Water-wind-light complementary system correlation scene generation method based on solar term division

Abstract

The invention provides a method for generating a correlation scene of a water-wind-solar complementary system based on solar term division. The method comprises the following steps: firstly, dividing historical data of tap power station runoff, wind power output and photovoltaic output by solar terms; then, generating a scene set considering the correlation of tap power station runoff, wind power output and photovoltaic output by using a C-vine copula method and a Monte Carlo sampling method; and finally, verifying the quality of the generated scene according to the fluctuation characteristics and the correlation of the generated scene. For the characteristic that the correlation among tap power station runoff, wind power output and photovoltaic output is relatively complex, the method can better capture the correlation among the tap power station runoff, the wind power output and the photovoltaic output and generate a scene set considering the correlation.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power systems, and particularly designs a method for generating correlation scenarios of a water-wind-solar complementary system based on solar term division. Background Art

[0002] The Lancang River Basin in southwest my country has abundant hydropower resources, and wind and solar energy resources are also relatively sufficient. Due to the complex terrain and changeable climate of the basin, hydropower output is significantly affected by seasonal precipitation, while wind and solar energy output is restricted by factors such as season and weather. In order to better dispatch and utilize these renewable energy sources, it is of great significance to study the correlation between them.

[0003] The patent document with application publication number CN110311420A discloses a method for generating a wind-solar combined output timing scenario. This method analyzes the correlation of wind-solar output in each period through the Copula function, and generates an output scenario that takes the correlation into account through the improved Latin hypercube sampling method. However, this method does not consider the impact of water inflow from the hydropower station. For a water-wind-solar complementary system, in order to generate a more realistic scheduling input scenario, it is necessary to consider the correlation between water inflow, wind power output, and photovoltaic output at the same time.

[0004] The patent document with application publication number CN117609586A discloses a method for generating long-term water, wind and solar scenes that considers temporal and spatial correlation. The method generates scenes that consider the temporal correlation and spatial correlation of water, wind and solar through Markov chain and C-vine copula respectively. However, the method generates scenes on a long-term time scale. For short-term scheduling of water, wind and solar systems, it is necessary to generate short-term water, wind and solar scenes that consider correlation. Summary of the invention

[0005] The technical problem that the present invention aims to solve is to provide a scenario generation method that takes into account the correlation between hydropower, wind power and photovoltaic power generation, so as to make the input scenario of the water-wind-solar complementary system scheduling more accurate and ensure that the scheduling result is closer to reality.

[0006] In order to realize the above-mentioned technical features, the purpose of the present invention is achieved as follows: a method for generating correlation scenarios of a water-wind-solar complementary system based on solar term division, firstly dividing the historical data of the runoff, wind power output and photovoltaic output of the Longtou Power Station according to the solar terms; then, using the C-vine copula method and the Monte Carlo sampling method, a scenario set is generated that considers the correlation of the runoff, wind power output and photovoltaic output of the Longtou Power Station; finally, the quality of the generated scenarios is verified by the fluctuation characteristics and correlation of the generated scenarios.

[0007] A method for generating correlation scenes of a water-wind-solar complementary system based on solar term division specifically comprises the following steps: Step 1: Obtain data on runoff, wind power output, and photovoltaic output of the leading power station: Collect the daily runoff of the Longtou Power Station in the lower reaches of the Lancang River in a certain year, as well as the wind speed of wind power stations within a certain range along the coast, and the temperature and radiation data of photovoltaic power stations, and calculate the wind power output and photovoltaic output respectively according to the wind power output formula and photovoltaic output formula; Step 2: Analyze the correlation between runoff, wind power output and photovoltaic output of the leading power station: The runoff, wind power output and photovoltaic output data of Longtou Power Station in the lower reaches of Lancang River in a certain year were divided according to the traditional 24 solar terms in China, and the runoff, wind power output and photovoltaic output data of Longtou Power Station in each solar term were analyzed using C-vine copula method to analyze the correlation between the runoff, wind power output and photovoltaic output of Longtou Power Station; Step 3: Generate scenarios of runoff, wind power output, and photovoltaic output of the leading power station taking into account the correlation: Taking the runoff, wind power output and photovoltaic output of the Longtou Power Station in each solar term as the objects, using the Monte Carlo sampling method and the conditional distribution function of the runoff, wind power output and photovoltaic output of the Longtou Power Station, a large number of scenarios that consider the correlation between the runoff, wind power output and photovoltaic output of the Longtou Power Station are generated; Step 4: Check the quality of the generated scene: Compare the variance between the generated scenario and the historical data to verify whether the generated scenario generally retains the fluctuation characteristics of the historical data; compare the relative error of the Kendall coefficient between the generated scenario and the historical data to verify whether the generated scenario generally retains the correlation between the historical data.

