Coordinated dispatching optimization method and system suitable for large power transmission and distribution network
Patent Information
- Application Number
- CN202610950873.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-29
- Publication Date
- 2026-09-15
Smart Images
Figure CN122763633A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of power system dispatching technology, specifically to a collaborative dispatching optimization method and system applicable to large-scale transmission and distribution networks. Background Technology
[0002] In modern large-scale power transmission and distribution networks, the high proportion of renewable energy sources such as wind and solar power is crucial for achieving the goal of green and low-carbon transformation. However, wind and solar resources are highly random, intermittent, and volatile (e.g., sudden changes in wind speed and a sharp drop in solar radiation due to cloud cover), posing significant challenges to the safe and stable operation of the power grid. Currently, the Particle Swarm Optimization (PSO) algorithm is widely used in power system optimization and scheduling due to its fast convergence speed and simple structure.
[0003] However, standard PSO (Particle Swarm Optimization) is typically used to handle deterministic mathematical models and struggles to directly address the probability distribution and random fluctuations in wind and solar power output. When faced with high-dimensional, nonlinear grid optimization problems, standard PSO is prone to premature convergence and cannot adapt to the dynamic changes in wind and solar resources. Specifically, the speed update mechanism of standard PSO lacks the ability to perceive the probabilistic characteristics of the solution space. When faced with drastic fluctuations in wind and solar power output, the particle swarm is prone to prematurely clustering at local extrema, causing the algorithm to lose its global search capability. Furthermore, due to the lack of an internal modeling mechanism for random variables, standard PSO is inadequate in handling probabilistic constraints such as "line power flow exceeding limits," making it difficult to find a precise balance between ensuring grid safety margins and maximizing renewable energy consumption. Summary of the Invention
[0004] In view of the above, it is necessary to provide a collaborative scheduling optimization method and system applicable to large-scale power transmission and distribution networks to solve the above problems.
[0005] The first aspect of this application provides a collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks, the method comprising: Based on real-time operating data of wind farms and photovoltaic power plants, neural networks are used to obtain wind power prediction power and photovoltaic prediction power respectively. Based on the predicted wind power and the predicted photovoltaic power, combined with the total installed capacity of the power station, and using historical zoning error statistics, a linear model of the standard deviation of wind power prediction error and a calculation model of the standard deviation of photovoltaic prediction error are established respectively. Using the obtained standard deviations of wind power prediction error and photovoltaic prediction error, as well as the complementary characteristics of wind and solar power output, a standard deviation of wind and solar power prediction output is constructed. Combined with the absolute change rate of predicted power and the coefficient of variation, a comprehensive fluctuation characteristic index is constructed. Using the comprehensive fluctuation characteristic index as a penalty factor, a multi-objective optimization function is constructed as a fitness function and solved using the particle swarm optimization algorithm to output the optimal scheduling scheme. The multi-objective optimization function is used to maximize the output of wind and solar power while using the comprehensive fluctuation characteristic index multiplied by the output of conventional units and multiplied by the penalty coefficient as a penalty term.
[0006] Preferably, the linear model for the standard deviation of the wind power prediction error is as follows: in: This represents the standard deviation of the wind power prediction error at time t; This represents the predicted wind power output at time t; Indicates the wind power output distribution coefficient; Indicates the installed capacity allocation coefficient; This indicates the total installed capacity of the wind farm.
[0007] Preferably, the specific process for obtaining the wind power output allocation coefficient and the installed capacity allocation coefficient is as follows: Collect historical data on predicted and actual wind power for the same period, and calculate the wind power prediction error at each time point. The range of the ratio between wind power prediction power and total installed wind power capacity is evenly divided into several intervals. The standard deviation of all wind power prediction errors in each interval is calculated to obtain data points with the median wind power prediction power in each interval as the independent variable and the standard deviation of wind power prediction error in the corresponding interval as the dependent variable. Using the median of the predicted wind power and the total installed wind power capacity as independent variables, and the standard deviation of the wind power prediction error as the dependent variable, a linear regression is performed to obtain the optimal wind power output allocation coefficient and installed capacity allocation coefficient.
