Output power value prediction method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides an output power value prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: inputting a first output power prediction value sequence and a current output power value into a pre-trained machine learning model for processing; and obtaining a first weight coefficient sequence corresponding to the first output power predicted value sequence and a second weight coefficient sequence corresponding to the current output power value. And on this basis, a second output power prediction value sequence of the new energy power supply equipment is determined, so that the prediction accuracy of the output power value of the new energy power supply equipment can be improved.

Description

Technical Field

[0001] This application relates to the field of new energy power supply technology, and in particular to a method, device, electronic device and storage medium for predicting output power value. Background Technology

[0002] When using renewable energy sources to power related load equipment (such as wind or solar power), the significant intermittency and volatility of renewable energy supply necessitate prediction of the output power of the renewable energy power supply equipment. However, due to the current low accuracy of predictions, problems such as low utilization rates of renewable energy sources may arise during subsequent scheduling.

[0003] Improving the accuracy of predicting the output power of new energy sources has become an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides an output power value prediction method, apparatus, electronic device, and storage medium, which can improve the prediction accuracy of output power values ​​of new energy power supply equipment.

[0005] In a first aspect, embodiments of this application provide an output power prediction method, comprising: acquiring a first output power prediction value sequence of a new energy power supply device within a first preset time period; acquiring the current output power value of the new energy power supply device; inputting the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value; determining a second output power prediction value sequence of the new energy power supply device based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence; wherein, the machine learning model employs a dynamic weighted fusion algorithm; and the training objective of the machine learning model is that the difference between the actual output power value and the predicted output power value of the new energy power supply device is less than a preset threshold.

[0006] Optionally, the method further includes: obtaining a third output power prediction value sequence of the new energy power supply equipment within a second preset time period; the step of inputting the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value includes: inputting the first output power prediction value sequence, the third output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence, a third weight coefficient sequence corresponding to the third output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value; Determining the second output power prediction sequence of the new energy power supply equipment based on the first output power prediction sequence, the first weighting coefficient sequence, the current output power value, and the second weighting coefficient sequence includes: determining the second output power prediction sequence of the new energy power supply equipment based on the first output power prediction sequence, the first weighting coefficient sequence, the current output power value, the second weighting coefficient sequence, the third output power prediction sequence, and the third weighting coefficient sequence.

[0007] Optionally, the pre-trained machine learning model is trained using the following method: obtaining the historical actual value and historical predicted value of the output power of the new energy power supply equipment; determining the corresponding weight coefficients based on the historical actual value and the historical predicted value of the output power; and training the initial machine learning model using the historical actual value, historical predicted value of the output power of the new energy power supply equipment, and the weight coefficients to obtain the trained machine learning model.

[0008] Optionally, determining the corresponding weighting coefficients based on the historical actual value of output power and the historical predicted value of output power includes: processing the historical actual value of output power and the historical predicted value of output power using a convex optimization algorithm or a grid search algorithm to determine the corresponding weighting coefficients; wherein the weighting coefficients minimize the error between the historical actual value of power and the historical predicted value of output power.

[0009] Optionally, obtaining the first output power prediction value sequence of the new energy power supply equipment within a first preset time period includes: obtaining relevant data affecting the output power value of the new energy power supply equipment; and determining the first output power prediction value sequence of the new energy power supply equipment within a first preset time period based on the relevant data.

[0010] Optionally, the method further includes: determining constraints on the output power value of the new energy power supply equipment based on the second output power prediction value sequence of the new energy power supply equipment; determining the output power value parameter of the new energy power supply equipment based on the constraints and a preset objective function; controlling the new energy power supply equipment to output power based on the output power value parameter; the optimization objective of the objective function is to maximize the output power value parameter of the new energy power supply equipment.

[0011] Optionally, the new energy power supply equipment is a wind power supply equipment or a solar power generation equipment.

[0012] Secondly, embodiments of this application provide an output power prediction device, comprising: a first acquisition module, configured to acquire a first output power prediction value sequence of a new energy power supply device within a first preset time period; a second acquisition module, configured to acquire the current output power value of the new energy power supply device; an input module, configured to input the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing, to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value; and a determination module, configured to determine a second output power prediction value sequence of the new energy power supply device based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence; wherein the machine learning model employs a dynamic weighted fusion algorithm; and the training objective of the machine learning model is that the difference between the actual output power value and the predicted output power value of the new energy power supply device is less than a preset threshold.

