Figure 6 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0076] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.
[0077] like Figure 1 The figure is a flow chart of a wind farm reliable capacity assessment method provided by an embodiment of the present application, the method comprising the following steps:
[0078] Step S101: using the Copula variational inference model to obtain the posterior distribution of the reliability of the wind farm units.
[0079] The Copula function is a statistical tool used to describe multivariate dependencies, which can accurately capture the complex relationship between wind speed, wind direction, and wind turbine reliability. This application uses the variational inference technology based on Copula to obtain the posterior distribution of wind turbine reliability without seriously underestimating the posterior variance, and accurately describe its uncertainty.
[0080] Preferably, Figure 2 As shown, step S101 may further include the following sub-steps:
[0081] Step S1011: pre-processing the historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, and the historical data includes wind direction D, wind speed W, and unit output power P.
[0082] First, historical data of the wind farm is collected, including wind direction D, wind speed W, and unit output power P, and then the collected historical data is preprocessed. The preprocessing here may include normalization processing and data cleaning and filling processing.
[0083] 1) Normalization
[0084] The min-max normalization method can be used for dimensionless processing to avoid the influence of different dimensions on model learning. The normalization formula is as follows:
[0085]
[0086] Among them, x nor is the normalized data.
[0087] Considering the physical meaning of wind power, the prediction results also need to be denormalized later, and the formula is expressed as:
[0088] x dnor =(x max -x min )x+x min ;
[0089] Among them, x dnor are the data after denormalization.
[0090] 2) Data cleaning and filling processing
[0091] There are data anomalies and missing data in the process of data storage and transmission. These abnormal data will have a certain impact on the prediction accuracy of the model. The quartile method is used to clean the abnormal data, and its formula is expressed as:
[0092] [F l ,F u ] = [Q1-1.5IQR, Q3-1.5IQR];
[0093] IQR = Q3-Q1;
[0094] Among them, F l , F u is the upper limit of the outlier, IQR is the interquartile range, Q1 and Q3 are the first and third quartiles of the data.
[0095] After cleaning the abnormal data, the missing data and the cleaned data are filled based on the interpolation method.
[0096] Step S1012: Use the preprocessed historical data as the observed data X(t), and use the unit reliability R(t) as the latent variable Z(t).
[0097] In this embodiment, it is assumed that the initial edge of the unit reliability follows the Weibull distribution, and the unit reliability refers to the reliability index of the wind turbine unit, such as failure rate, availability, etc. As a complex mechanical system, the reliability of the unit greatly affects the output power of the wind farm.
[0098] Weibull distribution is a commonly used probability distribution used to model the reliability, failure time, etc. of mechanical devices and other engineering systems. The assumed initial marginal distribution obeys the Weibull distribution, which means that when the model is initialized, this application assumes that the random variable R of the unit reliability follows the Weibull distribution. This assumption is based on practical experience and statistical analysis, and can better describe the reliability characteristics of the unit. Specifically, in the variational inference model, it is first assumed that the prior distribution of the unit reliability R is a Weibull distribution, and then by observing the data of wind direction (D), wind speed (W) and output power (P), inference and optimization are performed to finally obtain the posterior distribution of the unit reliability R. This process helps to capture the complex dependencies between latent variables and observed data, thereby more accurately describing the operating characteristics and reliability of wind farms.
[0099] Step S1013: Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t):
[0100] q(Z)=C(F(R));
[0101] In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable.
[0102] Step S1014: Combine the observed data X(t) to construct a joint distribution p(X, Z):
[0103] p(X,Z)=p(X|Z)p(Z);
[0104] In the above formula, p(X|Z) is the distribution of the observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t).
[0105] Step S1015: construct the objective function ELBO as follows:
[0106] ELBO=E q(Z)[logp(D,W,P,Z)-logq(Z)]
[0107] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable.
[0108] Step S1016: By maximizing ELBO using the gradient descent method, the Copula parameters and the marginal distribution parameters are optimized to obtain the posterior distribution q'(Z) of the optimal unit reliability R.
