A dispatching method for power consumption of a water-wind-solar power system coupled with meteorological downscaling

Abstract

The present invention belongs to the field of multi - energy complementary coordinated scheduling, and discloses a power generation and consumption scheduling method for a water - wind - light system coupled with meteorological downscaling. The support vector machine regression algorithm is used to identify different hydrological and meteorological variable data, and high - resolution spatial downscaling is achieved by establishing the statistical relationship between the observed data and meteorological factors; the wind - light power output process is calculated using the empirical formula of wind - light power generation; and a series of time - series peak - shaving modes are introduced to determine the peak - shaving and water - wind - light power consumption linkage equation, avoiding the over - estimation of energy consumption that may be caused by ignoring the impact of climate change and short - term power generation rules. Through the analysis of an engineering example composed of wind - light power stations in the Lancang River Basin and its surrounding areas in Yunnan, the results show that the present invention can effectively reduce the hydrological and meteorological downscaling error, and the water - wind - light power consumption linkage equation can more accurately describe the power generation law of water - wind - light energy, making the scheduling results show better accuracy and reliability.

Description

Technical Field

[0001] The present invention belongs to the field of multi - energy complementary coordinated scheduling, and relates to a power generation and consumption scheduling method for a water - wind - light system coupled with meteorological downscaling. Background Art

[0002] With the rapid transformation of China's energy structure towards green and low - carbon, the installed capacity ratio of new energy dominated by wind and light in the power grid is increasing, making the impact of weather on the operation of the power system increasingly prominent. In particular, the power generation of various clean energy sources such as water, wind, and light all severely depends on the changes of meteorological factors at the physical level. In this case, it is very important and necessary to finely evaluate the impact of complex climate and meteorological conditions on multi - energy complementary coordinated scheduling and new energy consumption.

[0003] At present, using downscaling technology for meteorological data to obtain high-resolution meteorological conditions is an important way to accurately reflect future climate change and improve data accuracy. The two main meteorological data downscaling techniques are: dynamic downscaling and statistical downscaling. Dynamic downscaling starts from the physical level to obtain complete climate variables with spatial continuity, but the complex model structure and huge computational amount limit its computational accuracy. Statistical downscaling realizes downscaling by establishing the statistical relationship between the observed data and the data to be downscaled. Its principle is simple and intuitive, and the computational amount is small, and it has been widely used (Wang S, Zhu J, Huang G, et al. Assessment of climate change impacts on energy capacity planning in Ontario, Canada using high-resolution regional climate model[J]. Journal of cleaner production, 2020, 274: 123026.). Among them, machine learning algorithms have the advantages of high accuracy, good applicability, and not being prone to mutations, and are widely used in large-scale hydrological simulation statistical downscaling, and can obtain information from high-dimensional climate variables without prior analysis of the local hydrological processes in a specific area (Hu Baojian, Li Wei, Chen Chuanfa, et al. Using spatial random forest method to improve the quality of GPM satellite remote sensing precipitation[J]. Journal of Remote Sensing, 2024, 28(02): 414-425.). In terms of new energy consumption assessment, the impact of climate change on the clean energy system is usually evaluated on a monthly or dekadal scale, and it is assumed that the inputs of the model, such as reservoir inflow and new energy output, remain unchanged during this period (Zhang Y, Cheng C, Yang T, et al. Assessment of climate change impacts on the hydro-wind-solar energy supply system[J]. Renewable and Sustainable Energy Reviews, 2022, 162: 112480.). However, the output of new energy has strong intermittency, volatility, and unpredictability within a day. These characteristics will not be able to accurately reflect the short-term energy consumption law, resulting in inaccurate assessment results and being difficult to apply.