[0013] Thirdly, embodiments of this application provide an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the electronic device implements any of the methods described in the first aspect.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the methods described in the first aspect.

[0015] This application provides an output power prediction method, apparatus, electronic device, and storage medium. A first output power prediction sequence and a current output power value are input into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction sequence and a second weight coefficient sequence corresponding to the current output power value. Based on this, a second output power prediction sequence for the new energy power supply equipment can be determined, thereby improving the prediction accuracy of the new energy power supply equipment's output power value. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic flowchart of an output power value prediction method provided for an embodiment of this application; Figure 2 Another schematic flowchart of the output power value prediction method provided for embodiments of this application; Figure 3 A flowchart illustrating a method for training a machine learning model provided in an embodiment of this application; Figure 4 A flowchart illustrating a scheduling method provided in an embodiment of this application; Figure 5 A schematic diagram of an output power prediction device provided for an embodiment of this application; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] In this article, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A exists alone, A and B exist simultaneously, and B exists alone. The symbol " / " in this article indicates that the related objects are in an "or" relationship; for example, A / B means A or B.

[0019] The terms "first" and "second," etc., used in the specification and claims herein are used to distinguish different objects, not to describe a specific order of objects. For example, "first response message" and "second response message," etc., are used to distinguish different response messages, not to describe a specific order of response messages.

[0020] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is 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. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more, for example, multiple processors means two or more processors, multiple elements means two or more elements, etc.

[0022] This application provides an output power value prediction method, apparatus, electronic device, and storage medium, which can improve the prediction accuracy of output power values.

[0023] Firstly, such as Figure 1 As shown, this application provides an output power prediction method, which can be executed by an electronic device, such as a smart interactive flat panel, laptop, desktop computer, mobile phone, tablet computer, or server, etc. This application does not limit this type of device. The output power prediction method specifically includes the following steps: S11, Obtain the first output power prediction value sequence of the new energy power supply equipment within the first preset time period.

[0024] In this step, new energy power supply equipment refers to equipment that generates electricity using renewable energy sources, characterized by cleanliness, low carbon emissions, and sustainability. Specifically, new energy power supply equipment can be solar power equipment, wind power equipment, or tidal power equipment.

[0025] However, compared to power supply from the power grid, the power output of renewable energy power supply equipment is typically intermittent and fluctuating. In order to rationally schedule the output power of renewable energy power supply equipment, current methods typically employ output power prediction models to predict multiple output power values ​​of renewable energy power supply equipment within a future period (i.e., the first preset time period), thereby generating a first power prediction value sequence.

[0026] In one example, the current time is 10:00, and the first preset time period is from 10:01 to 10:15. The output power prediction module can determine the sequence of predicted output power values ​​(i.e., the first power prediction value sequence) of the renewable energy power supply equipment within the aforementioned time period. For example, with a time interval of 1 minute, this first power prediction value sequence includes the predicted output power values ​​at 10:01, 10:02, 10:03 to 10:15, designated as P1, P2, P3 to P15 respectively. However, directly using the above method results in low prediction accuracy, which may lead to low renewable energy utilization or triggering State Grid penalties during subsequent scheduling. This application's embodiments make some improvements based on this, and the specific improvements will be described in detail below.

[0027] In one specific implementation, the electronic device can acquire relevant data affecting the output power value of the new energy power supply equipment. For example, when the new energy power supply equipment is a wind power supply equipment, the aforementioned relevant data may include one or more of the following: sky imaging images, data collected by radiometers, satellite cloud images, data collected by wind measurement towers, the geographical location of the wind turbine, data collected by wind speed radar, data collected by temperature and humidity measuring instruments, rainfall data, and weather forecasts. Then, the electronic device can determine a sequence of predicted first output power values ​​for the new energy power supply equipment within a first preset time period based on the aforementioned relevant data.

[0028] S12, obtain the current output power value of the new energy power supply equipment.

[0029] In this step, since the new energy power supply equipment can measure the current output power value Pnow in real time, it sends the current output power value Pnow to the aforementioned electronic equipment, thereby enabling the electronic equipment to obtain the current output power value Pnow.

[0030] For example, the new energy power supply equipment can be a wind power supply equipment, which may include components such as blades, gearboxes, generators, and pitch control systems. The wind power supply equipment can measure the output power value in real time, thereby obtaining the current output power value.