[0109] Step S102: Based on the posterior distribution, using I-Spline, construct a wind farm output model whose power output monotonically does not decrease when the wind speed increases.
[0110] Preferably, Figure 3 As shown, this step may further include the following sub-steps:
[0111] Step S1021: Generate a latent variable sample set {Z i}.
[0112] Step S1022: Combine the wind speed W, wind direction D, and unit output power P in the historical data, and use the I-Spline technology to fit the wind farm output model f(W, D, Z) to describe the relationship between the wind speed W, wind direction D, potential variable Z, and output power P:
[0113] P=f(W,D,Z).
[0114] Step S1023: In the process of fitting the wind farm output model f(W, D, Z) using the I-Spline technology, the coefficient β of the I-Spline is optimized by minimizing the fitting error through the least squares method. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows:
[0115]
[0116] In the above formula, S j Represents node-based basis functions.
[0117] Step S1024: Add the following constraints during the optimization process to ensure that the derivative is non-negative:
[0118]
[0119] Through the above steps, the I-Spline technology is used to build a wind farm output model in which the power output does not decrease monotonically when the wind speed increases. This can effectively capture the complex relationship between wind speed, wind direction, unit reliability and power output, and improve the accuracy and reliability of wind farm output prediction.
[0120] Step S103: Generate a joint probability distribution of wind farm output power using the wind farm output model.
[0121] The output power samples {P i} as random variables and construct their joint probability distribution.
[0122] The joint probability distribution can be estimated using nonparametric methods such as kernel density estimation (KDE), or discretized using methods such as histograms.
[0123] The joint probability distribution can be expressed as P(P1, P2, ..., P n ), where P i is the output power sample. This distribution describes the probability distribution of wind farm output power under different wind speed, wind direction and unit reliability conditions. Through this step, the wind farm output model is used to generate the joint probability distribution of wind farm output power, which can effectively describe the output power characteristics of the wind farm under different conditions.
[0124] Step S104: using the sequential Monte Carlo method to randomly sample the joint probability distribution to calculate the reliability index of the power transmission system.
[0125] First, it is necessary to select reliability indicators. In this embodiment, the selected reliability indicators include the power shortage probability (LOLP), the power shortage expectation (LOLE) and the unsupplied energy (EENS). Their formulas are as follows:
[0126]
[0127] in, P w,t is the wind turbine output power calculated based on the real-time wind speed of the wind farm, ρ is the air density, R is the radius of the wind turbine, and f N is the rated frequency of the power grid, β and w are the pitch angle and rotor speed of the corresponding wind turbine, G is the gearbox ratio, p is the number of pole pairs of the wind turbine, and C p is the characteristic function of the wind turbine.
[0128] Sequential Monte Carlo (SMC) is a numerical method for estimating complex probability distributions. It gradually approaches the target distribution by randomly sampling the probability distribution. The SMC method usually includes the following steps: initialization, importance sampling, resampling, and weight update. Using the resampled sample set, the reliability indicators of the transmission system are calculated using the above formula, namely the power shortage probability (LOLP), power shortage expectation (LOLE), and unpowered energy (EENS).
[0129] Step S105: Use the binary search method to search for credible capacity to ensure that the reliability index of the wind farm after access is equal to that before access.
[0130] Preferably, Figure 4 As shown, this step S105 may further include the following sub-steps:
[0131] Step S1051: Calculate the power shortage when the wind farm is connected to the grid, set a gradient value of load increase on the basis of ensuring that the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value.
[0132] Step S1052: Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold. If the EENS value is lower than the first preset threshold, continue to increase the load; if the EENS value is higher than the first preset threshold, stop increasing the load.
[0133] Step S1053: Determine two boundary load values of the trusted capacity search range based on the above process.
[0134] Step S1054: Take the average of the two boundary load values and calculate the EENS at this time. If the obtained EENS value is less than the second preset threshold, the average is updated to the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, the average is updated to the right endpoint of the search range.