[0004] In view of the above problems, the present invention proposes a power generation consumption scheduling method for a water-wind-solar system coupled with meteorological downscaling, and applies and tests it with an engineering example composed of a certain large river basin and its surrounding wind and solar power stations. The results show that the present invention can effectively reduce the hydrometeorological downscaling error, and the power generation law of water-wind-solar energy can be more accurately described by the power consumption linkage equation of water-wind-solar, making the scheduling results show better accuracy and reliability. Summary of the Invention

[0005] The present invention mainly solves the technical problem of power generation consumption scheduling for a water-wind-solar system coupled with meteorological downscaling. The purpose is to improve the accuracy and reliability of the consumption scheduling results by finely evaluating the impact of complex climate and meteorological conditions on the coordinated scheduling of multiple energy sources and the consumption of new energy.

[0006] The technical solution of the present invention:

[0007] A power generation consumption scheduling method for a water-wind-solar system coupled with meteorological downscaling, comprising the following steps:

[0008] (1) Use support vector machine regression (SVR) to perform spatial downscaling on hydrometeorological variable data to accurately reflect the influence of climate change on the output of each water-wind-solar power station.

[0009] (1.1) Select hydrometeorological variables as the original data set: use hourly precipitation, evaporation, surface air temperature, and soil water content (0-35 cm) to map the hourly inflow runoff changes of hydropower stations, and hourly 10m wind speed to predict the near-surface wind speed at wind and photovoltaic power stations, and hourly surface shortwave radiation and surface air temperature to predict the radiation received by the solar panels and the ambient temperature of photovoltaic power stations.

[0010] (1.2) Division of the original data set of hydrometeorological variables: Arrange the original data in chronological order and divide it into a training set, a validation set, and a test set according to a ratio of 6:2:2.

[0011] (1.3) Assume that the training set is {(x i , y i )}, i ∈ [1, N), where x i is the large-scale hydrometeorological variable data of different general circulation models (GCM), y i is the actual data corresponding to the time, and N is the size of the data set; represent the linear regression decision surface function of SVR with Equation (1), where ω is the weight vector and b is the bias; apply the non-linear transformation function to map the input space to a high-dimensional feature space:

[0012]

[0013] (1.4) Establish the insensitive loss function of the allowable prediction error ε for hydrometeorological variables:

[0014]

[0015] (1.5) Aiming at minimizing the structural risk of the prediction error of hydrometeorological variables, use the Python-sklearn program module to convert the prediction error minimization problem into an equivalent quadratic convex programming problem under the constraint (4) for solution;

[0016]

[0017] In the formula: ξ i and are the positive relaxation factors of the prediction error. ξ i represents the degree of relaxation when the predicted value is higher than the true value, represents the degree of relaxation when the predicted value is lower than the true value. C is the regularization penalty coefficient of the prediction error;

[0018] (2) Let p i,t be the predicted value of the hydrometeorological variable at the power station i at time period t; r i,t be the true observed value of the hydrometeorological variable at the power station i at time period t; be the maximum value of the true observed value of the hydrometeorological variable at the power station i at time period t; be the minimum value of the true observed value of the hydrometeorological variable at the power station i at time period t; Use the normalized root mean square error (NRMSE) and relative squared error (RSE) to characterize the difference between the observed and predicted time series of hydrometeorological variables on the test set to evaluate the prediction performance. The smaller the index value, the better the model effect; The specific calculation formula is as follows:

[0019]

[0020] In the formula: p i,t is the predicted value of the hydrometeorological variable at the power station i at time period t; r i,t is the true observed value of the hydrometeorological variable at the power station i at time period t; is the maximum value of the true observed value of the hydrometeorological variable at the power station i at time period t; is the minimum value of the true observed value of the hydrometeorological variable at the power station i at time period t;

[0021] (3) Input the hydrometeorological variables at the power station, construct the empirical formulas for wind power generation and photovoltaic power generation, use the Python programming language to import the downscaled data of future hydrometeorological variables from the Excel file, and use the Python-math library to solve the calculated downscaled data again to obtain the change process of the wind-solar power output rate; The specific formula is as follows:

[0022] Empirical formula for wind power generation:

[0023]