[0031] S13, the first output power prediction value sequence and the current output power value are input into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value.

[0032] This application includes machine learning models employing dynamic fusion algorithms. Specifically, the aforementioned machine learning models can be implemented using, but are not limited to, attention mechanisms, ensemble learning, autoregressive neural networks, and optimization search.

[0033] After the machine learning model is trained, the first output power prediction value sequence and the current output power value in step S11 are input into the machine learning model for processing, so as to obtain the weight coefficients (i.e., the first weight coefficients) corresponding to each output power prediction value in the first output power prediction value sequence and multiple weight coefficients (i.e., the second weight coefficients) corresponding to the current power value.

[0034] Specifically, following the previous example, after inputting the aforementioned first output power prediction value sequence (including 15 output power prediction values ​​from P1 to P15) and the current output power value Pnow into the trained machine learning model, we can obtain 15 weight coefficients corresponding to the output power prediction values ​​at the aforementioned 15 time points, as well as 15 weight coefficients corresponding to the current output power value.

[0035] Specifically, the 15 weighting coefficients corresponding to the predicted output power values ​​at the aforementioned 15 time points are weighting coefficient W1 for predicted output power value P1, weighting coefficient W2 for predicted output power value P2, and so on up to weighting coefficient W15 for predicted output power value P15. The 15 weighting coefficients corresponding to the current output power value at the aforementioned 15 time points are weighting coefficient Wn1, weighting coefficient Wn2, and so on up to weighting coefficient Wn15.

[0036] S14. Determine the second output power prediction value sequence of the new energy power supply equipment based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence.

[0037] In this step, continuing with the previous example, at 10:01, the predicted output power value (i.e., the second predicted output power value) P1B = W1*P1 + Wn1*Pnow can be determined based on the weighting coefficient W1 of P1 and the weighting coefficient Wn1 of Pnow. At 10:02, the predicted output power value (i.e., the second predicted output power value) P2B = W2*P2 + Wn2*Pnow can be determined based on the weighting coefficient W2 of P2 and the weighting coefficient Wn2 of Pnow. Similarly, at 10:15, the predicted power value (i.e., the second predicted output power value) P15B = W15*P15 + Wn15*Pnow can be determined based on the weighting coefficient W15 of P15 and the weighting coefficient Wn15 of Pnow. In this way, the second predicted power values ​​for each of the aforementioned times can be determined, and these values ​​together form the second predicted power value sequence. Since the machine learning model uses a dynamic weighted fusion algorithm, and the training objective of the machine learning model is that the difference between the actual output power value and the predicted output power value of the new energy power supply equipment is less than a preset threshold, the above method can make the second output power prediction value at each time point more accurate than the first output power prediction value.

[0038] In some embodiments of this application, such as Figure 2 As shown, before or after the aforementioned step S11, the output power value prediction method provided in this application embodiment may further include the following steps: S11A, obtain the sequence of predicted third output power values ​​of the new energy power supply equipment within a second preset time period.

[0039] Similar to step S11, the electronic device can acquire a sequence of predicted output power values ​​for the new energy power supply equipment within a second preset time period. For example, continuing with the previous example, the current time is 10:00, and the second preset time period is from 10:01 to 11:00. For instance, with a time interval of 1 minute, the second power prediction value sequence includes the predicted output power values ​​for 60 times from 10:01, 10:02, 10:03 to 11:00, which are P1SE, P2SE, P3SE up to P60SE, respectively.

[0040] It should be noted that the output power prediction model used in this step is different from the output power prediction model used in step S11.

[0041] Next, the electronic device can perform step S12. For details of step S12, please refer to the detailed description above, which will not be repeated here.

[0042] Then, the electronic device can perform the following steps: S13A, the first output power prediction value sequence, the third output power prediction value sequence, and the current output power value are input into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence, a third weight coefficient sequence corresponding to the third output power prediction value sequence, and a second weight coefficient sequence corresponding to the current output power value.

[0043] In this step, the embodiments of this application include a machine learning model employing a dynamic fusion algorithm. After the machine learning model is trained, the first output power prediction value sequence in step S11, the third output power prediction value sequence in step S11A, and the current output power value in step S12 are input into the machine learning model for processing. This yields the weight coefficients (i.e., the first weight coefficients) corresponding to each output power prediction value in the first output power prediction value sequence, the weight coefficients (i.e., the second weight coefficients) corresponding to the current power value, and the weight coefficients (i.e., the third weight coefficients) corresponding to each output power prediction value in the third output power prediction value sequence.

