Wind power prediction method, device, computer equipment and storage medium

Abstract

The present invention relates to the field of new energy technology, and discloses a wind power prediction method, device, computer equipment and storage medium. The method comprises: obtaining predicted meteorological data of a target wind farm; marking the predicted meteorological data for a power restriction state based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result; extracting the predicted meteorological data based on the power restriction marking result to obtain predicted power restriction meteorological data under a power restriction constraint state; inputting the predicted power restriction meteorological data into a pre-trained power restriction power prediction model to obtain a power prediction result; the present invention can fully consider the target power restriction power under a power restriction constraint state existing in a future external environment corresponding to the predicted meteorological data of the target wind farm, and the power prediction result can reflect the power restriction power actually output by the target wind farm under the power restriction constraint, and can adaptively perform power prediction according to the future power restriction situation of the predicted meteorological data.

Description

Technical Field

[0001] The present invention relates to the field of new energy technologies, and in particular to a wind power prediction method, device, computer equipment and storage medium. Background Art

[0002] Ultra-short-term prediction of wind power plays a key role in the dispatch and operation management of wind farms. Ultra-short-term prediction methods of wind power in related technologies usually only establish a wind power prediction model based on meteorological data such as wind speed and wind direction and collected historical power generation data, and then obtain theoretical power prediction results by inputting meteorological prediction data into the wind power prediction model.

[0003] However, for operating wind farms, on the one hand, wind resources are concentrated and large in scale, far away from load centers, the market scale of resource areas is small, and it is difficult to consume locally; on the other hand, wind power itself is random, volatile and intermittent, and wind power grid connection requires the construction of supporting peak-shaving power sources; in addition, insufficient cross-regional power transmission capacity has also become one of the important constraints. The above three reasons result in some or all wind turbines in wind farms often being in operation but in a power-limited state due to receiving grid dispatch instructions.

[0004] When power forecasting is performed without considering the power curtailment status, the power forecast results for the target forecast date with power curtailment instructions are often inaccurate. For example, for a target forecast date that includes a power curtailment dispatching instruction, the predicted meteorological data corresponding to that day is input into the pre-built wind power forecasting model to obtain the theoretical power forecast result, but since there is a power curtailment dispatching instruction on the target forecast date, the wind turbines in the wind farm may be partially or completely shut down, so the actual power generation on the target forecast date will be less than the theoretical power forecast result.

[0005] Therefore, the ultra-short-term wind power prediction method in the related art has the problem of inaccurate power prediction when the power curtailment of the target wind farm is not considered. Summary of the invention

[0006] In view of this, the present invention provides a wind power prediction method, device, computer equipment and storage medium to solve the problem of inaccurate power prediction in the ultra-short-term wind power prediction method in the related art without considering the power limitation of the target wind farm.

[0007] In a first aspect, the present invention provides a wind power prediction method, comprising: obtaining predicted meteorological data of a target wind farm; marking the predicted meteorological data for a power restriction state based on the time characteristics and / or time series variation characteristics of the predicted meteorological data to obtain a power restriction marking result; extracting the predicted meteorological data based on the power restriction marking result to obtain predicted power restriction meteorological data under a power restriction constraint state; inputting the predicted power restriction meteorological data into a pre-trained power restriction prediction model to obtain a power prediction result; wherein the power restriction power prediction model is trained by using historical power restriction meteorological data under a power restriction constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power restriction meteorological data and the historical power generation data is learned.

[0008] As an exemplary embodiment, the predicted meteorological data is marked as being in a power restriction state based on the time characteristics and / or time series variation characteristics of the predicted meteorological data to obtain a power restriction marking result, including: acquiring a historical meteorological data set, wind turbine parameter data, and historical power generation data of a target wind farm; calculating the historical theoretical power data of the target wind farm based on the historical meteorological data set and the wind turbine parameter data; marking the historical power generation data for power restriction based on the numerical difference between the historical theoretical power data and the historical power generation data aligned with timestamps to obtain a historical power restriction marking result; extracting the historical meteorological data set based on the historical power restriction marking result to obtain a historical data set under a historical power restriction constraint state. Power-limiting meteorological data; extracting time series dependencies from the historical power-limiting meteorological data and / or extracting time features from the power-limiting marking results to obtain historical time series variation features and / or historical time features as historical power-limiting features; constructing a power-limiting training data set based on the historical power-limiting meteorological data, the historical power-limiting features and the power-limiting constraint state; inputting the power-limiting training data set into a pre-constructed preset marking model for model training, and in the process of model training, learning the corresponding relationship between the power-limiting meteorological data, the historical power-limiting features and the power-limiting constraint state to obtain a power-limiting state marking model; inputting the predicted meteorological data into the power-limiting state marking model to obtain the power-limiting marking result.