[0024] Where: is the wind speed at the height z meters of the wind turbine in the wind farm, m / s; are the cut-in wind speed and cut-out wind speed of the wind turbine in the wind farm, respectively, m / s; is the near-surface wind speed at the location of the wind farm, m / s; z0 is the surface roughness length, taken as 0.0002 m; is the output rate of the wind farm (i.e., the ratio of output to installed capacity); is the power generation function coefficient; is for wind farm n w installed capacity; Δt m is the number of hours in month m;

[0025] Empirical formula for photovoltaic power generation:

[0026]

[0027] Where: is the output rate of photovoltaic power station n pv ; is the performance ratio of the solar panel; is for photovoltaic power station n pv surface radiation, W·m 2 ; rsds STC is the surface radiation under standard atmospheric pressure (101.325 kPa) (rsds STC = 1000 W*m 2 ); γ is taken as -0.005 °C -1 , is the calculation coefficient of the empirical formula; is the temperature of the solar cell, affected by temperature, radiation and wind speed, Tas STC is the ambient air temperature under standard atmospheric pressure (101.325 kPa) (Tas STC = 25 °C); is for photovoltaic power station n pv installed capacity;

[0028]

[0029] Where: Coefficient value is the ambient temperature at photovoltaic power station n pv , °C; is the surface wind speed at photovoltaic power station n pv , m / s;

[0030] (4) Using the hydrometeorological variables at each power station as characteristic inputs, a piecewise linear fitting method is adopted to construct a joint equation for the consumption of hydropower, wind power, and photovoltaic power, so as to extract the complementary consumption relationship between hydropower, wind power, and photovoltaic power. The specific expression is as follows:

[0031]

[0032] In the formula: is the joint function for the consumption of hydropower, wind power, and photovoltaic power, representing the quantitative relationship between the consumption of hydropower, wind power, and photovoltaic power under the scenario where the load in month m is Taking the Gurobi solver as the modeling platform, using the Python language to transform the above nonlinear model into a mixed-integer linear programming for solution, and determining the influence relationship between the hydropower generation and the consumption scale of wind power and photovoltaic power. The specific influence includes four main stages:

[0033] Stage 1: The regulation ability of hydropower is insufficient, restricting the consumption of wind power and photovoltaic power. However, as the hydropower output increases, the flexibility of hydropower increases, and the proportion of wind power and photovoltaic power consumption shows an upward trend.

[0034] Stage 2: The regulation ability of hydropower can completely suppress the fluctuations of wind power and photovoltaic power generation and respond to the peak shaving demand of the receiving-end power grid. Therefore, the wind power and photovoltaic resources can be completely consumed by the receiving-end power grid.

[0035] Stage 3: The channel capacity restricts the bundled transmission output of hydropower, wind power, and photovoltaic power. Therefore, the proportion of wind power and photovoltaic power consumption shows a downward trend as the hydropower output increases.

[0036] Stage 4: When the hydropower output continues to increase until it breaks through the channel capacity limit, the wind power and photovoltaic power generation cannot be consumed.

[0037] The beneficial effects of the present invention: Compared with the single-month scale scheduling method, the scheduling method that couples meteorological downscaling and energy consumption characteristics can effectively reduce the inaccuracy of energy consumption assessment, associate the impact of climate change on clean energy generation, showing better reliability, practicability, and reducing the scheduling risk. Meteorological downscaling essentially refers to the refined prediction of hydrometeorological variables at each power station. By downscaling, the required high-precision data is obtained to determine the impact on clean energy generation, so as to describe in detail the volatility of the output of wind power and photovoltaic power stations. At the same time, by introducing the joint equation for the consumption of hydropower, wind power, and photovoltaic power, the impact relationship between hydropower output and the consumption of wind power and photovoltaic power is established, avoiding the overestimation of new energy consumption by the traditional single-month scale scheduling method under the influence of climate change and improving the accuracy of consumption analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is the overall solution framework diagram of the method of the present invention;

[0039] Figure 2 is the schematic diagram of the support vector machine regression algorithm;

[0040] Figure 3 Schematic diagram of error relaxation of support vector machine regression algorithm;

[0041] Figure 4 It is a schematic diagram of the consumption law of water, wind and light energy;

[0042] Figure 5 It is a schematic diagram of the changes in power generation and consumption of a water-wind-light complementary system. Specific implementation manners

[0043] The following further describes the specific implementation manners of the present invention in conjunction with the accompanying drawings and technical solutions.

