Single-machine power prediction method, and full-field energy regulation and control method and system

Abstract

The invention provides a single-machine power prediction method and a full-field energy regulation and control method and system. The method comprises the following steps: a processing period comprises a plurality of time periods and observation periods which are arranged according to a time sequence; assigning a time weight to each time period of the processing cycle; the method comprises the following steps: S1, taking a previous processing period as a target processing period; s2, obtaining the average generated power of the target processing cycle in each time period; s3, according to each average generation power and the corresponding time weight, obtaining a predicted generation power under the observation section; s4, acquiring a difference value between the actual generated power of the observation section and the predicted generated power, if the difference value is greater than a preset threshold value, adjusting each time weight, and repeating the steps S3 to S4 until the difference value is lower than the preset threshold value; and S5, taking the current processing cycle as a target processing cycle, and repeating the steps S2 to S3. According to the method, the problem that the calculation error of the predicted generated power of the single wind generating set is relatively large in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of wind farm regulation, and in particular to a single-machine power prediction method, a full-field energy regulation method and a system. Background Art

[0002] In the process of full-field power regulation, the wind farm needs to consider the actual power generation capacity of the wind turbine under the current situation. At present, for the full-field control target energy, the power generation capacity of a single wind turbine is calculated using the designed power curve constant, wherein the designed power curve constant does not take into account factors such as the actual installation terrain of each wind turbine. Affected by wind energy, wind energy is limited by factors such as season, temperature, and air density, resulting in a deviation between the predicted power generation calculated according to the designed power curve and the actual power generation of the wind turbine, making it impossible to meet the technical standards in the functional control with high precision requirements. This is particularly prominent in wind turbines with low single-unit capacity.

[0003] Especially when the wind farm is regulating the power of the whole field, the energy management platform receives the target power value issued by the AGC or the primary frequency modulation device as the control target value, and uses the actual feedback wind speed of each wind turbine generator set in combination with the original active power curve to calculate the active power target value of the single machine. However, during the actual operation of the wind turbine generator set, the comprehensive influence of wind energy per unit time and the actual defects of the unit itself leads to certain errors in the energy management platform's calculation of the predicted power generation of a single wind turbine generator set. These deviations easily lead to poor accuracy of power regulation of the whole field and fail to meet technical standards. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a single-machine power prediction method, a full-field energy control method and a system, which solve the problem of large errors in the calculation of predicted power generation of a single wind turbine generator set in the prior art.

[0005] At least one embodiment of the present invention provides a single-machine power prediction method, comprising:

[0006] Setting a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order, and an observation period located after a preset duration of the last time period;

[0007] Assign a time weight to each time period of the processing cycle, and perform the following steps:

[0008] S1, taking the previous processing cycle of the current processing cycle as the target processing cycle;

[0009] S2. Obtaining the average power generation of the wind turbine generator set in each of the time periods within the target processing cycle;

[0010] S3, obtaining the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period;

[0011] S4, obtaining the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation, and if the difference is greater than a preset threshold, adjusting each of the time weights, and repeating steps S3 to S4 until the difference is lower than the preset threshold;

[0012] S5. Take the current processing cycle as a new target processing cycle, and repeat steps S2 to S3 until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

[0013] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0014] Through the above steps, the present invention combines the data of multiple historical processing cycles and continuously adjusts the time weight value to determine the impact of the local environment on the power generation of the observation section in different time periods, so that the predicted power generation of the wind turbine generator set no longer only considers the designed power curve constant value, but combines the impact of historical actual power generation in different time periods and different environments, and predicts the power generation of the observation section through the average power generation of the recent time period of the same processing cycle and the time weight adjusted by the historical processing cycle;

[0015] Obviously, the single-machine power prediction method provided by the present invention can take into account the impact of local actual environmental changes on the average power generation of the observation section, making the prediction effect of this method on the predicted power generation of the observation section more accurate.