[0008] Preferably, in step 1, the wind power output and photovoltaic output are calculated respectively using the wind power output formula, the photovoltaic output formula and the collected meteorological data, which specifically includes the following steps: Step 1-1: Collect the daily runoff, wind speed, temperature and radiation data of Longtou Power Station in the middle and lower reaches of Lancang River in a certain year, with an interval of 1 hour; Step 1-2: The wind turbine wind speed of the wind power station is calculated based on the measured wind speed, and the expression is as follows: (1); In the formula, It is t The wind speed at the wind turbine at the moment, in m / s; The measured height is The t The measured wind speed at a certain moment, in m / s; is the height of the wind turbine hub in meters; q is the stability coefficient; Step 1-3: Calculate wind power output according to the wind power output formula. The expression is as follows: (2); In the formula, It is t Wind power output at a given moment, in MW; is the wind turbine cut-in wind speed, in m / s; is the cut-out wind speed of the wind turbine, in m / s; is the full wind speed of the wind turbine, in m / s; is the installed capacity of wind power, in MW; It is t The wind speed at the wind turbine at the moment, in m / s; Step 1-4: Calculate the temperature of the solar panel based on the measured temperature data from the weather station. The expression is as follows: (3); In the formula, It is t The temperature of the solar panel at the moment, in °C; It is t The measured temperature at the weather station at the moment, in °C; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions; is the temperature of the solar panel during normal operation; is the air temperature under standard test conditions; Step 1-5: Calculate the photovoltaic output according to the photovoltaic output formula. The expression is as follows: (4); In the formula, It is t The photovoltaic output at a certain moment, in MW; is the installed photovoltaic capacity, in MW; is the air temperature-power conversion coefficient; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions; It is t The temperature of the solar panel at the moment, in °C; is the air temperature under standard test conditions.

[0009] Preferably, in step 2, the C-vine copula method is used to analyze the correlation between the runoff, wind power output, and photovoltaic output of the Longtou Power Station, which specifically includes the following steps: Step 2-1: Use the traditional 24 solar terms in China to divide the runoff, wind power output and photovoltaic output data of Longtou Power Station; the solar terms are from "Lichun" to "Dahan", and each solar term contains 45 days of data. The runoff, wind power output and photovoltaic output data of Longtou Power Station in the same hour are used as the analysis objects, with a total of 24 time periods; Step 2-2: Use the Kendall correlation coefficient method to select the main variable of the C-vine copula. The sum of the absolute values ​​of the Kendall correlation coefficients between the main variable and the other two variables, namely, the runoff of the leading power station, wind power output, and photovoltaic output, must be the largest. The expression is as follows: (5); In the formula, is the Kendall correlation coefficient; C , D are the number of consistent and inconsistent element pairs in the two variables respectively; n is the number of data; Step 2-3: Use the kernel density estimation method to fit the distribution function of the runoff, wind power output, and photovoltaic output of the Longtou Power Station. The expression is as follows: (6); In the formula, x is a variable, which is the runoff of the leading power station, wind power output or photovoltaic output; is the probability density function corresponding to the variable, which is the probability density function of the runoff, wind power output or photovoltaic output of the leading power station; n is the number of data; h is the bandwidth, which controls the width of the kernel function; It is i The value of the data; is the kernel function; is the distribution function corresponding to the variable, which is the distribution function of the runoff, wind power output or photovoltaic output of the leading power station; Step 2-4: Use t copula function, Gaussian copula function, Clayton copula function, Gumbel copula function and Frank copula function to fit the joint distribution function of runoff, wind power output and photovoltaic output respectively. The expressions are as follows: (7); In the formula, It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint probability distribution function of It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint copula distribution function of ; is the marginal distribution function of the runoff from Longtou Power Station; is the marginal distribution function of wind power output; is the marginal distribution function of PV output; Step 2-5: Use Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) to select the optimal copula function. The smaller the value of AIC or BIC, the better the copula fitting effect. The expression is as follows: (8); Where L is the likelihood function; k is the number of parameters of the copula function; n is the number of data; Step 2-6: Assuming that wind power output is the main variable, according to the fitted copula distribution function and the structure of C-vine copula, the multi-dimensional copula probability density function of the runoff of the leading power station, wind power output and photovoltaic output is decomposed as follows: (9); In the formula, is the joint copula probability density function of runoff, wind power output and photovoltaic output of the leading power station; is the joint copula probability density function of runoff and wind power output of Longtou Power Station; is the joint copula probability density function of PV output and wind power output; is the joint copula probability density function of runoff and PV output of the leading power station under given wind power output conditions; According to the C-vine copula method, the conditional distribution functions of runoff, wind power output and photovoltaic output of Longtou Power Station are calculated, and the expressions are as follows: (10); In the formula, is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; , are the distribution functions of photovoltaic output and wind power output respectively; is the conditional distribution function of runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of runoff from Longtou Power Station; is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; It is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions.