[0044] Specifically, following the previous example, after inputting the aforementioned first output power prediction value sequence (including 15 output power prediction values ​​from P1, P2 to P15), the third output power prediction value sequence (including 60 output power prediction values ​​from P1SE, P2SE to P60SE), and the current output power value Pnow into the trained machine learning model, we can obtain 15 weight coefficients corresponding to the output power prediction values ​​at 15 (this is an example, but other values, such as 10) times in the first output power prediction value sequence, 15 weight coefficients corresponding to the output power prediction values ​​at 15 (this is an example, but other values, such as 10) times in the third output power prediction value sequence, and 15 (this is an example, but other values, such as 10) weight coefficients corresponding to the current output power value.

[0045] Specifically, the 15 weighting coefficients corresponding to the output power prediction values ​​at 15 time points in the first output power prediction value sequence are weighting coefficient W1 for output power prediction value P1, weighting coefficient W2 for output power prediction value P2, and so on up to weighting coefficient W15 for output power prediction value P15. The 60 weighting coefficients corresponding to the output power prediction values ​​at 60 time points in the third output power prediction value sequence are weighting coefficient W1SE for output power prediction value P1SE, weighting coefficient W2SE for output power prediction value P2SE, and so on up to weighting coefficient W15SE for output power prediction value P15SE. The weighting coefficients for output power prediction values ​​P16SE to P60SE are all 0. The 15 weighting coefficients corresponding to the current output power value at the aforementioned 15 time points are weighting coefficients Wn1, Wn2, and so on up to weighting coefficient Wn15.

[0046] S14A, based on the first output power prediction value sequence, the first weighting coefficient sequence, the current output power value, the second weighting coefficient sequence, the third output power prediction value sequence, and the third weighting coefficient sequence, determine the second output power prediction value sequence of the new energy power supply equipment.

[0047] In this step, following the previous example, at 10:01, the predicted output power value (i.e. the second predicted output power value) P1B can be determined based on the weighting coefficients W1 of P1, W1SE of P1SE, and Wn1 of Pnow.

[0048] At 10:02, the predicted output power value (i.e., the second predicted output power value) P2B can be determined based on the weighting coefficients W2 and W2SE of P2 and Wn2 of Pnow: P2B = W2*P2 + W2SE*P2SE + Wn2*Pnow. Similarly, at 10:15, the predicted power value (i.e., the second predicted output power value) P15B can be determined based on the weighting coefficients W15 and W15SE of P15 and Wn15 of Pnow: P15B = W15*P15 + W15SE*P15SE + Wn15*Pnow. In this way, the second predicted power values ​​for each of the aforementioned times can be determined, and these values ​​together form the second predicted power value sequence.

[0049] This application embodiment fuses prediction sources at different time scales, taking into account near-real-time prediction (prediction over a first preset time period), short-term prediction (prediction over a second preset time period), and current output power values, to form a unified fused prediction power sequence (second prediction power value sequence). Since this machine learning model employs a dynamic weighted fusion algorithm, and its training objective is that the difference between the actual output power value and the predicted output power value of the new energy power supply equipment is less than a preset threshold, the above method can make the second output power prediction value at each time point more accurate than the aforementioned first output power prediction value.

[0050] It should be noted that the above content is only an illustrative example. The embodiments of this application can use any number (e.g., three or four) of output power prediction value sequences. The more output power prediction value sequences used, the higher the accuracy of the second predicted power value sequence of the new energy power supply equipment.

[0051] In one implementation, the aforementioned machine learning model can be trained using the following methods. For example... Figure 3 As shown, the training method for a machine learning model may include the following steps: S21, obtain the historical actual value and historical predicted value of the output power of the new energy power supply equipment.

[0052] In this step, the electronic device can obtain multiple historical actual values ​​of output power and corresponding multiple historical predicted values ​​of output power from the new energy power supply equipment.

[0053] S22, determine the corresponding weighting coefficients based on the historical actual value of output power and the historical predicted value of output power.