[0016] In a single-machine power prediction method provided by one embodiment of the present invention, the sum of the time weights is 1, and the predicted power generation of the wind turbine in the observation period is obtained according to the average power generation and the time weights corresponding to the time periods, including:

[0017] The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

[0018] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0019] By taking the product of each of the average generated power and the time weight corresponding to each of the time periods as the power impact value, and taking the sum of the power impact values ​​as the predicted generated power of the wind turbine in the observation period, the impact of each time period on the generated power of the observation period can be conveniently expressed simply by the product of the time weight and the average generated power.

[0020] In a single machine power prediction method provided in one embodiment of the present invention, step S2 includes:

[0021] Obtain the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle;

[0022] For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval.

[0023] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0024] For different wind speed intervals, the present invention divides the average power generation corresponding to the different wind speed intervals one by one to calculate them separately, so as to avoid the sudden change of wind speed affecting the accuracy of predicted power generation.

[0025] In a single-machine power prediction method provided in one embodiment of the present invention, obtaining the average power generation corresponding to different wind speed intervals in each time period of the target processing cycle includes:

[0026] For each time period of the target processing cycle, the ratio of the total power generation corresponding to different wind speed intervals to the number of occurrences of the corresponding different wind speed intervals is used as the average power generation corresponding to the different wind speed intervals.

[0027] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0028] In order to facilitate calculation and statistics, the present invention uses the ratio of the total power generation corresponding to different wind speed intervals to the number of times the corresponding different wind speed intervals appear as the average power generation corresponding to the different wind speed intervals.

[0029] In a single-machine power prediction method provided by one embodiment of the present invention, a processing cycle before a current processing cycle is used as a target processing cycle, including:

[0030] Based on the differentiated time value method, the previous processing cycle of the current processing cycle is used as the target processing cycle.

[0031] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0032] By differentiating the time value selection method, the correlation between two consecutive processing cycles is made stronger, making the final prediction result more reliable.

[0033] In a single-machine power prediction method provided by one of the embodiments of the present invention, the durations of the time periods in the same processing cycle are arranged in descending order.

[0034] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0035] By arranging the duration of each time period in descending order, we can try to balance the impact of each time period on the prediction segment, avoiding excessive differences in time weights that affect subsequent prediction results.

[0036] The present invention also provides a full-field energy control method, including performing full-field energy control based on the predicted power generation of the wind turbine in the current processing cycle under the observation section obtained by a single-machine power prediction method as mentioned above, combined with the target power value issued by the host computer.

[0037] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0038] By adopting the predicted power generation power of the target observation section obtained by a single-machine power prediction method as mentioned above, the whole-field energy control method has higher accuracy and faster control speed during whole-field energy control.

[0039] At least one embodiment of the present invention further provides a single-machine power prediction system, comprising:

[0040] A time weight processing module is used to set a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order and an observation period located after a preset time length of the last time period; and a time weight is assigned to each time period of the processing cycle;

[0041] The first execution model takes the previous processing cycle of the current processing cycle as the target processing cycle, and sends the target processing cycle to the average power generation acquisition module;

[0042] An average power generation acquisition module is used to acquire the average power generation of the wind turbine generator set in each time period within the target processing cycle and send it to the prediction module;

[0043] A prediction module, which obtains the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period, and sends it to the calibration module;

[0044] A calibration module, which obtains the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation. If the difference is greater than a preset threshold, the time weights are adjusted and the prediction module is returned until the difference is lower than the preset threshold and the second execution module is entered;

[0045] The second execution model sets the current processing cycle as a new target processing cycle and returns to the first execution model until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

[0046] In a single-machine power prediction system provided by one embodiment of the present invention, the sum of the time weights is 1, and the prediction module includes:

[0047] The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

[0048] In a single-machine power prediction system provided by one embodiment of the present invention, the average power generation acquisition module includes:

[0049] Obtain the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle;

[0050] For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of a single machine power prediction method according to the present invention;

[0052] Figure 2 A logic schematic diagram of a single machine power prediction method according to the present invention;

[0053] Figure 3 The present invention is a flow chart of a single-machine power prediction system. DETAILED DESCRIPTION