[0008] Preferably, the calculation model for the standard deviation of the photovoltaic prediction error is obtained by multiplying the photovoltaic power output allocation coefficient at each time point with the photovoltaic predicted power, and the photovoltaic power output allocation coefficient is determined by the ratio between the photovoltaic prediction error and the photovoltaic predicted power in the same historical period.
[0009] Preferably, the standard deviation of the wind and solar power output forecast is obtained by fusing the standard deviation of the wind power forecast error and the standard deviation of the photovoltaic forecast error through the variance synthesis principle in statistics.
[0010] Preferably, the formula for the comprehensive volatility characteristic index is: in, Indicators representing comprehensive volatility characteristics; This represents the standard deviation of the predicted wind and solar power output at time t; Indicates the number of moments in the scheduling cycle; The predicted wind and solar power at time t is obtained by summing the predicted wind power and solar power at time t. This represents the coefficient of variation of the predicted wind and solar power during the scheduling period corresponding to time t; , , All of these represent weighting coefficients determined based on historical data.
[0011] Preferably, the specific formula for constructing the multi-objective optimization function is as follows: In the above formula, max() represents the maximization function; This represents the active power of the wind turbine at time t; This represents the photovoltaic output power at time t; Indicators representing comprehensive volatility characteristics; This represents the output of the conventional generating units at time t; This represents the risk cost discount factor; Indicates the unit price of wind power; Indicates the unit price of photovoltaic power generation; This represents the number of moments in the scheduling cycle.
[0012] Preferably, in the process of solving the multi-objective optimization function, the comprehensive fluctuation characteristic index is used to adaptively correct the physical constraints of the power grid. Specifically, the upper limit of the line transmission power is dynamically adjusted, the product of the preset safety margin coefficient and the comprehensive fluctuation characteristic index is calculated, the difference between the line rated power capacity and the product is compared with the minimum basic transmission power lower limit for maintaining power grid stability, and the maximum value is taken as the dynamic transmission limit.
[0013] Preferably, the specific process of using the particle swarm optimization algorithm to solve the problem and output the optimal scheduling scheme is as follows: In each iteration of fitness evaluation, the transmission power limit of the transmission line is dynamically lowered using the comprehensive fluctuation characteristic index, and particles that violate this dynamic limit are penalized. Under this physical constraint, non-dominated sorting technique is used to maintain the Pareto optimal solution stored in the external archive set, and a mutation operator that is dynamically adjusted with the number of iterations is introduced to randomly perturb the position vector of the particles. The global optimal guide particle is selected from the archive set using the crowding distance to update the velocity and position of the particle swarm until the optimal scheduling scheme is output.
[0014] Secondly, embodiments of this application also provide a collaborative scheduling optimization system suitable for large-scale power transmission and distribution networks, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.
[0015] The beneficial effects of the above scheme are as follows: By constructing a comprehensive fluctuation characteristic index that includes the standard deviation of wind and solar power output forecasts and introducing it as a penalty factor into a multi-objective optimization function, the safe and stable operation capability of large-scale power transmission and distribution networks under high-proportion renewable energy access is significantly improved. This method can adaptively adjust the conservatism of the scheduling strategy according to the magnitude of the prediction error, and uses an improved particle swarm optimization algorithm to solve for the Pareto optimal solution set. This not only ensures a balance between economy and safety in the scheduling scheme, but also effectively avoids the risk of equipment overload or tripping due to excessive prediction deviations.
[0016] Furthermore, this application establishes differentiated error modeling mechanisms for wind power and photovoltaic power, and quantifies the complementary characteristics of wind and solar power output by combining Pearson correlation coefficient, making uncertainty modeling more consistent with actual physical laws. By dynamically correcting the upper limit of line transmission power, adaptive adjustment of physical constraints is achieved. Compared with the traditional fixed safety domain method, this not only improves the safety margin of the power grid, but also fully taps the transmission potential of the transmission channel, effectively solving the problem in existing technologies that it is difficult to balance maximizing wind and solar power output with system operational safety. Attached Figure Description
[0017] Figure 1 A flowchart illustrating the steps of a collaborative scheduling optimization method for large-scale power transmission and distribution networks, provided as an embodiment of this application. Detailed Implementation
[0018] In the description of the embodiments in this application, the words "exemplary," "or," and "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary," "or," and "for example" is intended to present the relevant concepts in a specific manner.