[0054] For example, 9:00 and 9:01 are past historical times. The electronic device can determine the optimal weighting coefficient based on the current power value at 9:00, the predicted power value for 9:01 at 9:00, and the actual power value at 9:01. This method can be used to process multiple sets of data to obtain the corresponding optimal weighting coefficients. In a specific implementation, the electronic device can use a convex optimization algorithm or a grid search algorithm to process the aforementioned historical actual and predicted output power values ​​to determine the optimal weighting coefficients. The optimal weighting coefficients minimize the error between the historical actual and historical predicted values.

[0055] S23, using the historical actual value and historical predicted value of the output power of the new energy power supply equipment and the weight coefficient, the initial machine learning model is trained to obtain a trained machine learning model.

[0056] In this step, the historical actual output power, historical predicted output power, and the optimal weight coefficients obtained in the preceding steps of the new energy power supply equipment can be used as training samples to train the initial machine learning model. When the training objective is achieved, the trained machine learning model is obtained. The training objective of the machine learning model can be that the error between the actual output power and the predicted output power of the new energy power supply equipment is less than a preset threshold.

[0057] In some embodiments of this application, such as Figure 4 As shown, the method further includes: S15, determine the constraint conditions for the output power value of the new energy power supply equipment based on the second output power prediction value sequence of the new energy power supply equipment.

[0058] In this step, following the previous example, the 15 predicted output power values ​​in the second output power prediction value sequence are P1B, P2B, P3B up to P15B. Therefore, under the constraints, the output power value P1S of the new energy power supply equipment at 10:01 can be less than or equal to P1B. Similarly, the output power value P2S of the new energy power supply equipment at 10:02 can be less than or equal to P2B, and so on, so that the output power value P15S of the new energy power supply equipment at 10:15 can be less than or equal to P15B.

[0059] In addition, users can configure other constraints according to actual needs. For example, the aforementioned constraints may also include one or more of the following constraints: Specifically, constraints may include: DC power flow constraints: Based on the grid topology and branch parameters, a linearized DC power flow model is used to constrain the relationship between node voltage and power flow, ensuring stable bus voltage and current within limits. Grid power supply constraints: Specific requirements such as upper and lower thresholds for grid power supply and upper and lower thresholds for power variation at grid connection points. Energy storage power supply or charging constraints: Must always be maintained within the set safe operating range to avoid overcharging or over-discharging, ensuring its service life and operating efficiency. Load constraints: Upper and lower thresholds for load and upper and lower thresholds for load change rate. Power supply constraints for thermal power units: Upper and lower thresholds for output power and upper and lower thresholds for output power change rate.

[0060] S16. Determine the output power value parameters of the new energy power supply equipment according to the constraints and the preset objective function.

[0061] In this step, the electronic device can determine the output power parameters of the new energy power supply equipment based on the aforementioned constraints and the preset objective function. The optimization objective of the objective function is to maximize the output power parameters of the new energy power supply equipment. In one example, the objective function can be as follows:

[0062] Continuing with the previous example, here N=15, t represents the time 10:00, and k can take any value from 1 to 15. When k=1, t+k represents the time 10:01; when k=2, t+k represents the time 10:02; and so on, so that when k=15, t+k represents the time 10:15. Let be the wind curtailment power at time t+k. Let be the power of the discarded light at time t+k. Let be the grid power supply at time t+k, 1- This represents the load utilization rate at time t+k. Let be the power output of the energy storage power supply device at time t+k. Min represents minimizing the value of the objective function, that is, maximizing the output power parameter of the new energy power supply device. Based on the aforementioned constraints and objective function, the corresponding scheduling data can be determined. The scheduling data may include the output power parameter of the new energy power supply device. This scheduling data can minimize the objective function while satisfying the aforementioned constraints.

[0063] Users can set it according to their actual needs. , , , , as well as The value. For example, if a user wants to prioritize energy storage over grid power when wind and solar power supplies are insufficient, then the value can be increased. The value. If priority is given to using grid power, the value can be increased. The value.

[0064] It should be noted that the above is only an illustrative description, and users can set the target function according to their actual needs.

[0065] S17, control the new energy power supply equipment to output power according to the output power value parameter.

[0066] In this step, the electronic device can control the power output of the new energy power supply equipment according to the output power value parameters calculated above, thereby maximizing the output power value of the new energy power supply equipment and thus maximizing the utilization efficiency of the new energy power supply equipment.

[0067] It should be noted that the device can update the second output power prediction value sequence and scheduling data every certain period of time (e.g., 1 minute), and control the power output of new energy power supply equipment, grid power supply and energy storage power supply according to the scheduling data, so as to ensure that grid power supply or energy storage power supply can respond quickly when the power supply of new energy power supply equipment changes rapidly.