[0009] As an exemplary embodiment, the historical power-limiting meteorological data includes wind speed-related data such as historical wind speed and historical wind direction, and the temporal dependency extraction of the historical power-limiting meteorological data to obtain historical temporal change characteristics includes: extracting the wind speed-related data within a first preset time length before the power-limiting constraint state as historical wind speed characteristic data in the historical meteorological data set; extracting the temporal dependency of the historical wind speed characteristic data to obtain a first temporal change characteristic within the first preset time length before the power-limiting constraint state as a first historical temporal change characteristic.

[0010] As an exemplary embodiment, the historical power-limiting meteorological data includes environmental-related data such as historical temperature and historical humidity, and the temporal dependency extraction of the historical power-limiting meteorological data to obtain historical temporal change characteristics also includes: in the historical meteorological data set, extracting environmental-related data within a second preset time length before the power-limiting constraint state as historical environmental characteristic data; wherein the second preset time length is greater than the first preset time length; and performing temporal dependency extraction on the historical environmental characteristic data to obtain a second temporal change characteristic within the second preset time length before the power-limiting constraint state as a second historical temporal change characteristic.

[0011] As an exemplary embodiment, extracting time features from the power restriction marking result to obtain historical time features includes: extracting the historical start time and the historical end time of the power restriction constraint state from the power restriction marking result as the historical time features.

[0012] As an exemplary embodiment, the power restriction status marking of the predicted meteorological data based on the time characteristics and / or time series change characteristics of the predicted meteorological data also includes: obtaining historical power restriction instruction data; extracting the time characteristics of the historical power restriction instructions to obtain a historical target power restriction time interval; determining the target power restriction time of the predicted meteorological data as a historical time feature based on the historical target power restriction time interval; and marking the power restriction status of the predicted meteorological data based on the historical time feature to obtain the power restriction marking result.

[0013] As an exemplary embodiment, the acquiring of historical power-limiting instruction data includes: acquiring historical dispatching plan data of the target wind farm station and extracting historical power-limiting dispatching instructions in the historical dispatching plan data as the historical power-limiting instruction data.

[0014] In a second aspect, the present invention provides a wind power prediction device, comprising: an acquisition module for acquiring predicted meteorological data of a target wind farm; a power restriction marking module for marking the predicted meteorological data for a power restriction state based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result; an extraction module for extracting the predicted meteorological data based on the power restriction marking result to obtain predicted power restriction meteorological data under a power restriction constraint state; a power prediction module for inputting the predicted power restriction meteorological data into a pre-trained power restriction power prediction model to obtain a power prediction result; wherein the power restriction power prediction model is trained by using historical power restriction meteorological data under a power restriction constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series change characteristics of the historical power restriction meteorological data and the historical power generation data is learned.

[0015] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the wind power prediction method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the wind power prediction method of the first aspect or any corresponding embodiment thereof.

[0017] The invention provides a method for predicting wind power, comprising: obtaining predicted meteorological data of a target wind farm; marking the predicted meteorological data for a power-limiting state based on the time characteristics and / or time series variation characteristics of the predicted meteorological data to obtain a power-limiting marking result; extracting the predicted meteorological data based on the power-limiting marking result to obtain predicted power-limiting meteorological data under a power-limiting constraint state; inputting the predicted power-limiting meteorological data into a pre-trained power-limiting power prediction model to obtain a power prediction result; wherein the power-limiting power prediction model is trained by using historical power-limiting meteorological data under a power-limiting constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power-limiting meteorological data and the historical power generation data is learned; in the training method of the above-mentioned power-limiting power prediction model, the data set selects historical meteorological data under a power-limiting constraint state and historical power generation data. The time series variation characteristics of the power generation data and the historical meteorological data can reflect the changes in wind resources and the environment for the resource market. In the process of model training, the correspondence between the meteorological changes and the power restriction under the power restriction constraint state caused by different external conditions such as wind speed and temperature can be determined by learning the correspondence between the time series variation characteristics and the historical power generation data, and the correspondence between the meteorological changes and the power restriction under the power restriction constraint state can be captured. The correspondence between the meteorological data and the power generation data under the power restriction constraint state can be captured. Based on this, the predicted power restriction meteorological data is input into the pre-trained power restriction prediction model to obtain the implementation method of the power prediction result, which can fully consider the target power restriction power under the power restriction constraint state under the future external environment corresponding to the predicted meteorological data of the target wind farm. The power prediction result can reflect the power restriction power actually output by the target wind farm under the power restriction constraint, and can adaptively perform power prediction according to the future power restriction situation of the predicted meteorological data. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 is a schematic flow chart of a wind power prediction method according to an embodiment of the present invention;

[0020] Figure 2 is a structural block diagram of a wind power prediction device according to an embodiment of the present invention;

[0021] Figure 3 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] In the first aspect, according to an embodiment of the present invention, a wind power prediction method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0024] In this embodiment, a wind power prediction method is provided. Figure 1 is a flow chart of a wind power prediction method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0025] Step S101, obtaining forecast meteorological data of a target wind farm.