[0044] Generally speaking, downscaling the meteorological data at the water-wind-light clean energy system can more accurately reflect the impact of climate change on the power generation of each energy source. To establish the statistical relationship between the observed data and the data to be downscaled for downscaling, hydrometeorological variables are selected: the hourly precipitation, evaporation, surface air temperature, and soil moisture content (0-35 cm) are used to map the hourly changes in the reservoir inflow of hydropower stations, and the hourly 10m wind speed is used to predict the near-surface wind speed at the wind and photovoltaic power stations, and the hourly surface shortwave radiation and surface air temperature are used to predict the radiation received by the solar panels and the ambient temperature at the photovoltaic power stations;

[0045] Then, the historical data is partitioned and divided into a training set, a validation set, and a test set in a ratio of 6:2:2 in chronological order. Among them, the training set is used to fit the model, the validation set is used to optimize the hyperparameters of the model, and the test set is used to evaluate the performance of the trained model.

[0046] Assume that the training set is {(x i ,y i ), i∈[1, N)}, where x i is the large-scale hydrometeorological variable data of different general circulation models (GCMs), y i is the actual data corresponding to the time, and N is the size of the data set; the linear regression decision surface function of SVR is represented by Equation (14), where ω is the weight vector and b is the bias; the input space is mapped to a high-dimensional feature space by applying the nonlinear transformation function :

[0047]

[0048] Subsequently, the hydrometeorological variable-allowing prediction error ε-insensitive loss function is introduced:

[0049]

[0050] Such as Figure 2As shown in the figure, aiming at minimizing the structural risk of prediction error, the Python-sklearn program module is used to convert the prediction error minimization problem into an equivalent quadratic convex programming problem under the constraint (17) for solution:

[0051]

[0052] In the formula: ξ i and are the positive relaxation factors of the prediction error. ξ i represents the relaxation degree when the predicted value is higher than the true value, while represents the relaxation degree when the predicted value is lower than the true value. C is the prediction error penalty, and N is the sample size.

[0053] To measure the accuracy of the combined prediction of different downscaling techniques and meteorological data sources, the normalized root mean square error (NRMSE) and relative squared error (RSE) are used to characterize the differences between the observed and predicted time series of hydrometeorological variables on the test set, so as to reflect the effectiveness of the downscaling technique:

[0054]

[0055] In the formula, p i,t is the predicted value of the hydrometeorological variable at hydropower station i at time t; r i,t is the true observed value of the hydrometeorological variable at hydropower station i at time t; is the maximum value of the true observed value of the hydrometeorological variable at hydropower station i at time t; is the minimum value of the true observed value of the hydrometeorological variable at hydropower station i at time t.

[0056] Using the empirical formula of wind-solar power generation, the downscaled data of future hydrometeorological variables are imported from an Excel file using the Python programming language, and the Python-math library is used to solve the calculated downscaled data again to obtain the change process of the wind-solar power output rate. Among them, wind power generation can be converted through the empirical formula. And it is reasonable to approximate the wind power generation power curve with a polynomial power curve. The hourly wind speed at the hub height is used as the independent variable and the wind speed increases approximately logarithmically with height. The wind power generation empirical formula is as follows:

[0057]

[0058] In the formula: is the wind speed at the height of z meters of the wind turbine of the wind power station, m / s; are the cut-in wind speed and cut-out wind speed of the wind turbine of the wind power station, m / s respectively; is the near-surface wind speed at the location of the wind power station, m / s; z0 is the surface roughness length, 0.0002m; is the output rate of the wind power station (i.e., the ratio of output to installed capacity); is the power generation function coefficient; is the wind power station n w 's installed capacity; Δt m is the number of hours in month m.

