A real-time regulation method and device for a high-proportion renewable energy power system

Abstract

The application discloses a real-time regulation method and device for a high-proportion renewable energy power system. The application obtains predicted output data and real-time output data of renewable energy in the power system according to a preset clearing cycle; uses the predicted output data as training data to train a pre-established output power prediction model to obtain a target output power prediction model; inputs the real-time output data as calibration data into the target output power prediction model to obtain output power corresponding to the calibration data; obtains a prediction error of the target output power prediction model according to the calibration data and the output power corresponding to the calibration data; determines a predicted output interval of the renewable energy according to the prediction error of the target output power prediction model; and performs market clearing based on a pre-established unit generation plan to obtain unit output power, and regulates the unit output power according to the predicted output interval, so that the predicted output interval of the renewable energy can be considered to perform real-time regulation on the power system.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, and in particular to a real-time control method and apparatus for a high-proportion renewable energy power system. Background Technology

[0002] A high proportion of renewable energy has become a prominent feature of the future development of power systems. Wind power and solar power are currently the most mature renewable energy generation technologies, but both types of power sources have strong volatility and randomness. In future power systems with a high proportion of renewable energy integrated into the grid, power source fluctuations may even exceed load fluctuations, becoming the main source of uncertainty in the power system. How to comprehensively consider this dual uncertainty of power source and load and carry out real-time regulation of the power system has become a core issue in power system planning and operation. Summary of the Invention

[0003] To overcome the shortcomings of the prior art, the present invention provides a real-time control method and apparatus for a high-proportion renewable energy power system, which can take into account the predicted output range of renewable energy to control the power system in real time, which is conducive to fully absorbing renewable energy and effectively ensuring the safe and stable operation of the power system.

[0004] To address the aforementioned technical problems, in a first aspect, an embodiment of the present invention provides a real-time control method for a high-proportion renewable energy power system, comprising: Based on the preset clearing cycle, obtain the predicted output data and real-time output data of renewable energy in the power system; The predicted power output data is used as training data to train the pre-established power output prediction model, thereby obtaining the target power output prediction model. The real-time output data is used as calibration data and input into the target output power prediction model to obtain the output power corresponding to the calibration data. Based on the calibration data and the corresponding power output, the prediction error of the target power output prediction model is obtained. Based on the prediction error of the target output power prediction model, the predicted output range of the renewable energy is determined; Market clearing is performed based on a pre-established unit power generation plan to obtain the unit output power, and the unit output power is adjusted according to the predicted output range.

[0005] Further, the step of obtaining the prediction error of the target output power prediction model based on the calibration data and the output power corresponding to the calibration data specifically involves: Calculate the absolute error between the calibration data and the output power corresponding to the calibration data to obtain the prediction error of the target output power prediction model.

[0006] Furthermore, determining the predicted output range of the renewable energy source based on the prediction error of the target output power prediction model specifically involves: The quantiles are extracted from the prediction error of the target output power prediction model as the prediction error estimation range, and the prediction output interval is determined based on the predicted output data and the prediction error range.

[0007] Furthermore, the process of market clearing based on a pre-determined unit power generation plan to obtain the unit output power is specifically as follows: Under the condition of satisfying the power system security constraints, the generating unit power generation plan is formulated with the goal of minimizing the power system power purchase cost; Based on the unit's power generation plan, the unit's output power is obtained by maximizing market net revenue clearing according to the predicted output data.

[0008] Furthermore, the step of adjusting the unit's output power according to the predicted output range specifically involves: The power output of the generator unit is replaced with the maximum and minimum power output data of the predicted power output range to determine power flow exceedance. When it is determined that a line or section exceeds the limit, based on the unit's power generation plan, a new market net profit maximization clearing is performed according to the maximum output data or the minimum output data to obtain the new unit output power.

[0009] Secondly, an embodiment of the present invention provides a real-time control device for a high-proportion renewable energy power system, comprising: The power output data acquisition module is used to acquire the predicted power output data and real-time power output data of renewable energy in the power system according to the preset clearing cycle. The prediction model training module is used to train a pre-established power output prediction model using the predicted power output data as training data to obtain the target power output prediction model. The prediction model calibration module is used to input the real-time power output data as calibration data into the target power output prediction model to obtain the power output corresponding to the calibration data. The model error prediction module is used to obtain the prediction error of the target output power prediction model based on the calibration data and the output power corresponding to the calibration data. The power output range prediction module is used to determine the predicted power output range of the renewable energy source based on the prediction error of the target power output prediction model. The power system control module is used to clear the market based on a pre-defined unit power generation plan, obtain the unit output power, and control the unit output power according to the predicted output range.