[0026] In this embodiment, the predicted meteorological data can be obtained from the external weather forecast interface, or it can be predicted based on the historical meteorological data of the target wind farm using a meteorological prediction model; wherein the predicted meteorological data may include wind speed, wind direction, temperature, humidity, etc., and its time resolution may be 15 minutes.

[0027] In one embodiment, the time scale of the predicted meteorological data may be 4 hours or 24 hours to achieve ultra-short-term power prediction of the target wind farm.

[0028] In one embodiment, for predicted meteorological data with a time scale greater than 24 hours, they are intercepted based on their timestamps to obtain sub-predicted meteorological data with a time scale of 4 hours or 24 hours, and ultra-short-term power prediction of the target wind farm is further achieved based on the sub-predicted meteorological data.

[0029] Step S102: marking the predicted meteorological data as being in a power-limiting state based on the time characteristics and / or time-series variation characteristics of the predicted meteorological data to obtain a power-limiting marking result.

[0030] For operating wind farms, on the one hand, wind resources are concentrated and large in scale, far away from load centers, and the market size of resource areas is small and difficult to consume locally. On the other hand, wind power itself is random, volatile and intermittent, and wind power grid connection requires the construction of supporting peak-shaving power sources; in addition, insufficient cross-regional power transmission capacity has also become one of the important constraints. The above three reasons result in some or all wind turbines in wind farms often being in a power-limited state due to receiving grid dispatch instructions although they are put into operation; the above power-limited state is usually associated with specific time characteristics or time series change characteristics.

[0031] Specifically, the power restriction status of a wind farm is usually correlated with changes in wind resources and load at the resource site. On the one hand, for scenarios where wind resources in the resource site market do not change significantly but electricity demand at the resource site decreases, in order to avoid the situation where the power supply at the resource site exceeds the load, the dispatching end usually sends a grid dispatching instruction to the target wind farm station to put the target wind farm station in a power restriction constraint state. For example, under the premise that wind speed conditions are relatively stable, if residents' electricity consumption fluctuates during this period, the electricity demand at the resource site will decrease, and the wind farm station will have a situation where the power supply at the resource site exceeds the load when there is no power restriction. The above-mentioned fluctuations are greatly affected by environmental characteristics and time characteristics alone or together. For example, at the turn of summer and autumn, the temperature and humidity before and after rainfall change in a specific time sequence, while residents Electricity consumption tends to decrease with the time series change; in winter, the temperature and humidity also show specific time series changes before and after snowfall, and residential electricity consumption tends to decrease with the time series change; in addition, in each time period with a day as the time unit, since the peak value and trough value of the residential electricity consumption side are usually different, there is a power restriction caused by the change of wind resources in the resource area and the power change on the load side in the above scenarios; on the other hand, for the situation where the wind resources in the resource area market change significantly but the power demand in the resource area is not obvious, when the wind speed shows a trend of changing from low to high, during the peak period of wind speed in the peak wind speed range, in order to avoid the situation where the power supply of the resource area exceeds the load, the dispatching end usually sends a power grid dispatching instruction to the target wind farm station to make the target wind farm station in a power restriction constraint state.

[0032] Therefore, in order to realize the power prediction of the target wind farm station under the above-mentioned power restriction constraint state, in this embodiment, the predicted meteorological data is marked with a power restriction state based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result; specifically, the historical data of the target wind farm station in the power restriction state can be feature extracted in advance to obtain the historical time characteristics and historical time series change characteristics of the target wind farm station in the power restriction constraint state, and further the predicted meteorological data is marked with a power restriction individually or collectively according to the time characteristics and time series change characteristics of the predicted meteorological data to obtain a power restriction marking result.

[0033] As a possible implementation method, the power restriction marking of the predicted meteorological data can be achieved by training the model through the historical data under the power restriction constraint state to obtain the power restriction state marking model; wherein the historical data can include.

[0034] In one embodiment, the power restriction status marking model may mark the predicted meteorological data for power restriction status based on the degree of matching between the timestamps of the time features in the predicted meteorological data and the timestamps of the historical time features to obtain the power restriction marking result of the predicted meteorological data.