[0059] The photovoltaic power station generates electricity through solar cells, and its output rate can be expressed as an empirical function with ambient temperature and solar radiation as independent variables. The photovoltaic power generation empirical formula is as follows:

[0060]

[0061]

[0062] In the formula: is the output rate of the photovoltaic power station n pv ; is the performance ratio of the solar panel; is the photovoltaic power station n pv 's surface radiation, W·m 2 ; rsds STC is the surface radiation under standard atmospheric pressure (101.325 kPa) (rsds STC = 1000 W*m 2 ); γ takes -0.005 °C -1 , is the calculation coefficient of the empirical formula; is the temperature of the solar cell, which is jointly affected by temperature, radiation, and wind speed, Tas STC is the ambient air temperature under standard atmospheric pressure (101.325 kPa) (Tas STC = 25 °C); is the photovoltaic power station n pv 's installed capacity.

[0063]

[0064] In the formula: Coefficient value is the ambient temperature at the location of the photovoltaic power station n pv , °C; is the surface wind speed at the location of the photovoltaic power station n pv , m / s.

[0065] As Figure 4 shown, by using the spatio-temporal complementary consumption characteristics between power sources, the law of water-wind-light consumption can be extracted, which can help improve the accuracy of the dispatching simulation for energy consumption assessment under climate change.

[0066] Construct a joint equation for the consumption of hydropower, wind power, and photovoltaic power. Using the hydrological and meteorological variables at each power station as characteristic inputs, determine the joint equation for the consumption of hydropower, wind power, and photovoltaic power by means of piecewise linear fitting. Taking the Gurobi solver as the modeling platform, use the Python language to transform the above nonlinear model into a mixed-integer linear programming for solution, and determine the influence relationship between the hydropower generation and the consumption scale of wind and photovoltaic power.

[0067] To sum up, the expression of the joint equation for the consumption of hydropower, wind power, and photovoltaic power is as follows:

[0068]

[0069] In the formula: is the joint function for the consumption of hydropower, wind power, and photovoltaic power, indicating the quantitative relationship between the consumption of hydropower, wind power, and photovoltaic power under the load of in the m-th month.

[0070] For the single-month scale scheduling, introducing the joint equation for the consumption of hydropower, wind power, and photovoltaic power means establishing the correlation between hydropower generation and the consumption scale of wind and photovoltaic power, thus avoiding the overestimation of new energy consumption in the single-month scale scheduling under the influence of climate change and improving the rationality of the scheduling analysis results.

[0071] Taking the two large-scale multi-year regulating cascade hydropower station groups of LXW and LNZD on the main stream of a certain large river basin and the surrounding wind and photovoltaic power stations as the actual engineering background for application verification, using the hourly-scale wind and photovoltaic power output and runoff data in 2023 as samples. The power grid load process, wind and photovoltaic power output process, and historical data used in the present invention all refer to the actual operation data of the power grid and power stations.

[0072] The calculation results are shown in Tables 1 to 4. Compared with other downscaling machine learning algorithms, no matter what GCM model output data is used, the support vector machine regression algorithm (SVR) always performs best in predicting future hydrological and meteorological variables. In addition, FGOALS-g3-SVR, MRI-ESM2-0-SVR, MIROC6-SVR, and CanESM5-SVR perform best in predicting solar radiation, ambient temperature, surface wind speed, and runoff respectively. It can be seen that different GCMs show specific accuracy and reliability in predicting different hydrological and meteorological variables. Therefore, combining SVR with the outputs of different GCMs in the prediction model can provide more accurate information about climate change.

[0073] When the peak shaving depth remains unchanged, the annual variation in the system's wind and solar power accommodation conforms to the natural variation trend of its resources. Moreover, in the complementary accommodation characteristics of different months, the accommodation volumes of wind and solar power both increase with the increase in hydropower generation. However, when the hydropower generation continues to increase, it will inversely inhibit the accommodation of wind and solar power. Because too much or too little hydropower will weaken the regulation ability of hydropower, resulting in the hydropower system being unable to provide sufficient flexibility support for the accommodation of intermittent wind and solar power.