[0010] Furthermore, the model error prediction module is specifically used to calculate the absolute error between the calibration data and the output power corresponding to the calibration data, so as to obtain the prediction error of the target output power prediction model.

[0011] Furthermore, the power output interval prediction module is specifically used to extract quantiles from the prediction error of the target power output prediction model as the prediction error estimation range, and to determine the predicted power output interval based on the predicted power output data and the prediction error range.

[0012] Furthermore, the power system control module includes: The planning unit is used to formulate the generating unit's power generation plan under the condition of meeting the power system security constraints and with the goal of minimizing the power system's electricity purchase cost; The market clearing unit is used to perform market clearing to maximize net market revenue based on the unit's power generation plan and the predicted output data, thereby obtaining the unit's output power.

[0013] Furthermore, the power system control module includes: The power flow limit violation determination unit is used to replace the unit's output power with the maximum and minimum output data of the predicted output range to determine the power flow limit violation. The power system control unit is used to, when it is determined that a line or section has exceeded its limit, re-perform market net revenue maximization clearing based on the unit's power generation plan and the maximum or minimum output data to obtain a new unit output power.

[0014] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: By acquiring predicted and real-time power output data of renewable energy sources within the power system according to a preset clearing cycle, the predicted power output data is used as training data to train a pre-established power output prediction model, resulting in a target power output prediction model. Real-time power output data is input as calibration data into the target power output prediction model to obtain the power output corresponding to the calibration data. The prediction error of the target power output prediction model is obtained based on the calibration data and the corresponding power output. The predicted power output range of renewable energy is determined based on the prediction error of the target power output prediction model. Market clearing is performed based on a pre-defined unit power generation plan to obtain the unit power output. The unit power output is then adjusted according to the predicted power output range. This approach allows for real-time control of the power system, taking into account the predicted power output range of renewable energy, which is beneficial for fully absorbing renewable energy and effectively ensuring the safe and stable operation of the power system. Attached Figure Description

[0015] Figure 1This is a flowchart illustrating a real-time control method for a high-proportion renewable energy power system according to the first embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a real-time control device for a high-proportion renewable energy power system according to the second embodiment of the present invention. Detailed Implementation

[0016] The technical solutions of this invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0017] It should be noted that the step numbers in this document are only for the convenience of explaining the specific embodiments and are not intended to limit the order in which the steps are executed. The method provided in this embodiment can be executed by relevant terminal devices, and the following description uses a processor as the execution subject.

[0018] like Figure 1 As shown, the first embodiment provides a real-time control method for a high-proportion renewable energy power system, including steps S1 to S6: S1. Based on the preset clearing cycle, obtain the predicted output data and real-time output data of renewable energy in the power system; S2. Use the predicted power output data as training data to train the pre-established power output prediction model and obtain the target power output prediction model. S3. Input the real-time output data as calibration data into the target output power prediction model to obtain the output power corresponding to the calibration data; S4. Based on the calibration data and the output power corresponding to the calibration data, obtain the prediction error of the target output power prediction model; S5. Determine the predicted output range of renewable energy based on the prediction error of the target output power prediction model; S6. Based on the pre-established unit power generation plan, market clearing is carried out to obtain the unit output power, and the unit output power is adjusted according to the predicted output range.

[0019] As an example, in step S1, considering that renewable energy has time-series characteristics, according to a preset clearing cycle, such as 15 minutes, the predicted output data of renewable energy in the power system is directly retrieved from the wind, water and solar new energy system of the power company, and the real-time output data of renewable energy is collected and extracted.

[0020] Assuming that the predicted output data retrieved from the power company's wind, solar and hydropower new energy systems is reliable, this embodiment will not repeatedly consider the impact of factors such as climate, wind, and temperature on the real-time output of renewable energy when determining the predicted output range of renewable energy.

[0021] In step S2, a power output prediction model is pre-established, and the retrieved predicted power output data is used as training data to train and fit the power output prediction model, thereby obtaining the target power output prediction model. The random forest regression algorithm can be used for model training.

[0022] In step S3, the real-time output data is used as calibration data and input into the target output power prediction model to predict the power output and obtain the output power corresponding to the calibration data.