[0035] In one embodiment, the power restriction status marking model can mark the predicted meteorological data for power restriction status based on the degree of matching between the time series change characteristics of the data in the predicted meteorological data over time and the time series change characteristics of the historical meteorological data over time, and obtain the power restriction marking result of the predicted meteorological data.

[0036] In one embodiment, the power restriction status marking model can mark the predicted meteorological data for power restriction status based on the degree of matching between the timestamps of the time features in the predicted meteorological data and the timestamps of the historical time features and the degree of matching between the time series change characteristics of the data in the predicted meteorological data over time and the time series change characteristics of the historical meteorological data over time, so as to obtain the power restriction marking result of the predicted meteorological data.

[0037] As an exemplary embodiment, the power-limiting marking result includes a power-limiting constraint state and a non-power-limiting constraint state.

[0038] Step S103 : extracting the predicted meteorological data based on the power restriction marking result to obtain predicted power restriction meteorological data under the power restriction constraint state.

[0039] Step S104, input the predicted power-limiting meteorological data into a pre-trained power-limiting power prediction model to obtain a power prediction result; wherein the power-limiting power prediction model is trained by using the historical power-limiting meteorological data under the power-limiting constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power-limiting meteorological data and the historical power generation data is learned.

[0040] In this embodiment, the power-limiting power prediction model is obtained by training the historical power-limiting meteorological data and the corresponding historical power generation data under the power-limiting constraint state; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical meteorological data and the historical power generation data is learned; in the training method of the above-mentioned power-limiting power prediction model, the data set selects the historical meteorological data and the historical power generation data under the power-limiting constraint state, and the time series variation characteristics of the historical meteorological data can reflect the wind resource changes and environmental changes for the resource market. In the process of model training, different wind The correspondence between the meteorological changes under the power-limiting constraint state caused by external conditions such as speed and temperature and the power-limiting power can capture the correspondence between the meteorological data and the power generation power data under the power-limiting constraint state; based on this, the predicted power-limiting meteorological data is input into a pre-trained power-limiting power prediction model to obtain an implementation method for the power prediction result, which can fully consider the target power-limiting power under the power-limiting constraint state under the future external environment corresponding to the predicted meteorological data of the target wind farm. The power prediction result can reflect the power-limiting power actually output by the target wind farm under the power-limiting constraint, and can adaptively perform power prediction according to the future power-limiting situation of the predicted meteorological data.

[0041] As an exemplary embodiment, the predicted meteorological data is marked as being in a power restriction state based on the time characteristics and / or time series variation characteristics of the predicted meteorological data to obtain a power restriction marking result, including: acquiring a historical meteorological data set, wind turbine parameter data, and historical power generation data of a target wind farm; calculating the historical theoretical power data of the target wind farm based on the historical meteorological data set and the wind turbine parameter data; marking the historical power generation data for power restriction based on the numerical difference between the historical theoretical power data and the historical power generation data aligned with timestamps to obtain a historical power restriction marking result; extracting the historical meteorological data set based on the historical power restriction marking result to obtain a historical power restriction constraint state. historical power-limiting meteorological data; extracting time series dependencies from the historical power-limiting meteorological data and / or extracting time features from the power-limiting marking results to obtain historical time series variation features and / or historical time features as historical power-limiting features; constructing a power-limiting training data set based on the historical power-limiting meteorological data and the historical power-limiting features; inputting the power-limiting training data set into a pre-constructed preset marking model for model training, and in the process of model training, learning the correspondence between the power-limiting meteorological data, the historical power-limiting features and the power-limiting constraint state to obtain the power-limiting state marking model; inputting the predicted meteorological data into the power-limiting state marking model to obtain the power-limiting marking result.

[0042] When the historical scheduling plan of the target wind farm is incomplete or missing, it is usually impossible to determine the historical moments when the target wind farm was in a power-rationing state during the historical period based on the historical scheduling plan; therefore, in this embodiment, the historical power generation is marked for power restriction by comparing the difference between the historical theoretical power data and the historical power generation data of the target wind farm.

[0043] Specifically, after obtaining the historical meteorological data set, wind turbine parameter data and historical power generation data of the target wind farm, the average wind speed of the entire wind farm is extracted from the historical meteorological data set, and the number of wind turbines, air density, swept area of ​​the wind rotor and power coefficient of the target wind farm are extracted from the wind turbine parameter data. The theoretical power of the target wind farm is further calculated using formula (1):

[0044] (1)

[0045] In formula (1), P represents historical theoretical power data, N Indicates the number of fans, ρ represents the air density, A represents the swept area of ​​the wind rotor, It represents the average wind speed of the aircraft head. C p Indicates the power factor.

