Power system scheduling method and system considering wind power plant output uncertainty risk

Abstract

The invention provides a power system scheduling method and system considering wind power plant output uncertainty risk, and relates to the technical field of power system scheduling, and the method specifically comprises the steps: collecting meteorological sensing data of each position in a wind power plant and high-resolution satellite data of a wind power plant region, and carrying out the local simulation through a numerical weather prediction model, high-precision meteorological data is obtained by combining data calculation; the method comprises the following steps: constructing a wind power plant output prediction model, collecting and processing meteorological data of each wind generating set in real time, inputting the processed high-precision meteorological data into the wind power plant output prediction model to obtain the total prediction output of the wind power plant, and dynamically adjusting the load distribution and reserve capacity of a power grid. According to the method, the high-resolution meteorological data and the microscopic meteorological model are integrated, the LSTM algorithm is used for real-time wind power output prediction, the power grid dispatching strategy is dynamically adjusted, the technical effect is better than that of the prior art, the accuracy of wind power output prediction is improved, and the risk caused by the uncertainty of wind power output is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and in particular to a power system dispatching method and system taking into account the uncertainty risk of wind farm output. Background Art

[0002] With the growth of global energy demand and the enhancement of environmental protection awareness, wind power generation, as an important renewable energy source, has received widespread attention. However, the randomness and uncertainty of wind farm output is one of the major challenges facing the power system. Wind power generation is highly dependent on meteorological conditions, especially the change of wind speed, which makes the output power of wind farms highly volatile and unpredictable. This volatility not only increases the complexity of power system scheduling, but also may pose a threat to the safe and stable operation of the power grid. Traditional power system scheduling methods usually rely on deterministic load forecasts and the operating characteristics of conventional generators to formulate scheduling strategies, but these methods show obvious shortcomings when facing the uncertainty of wind farm output, and it is difficult to adapt to modern power grids with an increasing proportion of wind power.

[0003] In the prior art, many wind power output prediction methods have been proposed, including methods based on statistical analysis and artificial intelligence algorithms. However, statistical analysis methods usually require a large amount of historical data support and have weak response capabilities to nonlinear and complex meteorological changes. Artificial intelligence algorithms, such as neural networks, although they have certain advantages in dealing with complex nonlinear problems, often rely on a large amount of high-quality data for training. Insufficient data accuracy and limited model generalization capabilities make these prediction models often face problems of insufficient accuracy and poor adaptability in practical applications. In addition, most existing methods ignore the high-precision simulation and analysis of real-time meteorological conditions and lack the ability to respond quickly to local meteorological changes, which further limits the accuracy of wind power output prediction and the dispatch efficiency of the power system. Therefore, improving the accuracy of wind farm output prediction and in-depth analysis of the impact of meteorological changes on wind farm output have become key technical issues that need to be solved urgently.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for dispatching a power system taking into account the risk of uncertainty in wind farm output, so as to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method and system for dispatching a power system considering the uncertainty risk of wind farm output, the specific steps include:

[0008] Step 1: Set up meteorological sensors at the location of each wind turbine within the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure;

[0009] Step 2: Use a numerical weather prediction model to perform local simulation, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure;

[0010] Step 3: retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. Use the historical time series high-precision meteorological data as the training set and the wind turbine output data as the label to train the model and obtain a trained wind farm output prediction model;

[0011] Step 4: Input the current high-precision meteorological data into the wind farm output prediction model to generate the local predicted output of the wind farm at the next moment in the future. The total predicted output of the wind farm is obtained by summing up and calculating. Based on the total predicted output of the wind farm, the load distribution and spare capacity of the power grid are dynamically adjusted, and the output prediction model and scheduling strategy are continuously optimized according to the actual situation.

[0012] Furthermore, a meteorological sensor is set at the location of each wind turbine in the wind farm and is numbered in sequence as 1, 2, 3, ..., M, that is, the numbers represent the locations of each wind turbine in the wind farm, and M represents the number of wind turbines in the wind farm.