[0135] Step S1055: When the difference between the left and right endpoints is less than the set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
[0136] The wind farm credible capacity assessment method proposed in the present invention can accurately capture the complex nonlinear relationship between wind speed, wind direction, unit reliability and power output by introducing Copula variational inference technology. Compared with traditional linear models, Copula variational inference can more effectively handle data uncertainty and dependencies, thereby improving the accuracy of output prediction. In addition, the I-Spline technology ensures the smoothness and monotonicity of the power curve, and can provide reliable prediction results even when the observation data is incomplete. By constructing the posterior distribution of wind turbine reliability, the performance differences and uncertainties between different wind turbines are fully considered. This method not only improves the accuracy of output prediction of a single unit, but also improves the reliability of the overall wind farm output assessment. Finally, by selecting appropriate reliability indicators and combining the posterior distribution, while considering wind resources and unit reliability, the power supply capacity of the wind farm is accurately assessed at different confidence levels, which helps to improve the accuracy of power supply reliability assessment.
[0137] like Figure 5 The figure shows a schematic diagram of the structure of a wind farm credible capacity assessment device provided by an embodiment of the present application, the device comprises: a posterior distribution acquisition unit 510, an output model construction unit 520, a probability distribution generation unit 530, a reliability calculation unit 540 and a credible capacity determination unit 550, which are sequentially connected. Among them:
[0138] A posterior distribution acquisition unit 510, used to obtain a posterior distribution of the reliability of the wind farm units using the variational inference model of the Copula;
[0139] An output model construction unit 520 is used to construct a wind farm output model whose power output does not decrease monotonically when the wind speed increases based on the posterior distribution and using I-Spline;
[0140] A probability distribution generating unit 530, configured to generate a joint probability distribution of wind farm output power using the wind farm output model;
[0141] A reliability calculation unit 540, configured to randomly sample the joint probability distribution using a sequential Monte Carlo method to calculate a reliability index of the power transmission system;
[0142] The credible capacity determination unit 550 is used to perform credible capacity search using a binary search method to ensure that the reliability index of the wind farm after access is equal to that before access.
[0143] Preferably, the posterior distribution acquisition unit 510 is specifically used for:
[0144] Preprocessing historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, wherein the historical data includes wind direction D, wind speed W, and unit output power P;
[0145] The preprocessed historical data is used as the observed data X(t), and the unit reliability R(t) is used as the latent variable Z(t);
[0146] Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t):
[0147] q(Z)=C(F(R));
[0148] In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable;
[0149] Combined with the observed data X(t), construct the joint distribution p(X,Z):
[0150] p(X,Z)=p(X|Z)p(Z);
[0151] In the above formula, p(X|Z) is the distribution of observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t);
[0152] The objective function ELBO is constructed as follows:
[0153] ELBO=E q(Z) [logp(D,W,P,Z)-logq(Z)]
[0154] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable.
[0155] By maximizing ELBO using the gradient descent method, optimizing the Copula parameters and marginal distribution parameters, the posterior distribution q'(Z) of the optimal unit reliability R is obtained.
[0156] Preferably, the output model building unit 520 is specifically used for:
[0157] Using the posterior distribution q'(Z), a latent variable sample set {Z i};
[0158] Combining the wind speed W, wind direction D, and unit output power P in the historical data, the I-Spline technology is used to fit the wind farm output model f(W, D, Z) to describe the relationship between wind speed W, wind direction D, potential variable Z and output power P:
[0159] P = f(W, D, Z);
[0160] In the process of fitting the wind farm output model f(W, D, Z) using I-Spline technology, the fitting error is minimized by the least squares method to optimize the coefficient β of I-Spline. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows:
[0161]
[0162] In the above formula, S j Represents node-based basis functions.
[0163] The following constraints are added during the optimization process to ensure that the derivative is non-negative:
[0164]
[0165] Preferably, the above reliability indicators include power shortage probability, power shortage expectation value and unpowered energy.