[0046] After the historical theoretical power data of the target wind farm is calculated, the historical power generation data is marked for power restriction based on the numerical difference between the historical theoretical power data and the historical power generation data with timestamp alignment, to obtain the historical power restriction marking result; specifically, the numerical difference between the historical theoretical power data and the historical power generation data is first calculated, and the historical power generation data whose numerical difference is greater than a preset value is further marked as a historical power restriction constraint state, and the preset value can be set to one percent of the numerical value of the historical theoretical power data corresponding to the current numerical difference.

[0047] In one embodiment, after obtaining the historical power restriction marking result, the historical time series variation feature is extracted from the power restriction marking result, and the extracted historical time series variation feature is used as the historical power restriction feature representing that the target wind farm is in a power restriction constraint state.

[0048] In one embodiment, after obtaining the power restriction marking result, the historical meteorological data set is extracted based on the power restriction marking result to obtain historical power restriction meteorological data; wherein the historical power restriction meteorological data may include wind speed-related data and environment-related data; specifically, the wind speed-related data may include historical wind speed, historical wind direction and other data, and the environment-related data may include historical temperature, historical humidity and other data.

[0049] Furthermore, the temporal dependency relationship of the historical power-limiting meteorological data is extracted, and the extracted historical temporal variation characteristics are used as the historical power-limiting characteristics that characterize the target wind farm being in a power-limiting constraint state.

[0050] In one embodiment, after obtaining historical power-limiting meteorological data, time-series dependency extraction is performed on the historical power-limiting meteorological data, and time features are extracted from the power-limiting marking results, and the extracted historical time-series change features and historical time features are used as historical power-limiting features.

[0051] Furthermore, a power-limiting training data set is constructed based on the historical power-limiting meteorological data, the historical power-limiting characteristics and the power-limiting constraint status, and the power-limiting training data set is input into a pre-constructed preset labeling model for model training. During the model training process, the correspondence between the power-limiting meteorological data, the historical power-limiting characteristics and the power-limiting constraint status is learned to obtain the power-limiting status labeling model.

[0052] As an exemplary embodiment, the preset labeling model may include a machine learning model and a deep learning model; specifically, the preset labeling model may be a decision tree, a support vector machine, or other model.

[0053] The above implementation method can match the historical power restriction feature with the power restriction constraint state representing the target wind farm by using the historical time series change feature and the historical time feature of the historical power restriction meteorological data alone or together as the form of the historical power restriction feature characterizing the power restriction constraint state.

[0054] After the model training is completed, the predicted meteorological data is input into the power restriction status marking model, which can determine the power restriction marking status of the target wind farm determined by the predicted meteorological data in the future period individually or collectively through the historical time series change characteristics or time characteristics of the predicted meteorological data.

[0055] In order to consider the change of wind resources of the target wind farm, as an exemplary embodiment, the historical power-limiting meteorological data includes wind speed-related data such as historical wind speed and historical wind direction, and the historical power-limiting meteorological data is subjected to time series dependency extraction to obtain historical time series change characteristics, including:

[0056] Extracting the wind speed related data within a first preset time period before the power restriction constraint state from the historical meteorological data set as historical wind speed characteristic data;

[0057] A time series dependency relationship is extracted from the historical wind speed characteristic data to obtain a first historical time series variation characteristic within the first preset time length before the power restriction constraint state as a first historical time series variation characteristic.

[0058] Due to the randomness of changes in wind resources, wind speed usually changes in a time dimension that is less than a day; based on this, in this embodiment, the time scale of the first preset time length is less than a day; in one embodiment, the time scale of the first preset time length may be the same as the collection resolution of historical meteorological data, which may be 15 minutes.

[0059] In one embodiment, the preset marking model may include a first time series feature extraction unit; the first time series feature extraction unit is used to extract the time series dependency relationship of the historical wind speed feature data to obtain the first historical time series change feature within the first preset time length before the power limit constraint state.

[0060] As an exemplary embodiment, the first temporal feature extraction unit may be implemented by at least one of a recurrent neural network (RNN), a long short-term memory neural network (LSTM), and a gated recurrent unit (GRU).

[0061] In this embodiment, when the power restriction status marking model inputs the predicted meteorological data, the first predicted time series feature of the predicted wind speed related data in the predicted meteorological data is extracted by the first time series feature extraction unit, and then the predicted meteorological data is marked for power restriction according to the correspondence between the first predicted time series change feature and the first historical time series change feature to obtain the power restriction marking result.