[0054] The principles and features of the present invention are described below. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0055] The present invention provides a single machine power prediction method, please refer to Figure 1 As shown, including:

[0056] Setting a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order, and an observation period located after a preset duration of the last time period;

[0057] Assign a time weight to each time period of the processing cycle, and perform the following steps:

[0058] S1, taking the previous processing cycle of the current processing cycle as the target processing cycle;

[0059] S2. Obtaining the average power generation of the wind turbine generator set in each of the time periods within the target processing cycle;

[0060] S3, obtaining the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period;

[0061] S4, obtaining the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation, and if the difference is greater than a preset threshold, adjusting each of the time weights, and repeating steps S3 to S4 until the difference is lower than the preset threshold;

[0062] S5. Take the current processing cycle as a new target processing cycle, and repeat steps S2 to S3 until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

[0063] Through the above steps, the present invention combines the data of multiple historical processing cycles and continuously adjusts the time weight value to determine the impact of the local environment on the power generation of the observation section in different time periods, so that the predicted power generation of the wind turbine generator set no longer only considers the designed power curve constant value, but combines the impact of historical actual power generation in different time periods and different environments, and predicts the power generation of the observation section through the average power generation of the recent time period of the same processing cycle and the time weight adjusted by the historical processing cycle;

[0064] Obviously, the single-machine power prediction method provided by the present invention can take into account the impact of local actual environmental changes on the average power generation of the observation section, making the prediction effect of this method on the predicted power generation of the observation section more accurate.

[0065] In another embodiment of the present invention, a single machine power prediction method is provided, comprising:

[0066] Set a processing period. In this embodiment, the processing period is 8 days. Of course, the period can be adjusted according to actual conditions and can be 1 day to 365 days.

[0067] The processing cycle includes a plurality of time periods arranged in chronological order, and an observation period located after the last time period of a preset duration;

[0068] Specifically, for each time period in the same processing cycle, the duration of each time period is arranged in descending order. By arranging the duration of each time period in descending order, the influence of each time period on the prediction segment can be balanced as much as possible, avoiding excessive differences in time weights that affect subsequent prediction results.

[0069] Taking a processing cycle of eight days as an example, a processing cycle can be divided into four time periods and one observation period, where the first to third days are the first time period, the fourth to fifth days are the second time period, the sixth to the first half of the seventh day are the third time period, the second half of the seventh day is the fourth time period, and the eighth day is the observation period;

[0070] Assign time weights to each of the time periods of the processing cycle. The initial time weights of the time periods can be set arbitrarily, as long as the sum of the time weights is 1. Then, perform the following steps:

[0071] S1, taking the previous processing cycle of the current processing cycle as the target processing cycle;

[0072] Specifically, based on the differentiated time value method, the previous processing cycle of the current processing cycle is used as the target processing cycle.

[0073] The differentiated value-taking method means that, for example, if the data in this embodiment is set to eight days, the data from one to eight days is a processing cycle. When it is the ninth day, the data of the ninth day is used as the last day of data, and the data of the first day is discarded, that is, the data from the second to the ninth day is the next processing cycle, and similarly, the data from the third to the tenth day is the next processing cycle.

[0074] By differentiating the time value selection method, the correlation between two consecutive processing cycles is made stronger, making the final prediction result more reliable.

[0075] S2. Obtaining the average power generation of the wind turbine generator set in each of the time periods within the target processing cycle;

[0076] Specifically, the average power generation corresponding to different wind speed intervals in each time period of the target processing cycle is obtained, that is, for each time period of the target processing cycle, the ratio of the total power generation corresponding to different wind speed intervals to the number of times the corresponding different wind speed intervals appear is used as the average power generation corresponding to the different wind speed intervals;

[0077] For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval.