[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this application's specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.
[0020] It should also be noted that the terms "first" and "second" in this application and its accompanying drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The methods disclosed in the embodiments of this application or the methods shown in the flowcharts include one or more steps for implementing the method. Without departing from the scope of protection of this application, the execution order of multiple steps can be interchanged, and some steps can also be deleted.
[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the collaborative dispatch optimization method and system applicable to large-scale power transmission and distribution networks provided in this application.
[0022] Please see Figure 1 The diagram illustrates a flowchart of a collaborative scheduling optimization method for large-scale power transmission and distribution networks according to an embodiment of this application. The method includes the following steps: The first step: Based on the real-time operating data of wind farms and photovoltaic power plants, a neural network is used to obtain the predicted power of wind power and the predicted power of photovoltaic power respectively.
[0023] Sensors are deployed at wind farms and photovoltaic power plants. For example, wind speed sensors are installed at the top of a wind measurement tower to acquire wind speed data; the real-time active power of the wind turbine is read through the main control PLC of the wind turbine. .
[0024] A thermopile solar intensity meter is installed on the photovoltaic array plane to measure global horizontal radiation (GHI) and tilted surface radiation (POA) to obtain irradiance data; multi-channel temperature and humidity transmitters are deployed to monitor the photovoltaic module backsheet temperature (directly affecting conversion efficiency) and ambient temperature; DC-side current and DC-side voltage are collected through a smart combiner box, and the AC-side photovoltaic output power is collected through an inverter. .
[0025] Since the collected raw data often contains noise and outliers, this application uses the 3σ criterion to remove gross errors: if a data point deviates from the mean by more than 3 times the standard deviation, it is judged as outlier data and removed. For short-term data loss (e.g., <5 minutes), linear interpolation is used to complete the data; for long-term data loss, it is marked as an invalid period and does not participate in subsequent optimization calculations. In the embodiments of this application, the sensors collect data at a frequency of 1Hz, which can be adjusted by the implementer according to the actual situation.
[0026] The second step is to establish a linear model for the standard deviation of wind power prediction error and a calculation model for the standard deviation of photovoltaic prediction error based on the predicted wind power and the predicted photovoltaic power, respectively, in conjunction with the total installed capacity of the power station, and by using historical zoning error statistics.
[0027] When the predicted wind power output is low, the actual output fluctuation range is usually small; however, as the predicted output increases, the actual power fluctuation range also expands, and this uncertainty exhibits an approximately linear growth trend with the predicted power. To accurately quantify this characteristic, this application establishes a linear model of the standard deviation of wind power output prediction error. This model, by introducing coefficient terms and constant terms related to the predicted power, effectively characterizes the linear growth of the prediction error standard deviation with the wind power output level, thus providing a more accurate probabilistic description for power system uncertainty management.
[0028] Taking time t as an example, this application obtains wind speed data and wind turbine active power data at time t and for a preset duration prior to it, and uses a long short-term memory network to obtain the wind power prediction at time t; the preset duration is 15 minutes.
[0029] Furthermore, the standard deviation of the prediction error for wind power output at time t. The following linear model can be used for estimation: in: This represents the standard deviation of the wind power prediction error at time t; This represents the predicted wind power output at time t; This represents the wind power output distribution coefficient, reflecting the characteristic that the larger the predicted power, the greater the absolute error. This represents the installed capacity allocation coefficient, which reflects a fundamental uncertainty related to the total capacity, even if the predicted output is very small, due to factors such as unit status. This represents the total installed capacity of the wind farm, specifically the sum of the rated power of all wind turbines. It is important to note that... , It is a non-negative real number.
[0030] It should be noted that, This represents the error component that is proportional to the predicted value. When the predicted wind power generation is large, the deviation range (standard deviation) between the actual value and the predicted value is usually larger due to the fluctuation of wind speed. This represents a basic error component. Even if the predicted power is zero (e.g., the wind speed is lower than the cut-in wind speed), the system still has a minimum uncertainty related to the scale of the wind farm because the wind turbine may start at any time or there may be measurement errors. Finally, the two are added together to form a comprehensive model that can reflect both the impact of the predicted scale and the basic uncertainty.