[0166] Preferably, the above-mentioned trusted capacity determination unit 550 is specifically used for:
[0167] Calculate the power shortage when the wind farm is connected to the grid, set the gradient value of load increase on the basis of ensuring the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value;
[0168] Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold value. If the EENS value is lower than the first preset threshold value, continue to increase the load; if the EENS value is higher than the first preset threshold value, stop increasing the load;
[0169] Based on the above process, two boundary load values of the trusted capacity search range are determined;
[0170] Taking the average of the two boundary load values and calculating the EENS at this time, if the obtained EENS value is less than the second preset threshold, updating the average as the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, updating the average as the right endpoint of the search range;
[0171] When the difference between the left and right endpoints is less than a set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
[0172] The wind farm reliable capacity assessment device proposed in the present invention can accurately capture the complex nonlinear relationship between wind speed, wind direction, unit reliability and power output by introducing Copula variational inference technology. Compared with the traditional linear model, Copula variational inference can more effectively handle the uncertainty and dependency of data, and improve the accuracy of output prediction. In addition, the I-Spline technology ensures the smoothness and monotonicity of the power curve, and can provide reliable prediction results even when the observation data is incomplete. By constructing the posterior distribution of wind turbine reliability, the performance differences and uncertainties between different wind turbines are fully considered. This method not only improves the accuracy of single unit output prediction, but also improves the reliability of the overall wind farm output assessment. Finally, by selecting appropriate reliability indicators and combining the posterior distribution, the wind resources and unit reliability are considered at the same time, and the power supply capacity of the wind farm is accurately evaluated at different confidence levels, which helps to improve the accuracy of power supply reliability assessment.
[0173] Figure 6 is a schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 6The electronic device shown is a general data processing device, which includes a general computer hardware structure, which includes at least a processor 801 and a memory 802. The processor 801 and the memory 802 are connected via a bus 803. The memory 802 is suitable for storing one or more instructions or programs executable by the processor 801. The one or more instructions or programs are executed by the processor 801 to implement the steps in the above-mentioned wind farm reliable capacity assessment method.
[0174] The processor 801 may be an independent microprocessor or a collection of one or more microprocessors. Thus, the processor 801 executes the commands stored in the memory 802 to execute the method flow of the embodiment of the present invention as described above to realize the processing of data and the control of other devices. The bus 803 connects the above-mentioned multiple components together, and at the same time connects the above-mentioned components to the display controller 804 and the display device and the input / output (I / O) device 805. The input / output (I / O) device 805 may be a mouse, a keyboard, a modem, a network interface, a touch input device, a somatosensory input device, a printer, and other devices known in the art. Typically, the input / output (I / O) device 805 is connected to the system via an input / output (I / O) controller 806.
[0175] The memory 802 may store software components such as an operating system, a communication module, an interaction module, and an application program. Each of the modules and applications described above corresponds to a set of executable program instructions that implement one or more functions and methods described in the embodiments of the invention.
[0176] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned wind farm reliable capacity assessment method are implemented.
[0177] From the above, it can be seen that the wind farm reliable capacity assessment method and device proposed in the present invention can accurately capture the complex nonlinear relationship between wind speed, wind direction, unit reliability and power output by introducing Copula variational inference technology. Compared with the traditional linear model, Copula variational inference can more effectively handle the uncertainty and dependency of data, and improve the accuracy of output prediction. In addition, the I-Spline technology ensures the smoothness and monotonicity of the power curve, and can provide reliable prediction results even when the observation data is incomplete. By constructing the posterior distribution of wind turbine reliability, the performance differences and uncertainties between different wind turbines are fully considered. This method not only improves the accuracy of single unit output prediction, but also improves the reliability of the overall wind farm output assessment. Finally, by selecting appropriate reliability indicators and combining the posterior distribution, while considering wind resources and unit reliability, the power supply capacity of the wind farm is accurately assessed at different confidence levels, which helps to improve the accuracy of power supply reliability assessment.