[0062] As an exemplary embodiment, the historical power-limiting meteorological data includes environmental-related data such as temperature and humidity, and the temporal dependency extraction of the historical power-limiting meteorological data to obtain historical temporal change characteristics also includes: in the historical meteorological data set, extracting environmental-related data within a second preset time length before the power-limiting constraint state as historical environmental characteristic data; wherein the second preset time length is greater than the first preset time length.

[0063] In this embodiment, environmental factors such as temperature and humidity usually change over a longer time dimension than wind speed, based on which the second preset time length is greater than the first preset time length. As an exemplary embodiment, the second preset time length may be 48 hours.

[0064] In one embodiment, the preset marking model may include a second time series feature extraction unit; the second time series feature extraction unit is used to extract the time series dependency relationship of the historical environmental feature data to obtain the second historical time series change characteristics within the second preset time length before the power limit constraint state.

[0065] As an exemplary embodiment, the second temporal feature extraction unit may be implemented by at least one of a recurrent neural network (RNN), a long short-term memory neural network (LSTM), and a gated recurrent unit (GRU).

[0066] In this embodiment, when the power restriction status marking model inputs the predicted meteorological data, the second predicted time series feature of the predicted environment-related data in the predicted meteorological data is extracted by the second time series feature extraction unit, and then the predicted meteorological data is marked for power restriction according to the correspondence between the second predicted time series change feature and the second historical time series change feature to obtain the power restriction marking result.

[0067] In one embodiment, the preset marking model may simultaneously include a first time series feature extraction unit and a second time series feature extraction unit; when the power limit state marking model inputs the predicted meteorological data, the first time series feature extraction unit performs time series dependency extraction on the historical wind speed feature data to obtain the first historical time series change feature within the first preset time length before the power limit constraint state, and the second time series feature extraction unit performs time series dependency extraction on the historical wind speed feature data to obtain the second historical time series change feature within the second preset time length before the power limit constraint state, and further performs power limit marking on the predicted meteorological data according to the corresponding relationship between the first predicted time series change feature and the first historical time series change feature and the corresponding relationship between the second predicted time series change feature and the second historical time series change feature to obtain the power limit marking result.

[0068] In one embodiment, when the predicted meteorological data is marked for power restrictions based on the correspondence between the first predicted time series change characteristic and the first historical time series change characteristic and the correspondence between the second predicted time series change characteristic and the second historical time series change characteristic, the first power restriction marking result can be obtained based on the correspondence between the first predicted time series change characteristic and the first historical time series change characteristic, and the second power restriction marking result can be obtained based on the correspondence between the second predicted time series change characteristic and the second historical time series change characteristic, and then the first power restriction marking result and the second power restriction marking result are weightedly fused to obtain the final power restriction marking result.

[0069] In one embodiment, the fusion weight when weighted fusion is performed on the first power limit marking result and the second power limit marking result is proportional to the similarity between each predicted time series change feature and the historical time series change feature.

[0070] As an exemplary embodiment, extracting time features from the power restriction marking result to obtain historical time features includes: extracting the historical start time and the historical end time of the power restriction constraint state from the power restriction marking result as the historical time features.

[0071] When the historical scheduling plan of the target wind farm is improved, the historical moments when the target wind farm was in a power-limited state in the historical period can be determined according to the historical scheduling plan; therefore, as an exemplary embodiment, the power-limited state marking of the predicted meteorological data based on the time characteristics and / or time series change characteristics of the predicted meteorological data also includes: obtaining historical power-limited instruction data; extracting time characteristics of the historical power-limited instructions to obtain a historical target power-limited time interval; determining the target power-limited time of the predicted meteorological data as a historical time feature based on the historical target power-limited time interval; marking the predicted meteorological data in a power-limited state based on the historical time characteristics to obtain the power-limited marking result.

[0072] In this embodiment, after obtaining the historical power-limiting instruction data, the timestamp of the historical power-limiting instruction can be extracted as a time feature to obtain the historical target power-limiting time interval, and the predicted meteorological data can be further marked with a power-limiting status through the inclusion relationship between the time period of the predicted meteorological data and the historical target power-limiting time interval.

[0073] As an exemplary embodiment, the acquiring of historical power-limiting instruction data includes: acquiring historical dispatching plan data of the target wind farm; and extracting historical power-limiting dispatching instructions from the historical dispatching plan data as the historical power-limiting instruction data.