[0031] It should be further explained that the wind power output allocation coefficient and the installed capacity allocation coefficient need to be obtained by fitting historical data. Specifically, this involves collecting historical data for the same period (the same period refers to the set of historical data that falls in the same season as the current forecast date; for example, if it is currently summer, historical forecast and actual power data from the past few summers should be retrieved to ensure that the fitted samples have similar meteorological distribution characteristics) of the predicted wind power. and actual wind power Data, and calculate the wind power prediction error at each time point. ; forecast wind power The range of the ratio between the wind power forecast and the total installed wind power capacity is evenly divided into several intervals; in this embodiment, the number of intervals is set to 10. The wind power prediction error for each interval is calculated. Standard deviation To obtain the values of each interval The median is the first dimension, and the standard deviation of the wind power forecast error for the corresponding interval is the second dimension of the data points; finally, the median is used as the second dimension. Median and As the independent variable, Using the least squares method for linear regression as the dependent variable, the optimal wind power output allocation coefficient and installed capacity allocation coefficient can be obtained.
[0032] Acquire light intensity data, ambient temperature data, and photovoltaic output power data at time t and for a preset duration prior to that time. Use a trained convolutional neural network (CNN) to obtain the photovoltaic predicted power at time t.
[0033] Standard deviation of prediction error for photovoltaic power output The standard deviation of the calculation model and the prediction error of wind power output Similarly, its model formula is: in, This represents the standard deviation of the photovoltaic prediction error at time t; Represents the predicted photovoltaic power at time t; This represents the photovoltaic power output allocation coefficient, which reflects the proportional relationship between photovoltaic prediction error and predicted power.
[0034] It should be understood that the uncertainty in photovoltaic (PV) output primarily stems from drastic fluctuations in solar irradiance (such as cloud cover). Since PV output is essentially linearly related to solar irradiance, when predicted solar intensity is high and predicted power is large, the power fluctuations caused by disturbances such as cloud cover are also greater. Therefore, the standard deviation of the prediction error is directly proportional to the predicted power. Because PV power generation is zero at night, its fundamental uncertainty is far less than that of wind power; therefore, such uncertainties can be ignored in many models. Items.
[0035] The method for obtaining the photovoltaic power output allocation coefficient is similar to that for wind power output allocation coefficient, specifically: collecting historical photovoltaic power forecasts for the same period. and actual photovoltaic power Data; Calculate the photovoltaic prediction error at each time point. ;Predicted photovoltaic power The value range is divided into several intervals, and in this embodiment, the number of intervals is 10; for all photovoltaic prediction errors within each interval... Standard deviation To obtain the values of each interval The median is the first dimension, and the standard deviation of the photovoltaic prediction error in the corresponding interval is the data point of the second dimension; the photovoltaic power output allocation coefficient is obtained by linear regression.
[0036] The third step is to use the obtained standard deviations of wind power prediction error and photovoltaic prediction error, as well as the complementary characteristics of wind and solar power output, to construct the standard deviation of wind and solar power prediction output, and to combine the absolute change rate of predicted power and the coefficient of variation to construct a comprehensive fluctuation characteristic index.
[0037] Based on the obtained standard deviation of wind power prediction error and the standard deviation of photovoltaic prediction error By combining the complementary characteristics of wind and solar power output, a standard deviation of wind and solar power output forecast is constructed. Furthermore, considering the randomness of wind and solar energy, a comprehensive fluctuation characteristic index is constructed to quantify the total fluctuation level of the wind and solar combined power generation system at time t. This is a core indicator for measuring the impact of the system on the power grid and is used to dynamically adjust the renewable energy consumption strategy in the subsequent optimization process to prevent power grid fluctuations caused by over-prediction.
[0038] Let the standard deviation of the wind and solar power forecast output at time t be denoted as . The specific formula is as follows: in, This represents the standard deviation of the wind power prediction error at time t; This represents the standard deviation of the photovoltaic prediction error at time t; This represents the Pearson correlation coefficient of the wind power and photovoltaic prediction errors in the scheduling cycle corresponding to time t; in this embodiment, the duration of the scheduling cycle is 24 hours, and the scheduling cycle corresponding to time t is specifically 15 minutes past time t.