[0013] Furthermore, the process of using the numerical weather forecast model to perform local simulation specifically includes:

[0014] a. Data preparation

[0015] Initial conditions: Obtain regional meteorological data from the GFS model, including wind speed, temperature, and air pressure;

[0016] Boundary conditions: high-resolution wind speed, temperature and air pressure data are obtained from satellite data, and wind speed, temperature and air pressure data are obtained from ground observation stations;

[0017] Topographic data: Obtain high-resolution topographic data of the wind farm and its surrounding areas from DEM data;

[0018] b. Model settings

[0019] Grid division: Set the horizontal grid resolution to 2 kilometers and the vertical grid resolution to 100 meters;

[0020] Physical parameterization scheme: WSM6 scheme is selected for cloud microphysical process, RRTMG scheme for radiation transfer, and YSU scheme for boundary layer process;

[0021] Time step: Set the time step to 10 seconds;

[0022] c. Model operation

[0023] Initialization: Input the initial conditions and boundary conditions into the WRF model for initialization;

[0024] Simulation operation: Start the WRF model to perform simulation operation and calculate the meteorological field of the wind farm and its surrounding areas;

[0025] Output results: Output simulation results every hour, including wind speed, temperature and air pressure;

[0026] d. Result analysis

[0027] Data processing: Interpolate and smooth the simulation results to obtain high-precision meteorological data for each point in the wind farm, including: local temperature T at location j in the wind farm at time t local (j, t); local air pressure P at position j in the wind farm at time t local (j, t), where t represents the time variable.

[0028] Furthermore, the formula for generating the local wind speed is as follows:

[0029] v local (j, t) = α*v global (t)+β*T local (j,t)+γ*P local (j,t)+C1

[0030] Among them, v locai (j, t) is the local wind speed at position j in the wind farm at time t, v global is the regional wind speed at time t, the regional wind speed is the wind speed data in the high-resolution satellite data, α is its preset proportional coefficient, T local (j, t) is the local temperature of position j in the wind farm at time t, β is its preset proportionality coefficient, P local (j, t) is the local air pressure at position j in the wind farm at time t, γ is its preset proportional coefficient, and α>β>γ>0, where t represents the time variable and C1 is the constant correction index.

[0031] Furthermore, the process of preprocessing the collected data specifically includes:

[0032] Remove data outliers: Calculate the standard deviation of the data and set 3 times the standard deviation as the threshold. Data points outside this range are considered outliers.

[0033] The formula used to calculate the standard deviation of the data is:

[0034]

[0035] Where σ is the standard deviation, N is the total number of data points, and x i is the i-th data point, μ is the mean of the data;

[0036] The formula for calculating the mean of the data is:

[0037]

[0038] Where μ is the mean of the data, N is the total number of data points, and x i is the i-th data point;

[0039] Data interpolation: For time series meteorological data, linear interpolation is used to fill missing values: assuming that the data changes linearly between two known points, the missing values ​​are estimated by calculating the linear relationship between the two points.

[0040] Furthermore, the process of constructing the wind farm output prediction model specifically includes:

[0041] Based on the long short-term memory network algorithm, a wind farm output prediction model is constructed. The retrieved historical time series high-precision meteorological data is used as the training set, and the corresponding wind turbine output data at the next moment is used as the label. The model is trained to obtain a trained wind farm output prediction model.

[0042] The process of generating the predicted output of the wind farm at the next moment in the future specifically includes:

[0043] The high-precision meteorological data collected in real time at the j position in the wind farm at time t0 is input into the wind farm output prediction model, and the predicted output W of the j position in the wind farm at time t0+1 is obtained as output. j (t0+1), that is:

[0044] W j (t0+1)=LSTM(v local (j,t0)T local (j, t0), P local (j, t0))

[0045] Among them, t0 represents the current time, W j(t0+1) is the predicted output of the jth wind turbine generator at time t0+1, v local (j,t0) is the local wind speed at position j in the wind farm at time t0, T local (j, t0) is the local temperature at position j in the wind farm at time t0; P local (j, t0) is the local air pressure at position j in the wind farm at time t0;

[0046] Therefore, the total predicted output of the wind farm at time t0+1 is:

[0047]

[0048] Where t0 represents the current time, P(t0+1) is the total predicted output of the wind farm at time t0+1, M is the number of wind turbines in the wind farm, and W j (t0+1) is the predicted output of the jth wind turbine at time t0+1.

[0049] Furthermore, the process of dynamically adjusting the load distribution and spare capacity of the power grid according to the real-time predicted wind farm output specifically includes:

[0050] R(t0+1)=max(0,D(t0)-P(t0+1))+S(t0)

[0051] Among them, t0 represents the current time, R(t0+1) is the backup capacity required by the wind farm at t0+1, D(t) is the grid load requirement of the wind farm at t0, P(t0+1) is the predicted output of the wind farm at t0+1, and S(t) is the system reserve capacity at t0.