[0023] In step S4, error analysis is performed based on the calibration data and the output power corresponding to the calibration data to obtain the prediction error of the target output power prediction model, so as to determine the quality of the target output power prediction model.

[0024] In step S5, the predicted output range of renewable energy is determined based on the prediction error of the target output power prediction model and the retrieved predicted output data.

[0025] Understandably, due to the time-series characteristics of renewable energy, the Python library MAPIE, used to obtain interval predictions, is employed to determine the predicted output range of renewable energy. By introducing a time-series-based output range prediction method, the predicted output data of renewable energy is segmented according to the time series, with each segment divided into 15-minute intervals. This allows the output power prediction model to learn from the errors generated by the training and calibration data, fitting a target output power prediction model. The prediction error generated by the target output power prediction model on the calibration data is analyzed, and an upper and lower limit for predicted output based on variable time intervals is established.

[0026] In step S6, for all generator units in the power system, a generator power generation plan is pre-formulated based on the actual power generation demand. Market clearing is carried out based on the generator power generation plan to obtain the generator output power. The generator output power is then adjusted according to the predicted output range to achieve real-time control of the power system.

[0027] This embodiment can take into account the predicted output range of renewable energy and make real-time control of the power system, which is conducive to fully absorbing renewable energy and effectively ensuring the safe and stable operation of the power system.

[0028] In a preferred embodiment, the step of obtaining the prediction error of the target output power prediction model based on the calibration data and the output power corresponding to the calibration data specifically involves: calculating the absolute error of the calibration data and the output power corresponding to the calibration data to obtain the prediction error of the target output power prediction model.

[0029] As an example, error analysis is performed by calculating the absolute error of the calibration data and the output power corresponding to the calibration data to obtain the prediction error of the target output power prediction model, so as to judge the quality of the target output power prediction model.

[0030] Understandably, this embodiment divides the data into two parts: one part is used as training data to train the model, and the other part is used as calibration data to test the model's error. Because the two parts of data are different, the estimated generalization error is closer to the actual model performance. However, in practical applications, only a limited amount of data is often available, requiring data reuse and multiple data splits to ultimately obtain better model analysis results.

[0031] This embodiment employs a day-ahead chain time series cross-validation method. The basic idea is to perform multiple training / test splits on the same dataset and then calculate the average error across these splits. Using this method, daily data can be used as the test set, and all previous data can be allocated to the training set. For example, assuming the dataset has N days, N-2 different training and test splits will be randomly generated based on the time series, ensuring that at least one day's training and calibration data is available. This method generates N-2 different training / test splits, each of which will yield an independent root mean square error (RMSE). The RMSE across each split set is averaged to calculate a robust estimate of the model error.

[0032] This embodiment performs error analysis by calculating the absolute error of the calibration data and the output power corresponding to the calibration data, which can accurately determine the quality of the target output power prediction model.

[0033] In a preferred embodiment, determining the predicted output range of renewable energy based on the prediction error of the target output power prediction model specifically involves: extracting quantiles from the prediction error of the target output power prediction model as the prediction error estimation range, and determining the predicted output range based on the predicted output data and the prediction error range.

[0034] As an example, in absolute error analysis, to set the tolerance for the prediction error of the target power output prediction model, an α quantile is introduced as the acceptable prediction error range. The 1-α quantile is extracted from the prediction error of the target power output prediction model, i.e., the absolute error distribution, as the prediction error estimation range. Here, the α quantile is a pre-set value, ranging from 1% to 10%, used to improve prediction accuracy when the estimated deviation of the predicted power output range is large under extreme weather conditions. For example, normally, the α quantile can be set to 5%, meaning a 95% accuracy rate of prediction is acceptable. When the real-time prediction deviation is too large, it indicates that a 5% tolerance is insufficient for the model fitting effect, and the α quantile can be adjusted to 10%, meaning a 90% accuracy rate of prediction is acceptable, thus improving the model fitting effect. Based on the prediction error range, errors can be added on both sides of the predicted power output data point to determine the predicted power output range.

[0035] This embodiment determines the predicted power output range by extracting quantiles from the prediction error of the target power output prediction model as the prediction error estimation range, which can better fit the target power output prediction model and accurately estimate the predicted power output range.