[0010] Preferably, in step 3, the Monte Carlo sampling method is used to combine the conditional distribution functions of the runoff, wind power output, and photovoltaic output of the leading power station to generate a scene set of the runoff, wind power output, and photovoltaic output of the leading power station that takes into account the correlation, which specifically includes the following steps: Step 3-1: Generate three uniformly distributed random numbers , solve the samples of runoff, wind power output and photovoltaic output of Longtou Power Station , first let Equal to the sample value of wind power output , the expression is as follows: (11); In the formula, is the distribution function of wind power output; is the sample value of wind power output; is a generated random number; Step 3-2: Order Equal to the conditional distribution function value of photovoltaic output, and solve the sample value of photovoltaic output according to the following formula , the expression is as follows: (12); In the formula, It is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; is the distribution function of wind power output; is the sample value of PV output; is a generated random number; Step 3-3: Order Equal to the conditional distribution function value of the runoff of Longtou Power Station, and solve the sample value of the runoff of Longtou Power Station according to the following formula , the expression is as follows: (13); In the formula, is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; It is the conditional distribution function of photovoltaic output under given wind power output conditions; is a generated random number; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of wind power output; is the sample value of runoff; Step 3-4: For the obtained sample After inverse transformation, we get a set of output scenarios, which are expressed as follows: (14); In the formula, x This is the generated runoff scenario for Longtou Power Station; y The wind power output scenario is generated; z The photovoltaic output scenario is generated; is the inverse function of the runoff distribution function of Longtou Power Station; is the inverse function of wind power output distribution function; is the inverse function of the photovoltaic output distribution function; Step 3-5: Repeat steps 3-1 to 3-4 for a total of 24 times to obtain the runoff, wind power output, and photovoltaic output scenarios of the leading power station for 24 hours a day; Step 3-6: Repeat steps 3-5 for a total of 45 times to obtain the runoff, wind power output, and photovoltaic output scenario set of the leading power station in a solar term. The expression is as follows: (15); In the formula, set is the scene set; It is t Runoff scenarios of Longtou Power Station generated in different time periods; It is t Wind power output scenarios generated in each time period; It is t PV output scenarios generated in each time period; t is the number of moments, ranging from 1 to 24.

[0011] Preferably, in step 4, the quality of the scene is verified using the variance and Kendall coefficient relative error of the generated scene, which specifically includes the following steps: Step 4-1: Use the variance of the generated scenario to verify whether the generated scenario retains the volatility characteristics of historical data. The expression is as follows: (16); In the formula, is the variance of the generated scene; To generate the scene t The value at a moment; is the mean of the generated scenarios; is the variance of historical data; For historical data t The value at a moment; is the mean of historical data; T is the total number of time periods; Step 4-2: Use the Kendall coefficient relative error of the generated scenario to verify whether the generated scenario retains the correlation between historical data. The expression is as follows: (17); In the formula, It is m Group Kendall coefficient relative error; It is m Kendall coefficient between group-generated scenarios; It is m Kendall coefficient between groups of historical data; m is the scene set number.

[0012] Beneficial effects of the present invention: 1. The present invention uses the C-vine copula method to analyze the dependence of runoff, wind power output and photovoltaic output of the leading power station in each period, and uses the Monte Carlo sampling method to generate scenarios that conform to the actual correlation. Compared with the previous scenario generation methods that only consider the correlation between feng shui or wind and photovoltaic, the method in this paper is closer to the actual system operation characteristics, ensuring that the data has higher accuracy and stability.

[0013] 2. The present invention divides historical data by solar terms, so that the runoff, wind power output and photovoltaic output scenarios of Longtou Power Station in each solar term have strong seasonal characteristics. The temporal consistency and seasonal regularity of the data are significantly improved, and the instability caused by the fluctuation differences of the annual data in different seasons is avoided. In the same solar term, the runoff, wind power and photovoltaic output of Longtou Power Station change less, the data is more concentrated, and the accuracy of correlation analysis is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0015] Figure 1 It is the overall flow chart of the present invention.

[0016] Figure 2 is the sum of the Kendall coefficients of different main variables in the embodiment of the present invention.

[0017] Figure 3 is a structural diagram of a C-vine copula in an embodiment of the present invention.

[0018] Figure 4 This is a water-wind-light edge distribution map fitted by the kernel density estimation method in an embodiment of the present invention.

[0019] Figure 5It is a copula surface diagram of wind power output-photovoltaic output and wind power output-Longtou Power Station runoff during the 'Lichun' solar term in an embodiment of the present invention.

[0020] Figure 6 It is a conditional copula surface diagram of the photovoltaic output during the 'Lichun' solar term and the runoff of the Longtou Power Station in an embodiment of the present invention.

[0021] Figure 7 It is a water scenery scene map generated in an embodiment of the present invention.

[0022] Figure 8 It is a relative error distribution diagram of the Kendall coefficient of the generated scene in the embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention more clear, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Embodiment 1: like Figure 1-8 As shown, a method for generating correlation scenes of a water-wind-solar complementary system based on solar term division is characterized in that it specifically includes the following steps: Step 1: Obtain data on runoff, wind power output, and photovoltaic output of the leading power station: The daily runoff from Longtou Power Station in the lower reaches of the Lancang River, the wind speed of wind power stations within 60 km along the coast, and the temperature and radiation data of photovoltaic power stations were collected from 2011 to 2013, and the wind power output and photovoltaic output were calculated according to the wind power output formula and photovoltaic output formula respectively.