[0052] Furthermore, the process of continuously optimizing the prediction model and scheduling strategy according to actual conditions specifically includes:

[0053] Collect the error data between wind farm output and predicted value in real time, calculate the prediction error, and regularly update and retrain the LSTM model based on the error data to improve the prediction accuracy. Retrain the model using the latest meteorological data and output data, analyze the adjusted grid operation data, evaluate the effect of the dispatch strategy, and optimize the dispatch strategy based on the actual operation and prediction error;

[0054] The calculation of the forecast error is based on the following formula:

[0055] E(t0)=P actual (t0)-P predicted (t0)

[0056] Among them, t0 represents the current time, E(t0) is the prediction error at time t0, and P actual(t0) is the actual wind power output at time t0, P predicted (t0) is the predicted wind power output at time t0.

[0057] The present invention further provides a power dispatching system considering the risk of uncertainty in wind farm output, wherein the power dispatching system considering the risk of uncertainty in wind farm output is used to execute the above-mentioned power dispatching method considering the risk of uncertainty in wind farm output, comprising:

[0058] The data acquisition module is used to set up meteorological sensors at the location of each wind turbine generator set in the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure;

[0059] A local simulation module is used to perform local simulation using a numerical weather prediction model, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure;

[0060] The model building module is used to retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. The historical time series high-precision meteorological data is used as a training set, and the wind turbine output data is used as a label to train the model to obtain a trained wind farm output prediction model;

[0061] The system dispatch module is used to input the current high-precision meteorological data into the wind farm output prediction model, generate the local predicted output of the wind farm at the next moment in the future, obtain the total predicted output of the wind farm by summing up, and dynamically adjust the load distribution and spare capacity of the power grid according to the total predicted output of the wind farm, and continuously optimize the output prediction model and dispatch strategy according to actual conditions.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] The present invention can obtain high-precision meteorological data at key locations in the wind farm by installing multiple meteorological sensors inside the wind farm and using high-resolution satellite data, combined with a numerical weather forecast model for local simulation. Based on these data, a long short-term memory network algorithm is used to construct a wind farm output prediction model, which significantly improves the accuracy and timeliness of the prediction; the input of real-time meteorological data enables the prediction model to quickly respond to changes in wind speed and direction, thereby dynamically adjusting the load distribution and spare capacity of the power grid, effectively reducing the risk of power grid operation caused by wind energy fluctuations. In addition, continuous monitoring and model optimization ensure the long-term stability and reliability of the prediction model, providing strong support for the stable operation of wind farms and the safe dispatch of power grids. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0065] Figure 2 It is a schematic diagram of the overall system module of the present invention. DETAILED DESCRIPTION

[0066] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0067] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0068] Example:

[0069] See also Figure 1 , the present invention provides a technical solution:

[0070] The power system dispatching method considering the uncertainty risk of wind farm output includes the following specific steps:

[0071] Step 1: Set up meteorological sensors at the location of each wind turbine within the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure;

[0072] In this embodiment, a meteorological sensor is set at the location of each wind turbine in the wind farm, and is numbered in sequence as 1, 2, 3, ..., M, that is, the numbers represent the locations of each wind turbine in the wind farm, and M represents the number of wind turbines in the wind farm.

[0073] Step 2: Use a numerical weather prediction model to perform local simulation, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure;

[0074] In this embodiment, the process of using the numerical weather forecast model to perform local simulation specifically includes:

[0075] a. Data preparation

[0076] Initial conditions: Obtain regional meteorological data from the GFS model, including wind speed, temperature, and air pressure;

[0077] Boundary conditions: high-resolution wind speed, temperature and air pressure data are obtained from satellite data, and wind speed, temperature and air pressure data are obtained from ground observation stations;

[0078] Topographic data: Obtain high-resolution topographic data of the wind farm and its surrounding areas from DEM data;

[0079] b. Model settings

[0080] Grid division: Set the horizontal grid resolution to 2 kilometers and the vertical grid resolution to 100 meters;

[0081] Physical parameterization scheme: WSM6 scheme is selected for cloud microphysical process, RRTMG scheme for radiation transfer, and YSU scheme for boundary layer process;

[0082] Time step: Set the time step to 10 seconds;

[0083] c. Model operation

[0084] Initialization: Input the initial conditions and boundary conditions into the WRF model for initialization;

[0085] Simulation operation: Start the WRF model to perform simulation operation and calculate the meteorological field of the wind farm and its surrounding areas;

[0086] Output results: Output simulation results every hour, including wind speed, temperature and air pressure;

[0087] d. Result analysis

[0088] Data processing: Interpolate and smooth the simulation results to obtain high-precision meteorological data for each point in the wind farm, including: local temperature T at location j in the wind farm at time t local (j,t); local air pressure P at position j in the wind farm at time t local (j,t), where t represents the time variable.