[0178] The preferred embodiments of the present invention are described above with reference to the accompanying drawings. Many features and advantages of these embodiments are clear from this detailed description, and therefore the claims are intended to cover all such features and advantages of these embodiments that fall within their true spirit and scope. In addition, since many modifications and changes are easily conceivable to those skilled in the art, it is not intended to limit the embodiments of the present invention to the precise structure and operation illustrated and described, but all suitable modifications and equivalents falling within its scope may be covered.
[0179] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0180] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.
[0181] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0183] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for evaluating the credible capacity of a wind farm, characterized in that: The method comprises: The Copula variational inference model is used to obtain the posterior distribution of unit reliability in wind farms. Based on the posterior distribution, an I-Spline is used to construct a wind farm output model in which the power output does not decrease monotonically when the wind speed increases; Generating a joint probability distribution of wind farm output power using the wind farm output model; The joint probability distribution is randomly sampled using a sequential Monte Carlo method to calculate a reliability index of the power transmission system; The binary search method is used to search for credible capacity to ensure that the reliability index of the wind farm after access is equal to that before access.
2. The wind farm reliable capacity assessment method according to claim 1, characterized in that: The method of using the variational inference model of the Copula to obtain the posterior distribution of the reliability of the wind farm units includes: Preprocessing historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, wherein the historical data includes wind direction D, wind speed W, and unit output power P; The preprocessed historical data is used as the observed data X(t), and the unit reliability R(t) is used as the latent variable Z(t); Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t): q(Z)=C(F(R)); In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable; Combined with the observed data X(t), construct the joint distribution p(X,Z): p(X,Z)=p(X|Z)p(Z); In the above formula, p(X|Z) is the distribution of observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t); The objective function ELBO is constructed as follows: ELBO=E q(Z) [logp(D,W,P,Z)-logq(Z)] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable; By maximizing ELBO using the gradient descent method, optimizing the Copula parameters and marginal distribution parameters, the posterior distribution q'(Z) of the optimal unit reliability R is obtained.
3. The wind farm reliable capacity assessment method according to claim 2, characterized in that: The wind farm output model constructed based on the posterior distribution and using I-Spline to achieve a monotonically non-decreasing power output when the wind speed increases includes: Using the posterior distribution q'(Z), a latent variable sample set {Z i }; Combining the wind speed W, wind direction D, and unit output power P in the historical data, the I-Spline technology is used to fit the wind farm output model f(W, D, Z) to describe the relationship between wind speed W, wind direction D, potential variable Z and output power P: P = f(W, D, Z); In the process of fitting the wind farm output model f(W, D, Z) using I-Spline technology, the fitting error is minimized by the least squares method to optimize the coefficient β of I-Spline. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows: In the above formula, S j represents the node-based basis function; The following constraints are added during the optimization process to ensure that the derivative is non-negative:
4. The wind farm reliable capacity assessment method according to claim 1, characterized in that: The reliability indicators include power shortage probability, power shortage expectation value and unpowered energy.
5. The wind farm reliable capacity assessment method according to claim 1, characterized in that: The method of searching for credible capacity by using the dichotomy method to ensure that the reliability index of the wind farm after access is equal to that before access includes: Calculate the power shortage when the wind farm is connected to the grid, set the gradient value of load increase on the basis of ensuring the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value; Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold value. If the EENS value is lower than the first preset threshold value, continue to increase the load; if the EENS value is higher than the first preset threshold value, stop increasing the load; Based on the above process, two boundary load values of the trusted capacity search range are determined; Taking the average of the two boundary load values and calculating the EENS at this time, if the obtained EENS value is less than the second preset threshold, updating the average as the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, updating the average as the right endpoint of the search range; When the difference between the left and right endpoints is less than a set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
6. A wind farm reliable capacity assessment device, characterized in that: The device comprises: A posterior distribution acquisition unit is used to obtain the posterior distribution of the reliability of the wind farm units using the variational inference model of Copula; An output model construction unit, configured to construct, based on the posterior distribution, a wind farm output model using I-Spline, in which the power output does not decrease monotonically when the wind speed increases; A probability distribution generating unit, used to generate a joint probability distribution of wind farm output power using the wind farm output model; A reliability calculation unit, used for randomly sampling the joint probability distribution using a sequential Monte Carlo method to calculate a reliability index of the power transmission system; The credible capacity determination unit is used to search for credible capacity using a dichotomy method to ensure that the reliability index of the wind farm after access is equal to that before access.