[0078] That is, when performing calculations, the present invention only considers the corresponding power generation under one wind speed interval at a time. For example, the wind speed interval is divided into 2m / s-2.5m / s, 2.5m / s-3m / s, 3m / s-3.5m / s, etc. Each time a single-machine power prediction method is executed, only one wind speed interval is considered for prediction;

[0079] In order to facilitate calculation and statistics, the present invention takes the ratio of the total power generation corresponding to different wind speed intervals to the number of times the corresponding different wind speed intervals appear as the average power generation corresponding to different wind speed intervals. For different wind speed intervals, the present invention divides the average power generation corresponding to different wind speed intervals one by one for separate calculation to avoid the sudden change in wind speed from affecting the accuracy of predicted power generation.

[0080] S3, obtaining the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period;

[0081] Specifically, since the sum of the time weights is 1, the present invention obtains the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weights corresponding to the time periods, including:

[0082] The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

[0083] By taking the product of each of the average generated power and the time weight corresponding to each of the time periods as the power impact value, and taking the sum of the power impact values ​​as the predicted generated power of the wind turbine in the observation period, the impact of each time period on the generated power of the observation period can be conveniently expressed simply by the product of the time weight and the average generated power.

[0084] S4, obtaining the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation, and if the difference is greater than a preset threshold, adjusting each of the time weights, and repeating steps S3 to S4 until the difference is lower than the preset threshold;

[0085] S5. Take the current processing cycle as a new target processing cycle, and repeat steps S2 to S3 until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

[0086] In summary:

[0087] According to the active dispatch control method in the energy management platform, after receiving the target power value sent by the AGC, primary frequency modulation device and other host computers, the energy management platform calculates the difference between the current actual power value and the power target value, and the power difference is used as the upper limit of energy adjustment. At the same time, the adjustable upper limit of a single wind turbine generator set is calculated based on the feedback status of the single wind turbine generator set and the set power curve, and the target power value of the single wind turbine generator set is sent according to the upper limit.

[0088] This patent is based on the long short-term memory network and introduces the concept of deep residual network. By differentiating the deviation results of actual wind speed, power and theoretical power under different time values, it proposes the integration of actual power curves and active power control technology of wind turbines under time weighting according to time division, improves the accuracy of active power control of different wind turbines in field group control, and reduces the active power distribution error of wind turbines.

[0089] Based on this, please refer to Figure 2 As shown, the present invention can first obtain the theoretical power curve and initial time weight ratio of the wind turbine generator set;

[0090] Subsequently, the actual wind speed, pitch angle, generator speed, converter torque, active power and other operating parameters of the wind turbine generator set are obtained to calculate the actual power curve, where the length of time for obtaining data can be set according to the actual situation. For example, if the data is set to eight days, when the ninth day comes, the ninth day's data is used as the last day's data, and the first day's data is discarded.

[0091] The average power generation under different wind speed ranges = the total power generation recorded under the wind speed range / the number of times the wind speed range appears and is recorded.

[0092] Subsequently, the average power generation is used to learn and predict the power generation of the wind turbine in different wind speed ranges during the observation period according to the set time weight ratio.

[0093] The predicted power generation power of the observation section = (the average power generation power corresponding to different wind speed intervals * the time weight ratio a divided according to time) + (the average power generation power corresponding to different wind speed intervals * the time weight ratio b divided according to time) +..., where a+b+...=1.

[0094] Subsequently, the predicted power generation calculated using the initial time weight ratio is compared with the average power generation during actual operation to calculate the difference ratio.

[0095] Subsequently, when the difference ratio is lower than the current preset threshold, the predicted generated power calculated under the initial time weight ratio is used to allocate the target power.

[0096] If the difference ratio is higher than the current preset threshold, the time weight ratio at different times is optimized and adjusted according to the part exceeding the preset value.

[0097] Then, the above steps are performed again using the optimized time weights at different times until the difference ratio is less than a preset threshold.