[0089] The formula used to generate the local wind speed is as follows:

[0090] v local (j,t)=α*v global (t)+β*T local (j,t)+γ*P local (j,t)+C1

[0091] Among them, v local (j,t) is the local wind speed at position j in the wind farm at time t, v global is the regional wind speed at time t, the regional wind speed is the wind speed data in the high-resolution satellite data, α is its preset proportional coefficient, T local (j, t) is the local temperature of position j in the wind farm at time t, β is its preset proportionality coefficient, P local (j, t) is the local air pressure at position j in the wind farm at time t, γ is its preset proportionality coefficient, and α>β>γ>0, because the regional wind speed usually has a large weight. The formation of local wind speed is mainly determined by the large-scale regional wind speed and local terrain conditions. Local temperature and air pressure also affect wind speed, but usually these factors work on a microscopic scale and the impact is relatively small. Among them; t represents the time variable, and C1 is the constant correction index.

[0092] When the regional wind speed v at time t global When the local wind speed v local (j, t) will increase accordingly. When the local temperature rises, the surface temperature rises, the warm air rises, and the cold air falls. This vertical movement will affect the local wind speed, and in areas with strong convection, the wind speed will increase, so v local (j, t) increases; when the local air pressure increases, the pressure gradient force also increases, so the wind speed increases, so v local (j,t) increases; that is, v global (t), T local (j,t),P local (j,t) and v local (j, t) are positively correlated.

[0093] Step 3: retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. Use the historical time series high-precision meteorological data as the training set and the wind turbine output data as the label to train the model and obtain a trained wind farm output prediction model;

[0094] In this embodiment, the process of preprocessing the collected data specifically includes:

[0095] Remove data outliers: Calculate the standard deviation of the data and set 3 times the standard deviation as the threshold. Data points outside this range are considered outliers.

[0096] The formula used to calculate the standard deviation of the data is:

[0097]

[0098] Where σ is the standard deviation, N is the total number of data points, and x i is the i-th data point, μ is the mean of the data;

[0099] The formula for calculating the mean of the data is:

[0100]

[0101] Where μ is the mean of the data, N is the total number of data points, and x i is the i-th data point;

[0102] Data interpolation: For time series meteorological data, linear interpolation is used to fill missing values: assuming that the data changes linearly between two known points, the missing values ​​are estimated by calculating the linear relationship between the two points.

[0103] The process of constructing the wind farm output prediction model specifically includes:

[0104] Based on the long short-term memory network algorithm, a wind farm output prediction model is constructed. The collected historical time series meteorological data is used as the training set, and the wind farm output data is used as the label to train the model to obtain a trained wind farm output prediction model.

[0105] Deep learning algorithms can process and identify patterns in large amounts of complex data, providing more accurate prediction results than traditional machine learning methods, especially when processing data with time series characteristics. The model can automatically learn and extract useful features from the data, avoiding the complexity and limitations of manual feature engineering and improving the generalization ability of the model. Since there is a nonlinear relationship between wind power output and wind speed, temperature and air pressure, deep learning can effectively capture these nonlinear relationships and provide more accurate predictions. The wind farm output prediction model includes an input layer, a fully connected layer, an activation function layer and an output layer. The input layer is used to receive processed meteorological data; the fully connected layer is used to integrate the features of the previous layer to make the final prediction. When predicting the output, the fully connected layer maps the features to the prediction results; the activation function layer is used to introduce nonlinearity; and the output layer generates the final prediction results.