[0025] Step 2: Analyze the correlation between runoff, wind power output and photovoltaic output of the leading power station: The runoff, wind power output and photovoltaic output data of Longtou Power Station in the lower reaches of Lancang River from 2011 to 2013 were divided according to the traditional 24 solar terms in China. The runoff, wind power output and photovoltaic output data of Longtou Power Station in each solar term were analyzed using the C-vinecopula method.

[0026] Step 3: Generate scenarios of runoff, wind power output, and photovoltaic output of the leading power station taking into account the correlation: Taking the runoff, wind power output and photovoltaic output of the leading power station in each solar term as the objects, the Monte Carlo sampling method and the conditional distribution functions of the runoff, wind power output and photovoltaic output of the leading power station are used to generate a large number of scenarios that consider the correlation between the runoff, wind power output and photovoltaic output of the leading power station.

[0027] Step 4: Check the quality of the generated scene: Compare the variance between the generated scenario and the historical data to verify whether the generated scenario generally retains the fluctuation characteristics of the historical data; compare the relative error of the Kendall coefficient between the generated scenario and the historical data to verify whether the generated scenario generally retains the correlation between the historical data.

[0028] Furthermore, in step 1, the wind power output and photovoltaic output are calculated respectively using the wind power output formula, the photovoltaic output formula and the collected meteorological data, which specifically includes the following steps: Step 1-1: Collect daily runoff, wind speed, temperature and radiation data from Longtou Power Station in the middle and lower reaches of the Lancang River from 2011 to 2013, with an interval of 1 hour.

[0029] Step 1-2: The wind turbine wind speed of the wind power station is calculated based on the measured wind speed, and the expression is as follows: (1); In the formula, It is t The wind speed at the wind turbine at the moment, in m / s; The measured height is The t The measured wind speed at a certain moment, in m / s; is the height of the wind turbine hub, in meters; q is the stability coefficient, usually 0.18-0.25; Step 1-3: Calculate wind power output according to the wind power output formula. The expression is as follows: (2); In the formula, It is t Wind power output at a given moment, in MW; is the wind turbine cut-in wind speed, in m / s; is the cut-out wind speed of the wind turbine, in m / s; is the full wind speed of the wind turbine, in m / s; is the installed capacity of wind power, in MW; It is t The wind speed at the wind turbine at the moment, in m / s; Step 1-4: Calculate the temperature of the solar panel based on the measured temperature data from the weather station. The expression is as follows: (3); In the formula, It is t The temperature of the solar panel at the moment, in °C; It is t The measured temperature at the weather station at the moment, in °C; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions, usually 1000 W / m 2 ; The temperature of the solar panel during normal operation, usually 46℃-50℃; It is the air temperature under standard test conditions, usually 25 ℃.

[0030] Step 1-5: Calculate the photovoltaic output according to the photovoltaic output formula. The expression is as follows: (4); In the formula, It is t The photovoltaic output at a certain moment, in MW; is the installed photovoltaic capacity, in MW; is the air temperature-power conversion coefficient; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions; It is t The temperature of the solar panel at the moment, in °C; is the air temperature under standard test conditions.

[0031] Furthermore, in step 2, the C-vine copula method is used to analyze the correlation between the runoff, wind power output, and photovoltaic output of the Longtou Power Station, which specifically includes the following steps: Step 2-1: Use the traditional 24 Chinese solar terms to divide the Longtou Power Station runoff, wind power output and photovoltaic output data. The solar terms are from 'Lichun' to 'Dahan', and each solar term contains 45 days of data. The Longtou Power Station runoff, wind power output, and photovoltaic output data of the same hour are used as analysis objects, for a total of 24 time periods; Step 2-2: Use the Kendall correlation coefficient method to select the main variable of the C-vine copula. The sum of the absolute values ​​of the Kendall correlation coefficients of the main variable and the other two variables, including runoff, wind power output, and photovoltaic output, must be the largest. The expression is as follows: (5); In the formula, is the Kendall correlation coefficient; C , D are the number of consistent and inconsistent element pairs in the two variables respectively; n is the number of data; Step 2-3: Use the kernel density estimation method to fit the distribution function of the runoff, wind power output, and photovoltaic output of the Longtou Power Station. The expression is as follows: (6); In the formula, x is a variable, which can be the runoff of the leading power station, wind power output or photovoltaic output; is the probability density function corresponding to the variable, which can be the probability density function of the runoff, wind power output or photovoltaic output of the leading power station; n is the number of data; h is the bandwidth, which controls the width of the kernel function; It is i The value of the data; is the kernel function; It is the distribution function corresponding to the variable, which can be the distribution function of the runoff, wind power output or photovoltaic output of the leading power station.