[0074] To verify the accuracy of the method of the present invention for evaluating the accommodation of water, wind, and solar energy, two complementary operation modes are constructed. Considering the current actual peak shaving requirements, the simulation results of the separate operation models of water, wind, and solar power and the two complementary operation models are compared. The traditional long-term scheduling model without nesting the short-term complementary accommodation characteristic curve of base energy is the complementary operation mode 1, abbreviated as mode 1; the long-term scheduling model nested with the short-term complementary accommodation characteristic curve is the complementary operation mode 2, abbreviated as mode 2. The results of the two modes are as Figure 5 shown. The overall power generation level of mode 1 is higher than that of mode 2. Among them, under the SSP119 climate change scenario, the total energy accommodation of mode 2 only accounts for 72% of that of mode 1, and the average value of the energy accommodation ratio under different climate changes is about 79%. This means that in the traditional single monthly scale framework, when evaluating future energy generation considering the impact of climate change in a basin, if the short-term peak shaving requirements of the power grid are ignored, the expected benefits may ultimately not be achieved. It can be seen that ignoring the short-term complementary characteristics of the water-wind-solar complementary system in such a framework will result in an overly optimistic accommodation evaluation result, and this result may be exacerbated under the combined influence of different climate changes.

[0075] Through the comparative analysis of different algorithms and different schemes, it is verified that the method for dispatching the power generation and accommodation of the water-wind-solar system proposed by the present invention coupling meteorological downscaling is applicable to the single monthly scale framework, and the results have high accuracy and strong applicability, and can effectively evaluate the accommodation level of clean energy under the influence of climate and meteorological changes.

[0076] Table 1 Predictions of solar radiation by different combinations of GCM model output data and downscaling techniques

[0077]

[0078] Table 2 Predictions of ambient temperature by different combinations of GCM model output data and downscaling techniques

[0079]

[0080] Table 3 Predictions of surface wind speed by different combinations of GCM model output data and downscaling techniques

[0081]

[0082]

[0083] Table 4 Prediction of runoff by the combined output data of different GCM models and downscaling techniques

[0084]