[0036] In a preferred embodiment, the step of clearing the market based on a pre-defined unit power generation plan to obtain the unit output power specifically involves: under the condition of meeting power system security constraints, formulating a unit power generation plan with the goal of minimizing the power system's electricity purchase cost; and based on the unit power generation plan, clearing the market to maximize net revenue according to the predicted output data to obtain the unit output power.

[0037] As an example, an objective function is constructed with the goal of minimizing the power system's electricity purchase cost. Constraints are then constructed based on power system security constraints. Combining the objective function and constraints, a generating unit production plan is generated. These power system security constraints include systemic constraints related to all generating units, such as power system power balance constraints and power system reserve constraints, as well as individual unit constraints, such as upper and lower limits for unit output power, unit output power ramp-up rate constraints, and minimum operating and downtime constraints.

[0038] Based on the unit's power generation plan, the market net revenue maximization clearing is performed according to the predicted output data to obtain the unit's output power. The market net revenue maximization is shown in equation (1): (1); In equation (1), n ​​is the total number of nodes in the power system; Let i be the net market revenue of the i-th node. , , Let be the electricity revenue and generation cost of the i-th node, respectively. , These represent the load demand and power generation of the i-th node, respectively.

[0039] This embodiment maximizes market economic benefits by clearing out the power generation plan to maximize net market revenue. It can ensure the maximization of market economic benefits during the real-time regulation of the power system in the predicted output range of renewable energy.

[0040] In a preferred embodiment, the step of adjusting the unit output power according to the predicted output range specifically involves: replacing the unit output power with the maximum and minimum output data of the predicted output range to determine the power flow limit exceedance; when it is determined that a line or section exceeds the limit, based on the unit power generation plan, the market net profit maximization clearing is performed again according to the maximum or minimum output data to obtain a new unit output power.

[0041] As an example, the unit output power is replaced with the maximum and minimum output data of the predicted output range, respectively, and then substituted into the power system power flow model to calculate the power flow distribution.

[0042] Power flow calculation is the process of solving for unknown quantities from known quantities in a power system using a program. The known quantities are various parameters in the network, including the resistance and reactance of each branch, the susceptance to ground, and the active or reactive power injected into some nodes. The unknown quantities include the active and reactive power injected into each node, the magnitude and phase of the voltage at each node, etc., as shown in equation (2): (2); In equation (2), n is the total number of nodes in the power system; P i Q i Let U be the active power and reactive power of the i-th node, respectively; i U j Let G be the voltages of the i-th node and the j-th node, respectively; ij B is the reactance from node j to node i. ij Let δ be the susceptance from the j-th node to the i-th node. ij It is represented as the phase difference between the j-th node and the i-th node.

[0043] The i-th node in the power system has four variables, namely active power P. i Reactive power Q i Node voltage U i Phase δ i The above equations can be written separately for the net active and reactive power injected into the network by the generating units and loads of the entire power system.

[0044] By comparing the predicted output data of renewable energy and the impact of the predicted output range on power flow distribution, and based on the calculated active and reactive power power flow results of each node after substituting the maximum and minimum output data of the predicted output range, a power flow exceeding limit judgment is made, that is, judging the following two conditions: 1. Under the two conditions of substituting the maximum and minimum output data of the predicted output range, is there any active power flow exceeding the limit of the line, that is, does the active power exceed the maximum allowable transmission power of the line? 2. Whether the power flow control section exceeds the limit, that is, whether the calculated active power section exceeds the maximum allowable transmission power of the transmission section.

[0045] If either of the above two conditions is met, it means that once the real-time output of renewable energy generates electricity according to the maximum or minimum output data of the predicted output range, a power flow over-limit will occur, causing the power system power flow to break through the power system safety constraints, affecting the safe and stable operation of the power system, and requiring a new clearing of the unit generation plan.

[0046] When either of the above two conditions is met, the predicted output data retrieved from the power company's wind, solar and hydropower new energy system will be replaced with the maximum or minimum output data within the predicted output range. Based on the unit's power generation plan, the market net profit maximization clearing will be performed again according to the maximum or minimum output data to obtain the new unit output power.

[0047] For example, if the maximum output data of the predicted output range makes condition 1 true, then the predicted output data in the unit's power generation plan is replaced with the maximum output data of the predicted output range, while other data remain unchanged, and the market net profit maximization clearing is performed again.