[0032] Step 2-4: Use t copula function, Gaussian copula function, Clayton copula function, Gumbel copula function and Frank copula function to fit the joint distribution function of runoff, wind power output and photovoltaic output respectively. The expressions are as follows: (7); In the formula, It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint probability distribution function of It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint copula distribution function of ; is the marginal distribution function of the runoff from Longtou Power Station; is the marginal distribution function of wind power output; is the marginal distribution function of PV output; Step 2-5: Use Akaike information criterion (AIC) and Bayesian information criterion (BIC) to select the optimal copula function. The smaller the value of AIC or BIC, the better the copula fitting effect. The expression is as follows: (8); Where L is the likelihood function; k is the number of parameters of the copula function; n is the number of data; Step 2-6: Assuming that wind power output is the main variable, according to the fitted copula distribution function and the structure of C-vine copula, the multi-dimensional copula probability density function of the runoff of the leading power station, wind power output and photovoltaic output is decomposed as follows: (9); In the formula, is the joint copula probability density function of runoff, wind power output and photovoltaic output of the leading power station; is the joint copula probability density function of runoff and wind power output of Longtou Power Station; is the joint copula probability density function of PV output and wind power output; is the joint copula probability density function of runoff and PV output of the leading power station under given wind power output conditions; According to the C-vine copula method, the conditional distribution functions of runoff, wind power output and photovoltaic output of Longtou Power Station are calculated, and the expressions are as follows: (10); In the formula, is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; , are the distribution functions of photovoltaic output and wind power output respectively; is the conditional distribution function of runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of runoff from Longtou Power Station; is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; It is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions.

[0033] Furthermore, in step 3, the Monte Carlo sampling method is used to combine the conditional distribution functions of the runoff, wind power output, and photovoltaic output of the leading power station to generate a scene set of the runoff, wind power output, and photovoltaic output of the leading power station that takes into account the correlation, which specifically includes the following steps: Step 3-1: Generate three uniformly distributed random numbers , solve the samples of runoff, wind power output and photovoltaic output of Longtou Power Station , first let Equal to the sample value of wind power output , the expression is as follows: (11); In the formula, is the distribution function of wind power output; is the sample value of wind power output; is a generated random number; Step 3-2: Order Equal to the conditional distribution function value of photovoltaic output, and solve the sample value of photovoltaic output according to the following formula , the expression is as follows: (12); In the formula, It is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; is the distribution function of wind power output; is the sample value of PV output; is a generated random number; Step 3-3: Order Equal to the conditional distribution function value of the runoff of Longtou Power Station, and solve the sample value of the runoff of Longtou Power Station according to the following formula , the expression is as follows: (13); In the formula, is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; It is the conditional distribution function of photovoltaic output under given wind power output conditions; is a generated random number; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of wind power output; is the sample value of runoff; Step 3-4: For the obtained sample After inverse transformation, we get a set of output scenarios, which are expressed as follows: (14); In the formula, xThis is the generated runoff scenario for Longtou Power Station; y The wind power output scenario is generated; z The photovoltaic output scenario is generated; is the inverse function of the runoff distribution function of Longtou Power Station; is the inverse function of wind power output distribution function; is the inverse function of the photovoltaic output distribution function; Step 3-5: Repeat steps 3-1 to 3-4 for a total of 24 times to obtain the runoff, wind power output, and photovoltaic output scenarios of the leading power station for 24 hours a day; Step 3-6: Repeat steps 3-5 for a total of 45 times to obtain the runoff, wind power output, and photovoltaic output scenario set of the leading power station in a solar term. The expression is as follows: (15); In the formula, set is the scene set; It is t Runoff scenarios of Longtou Power Station generated in different time periods; It is t Wind power output scenarios generated in each time period; It is t PV output scenarios generated in each time period; t is the number of moments, ranging from 1 to 24.

[0034] Furthermore, in step 4, the quality of the scene is verified using the variance and Kendall coefficient relative error of the generated scene, which specifically includes the following steps: Step 4-1: Use the variance of the generated scenario to verify whether the generated scenario retains the volatility characteristics of historical data. The expression is as follows: (16); In the formula, is the variance of the generated scene; To generate the scene t The value at a moment; is the mean of the generated scenarios; is the variance of historical data; For historical data t The value at a moment; is the mean of historical data; T is the total number of time periods; Step 4-2: Use the Kendall coefficient relative error of the generated scenario to verify whether the generated scenario retains the correlation between historical data. The expression is as follows: (17); In the formula, It is m Group Kendall coefficient relative error; It is m Kendall coefficient between group-generated scenarios; It is m Kendall coefficient between groups of historical data; m is the scene set number.

[0035] Embodiment 2: The cascade hydropower stations in the middle and lower reaches of a river basin in southwest my country and the wind and solar power stations within 60 km around it from 2011 to 2013 were taken as the analysis objects. The installed capacity of wind power and photovoltaic power was set to 6000MW, and the meteorological data and runoff were sampled every 1 hour.