[0039] It should be understood that wind power and solar power often have "anti-peak shaving" or "complementary" characteristics (for example, strong winds and weak solar power at night, and strong solar power and weak winds during the day); when When (negative correlation), the cross term in the formula is negative, making It quantifies the smoothing effect brought about by wind-solar complementarity, that is, the fluctuation of the combined output is less than the sum of the fluctuations of the two individually.
[0040] Furthermore, substituting the standard deviation of the wind and solar power forecast into the following formula, we obtain the comprehensive fluctuation characteristic index. : in, This represents a comprehensive fluctuation characteristic index. The larger the value, the stronger the uncertainty of wind and solar power output, and the more reserve capacity should be reserved in the dispatch strategy; the smaller the value, the more stable the output, and the more fully it can be absorbed. This represents the standard deviation of the predicted wind and solar power output at time t; This represents the number of moments in the scheduling cycle. In this embodiment, the suggested duration of the scheduling cycle is 24 hours, the suggested time step is 15 minutes, and the value of T is 96. By averaging and aggregating the collected high-frequency data over 15 minutes, algorithm input data corresponding to each time step is generated. The predicted wind and solar power at time t is obtained by summing the predicted wind power and solar power at time t. This represents the coefficient of variation of the predicted wind and solar power at time t within the corresponding scheduling period. When the value is less than a preset threshold, a very small positive real number (in this embodiment, the value is taken as ) is added to the denominator of the CV. ); , , All of these represent weighting coefficients calibrated based on historical data, specifically based on the error distribution of the historical data, used to balance the impact of dispersion, rate of change, and relative fluctuation on the scheduling strategy.
[0041] It should be noted that before weighted summation, the standard deviation term is divided by the total installed capacity of wind and solar power, and the rate of change term is divided by the historical maximum rate of change, so that all sub-items become dimensionless data before being multiplied by the weighting coefficient.
[0042] The above formula, after dimensionlessly processing the absolute error risk, the degree of fluctuation over time, and the relative dispersion of the combined wind and solar power forecast, performs a weighted summation to quantify the overall uncertainty level of the system. Among these, the absolute deviation term... The standard deviation of the joint predicted power output is calculated by introducing the correlation coefficient between wind and solar power, and then divided by the total installed capacity of wind and solar power to achieve dimensionlessness, thus characterizing the absolute error risk after complementarity; the rate of change term : Calculate the average absolute rate of change of predicted power within the scheduling period, divide by the historical maximum rate of change to achieve dimensionlessness, characterizing the severity of the ramp / ramp dimension; relative fluctuation term Calculate the coefficient of variation of the predicted power (adding a very small positive number to the denominator of the mean to prevent division by zero), and use it directly as a dimensionless term to characterize the relative dispersion; multiply the above three dimensionless indices by the weighting coefficients α, β, and γ respectively, and then sum them to generate a comprehensive fluctuation characteristic index. .
[0043] The fourth step: Using the comprehensive fluctuation characteristic index as a penalty factor, construct a multi-objective optimization function as a fitness function and solve it using the particle swarm optimization algorithm to output the optimal scheduling scheme. The multi-objective optimization function is used to maximize wind and solar power output.
[0044] This application will incorporate comprehensive volatility characteristic indicators. As a penalty factor, in maximizing the output of the scenery ( At the same time, utilize Multiply by the output of the conventional unit and multiply by the penalty factor As a penalty term, the following multi-objective optimization function is constructed: In the above formula, max() represents the maximization function; This represents the active power of the wind turbine at time t; This represents the photovoltaic output power at time t; Indicators representing comprehensive volatility characteristics; This represents the output of the conventional generating units at time t; This represents the risk cost discount factor, giving the penalty item "the backup economic cost brought about by uncertainty" a clear physical meaning. In this embodiment, the value is taken as 1.5 times the average on-grid electricity price of wind and solar power. Indicates the unit price of wind power; Indicates the unit price of photovoltaic power generation; This represents the number of moments within the scheduling cycle. It should be noted that... and The physical dimension is currency / unit of electricity (i.e., the benchmark on-grid electricity price).