7. The wind farm reliable capacity assessment device according to claim 6, characterized in that: The posterior distribution acquisition unit is specifically used for: Preprocessing historical data of the wind farm, wherein the historical data includes wind direction, wind speed and unit output power, wherein the historical data includes wind direction D, wind speed W, and unit output power P; The preprocessed historical data is used as the observed data X(t), and the unit reliability R(t) is used as the latent variable Z(t); Select Gaussian Copula function C Gaussian Define the joint distribution q(Z) of the latent variables Z(t): q(Z)=C(F(R)); In the above formula, F(R) is the cumulative distribution function of Weibull distribution, R represents unit reliability, and Z represents the latent variable; Combined with the observed data X(t), construct the joint distribution p(X,Z): p(X,Z)=p(X|Z)p(Z); In the above formula, p(X|Z) is the distribution of observed data X(t) under the given latent variable Z(t), and p(Z) is the prior distribution of the latent variable Z(t); The objective function ELBO is constructed as follows: ELBO=E q(Z) [logp(D,W,P,Z)-logq(Z)] E represents expected value, D represents wind direction data, W represents wind speed data, P represents output power data, and Z represents latent variable; By maximizing ELBO using the gradient descent method, optimizing the Copula parameters and marginal distribution parameters, the posterior distribution q'(Z) of the optimal unit reliability R is obtained.
8. The wind farm reliable capacity assessment device according to claim 7, characterized in that: The output model building unit is specifically used for: Using the posterior distribution q(Z), a latent variable sample set {Z i }; Combining the wind speed W, wind direction D, and unit output power P in the historical data, the I-Spline technology is used to fit the wind farm output model f(W, D, Z) to describe the relationship between wind speed W, wind direction D, potential variable Z and output power P: P = f(W, D, Z); In the process of fitting the wind farm output model f(W, D, Z) using I-Spline technology, the fitting error is minimized by the least squares method to optimize the coefficient β of I-Spline. j , to ensure that the power output of the fitted power model does not decrease monotonically when the wind speed increases. The optimization objective function is as follows: In the above formula, S j represents the node-based basis function; The following constraints are added during the optimization process to ensure that the derivative is non-negative:
9. The wind farm reliable capacity assessment device according to claim 6, characterized in that: The reliability indicators include power shortage probability, power shortage expectation value and unpowered energy.
10. The wind farm reliable capacity assessment device according to claim 6, characterized in that: The trusted capacity determination unit is specifically used for: Calculate the power shortage when the wind farm is connected to the grid, set the gradient value of load increase on the basis of ensuring the wind farm is connected to the grid, and gradually increase the grid load based on the gradient value; Calculate the EENS after each grid load increase, and compare the calculation result with the preset threshold value. If the EENS value is lower than the first preset threshold value, continue to increase the load; if the EENS value is higher than the first preset threshold value, stop increasing the load; Based on the above process, two boundary load values of the trusted capacity search range are determined; Taking the average of the two boundary load values and calculating the EENS at this time, if the obtained EENS value is less than the second preset threshold, updating the average as the left endpoint of the search range; if the EENS value is greater than or equal to the second preset threshold, updating the average as the right endpoint of the search range; When the difference between the left and right endpoints is less than a set third preset threshold, the search process ends, and the load increase determined at this time is used as the credible capacity of the wind farm.
11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.
12. 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 according to any one of claims 1 to 5 are implemented.