[0106] Step 4: Input the current high-precision meteorological data into the wind farm output prediction model to generate the local predicted output of the wind farm at the next moment in the future. The total predicted output of the wind farm is obtained by summing up and dynamically adjusting the load distribution and spare capacity of the power grid according to the total predicted output of the wind farm. The output prediction model and dispatching strategy are continuously optimized according to the actual situation:

[0107] In this embodiment, the process of constructing the wind farm output prediction model specifically includes:

[0108] Based on the long short-term memory network algorithm, a wind farm output prediction model is constructed. The retrieved historical time series high-precision meteorological data is used as the training set, and the corresponding wind turbine output data at the next moment is used as the label. The model is trained to obtain a trained wind farm output prediction model.

[0109] The process of generating the predicted output of the wind farm at the next moment in the future specifically includes:

[0110] The high-precision meteorological data collected in real time at the j position in the wind farm at time t0 is input into the wind farm output prediction model, and the predicted output W of the j position in the wind farm at time t0+1 is obtained as output. j (t0+1), that is:

[0111] W j (t0+1)=LSTM(v local (j,t0),T local (j,t0),P local (j,t0))

[0112] Among them, t0 represents the current time, W j (t0+1) is the predicted output of the jth wind turbine generator at time t0+1, v local(j,t0) is the local wind speed at position j in the wind farm at time t0, T local (j, t0) is the local temperature at position j in the wind farm at time t0; P local (j, t0) is the local air pressure at position j in the wind farm at time t0;

[0113] Therefore, the total predicted output of the wind farm at time t0+1 is:

[0114]

[0115] Where t0 represents the current time, P(t0+1) is the total predicted output of the wind farm at time t0+1, M is the number of wind turbines in the wind farm, and W j (t0+1) is the predicted output of the jth wind turbine at time t0+1.

[0116] The process of dynamically adjusting the load distribution and spare capacity of the power grid according to the real-time predicted wind farm output specifically includes:

[0117] R(t0+1)=max(0,D(t0)-P(t0+1))+S(t0)

[0118] Among them, t0 represents the current time, R(t0+1) is the backup capacity required by the wind farm at t0+1, D(t) is the grid load requirement of the wind farm at t0, P(t0+1) is the predicted output of the wind farm at t0+1, and S(t) is the system reserve capacity at t0.

[0119] The process of continuously optimizing the prediction model and scheduling strategy according to actual conditions specifically includes:

[0120] Collect the error data between wind farm output and predicted value in real time, calculate the prediction error, and regularly update and retrain the LSTM model based on the error data to improve the prediction accuracy. Retrain the model using the latest meteorological data and output data, analyze the adjusted grid operation data, evaluate the effect of the dispatch strategy, and optimize the dispatch strategy based on the actual operation and prediction error;

[0121] The calculation of the forecast error is based on the following formula:

[0122] E(t0)=P actual (t0)-P predicted (t0)

[0123] Among them, t0 represents the current time, E(t0) is the prediction error at time t0, and P actual (t0) is the actual wind power output at time t0, P predicted (t0) is the predicted wind power output at time t0.

[0124] By collecting the error data between wind farm output and predicted values ​​in real time and regularly updating and retraining the LSTM model, the prediction accuracy can be significantly improved. The LSTM network has good processing capabilities for time series data and can capture the complex dynamic changes in wind farm output.

[0125] See also Figure 2 , the power dispatching system considering the uncertainty risk of wind farm output includes:

[0126] The data acquisition module is used to set up meteorological sensors at the location of each wind turbine generator set in the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure;

[0127] A local simulation module is used to perform local simulation using a numerical weather prediction model, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure;

[0128] The model building module is used to retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. The historical time series high-precision meteorological data is used as a training set, and the wind turbine output data is used as a label to train the model to obtain a trained wind farm output prediction model;

[0129] The system dispatch module is used to input the current high-precision meteorological data into the wind farm output prediction model, generate the local predicted output of the wind farm at the next moment in the future, obtain the total predicted output of the wind farm by summing up, and dynamically adjust the load distribution and spare capacity of the power grid according to the total predicted output of the wind farm, and continuously optimize the output prediction model and dispatch strategy according to actual conditions.

[0130] The specific values ​​of α, β and γ in the formula are generally determined by technical personnel in this field according to actual conditions. The technical personnel in this field collect multiple groups of sample data, and set corresponding preset proportional coefficients for each group of sample data. The preset proportional coefficients and the collected sample data are substituted into the formula. Through repeated experiments and parameter adjustments, the accuracy of the model output and the rationality of the results are observed, and these factor coefficients are gradually adjusted. The performance and effect of the model under different parameter settings are compared to find the optimal coefficient combination, and the calculated factor coefficients are screened and averaged to obtain the values ​​of α, β and γ.