[0068] Secondly, such as Figure 5 As shown in the figure, this application embodiment provides an output power prediction device 5, including: a first acquisition module 51, used to acquire a first output power prediction value sequence of a new energy power supply device within a first preset time period; a second acquisition module 52, used to acquire the current output power value of the new energy power supply device; an input module 53, used to input the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing, so as to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value; and a determination module 54, used to determine a second output power prediction value sequence of the new energy power supply device based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence; wherein, the machine learning model adopts a dynamic weighted fusion algorithm; and the training objective of the machine learning model is that the difference between the actual output power value and the output power prediction value of the new energy power supply device is less than a preset threshold.

[0069] Optionally, the output power prediction device 5 may further include: a third acquisition module, used to acquire a third output power prediction value sequence of the new energy power supply equipment within a second preset time period; an input module 53 specifically used to: input the first output power prediction value sequence, the third output power prediction value sequence, and the current output power value into a pre-trained machine learning model for processing, to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence, a third weight coefficient sequence corresponding to the third output power prediction value sequence, and a second weight coefficient sequence corresponding to the current output power value; and a determination module 54 specifically used to: determine a second output power prediction value sequence of the new energy power supply equipment based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, the second weight coefficient sequence, the third output power prediction value sequence, and the third weight coefficient sequence.

[0070] Optionally, the output power prediction device 5 may further include: a fourth acquisition module, used to acquire the historical actual value and historical predicted value of the output power of the new energy power supply equipment; a second determination module, used to determine the corresponding weight coefficients based on the historical actual value and the historical predicted value of the output power; and a fifth acquisition module, used to train the initial machine learning model using the historical actual value, the historical predicted value of the output power of the new energy power supply equipment, and the weight coefficients, to obtain a trained machine learning model.

[0071] Optionally, the second determining module is specifically used to: process the historical actual value of output power and the historical predicted value of output power using a convex optimization algorithm or a grid search algorithm to determine the corresponding weighting coefficients; wherein the weighting coefficients minimize the error between the historical actual value of power and the historical predicted value of output power.

[0072] Optionally, the first acquisition module 51 is specifically used to: acquire relevant data affecting the output power value of the new energy power supply equipment; and determine a first output power prediction value sequence of the new energy power supply equipment within a first preset time period based on the relevant data.

[0073] Optionally, the output power prediction device 5 may further include: a third determining module, used to determine the constraint conditions of the output power value of the new energy power supply equipment based on the second output power prediction value sequence of the new energy power supply equipment; a fourth determining module, used to determine the output power value parameters of the new energy power supply equipment based on the constraint conditions and a preset objective function; and a control module, used to control the new energy power supply equipment to output power based on the output power value parameters; wherein the optimization objective of the objective function is to maximize the output power value parameters of the new energy power supply equipment.

[0074] Optionally, the new energy power supply equipment is a wind power supply equipment or a solar power generation equipment.

[0075] The output power prediction device 5 provided in this application embodiment can execute the method executed by the electronic device in the above embodiment. Its implementation principle and technical effect are similar, and will not be described again here.

[0076] like Figure 6 As shown in the embodiments of this application, an electronic device is also provided. The electronic device includes a memory and a processor. The memory is used to store a computer program. When the computer program is executed by the processor, the output power value prediction method described above can be implemented. For details, please refer to the description of the foregoing embodiments.

[0077] Specifically, at the hardware level, the electronic device may include a processor, an internal bus, and memory. The memory may include main memory and non-volatile memory. The processor reads the corresponding computer program from the non-volatile memory into main memory and then executes it. Those skilled in the art will understand that... Figure 6 The structure shown is for illustrative purposes only and does not limit the structure of the electronic device described above. For example, the electronic device may also include components that are larger than... Figure 6 The components shown may include more or fewer components, such as other processing hardware like a GPU (Graphics Processing Unit) or external communication ports. Of course, this application does not exclude other implementation methods besides software implementations, such as logic devices or a combination of hardware and software.

[0078] In this embodiment, the processor may include a central processing unit (CPU) or a graphics processing unit (GPU), and may also include other microcontrollers, logic gates, integrated circuits, or appropriate combinations thereof with logic processing capabilities. The memory described in this embodiment can be a storage device for storing information. In digital systems, a device capable of storing binary data can be a memory; in integrated circuits, a circuit without physical form but with storage function can also be a memory, such as RAM or FIFO; in a system, a storage device with physical form can also be called a memory. In implementation, this memory can also be implemented using a cloud storage method; the specific implementation method is not limited in this specification.