[0074] This embodiment provides a wind power prediction device, such as Figure 2 As shown, including:

[0075] An acquisition module 501 is used to acquire the forecasted meteorological data of the target wind farm;

[0076] A power restriction marking module 502 is used to mark the predicted meteorological data for power restriction based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result;

[0077] An extraction module 503, configured to extract the predicted meteorological data based on the power restriction marking result to obtain the predicted power restriction meteorological data under the power restriction constraint state;

[0078] The power prediction module 504 is used to input the predicted power-limiting meteorological data into a pre-trained power-limiting power prediction model to obtain a power prediction result; wherein the power-limiting power prediction model is trained by the historical power-limiting meteorological data under the power-limiting constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power-limiting meteorological data and the historical power generation data is learned.

[0079] It should be noted here that the examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments.

[0080] It should be noted that the above modules as part of the device can be implemented by software or hardware, wherein the hardware environment includes a network environment.

[0081] An embodiment of the present invention also provides a computer device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus, the memory is used to store a computer program; the processor is used to execute a method in any of the above embodiments by running the computer program stored in the memory.

[0082] Figure 3 is a structural block diagram of an optional computer device according to an embodiment of the present application, such as Figure 3 As shown, it includes a processor 10, a communication interface 20, a memory 30 and a communication bus 40, wherein the processor 10, the communication interface 20 and the memory 30 communicate with each other through the communication bus 40, wherein:

[0083] A memory 30, for storing computer programs;

[0084] The processor 10 is used to implement the method for constructing a disease database as described in any of the above embodiments when executing the computer program stored in the memory 30.

[0085] Optionally, in this embodiment, the communication bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0086] The communication interface is used for communication between the above-mentioned computer device and other devices.

[0087] The memory may include RAM, or may include non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0088] The above-mentioned processor can be a general-purpose processor, which can include but not be limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processing), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0089] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0090] It can be understood by those skilled in the art that Figure 3 The structure shown is for illustration only. The device for implementing any one of the methods in the above embodiments may be a terminal device, which may be a smart phone (such as an Android phone, an IOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 3 It does not limit the structure of the above electronic device. For example, the terminal device may also include Figure 3 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 3 Different configurations shown.

[0091] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, which can include: a flash drive, ROM, RAM, a magnetic disk or an optical disk, etc.

[0092] As an exemplary embodiment, the present application also provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute any one of the method steps of the present embodiment when running.

[0093] Optionally, in this embodiment, the above-mentioned storage medium can be used to execute the program code of the method steps of the embodiment of the present application.

[0094] Optionally, in this embodiment, the storage medium may be located on at least one network device among a plurality of network devices in the network shown in the above embodiment.

[0095] Optionally, in this embodiment, the storage medium is configured to store data for executing the method in the above embodiment.

[0096] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, which will not be described in detail in this embodiment.

[0097] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a ROM, a RAM, a mobile hard disk, a magnetic disk, or an optical disk.

[0098] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0099] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers or network devices, etc.) to execute all or part of the steps of the method in the above embodiments.

[0100] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0101] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution provided in this embodiment.

[0102] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0103] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0104] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A wind power prediction method, characterized in that: include: Obtain forecast meteorological data for the target wind farm; Marking the predicted meteorological data for power restriction status based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result; It specifically includes: acquiring a historical meteorological data set, wind turbine parameter data and historical power generation data of a target wind farm; calculating the historical theoretical power data of the target wind farm based on the historical meteorological data set and the wind turbine parameter data; marking the historical power generation data for power restriction based on the numerical difference between the historical theoretical power data and the historical power generation data aligned with timestamps, and obtaining a historical power restriction marking result; extracting the historical meteorological data set based on the historical power restriction marking result, and obtaining historical power restriction meteorological data under a historical power restriction constraint state; wherein the historical power restriction meteorological data includes wind speed-related data of historical wind speed and historical wind direction; extracting time series dependencies from the historical power restriction meteorological data and / or extracting time features from the power restriction marking result, and obtaining historical time series change features and / or historical time features as historical power restriction features. ; It specifically includes: in the historical meteorological data set, extracting the wind speed related data within the first preset time length before the power restriction constraint state as historical wind speed feature data; extracting the time series dependency relationship of the historical wind speed feature data, and obtaining the first time series change feature within the first preset time length before the power restriction constraint state as the first historical time series change feature; constructing a power restriction training data set based on the historical power restriction meteorological data, the historical power restriction feature and the historical power restriction constraint state; inputting the power restriction training data set into a pre-constructed preset marking model for model training, and in the process of model training, learning the correspondence between the power restriction meteorological data, the historical power restriction feature and the historical power restriction constraint state to obtain a power restriction state marking model; inputting the predicted meteorological data into the power restriction state marking model to obtain the power restriction marking result; Extracting the predicted meteorological data based on the power restriction marking result to obtain predicted power restriction meteorological data under the power restriction constraint state; The predicted power-limiting meteorological data is input into a pre-trained power-limiting power prediction model to obtain a power prediction result; wherein, the power-limiting power prediction model is trained by the historical power-limiting meteorological data under the power-limiting constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power-limiting meteorological data and the historical power generation data is learned.