[0048] If it is found that the generator power generation plan after the re-clearing still causes condition 1 or condition 2 to be met, it means that the conventional generator adjustment methods have been exhausted and the system algorithm no longer has the ability to remove condition 1 or condition 2. At this time, the system will still clear according to the current cleared generator power generation plan, and the system will issue an alarm indicating that this clearing will continue to cause condition 1 or condition 2 to be met, and list the affected power flow lines or sections.

[0049] This embodiment effectively ensures the safe and stable operation of the power system by performing power flow over-limit judgment during the process of adjusting the unit output power according to the predicted output range.

[0050] Based on the same inventive concept as the first embodiment, the second embodiment provides as follows: Figure 2The real-time control device for a high-proportion renewable energy power system, as shown, includes: an output data acquisition module 21, used to acquire predicted output data and real-time output data of renewable energy in the power system according to a preset clearing cycle; a prediction model training module 22, used to train a pre-established output power prediction model using the predicted output data as training data to obtain a target output power prediction model; a prediction model calibration module 23, used to input real-time output data as calibration data into the target output power prediction model to obtain the output power corresponding to the calibration data; a model error prediction module 24, used to obtain the prediction error of the target output power prediction model based on the calibration data and the output power corresponding to the calibration data; an output range prediction module 25, used to determine the predicted output range of renewable energy based on the prediction error of the target output power prediction model; and a power system control module 26, used to perform market clearing based on a pre-established unit power generation plan to obtain the unit output power, and to control the unit output power according to the predicted output range.

[0051] In a preferred embodiment, the model error prediction module 24 is specifically used to calculate the absolute error of the calibration data and the output power corresponding to the calibration data, so as to obtain the prediction error of the target output power prediction model.

[0052] In a preferred embodiment, the power output range prediction module 25 is specifically used to extract quantiles from the prediction error of the target power output prediction model as the prediction error estimation range, and determine the predicted power output range based on the predicted power output data and the prediction error range.

[0053] In a preferred embodiment, the power system control module 26 includes: a planning unit, used to formulate a generating unit plan with the goal of minimizing the power system's electricity purchase cost, under the condition of meeting power system security constraints; and a market clearing unit, used to perform market clearing to maximize net market revenue based on the generating unit plan and predicted output data, thereby obtaining the generating unit's output power.

[0054] In a preferred embodiment, the power system control module 26 includes: a power flow limit violation judgment unit, used to replace the unit output power with the maximum output data and minimum output data of the predicted output range to perform power flow limit violation judgment; and a power system control unit, used to, when it is determined that a line or section has exceeded the limit, re-perform market net profit maximization clearing based on the unit power generation plan and according to the maximum output data or minimum output data to obtain a new unit output power.

[0055] In summary, implementing the embodiments of the present invention has the following beneficial effects: By acquiring predicted and real-time power output data of renewable energy sources within the power system according to a preset clearing cycle, the predicted power output data is used as training data to train a pre-established power output prediction model, resulting in a target power output prediction model. Real-time power output data is input as calibration data into the target power output prediction model to obtain the power output corresponding to the calibration data. The prediction error of the target power output prediction model is obtained based on the calibration data and the corresponding power output. The predicted power output range of renewable energy is determined based on the prediction error of the target power output prediction model. Market clearing is performed based on a pre-defined unit power generation plan to obtain the unit power output. The unit power output is then adjusted according to the predicted power output range. This approach allows for real-time control of the power system, taking into account the predicted power output range of renewable energy, which is beneficial for fully absorbing renewable energy and effectively ensuring the safe and stable operation of the power system.

[0056] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

[0057] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

Claims

1. A real-time control method for a high-proportion renewable energy power system, characterized in that, include: Based on the preset clearing cycle, obtain the predicted output data and real-time output data of renewable energy in the power system; The predicted power output data is used as training data to train the pre-established power output prediction model, thereby obtaining the target power output prediction model. The real-time output data is used as calibration data and input into the target output power prediction model to obtain the output power corresponding to the calibration data. Based on the calibration data and the corresponding power output, the prediction error of the target power output prediction model is obtained. Based on the prediction error of the target output power prediction model, the predicted output range of the renewable energy is determined; Market clearing is performed based on a pre-established unit power generation plan to obtain the unit output power, and the unit output power is adjusted according to the predicted output range. The step of determining the predicted output range of the renewable energy source based on the prediction error of the target output power prediction model includes: The Python library MAPIE, used for interval prediction, is employed to determine the predicted output range of renewable energy. By introducing a time-series-based output range prediction method, the predicted output data of renewable energy is segmented according to the time series, with each segment divided into 15-minute intervals. This allows the output power prediction model to learn from the errors generated by the training and calibration data, and fit to obtain the target output power prediction model. The prediction error of the target output power prediction model on the calibration data is analyzed, and upper and lower limits of the predicted output based on variable time intervals are established. The specific steps of adjusting the unit's output power according to the predicted output range are as follows: The power output of the generator unit is replaced with the maximum and minimum power output data of the predicted power output range to determine power flow exceedance. When it is determined that a line or section exceeds the limit, based on the unit's power generation plan, a new market net profit maximization clearing is performed according to the maximum output data or the minimum output data to obtain the new unit output power.