[0036] Figure 2 The sum of the absolute values ​​of the Kendall coefficients when the main variables are the runoff of the Longtou Power Station, photovoltaic output, and wind power output is shown. It can be seen that when the main variable is wind power output, the maximum sum of the absolute values ​​of the Kendall coefficients is 1.01. Therefore, wind power output is selected as the main variable of C-vine copula. Figure 3 A structural diagram of the C-vine copula is shown. Figure 4 The kernel density estimation method is used to fit the marginal distribution of runoff, wind power output, and photovoltaic output of the Longtou Power Station. From the figure, it can be seen that the marginal distribution fitted by the kernel density estimation method is almost the same as the empirical distribution of each variable, which indicates that the kernel density estimation method has a higher accuracy in fitting the marginal distribution of variables.

[0037] Table 1 shows the AIC and BIC values ​​of the joint distribution function of wind power output, photovoltaic output, and runoff from the leading power station fitted by five copula functions. It can be seen that the AIC and BIC values ​​of the Frank copula fitting the joint distribution function of wind power output-photovoltaic output and wind power output-runoff from the leading power station are the lowest, and the AIC and BIC values ​​of the Gumbel copula fitting the conditional distribution function of photovoltaic output-runoff from the leading power station under the given wind power output are the lowest. Therefore, the Frank copula is the optimal copula function for fitting wind power output-photovoltaic output and wind power output-runoff from the leading power station, and the Gumbel copula is the optimal copula function for fitting the conditional distribution of photovoltaic output-runoff from the leading power station under the given wind power output. Figure 5Taking the "Lichun" solar term as an example, the copula distribution function and density function of Frank copula analysis of wind power output and photovoltaic output, wind power output and Longtou power station runoff are shown. It can be seen that the peak of the copula density function of wind power output-photovoltaic output appears at (1,0) and (0,1), and the copula distribution function is concave on the diagonal, which indicates that there is a negative correlation between wind and solar output. The peak of the copula density function of wind power output-Longtou power station runoff appears at (0,0) and (1,1), and the copula distribution function is convex on the diagonal, indicating that there is a positive correlation between wind power output and runoff. Figure 6 Taking the "Lichun" solar term as an example, the Gumbel copula is used to fit the conditional distribution function of photovoltaic output and runoff from the leading power station under given wind power output. It can be seen from the figure that the value of the copula density function at (1,1) is much higher than the value at (0,0), indicating that photovoltaic output and runoff from the leading power station have upper tail dependence under given wind power output.

[0038] Figure 7 The generated scenarios of runoff, wind power output and photovoltaic output considering correlation are shown in Table 2. The fluctuation characteristics of historical data and generated scenarios are shown in Table 2. It can be seen that the mean curve of the generated scenario and the mean curve of historical data are basically coincident, and the mean variance of the generated scenario and the variance of the mean of historical data are in the same order of magnitude. Among them, the variance of photovoltaic output is the largest, which is 10 6 The variance of wind power output is relatively large, about 10 4 -10 5 The variance of runoff from Longtou Power Station is the smallest, both within 10 3 And below. This is consistent with the characteristics of relatively stable runoff under natural conditions, while the wind and solar output fluctuates greatly. Due to climate factors in this region, the wind and solar output fluctuates greatly in winter and spring. This proves that the scenarios generated by the proposed method can effectively characterize the fluctuation characteristics of historical data.

[0039] Figure 8 The distribution diagram of the relative error of the Kendall coefficient of the generated scenario is shown in Figure 2. It can be seen that the relative errors of the Kendall coefficient of the generated scenario are mostly less than 30%, and the medians of the relative errors of the Kendall coefficient of wind power-photovoltaic, wind power-leading power station runoff, and photovoltaic-leading power station runoff are 28.25%, 21.77%, and 32.23%, respectively. This shows that the proposed method can better preserve the correlation between historical data.

[0040] Table 1 AIC and BIC values ​​of different copula functions

[0041] Table 2 Fluctuation characteristics of water scenery scenes

Claims

1. A method for generating correlation scenes of a water-wind-solar complementary system based on solar term division, characterized in that: Firstly, the historical data of runoff, wind power output and photovoltaic output of Longtou Power Station were divided according to solar terms. Then, the C-vine copula method and Monte Carlo sampling method were used to generate a set of scenarios that took into account the correlation between runoff, wind power output and photovoltaic output of Longtou Power Station. Finally, the quality of the generated scenarios was verified by the fluctuation characteristics and correlation of the generated scenarios.