[0131] In addition, the size of the preset factor coefficient is to quantify each parameter to obtain a specific value. In order to facilitate subsequent comparison, the size of the coefficient depends on the amount of sample data and the preset proportional coefficient initially set by technical personnel in this field for each set of sample data. It is not unique as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0132] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0133] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0134] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0135] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A power system dispatching method considering the uncertainty risk of wind farm output, characterized in that: The specific steps include: Step 1: Set up meteorological sensors at the location of each wind turbine within the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure; Step 2: Use a numerical weather prediction model to perform local simulation, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure; Step 3: retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. Use the historical time series high-precision meteorological data as the training set and the wind turbine output data as the label to train the model and obtain a trained wind farm output prediction model; Step 4: Input the current high-precision meteorological data into the wind farm output prediction model to generate the local predicted output of the wind farm at the next moment in the future. The total predicted output of the wind farm is obtained by summing up and calculating. Based on the total predicted output of the wind farm, the load distribution and spare capacity of the power grid are dynamically adjusted, and the output prediction model and scheduling strategy are continuously optimized according to the actual situation.

2. The power system dispatching method considering the uncertainty risk of wind farm output according to claim 1 is characterized in that: The meteorological sensors are arranged at the position of each wind turbine generator set in the wind farm and are numbered in sequence as 1, 2, 3, ..., M, that is, the numbers represent the positions of the wind turbine generator sets in the wind farm, and M represents the number of wind turbine generator sets in the wind farm.

3. The power system dispatching method considering the risk of wind farm output uncertainty according to claim 1 is characterized in that: The process of using the numerical weather forecast model to perform local simulation specifically includes: a. Data preparation Initial conditions: Obtain regional meteorological data from the GFS model, including wind speed, temperature, and air pressure; Boundary conditions: high-resolution wind speed, temperature and air pressure data are obtained from satellite data, and wind speed, temperature and air pressure data are obtained from ground observation stations; Topographic data: Obtain high-resolution topographic data of the wind farm and its surrounding areas from DEM data; b. Model settings Grid division: Set the horizontal grid resolution to 2 kilometers and the vertical grid resolution to 100 meters; Physical parameterization scheme: WSM6 scheme is selected for cloud microphysical process, RRTMG scheme for radiation transfer, and YSU scheme for boundary layer process; Time step: Set the time step to 10 seconds; c. Model operation Initialization: Input the initial conditions and boundary conditions into the WRF model for initialization; Simulation operation: Start the WRF model to perform simulation operation and calculate the meteorological field of the wind farm and its surrounding areas; Output results: Output simulation results every hour, including wind speed, temperature and air pressure; d. Result analysis Data processing: Interpolate and smooth the simulation results to obtain high-precision meteorological data for each point in the wind farm, including: local temperature T at location j in the wind farm at time t local (j,t); local air pressure P at position j in the wind farm at time t local (j,t), where t represents the time variable.

4. The power system dispatching method considering the risk of wind farm output uncertainty according to claim 3 is characterized by: The formula used to generate the local wind speed is as follows: v local (j,t)=α*v global (t)+β*T local (j,t)+γ*P local (j,t)+C1 Among them, v local (j,t) is the local wind speed at position j in the wind farm at time t, v global is the regional wind speed at time t, the regional wind speed is the wind speed data in the high-resolution satellite data, α is its preset proportional coefficient, T local (j, t) is the local temperature of position j in the wind farm at time t, β is its preset proportionality coefficient, P local (j, t) is the local air pressure at position j in the wind farm at time t, γ is its preset proportional coefficient, and α>β>γ>0, where t represents the time variable and C1 is the constant correction index.

5. The power system dispatching method considering the risk of wind farm output uncertainty according to claim 1 is characterized in that: The process of preprocessing the collected data specifically includes: Remove data outliers: Calculate the standard deviation of the data and set 3 times the standard deviation as the threshold. Data points outside this range are considered outliers. The formula used to calculate the standard deviation of the data is: Where σ is the standard deviation, N is the total number of data points, and x i is the i-th data point, μ is the mean of the data; The formula for calculating the mean of the data is: Where μ is the mean of the data, N is the total number of data points, and x i is the i-th data point; Data interpolation: For time series meteorological data, linear interpolation is used to fill missing values: assuming that the data changes linearly between two known points, the missing values ​​are estimated by calculating the linear relationship between the two points.