2. The real-time control method for a high-proportion renewable energy power system as described in claim 1, characterized in that, The step of obtaining the prediction error of the target output power prediction model based on the calibration data and the corresponding output power is as follows: Calculate the absolute error between the calibration data and the output power corresponding to the calibration data to obtain the prediction error of the target output power prediction model.

3. The real-time control method for a high-proportion renewable energy power system as described in claim 1, characterized in that, The step of determining the predicted output range of the renewable energy source based on the prediction error of the target output power prediction model specifically involves: The quantiles are extracted from the prediction error of the target output power prediction model as the prediction error estimation range, and the prediction output interval is determined based on the predicted output data and the prediction error range.

4. The real-time control method for a high-proportion renewable energy power system as described in claim 1, characterized in that, The process of market clearing based on a pre-determined unit power generation plan to obtain the unit output power is as follows: Under the condition of satisfying the power system security constraints, the generating unit power generation plan is formulated with the goal of minimizing the power system power purchase cost; Based on the unit's power generation plan, the unit's output power is obtained by maximizing market net revenue clearing according to the predicted output data.

5. A real-time control device for a high-proportion renewable energy power system, characterized in that, include: The power output data acquisition module is used to acquire the predicted power output data and real-time power output data of renewable energy in the power system according to the preset clearing cycle. The prediction model training module is used to train a pre-established power output prediction model using the predicted power output data as training data to obtain the target power output prediction model. The prediction model calibration module is used to input the real-time power output data as calibration data into the target power output prediction model to obtain the power output corresponding to the calibration data. The model error prediction module is used to obtain the prediction error of the target output power prediction model based on the calibration data and the output power corresponding to the calibration data. The power output range prediction module is used to determine the predicted power output range of the renewable energy source based on the prediction error of the target power output prediction model. The power system control module is used to clear the market based on a pre-defined unit power generation plan, obtain the unit output power, and control the unit output power according to the predicted output range. The step of determining the predicted output range of the renewable energy source based on the prediction error of the target output power prediction model includes: The Python library MAPIE, used for interval prediction, is employed to determine the predicted output range of renewable energy. By introducing a time-series-based output range prediction method, the predicted output data of renewable energy is segmented according to the time series, with each segment divided into 15-minute intervals. This allows the output power prediction model to learn from the errors generated by the training and calibration data, and fit to obtain the target output power prediction model. The prediction error of the target output power prediction model on the calibration data is analyzed, and upper and lower limits of the predicted output based on variable time intervals are established. The power system control module includes: The power flow limit violation determination unit is used to replace the unit's output power with the maximum and minimum output data of the predicted output range to determine the power flow limit violation. The power system control unit is used to, when it is determined that a line or section has exceeded its limit, re-perform market net revenue maximization clearing based on the unit's power generation plan and the maximum or minimum output data to obtain a new unit output power.

6. The real-time control device for a high-proportion renewable energy power system as described in claim 5, characterized in that, The model error prediction module is specifically used to calculate the absolute error of the calibration data and the output power corresponding to the calibration data, so as to obtain the prediction error of the target output power prediction model.

7. The real-time control device for a high-proportion renewable energy power system as described in claim 5, characterized in that, The output range prediction module is specifically used to extract quantiles from the prediction error of the target output power prediction model as the prediction error estimation range, and to determine the predicted output range based on the predicted output data and the prediction error range.

8. The real-time control device for a high-proportion renewable energy power system as described in claim 5, characterized in that, The power system control module includes: The planning unit is used to formulate the generating unit's power generation plan under the condition of meeting the power system security constraints and with the goal of minimizing the power system's electricity purchase cost; The market clearing unit is used to perform market clearing to maximize net market revenue based on the unit's power generation plan and the predicted output data, thereby obtaining the unit's output power.

Images