[0045] When used to represent a comprehensive volatility characteristic indicator At higher levels, the algorithm tends to increase the output of conventional units to ensure safety, sacrificing some green energy; when... When the energy level is low, the algorithm will minimize the output of conventional generating units to achieve "full consumption" of green energy.
[0046] This application employs a particle swarm optimization algorithm to solve the aforementioned optimization problem. The solution process begins with the initialization of the solution space and particle encoding. The algorithm first constructs a multi-dimensional search space containing all decision variables, where each particle represents a complete 24-hour scheduling scheme. The particle's position vector is encoded as time-series data, specifically covering the wind turbine output, energy storage system charging and discharging power, and the curtailment rate of wind and solar power generation at each moment. After random initialization of the population, each particle is assigned an initial velocity and begins iterative search within the solution space, aiming to find the optimal operating trajectory that balances economy and safety.
[0047] In each iteration, the core of the algorithm lies in dynamic constraint handling and fitness evaluation. The system first calculates a comprehensive fluctuation characteristic index based on the scheduling scheme represented by the current particle and combined with wind and solar forecast data. This indicator quantifies the prediction error risk faced by the current approach. Subsequently, the algorithm utilizes... Adaptive adjustment to the physical constraints of the power grid: when fluctuation indicators are large, the upper limit of line transmission power. It will be forcibly reduced, that is ,in, This indicates a dynamic transmission limit, used to represent the maximum allowable actual transmission power of a line; The rated power capacity of the line represents the maximum power limit that the transmission line or transformer is allowed to operate stably for a long period of time in terms of physical design; k represents the safety margin coefficient, which is set to 0.1 in this embodiment. It is used to control the degree of conservatism in the face of risks, meaning that when the comprehensive fluctuation characteristic index reaches the maximum value of 1, the system will actively reduce the rated capacity of the line by 10% as a safety buffer space to cope with sudden fluctuations in wind and solar power output; thus reserving the necessary safety margin for prediction errors. This represents the minimum baseline transmission power required to maintain grid stability. Based on this, the algorithm calculates a comprehensive fitness value that includes absorption benefits and conventional unit penalties. If the power flow distribution of a particle violates this dynamically contracting boundary, it will be subject to a severe penalty function correction to ensure that the search direction always converges towards the safe region.
[0048] Finally, the algorithm updates and outputs the solution set through a multi-objective optimization mechanism. Utilizing non-dominated sorting techniques, the algorithm maintains an external archive set to store Pareto optimal solutions and selects the globally optimal guiding particle from the archive set based on crowding distance. ), combined with the particle's own historical best ( The algorithm updates velocity and position using a variable frequency drive (VFD). To prevent the algorithm from getting trapped in local optima, a mutation operator that dynamically adjusts with the number of iterations is introduced to randomly perturb some particles. After a preset number of iterations, the algorithm terminates the search and outputs a Pareto optimal solution set that includes wind power, energy storage, and curtailment strategies, allowing decision-makers to select the final scheduling scheme based on actual operational preferences.
[0049] Based on the same inventive concept as the above methods, embodiments of this application also provide a collaborative scheduling optimization system suitable for large-scale power transmission and distribution networks, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described collaborative scheduling optimization methods suitable for large-scale power transmission and distribution networks.
[0050] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description; sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.
[0051] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from its essential characteristics. Therefore, the embodiments described above should be considered exemplary and non-limiting in all respects; modifications to the technical solutions described in the foregoing embodiments, or equivalent substitutions of some technical features, without causing the essence of the corresponding technical solutions to deviate from the scope of the technical solutions in the embodiments of this application, should all be included within the protection scope of this application.
Claims
1. A collaborative dispatch optimization method applicable to large-scale power transmission and distribution networks, characterized in that, The method includes the following steps: Based on real-time operating data of wind farms and photovoltaic power plants, neural networks are used to obtain wind power prediction power and photovoltaic prediction power respectively. Based on the predicted wind power and the predicted photovoltaic power, combined with the total installed capacity of the power station, and using historical zoning error statistics, a linear model of the standard deviation of wind power prediction error and a calculation model of the standard deviation of photovoltaic prediction error are established respectively. Using the obtained standard deviations of wind power prediction error and photovoltaic prediction error, as well as the complementary characteristics of wind and solar power output, a standard deviation of wind and solar power prediction output is constructed. Combined with the absolute change rate of predicted power and the coefficient of variation, a comprehensive fluctuation characteristic index is constructed. Using the comprehensive fluctuation characteristic index as a penalty factor, a multi-objective optimization function is constructed as a fitness function and solved using the particle swarm optimization algorithm to output the optimal scheduling scheme. The multi-objective optimization function is used to maximize wind and solar power output.
2. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The linear model for the standard deviation of wind power prediction error is as follows: in: This represents the standard deviation of the wind power prediction error at time t; This represents the predicted wind power output at time t; Indicates the wind power output distribution coefficient; Indicates the installed capacity allocation coefficient; This indicates the total installed capacity of the wind farm.
3. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 2, characterized in that, The specific process for obtaining the wind power output allocation coefficient and the installed capacity allocation coefficient is as follows: Collect historical data on predicted and actual wind power for the same period, and calculate the wind power prediction error at each time point. The range of the ratio between wind power prediction power and total installed wind power capacity is evenly divided into several intervals. The standard deviation of all wind power prediction errors in each interval is calculated to obtain data points with the median wind power prediction power in each interval as the independent variable and the standard deviation of wind power prediction error in the corresponding interval as the dependent variable. Using the median of the predicted wind power and the total installed wind power capacity as independent variables, and the standard deviation of the wind power prediction error as the dependent variable, a linear regression is performed to obtain the optimal wind power output allocation coefficient and installed capacity allocation coefficient.
4. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The calculation model for the standard deviation of the photovoltaic prediction error is obtained by multiplying the photovoltaic power distribution coefficient at each moment with the photovoltaic power prediction power. The photovoltaic power distribution coefficient is determined by the ratio between the photovoltaic prediction error and the photovoltaic power prediction power in the same historical period.
5. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The standard deviation of the predicted wind and solar power output is obtained by fusing the standard deviation of the wind power prediction error and the standard deviation of the photovoltaic prediction error using the variance synthesis principle in statistics.
6. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The formula for the comprehensive volatility characteristic index is: in, Indicators representing comprehensive volatility characteristics; This represents the standard deviation of the predicted wind and solar power output at time t; Indicates the number of moments in the scheduling cycle; The predicted wind and solar power at time t is obtained by summing the predicted wind power and solar power at time t. This represents the coefficient of variation of the predicted wind and solar power during the scheduling period corresponding to time t; , , All of these represent weighting coefficients determined based on historical data.
7. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The specific formula for constructing the multi-objective optimization function is as follows: In the above formula, max() represents the maximization function; This represents the active power of the wind turbine at time t; This represents the photovoltaic output power at time t; Indicators representing comprehensive volatility characteristics; This represents the output of the conventional generating units at time t; This represents the risk cost discount factor; Indicates the unit price of wind power; Indicates the unit price of photovoltaic power generation; This represents the number of moments in the scheduling cycle.
8. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 7, characterized in that, In the process of solving the multi-objective optimization function, the comprehensive fluctuation characteristic index is used to adaptively correct the physical constraints of the power grid. Specifically, the upper limit of the line transmission power is dynamically adjusted, the product of the preset safety margin coefficient and the comprehensive fluctuation characteristic index is calculated, the difference between the line rated power capacity and the product is compared with the minimum basic transmission power lower limit for maintaining power grid stability, and the maximum value is taken as the dynamic transmission limit.
9. The collaborative scheduling optimization method applicable to large-scale power transmission and distribution networks as described in claim 1, characterized in that, The specific process of using the particle swarm optimization algorithm to solve the problem and output the optimal scheduling scheme is as follows: In each iteration of fitness evaluation, the transmission power limit of the transmission line is dynamically lowered using the comprehensive fluctuation characteristic index, and particles that violate this dynamic limit are penalized. Under this physical constraint, non-dominated sorting technique is used to maintain the Pareto optimal solution stored in the external archive set, and a mutation operator that is dynamically adjusted with the number of iterations is introduced to randomly perturb the position vector of the particles. The global optimal guide particle is selected from the archive set using the crowding distance to update the velocity and position of the particle swarm until the optimal scheduling scheme is output.
10. A collaborative dispatch optimization system suitable for large-scale power transmission and distribution networks, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.