2. According to the method for generating correlation scenes of a water-wind-solar complementary system based on solar term division according to claim 1, it is characterized in that: The specific steps include: Step 1: Obtain data on runoff, wind power output, and photovoltaic output of the leading power station: Collect the daily runoff of the Longtou Power Station in the lower reaches of the Lancang River in a certain year, as well as the wind speed of wind power stations within a certain range along the coast, and the temperature and radiation data of photovoltaic power stations, and calculate the wind power output and photovoltaic output respectively according to the wind power output formula and photovoltaic output formula; Step 2: Analyze the correlation between runoff, wind power output and photovoltaic output of the leading power station: The runoff, wind power output and photovoltaic output data of Longtou Power Station in the lower reaches of Lancang River in a certain year were divided according to the traditional 24 solar terms in China, and the runoff, wind power output and photovoltaic output data of Longtou Power Station in each solar term were analyzed using C-vine copula method to analyze the correlation between the runoff, wind power output and photovoltaic output of Longtou Power Station; Step 3: Generate scenarios of runoff, wind power output, and photovoltaic output of the leading power station taking into account the correlation: Taking the runoff, wind power output and photovoltaic output of the Longtou Power Station in each solar term as the objects, using the Monte Carlo sampling method and the conditional distribution function of the runoff, wind power output and photovoltaic output of the Longtou Power Station, a large number of scenarios that consider the correlation between the runoff, wind power output and photovoltaic output of the Longtou Power Station are generated; Step 4: Check the quality of the generated scene: Compare the variance between the generated scenario and the historical data to verify whether the generated scenario generally retains the fluctuation characteristics of the historical data; compare the relative error of the Kendall coefficient between the generated scenario and the historical data to verify whether the generated scenario generally retains the correlation between the historical data.

3. The method for generating correlation scenes of a water-wind-solar complementary system based on solar term division according to claim 2 is characterized in that: In step 1, the wind power output and photovoltaic output are calculated respectively using the wind power output formula, the photovoltaic output formula and the collected meteorological data, which specifically includes the following steps: Step 1-1: Collect the daily runoff, wind speed, temperature and radiation data of Longtou Power Station in the middle and lower reaches of Lancang River in a certain year, with an interval of 1 hour; Step 1-2: The wind turbine wind speed of the wind power station is calculated based on the measured wind speed, and the expression is as follows: (1)； In the formula, It is t The wind speed at the wind turbine at the moment, in m / s; The measured height is The t The measured wind speed at a certain moment, in m / s; is the height of the wind turbine hub in meters; q is the stability coefficient; Step 1-3: Calculate wind power output according to the wind power output formula. The expression is as follows: (2)； In the formula, It is t Wind power output at a given moment, in MW; is the wind turbine cut-in wind speed, in m / s; is the cut-out wind speed of the wind turbine, in m / s; is the full wind speed of the wind turbine, in m / s; is the installed capacity of wind power, in MW; It is t The wind speed at the wind turbine at the moment, in m / s; Step 1-4: Calculate the temperature of the solar panel based on the measured temperature data from the weather station. The expression is as follows: (3)； In the formula, It is t The temperature of the solar panel at the moment, in °C; It is t The measured temperature at the weather station at the moment, in °C; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions; is the temperature of the solar panel during normal operation; is the air temperature under standard test conditions; Step 1-5: Calculate the photovoltaic output according to the photovoltaic output formula. The expression is as follows: (4)； In the formula, It is t The photovoltaic output at a certain moment, in MW; is the installed photovoltaic capacity, in MW; is the air temperature-power conversion coefficient; It is t Solar radiation at the moment, in W / m 2 ; is the solar radiation under standard test conditions; It is t The temperature of the solar panel at the moment, in °C; is the air temperature under standard test conditions.

4. The method for generating correlation scenes of a water-wind-solar hybrid system based on solar term division according to claim 3 is characterized in that: In step 2, the C-vine copula method is used to analyze the correlation between the runoff, wind power output, and photovoltaic output of the Longtou Power Station, which specifically includes the following steps: Step 2-1: Use the traditional 24 solar terms in China to divide the runoff, wind power output and photovoltaic output data of Longtou Power Station; the solar terms are from "Lichun" to "Dahan", and each solar term contains 45 days of data. The runoff, wind power output and photovoltaic output data of Longtou Power Station in the same hour are used as the analysis objects, with a total of 24 time periods; Step 2-2: Use the Kendall correlation coefficient method to select the main variable of the C-vine copula. The sum of the absolute values ​​of the Kendall correlation coefficients between the main variable and the other two variables, namely, the runoff of the leading power station, wind power output, and photovoltaic output, must be the largest. The expression is as follows: (5)； In the formula, is the Kendall correlation coefficient; C , D are the number of consistent and inconsistent element pairs in the two variables respectively; n is the number of data; Step 2-3: Use the kernel density estimation method to fit the distribution function of the runoff, wind power output, and photovoltaic output of the Longtou Power Station. The expression is as follows: (6)； In the formula, x is a variable, which is the runoff of the leading power station, wind power output or photovoltaic output; is the probability density function corresponding to the variable, which is the probability density function of the runoff, wind power output or photovoltaic output of the leading power station; n is the number of data; h is the bandwidth, which controls the width of the kernel function; It is i The value of the data; is the kernel function; is the distribution function corresponding to the variable, which is the distribution function of the runoff, wind power output or photovoltaic output of the leading power station; Step 2-4: Use t copula function, Gaussian copula function, Clayton copula function, Gumbel copula function and Frank copula function to fit the joint distribution function of runoff, wind power output and photovoltaic output respectively. The expressions are as follows: (7)； In the formula, It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint probability distribution function of It is the runoff from Longtou Power Station , wind power output , photovoltaic output The joint copula distribution function of ; is the marginal distribution function of the runoff from Longtou Power Station; is the marginal distribution function of wind power output; is the marginal distribution function of PV output; Step 2-5: Use Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) to select the optimal copula function. The smaller the value of AIC or BIC, the better the copula fitting effect. The expression is as follows: (8)； Where L is the likelihood function; k is the number of parameters of the copula function; n is the number of data; Step 2-6: Assuming that wind power output is the main variable, according to the fitted copula distribution function and the structure of C-vine copula, the multi-dimensional copula probability density function of the runoff of the leading power station, wind power output and photovoltaic output is decomposed as follows: (9)； In the formula, is the joint copula probability density function of runoff, wind power output and photovoltaic output of the leading power station; is the joint copula probability density function of runoff and wind power output of Longtou Power Station; is the joint copula probability density function of PV output and wind power output; is the joint copula probability density function of runoff and PV output of the leading power station under given wind power output conditions; According to the C-vine copula method, the conditional distribution functions of runoff, wind power output and photovoltaic output of Longtou Power Station are calculated, and the expressions are as follows: (10)； In the formula, is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; , are the distribution functions of photovoltaic output and wind power output respectively; is the conditional distribution function of runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of runoff from Longtou Power Station; is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; It is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions.