6. The power system dispatching method considering the risk of wind farm output uncertainty according to claim 5 is characterized in that: The process of constructing the wind farm output prediction model specifically includes: Based on the long short-term memory network algorithm, a wind farm output prediction model is constructed. The retrieved historical time series high-precision meteorological data is used as the training set, and the corresponding wind turbine output data at the next moment is used as the label. The model is trained to obtain a trained wind farm output prediction model. The process of generating the predicted output of the wind farm at the next moment in the future specifically includes: The high-precision meteorological data collected in real time at the j position in the wind farm at time t0 is input into the wind farm output prediction model, and the predicted output W of the j position in the wind farm at time t0+1 is obtained as output. j (t0+1), that is: W j (t0+1)=LSTM(v local (j,t0),T local (j,t0),P local (j,t0)) Among them, t0 represents the current time, W j (t0+1) is the predicted output of the jth wind turbine generator at time t0+1, v local (j,t0) is the local wind speed at position j in the wind farm at time t0, T local (j, t0) is the local temperature at position j in the wind farm at time t0; P local (j, t0) is the local air pressure at position j in the wind farm at time t0; Therefore, the total predicted output of the wind farm at time t0+1 is: Among them, t0 represents the current time, P(t0+1) is the total predicted output of the wind farm at time t0+1, and M is the number of wind turbines in the wind farm.

7. The power system dispatching method considering the risk of wind farm output uncertainty according to claim 6 is characterized by: The process of dynamically adjusting the load distribution and spare capacity of the power grid according to the real-time predicted wind farm output specifically includes: R(t0+1)=max(0,D(t0)-P(t0+1))+S(t0) Among them, t0 represents the current time, R(t0+1) is the backup capacity required by the wind farm at t0+1, D(t) is the grid load requirement of the wind farm at t0, P(t0+1) is the predicted output of the wind farm at t0+1, and S(t) is the system reserve capacity at t0.

8. The method for dispatching a power system considering the risk of wind farm output uncertainty according to claim 1, characterized in that: The process of continuously optimizing the prediction model and scheduling strategy according to actual conditions specifically includes: Collect the error data between wind farm output and predicted value in real time, calculate the prediction error, regularly update and retrain the LSTM model based on the error data to improve the prediction accuracy, retrain the model with the latest meteorological data and output data, analyze the adjusted power grid operation data, evaluate the effect of the dispatch strategy, and optimize the dispatch strategy according to the actual operation situation and prediction error; The calculation of the forecast error is based on the following formula: E(t0)=P actual (t0)-P predicted (t0) Among them, t0 represents the current time, E(t0) is the prediction error at time t0, and P actual (t0) is the actual wind power output at time t0, P predicted (t0) is the predicted wind power output at time t0.

9. A power system dispatching system considering the uncertainty risk of wind farm output, characterized by: The power dispatching system considering the risk of uncertainty in wind farm output is used to execute the power system dispatching method considering the risk of uncertainty in wind farm output as described in any one of claims 1 to 8, comprising: The data acquisition module is used to set up meteorological sensors at the location of each wind turbine generator set in the wind farm to collect meteorological sensor data in real time, including wind speed, temperature and air pressure; collect high-resolution satellite data of the wind farm area in real time, including wind speed, temperature and air pressure; A local simulation module is used to perform local simulation using a numerical weather prediction model, input meteorological sensor data and high-resolution satellite data into the numerical weather prediction model, and output high-precision meteorological data at the location of each wind turbine in the wind farm. The high-precision meteorological data includes local temperature, local wind speed and local air pressure; The model building module is used to retrieve the historical time series high-precision meteorological data and wind turbine output data of each wind turbine generator set, pre-process the data, and build a wind farm output prediction model based on the long short-term memory network algorithm. The historical time series high-precision meteorological data is used as a training set, and the wind turbine output data is used as a label to train the model to obtain a trained wind farm output prediction model; The system dispatch module is used to input the current high-precision meteorological data into the wind farm output prediction model, generate the local predicted output of the wind farm at the next moment in the future, obtain the total predicted output of the wind farm by summing up, and dynamically adjust the load distribution and spare capacity of the power grid according to the total predicted output of the wind farm, and continuously optimize the output prediction model and dispatch strategy according to actual conditions.

Images