Claims

1. A power generation and consumption scheduling method for a water-wind-solar system coupled with meteorological downscaling, characterized in that It includes the following steps: (1) Use support vector machine regression to perform spatial downscaling on hydrometeorological variable data to accurately reflect the impact of climate change on the power generation of each hydropower, wind power, and photovoltaic power station; (1.1) Select hydrometeorological variables as the original data set: Use hourly precipitation, evaporation, surface air temperature, and soil water content to map the hourly changes in the inflow of hydropower stations. The hourly 10m wind speed is used to predict the near-surface wind speed at wind power and photovoltaic power stations hour by hour. The hourly surface shortwave radiation and surface air temperature are used to predict the radiation received by the solar panels and the ambient temperature at photovoltaic power stations hour by hour; (1.2) Division of the original data set of hydrometeorological variables: According to the time sequence, the original data is divided into a training set, a validation set, and a test set in a ratio of 6:2:2; (1.3) Suppose the training set is {(x i , y i ), i ∈ [1, N)}, where x i is the large-scale hydrometeorological variable data of different general circulation models, y i is the actual data at the corresponding time, and N is the size of the dataset; the linear regression decision surface function of SVR is represented by Equation (1), where ω is the weight vector and b is the bias; apply the non-linear transformation function to map the input space to a high-dimensional feature space: (1.4) Establish an insensitive loss function for the allowable prediction error ε of hydrometeorological variables: (1.5) With the goal of minimizing the structural risk of the prediction error of hydrometeorological variables, use the Python-sklearn program module to convert the prediction error minimization problem into an equivalent quadratic convex programming problem under the constraint (4) for solution; Where: ξ i and are positive relaxation factors of the prediction error. ξ i represents the degree of relaxation when the predicted value is higher than the true value, represents the degree of relaxation when the predicted value is lower than the true value, and C is the regularization penalty coefficient of the prediction error; (2) Let p i,t be the predicted value of the hydro-meteorological variable at power station i at time period t; r i,t be the true observed value of the hydro-meteorological variable at power station i at time period t; be the maximum value of the true observed value of the hydro-meteorological variable at power station i at time period t; r i,t be the minimum value of the true observed value of the hydro-meteorological variable at power station i at time period t; The normalized root mean square error NRMSE and the relative squared error RSE are used to characterize the difference between the observed and predicted time series of the hydro-meteorological variable on the test set to evaluate the prediction performance. The smaller the index value, the better the model performance; the specific calculation formula is as follows: Where: p i,t is the predicted value of the hydro-meteorological variable of power station i at time period t; r i,t is the true observed value of the hydro-meteorological variable of power station i at time period t; is the maximum value of the true observed value of the hydro-meteorological variable of power station i at time period t; r i,t is the minimum value of the true observed value of the hydro-meteorological variable of power station i at time period t; (3) Input the hydrometeorological variables at the power station, construct empirical formulas for wind power generation and photovoltaic power generation, use the Python programming language to import the downscaled data of future hydrometeorological variables from an Excel file, and use the Python-math library to solve the calculated downscaled data again to obtain the change process of the wind and light output rates; The specific formulas are as follows: Empirical formula for wind power generation: In the formula: is the wind speed at the height z meters of the wind turbine in the wind farm, m / s; are respectively the cut-in wind speed and cut-out wind speed of the wind turbine in the wind farm, m / s; is the near-surface wind speed at the location of the wind farm, m / s; z0 is the surface roughness length, taken as 0.0002 m; is the output rate of the wind farm, that is, the ratio of the output to the installed capacity; is the power generation function coefficient; is for wind farm n w installed capacity; Δt m is the number of hours in month m; Empirical formula for photovoltaic power generation: Where: is the output rate of photovoltaic power station n pv ; is the performance ratio of solar panels is the surface radiation of photovoltaic power station n pv , W·m 2 ; rsds STC is the surface radiation under standard atmospheric pressure, rsds STC = 1000W*m 2 ; γ takes -0.005°C -1 , which is the calculation coefficient of the empirical formula is the temperature of the solar cell, which is jointly affected by temperature, radiation and wind speed, Tas STC is the ambient air temperature under standard atmospheric pressure, Tas STC = 25°C; is the installed capacity of photovoltaic power station n pv ; Where: Coefficient value is the ambient temperature at photovoltaic power station n pv , in °C; is the ground surface wind speed at photovoltaic power station n pv , in m / s; (4) Using the hydrometeorological variables at each power station as the characteristic input, adopt a piecewise linear fitting method to construct a coupled equation for the consumption of hydropower, wind power, and photovoltaic power to realize the disjunction of the complementary consumption relationship of hydropower, wind power, and photovoltaic power. The specific expression is as follows: In the formula: is the water-wind-solar power consumption linkage function, representing the quantitative relationship between the water-wind-solar power consumption under the load of in the m-th month; Taking the Gurobi solver as the modeling platform, use the Python language to convert the above nonlinear model into a mixed-integer linear programming for solution, and determine the influence relationship between the hydropower generation amount and the consumption scale of wind and photovoltaic power. The specific influence includes four stages: Stage 1: The regulation capacity of hydropower is insufficient, restricting the consumption of wind and light; as the hydropower output increases, the flexibility of hydropower increases, and the proportion of wind and light consumption shows an upward trend; Stage 2: The regulation capacity of hydropower can completely suppress the fluctuations of wind and photovoltaic power generation and respond to the peak shaving demand of the receiving-end power grid; thus, the wind and light resources can be completely consumed by the receiving-end power grid; Stage 3: The channel capacity limits the bundled external output of hydropower, wind power, and photovoltaic power; thus, the proportion of wind and light consumption shows a downward trend as the hydropower output increases; Stage 4: The hydropower output continues to increase until it breaks through the channel capacity limit, and the wind and photovoltaic power generation cannot be consumed.

Images