5. The method for generating correlation scenes of a water-wind-solar hybrid system based on solar term division according to claim 4 is characterized in that: In step 3, the Monte Carlo sampling method is used to combine the conditional distribution functions of the runoff, wind power output, and photovoltaic output of the leading power station to generate a scene set of the runoff, wind power output, and photovoltaic output of the leading power station that takes into account the correlation, which specifically includes the following steps: Step 3-1: Generate three uniformly distributed random numbers , solve the samples of runoff, wind power output and photovoltaic output of Longtou Power Station , first let Equal to the sample value of wind power output , the expression is as follows: (11)； In the formula, is the distribution function of wind power output; is the sample value of wind power output; is a generated random number; Step 3-2: Order Equal to the conditional distribution function value of photovoltaic output, and solve the sample value of photovoltaic output according to the following formula , the expression is as follows: (12)； In the formula, It is the conditional distribution function of photovoltaic output under given wind power output conditions; is the joint copula distribution function of PV output and wind power output; is the distribution function of wind power output; is the sample value of PV output; is a generated random number; Step 3-3: Order Equal to the conditional distribution function value of the runoff of Longtou Power Station, and solve the sample value of the runoff of Longtou Power Station according to the following formula , the expression is as follows: (13)； In the formula, is the conditional distribution function of the runoff from the Longtou Power Station under given wind power output and photovoltaic output conditions; is the joint copula distribution function of the runoff and photovoltaic output of the leading power station under given wind power output conditions; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; It is the conditional distribution function of photovoltaic output under given wind power output conditions; is a generated random number; is the conditional distribution function of the runoff from Longtou Power Station under given wind power output conditions; is the joint copula distribution function of runoff and wind power output of Longtou Power Station; is the distribution function of wind power output; is the sample value of runoff; Step 3-4: For the obtained sample After inverse transformation, we get a set of output scenarios, which are expressed as follows: (14)； In the formula, x This is the generated runoff scenario for Longtou Power Station; y The wind power output scenario is generated; z The photovoltaic output scenario is generated; is the inverse function of the runoff distribution function of Longtou Power Station; is the inverse function of wind power output distribution function; is the inverse function of the photovoltaic output distribution function; Step 3-5: Repeat steps 3-1 to 3-4 for a total of 24 times to obtain the runoff, wind power output, and photovoltaic output scenarios of the leading power station for 24 hours a day; Step 3-6: Repeat steps 3-5 for a total of 45 times to obtain the runoff, wind power output, and photovoltaic output scenario set of the leading power station in a solar term. The expression is as follows: (15)； In the formula, set is the scene set; It is t Runoff scenarios of Longtou Power Station generated in different time periods; It is t Wind power output scenarios generated in each time period; It is t PV output scenarios generated in each time period; t is the number of moments, ranging from 1 to 24.

6. The method for generating correlation scenes of a water-wind-solar hybrid system based on solar term division according to claim 5 is characterized in that: In step 4, the quality of the scene is verified using the variance of the generated scene and the relative error of the Kendall coefficient, which specifically includes the following steps: Step 4-1: Use the variance of the generated scenario to verify whether the generated scenario retains the volatility characteristics of historical data. The expression is as follows: (16)； In the formula, is the variance of the generated scene; To generate the scene t The value at a moment; is the mean of the generated scenarios; is the variance of historical data; For historical data t The value at a moment; is the mean of historical data; T is the total number of time periods; Step 4-2: Use the Kendall coefficient relative error of the generated scenario to verify whether the generated scenario retains the correlation between historical data. The expression is as follows: (17)； In the formula, It is m Group Kendall coefficient relative error; It is m Kendall coefficient between group-generated scenarios; It is m Kendall coefficient between groups of historical data; m is the scene set number.

Images