2. The wind power prediction method according to claim 1, characterized in that: The historical power-limiting meteorological data includes environmental data related to historical temperature and historical humidity, and extracting the time series dependency relationship of the historical power-limiting meteorological data to obtain historical time series change characteristics also includes: In the historical meteorological data set, environment-related data within a second preset time period before the power restriction constraint state is extracted as historical environmental characteristic data; wherein the second preset time period is greater than the first preset time period; A time series dependency relationship is extracted from the historical environmental feature data to obtain a second time series variation feature within the second preset time period before the power restriction constraint state as a second historical time series variation feature.

3. The wind power prediction method according to claim 2, characterized in that: The extracting time features of the power restriction marking result to obtain historical time features includes: In the power restriction marking result, the historical start time and the historical end time of the power restriction constraint state are extracted as the historical time feature.

4. The wind power prediction method according to claim 1, characterized in that: The marking of the power-limiting state of the predicted meteorological data based on the time characteristics and / or time series variation characteristics of the predicted meteorological data further includes: Obtain historical power restriction instruction data; Extracting time features of the historical power-limiting instructions to obtain a historical target power-limiting time interval; Determining the target power-cut time of the predicted meteorological data as a historical time feature based on the historical target power-cut time interval; The predicted meteorological data is marked with a power restriction state based on the historical time feature to obtain the power restriction marking result.

5. The wind power prediction method according to claim 4, characterized in that: The obtaining of historical power-limiting instruction data includes: Acquire historical dispatch plan data of the target wind farm; The historical power-limiting scheduling instructions in the historical scheduling plan data are extracted as the historical power-limiting instruction data.

6. A wind power prediction device, characterized in that: The device comprises: An acquisition module is used to acquire the forecasted meteorological data of the target wind farm station; A power restriction marking module is used to mark the predicted meteorological data for power restriction based on the time characteristics and / or time series change characteristics of the predicted meteorological data to obtain a power restriction marking result; it specifically includes: obtaining a historical meteorological data set, wind turbine parameter data and historical power generation data of a target wind farm; calculating the historical theoretical power data of the target wind farm based on the historical meteorological data set and the wind turbine parameter data; marking the historical power generation data for power restriction based on the numerical difference between the historical theoretical power data and the historical power generation data aligned with timestamps to obtain a historical power restriction marking result; extracting the historical meteorological data set based on the historical power restriction marking result to obtain historical power restriction meteorological data under a historical power restriction constraint state; wherein the historical power restriction meteorological data includes wind speed-related data of historical wind speed and historical wind direction; extracting time series dependencies of the historical power restriction meteorological data and / or performing time feature analysis on the power restriction marking result. Extraction, to obtain historical time series change characteristics and / or historical time characteristics as historical power restriction characteristics; which specifically includes: in the historical meteorological data set, extracting the wind speed related data within the first preset time length before the power restriction constraint state as historical wind speed characteristic data; performing time series dependency extraction on the historical wind speed characteristic data, to obtain the first time series change characteristics within the first preset time length before the power restriction constraint state as the first historical time series change characteristics; constructing a power restriction training data set based on the historical power restriction meteorological data, the historical power restriction characteristics and the historical power restriction constraint state; inputting the power restriction training data set into a pre-constructed preset marking model for model training, and in the process of model training, learning the corresponding relationship between the power restriction meteorological data, the historical power restriction characteristics and the historical power restriction constraint state to obtain a power restriction state marking model; inputting the predicted meteorological data into the power restriction state marking model to obtain the power restriction marking result; An extraction module, configured to extract the predicted meteorological data based on the power restriction marking result to obtain the predicted power restriction meteorological data under the power restriction constraint state; A power prediction module is used to input the predicted power-limiting meteorological data into a pre-trained power-limiting power prediction model to obtain a power prediction result; wherein the power-limiting power prediction model is trained by the historical power-limiting meteorological data under the power-limiting constraint state and its corresponding historical power generation data; in the process of model training, the corresponding relationship between the time series variation characteristics of the historical power-limiting meteorological data and the historical power generation data is learned.

7. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the wind power prediction method according to any one of claims 1 to 5 by executing the computer instructions.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the wind power prediction method according to any one of claims 1 to 7.

Images