[0079] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the output power value prediction method described above.

[0080] This application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the output power value prediction method as described above.

[0081] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting output power, characterized in that, include: Obtain the sequence of predicted first output power values ​​of new energy power supply equipment within a first preset time period; Obtain the current output power value of the new energy power supply equipment; The first output power prediction value sequence and the current output power value are input into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value. The second output power prediction sequence of the new energy power supply equipment is determined based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence. The machine learning model employs a dynamic weighted fusion algorithm; the training objective of the machine learning model is that the difference between the actual output power and the predicted output power of the new energy power supply equipment is less than a preset threshold.

2. The output power prediction method according to claim 1, characterized in that, The method further includes: obtaining a sequence of predicted third output power values ​​of the new energy power supply equipment within a second preset time period; The step of inputting the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value includes: The first output power prediction value sequence, the third output power prediction value sequence, and the current output power value are input into a pre-trained machine learning model for processing to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence, a third weight coefficient sequence corresponding to the third output power prediction value sequence, and a second weight coefficient sequence corresponding to the current output power value. Determining the second output power prediction sequence of the new energy power supply equipment based on the first output power prediction sequence, the first weighting coefficient sequence, the current output power value, and the second weighting coefficient sequence includes: The second output power prediction sequence of the new energy power supply equipment is determined based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, the second weight coefficient sequence, the third output power prediction value sequence, and the third weight coefficient sequence.

3. The output power prediction method according to claim 1, characterized in that, The pre-trained machine learning model was trained using the following method: Obtain the historical actual value and historical predicted value of the output power of the new energy power supply equipment; Based on the historical actual value of output power and the historical predicted value of output power, determine the corresponding weighting coefficients; The initial machine learning model is trained using the historical actual value of the output power of the new energy power supply equipment, the historical predicted value of the output power, and the weighting coefficients to obtain a trained machine learning model.

4. The output power prediction method according to claim 3, characterized in that, The step of determining the corresponding weighting coefficients based on the historical actual value of output power and the historical predicted value of output power includes: The historical actual value of output power and the historical predicted value of output power are processed using a convex optimization algorithm or a grid search algorithm to determine the corresponding weighting coefficients; wherein the weighting coefficients minimize the error between the historical actual value of power and the historical predicted value of output power.

5. The output power prediction method according to claim 1, characterized in that, The acquisition of the first output power prediction value sequence of the new energy power supply equipment within a first preset time period includes: Obtain relevant data affecting the output power value of the new energy power supply equipment; Based on the relevant data, a sequence of predicted first output power values ​​for the new energy power supply equipment within a first preset time period is determined.

6. The output power prediction method according to claim 1, characterized in that, The method further includes: Based on the second output power prediction value sequence of the new energy power supply equipment, determine the constraint conditions for the output power value of the new energy power supply equipment; Based on the constraints and the preset objective function, determine the output power parameters of the new energy power supply equipment; Based on the output power value parameter, control the new energy power supply equipment to output power; The objective function aims to maximize the output power value parameter of the new energy power supply equipment.

7. The output power value prediction method according to any one of claims 1 to 6, characterized in that, The new energy power supply equipment is either wind power supply equipment or solar power generation equipment.

8. An output power value prediction device, characterized in that, include: The first acquisition module is used to acquire the first output power prediction value sequence of the new energy power supply equipment within a first preset time period; The second acquisition module is used to acquire the current output power value of the new energy power supply equipment; The input module is used to input the first output power prediction value sequence and the current output power value into a pre-trained machine learning model for processing, so as to obtain a first weight coefficient sequence corresponding to the first output power prediction value sequence and a second weight coefficient sequence corresponding to the current output power value. The determining module is configured to determine the second output power prediction value sequence of the new energy power supply equipment based on the first output power prediction value sequence, the first weight coefficient sequence, the current output power value, and the second weight coefficient sequence. The machine learning model employs a dynamic weighted fusion algorithm; the training objective of the machine learning model is that the difference between the actual output power and the predicted output power of the new energy power supply equipment is less than a preset threshold.

9. An electronic device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it causes the electronic device to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images