[0098] The following uses a wind speed range (4m / s-4.5m / s) as an example for explanation. The following data are for illustration only and do not represent actual data:

[0099] 1. The theoretical power curve shows that the theoretical average power generation in this wind speed range is 200kW;

[0100] 2. Get the operating data of the previous 7 days (time is adjustable), divide the 7-day operating data into 4 time periods, set the initial time weight ratio for each time period (the initial time weight ratio is set according to the impact of time on subsequent situations), the first time period is the 7th to 4th day, with a time weight ratio of 0.1, the second time period is the 4th to 2nd day, with a time weight ratio of 0.3, the third time period is the 2nd to 0.5th day, with a time weight ratio of 0.4, and the fourth time period is the 0.5th to the present, with a time weight ratio of 0.2.

[0101] 3. The average power generation in the wind speed interval obtained in the first time period is 180kW, the average power generation in the wind speed interval obtained in the second time period is 210kW, the average power generation in the wind speed interval obtained in the third time period is 196kW, and the average power generation in the wind speed interval obtained in the fourth time period is 220kW. According to the time weight ratios in different time periods, the recent power generation capacity of the wind turbine generator set at the wind speed is = (180*0.1) + (210*0.3) + (196*0.4) + (220*0.2) = 203.4kW.

[0102] 4. When the wind turbine generator set is running in the current (today) observation period, the actual power in this wind speed range is 215kw, which is 15kW different from the initial theory and 11.6kW different from the time weight ratio. If the preset threshold is set to 5kW, the time weight ratio in different time periods will be adjusted (according to the idea that the closer the time is, the higher the time weight ratio is, increase or decrease by 0.1 each time, compare the adjusted power with the actual situation, and continue to increase or decrease the time weight ratio if the difference is greater than the preset threshold. If the difference is less than the threshold, the self-learning process is completed).

[0103] 5. The adjusted time weight ratio is 0.08, 0.15, 0.28, 0.5, and the power is 210.78KW. The difference is 4.22KW, which is less than the set 5kW. It is considered that the actual average power generation in this wind speed interval at this time is 210.78kW, and the different zones are summarized as a whole to form the actual operating power curve of the unit.

[0104] 6. When power dispatch is required, the actual operating power curve that has been adjusted and is less than the threshold is used to accurately allocate the power of each wind turbine generator set, thereby achieving precise control of the energy of the entire field.

[0105] The present invention also provides a full-field energy control method, including performing full-field energy control based on the predicted power generation of the wind turbine in the current processing cycle under the observation section obtained by a single-machine power prediction method as mentioned above, combined with the target power value issued by the host computer.

[0106] By adopting the predicted power generation power of the target observation section obtained by a single-machine power prediction method as mentioned above, the whole-field energy control method has higher accuracy and faster control speed during whole-field energy control.

[0107] The present invention also provides a single machine power prediction system, please refer to Figure 3 As shown, including:

[0108] A time weight processing module is used to set a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order and an observation period located after a preset time length of the last time period; and a time weight is assigned to each time period of the processing cycle;

[0109] The first execution model takes the previous processing cycle of the current processing cycle as the target processing cycle, and sends the target processing cycle to the average power generation acquisition module;

[0110] An average power generation acquisition module is used to acquire the average power generation of the wind turbine generator set in each time period within the target processing cycle and send it to the prediction module;

[0111] A prediction module, which obtains the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period, and sends it to the calibration module;

[0112] A calibration module, which obtains the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation. If the difference is greater than a preset threshold, the time weights are adjusted and the prediction module is returned until the difference is lower than the preset threshold and the second execution module is entered;

[0113] The second execution model sets the current processing cycle as a new target processing cycle and returns to the first execution model until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

[0114] Specifically, the sum of the time weights is 1, and the prediction module includes:

[0115] The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

[0116] Specifically, the average power generation acquisition module includes:

[0117] Obtain the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle;

[0118] For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval.

[0119] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0120] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0121] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0122] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A single machine power prediction method, characterized in that: include: Setting a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order, and an observation period located after a preset duration of the last time period; Assign a time weight to each time period of the processing cycle, and perform the following steps: S1, taking the previous processing cycle of the current processing cycle as the target processing cycle; S2. Obtaining the average power generation of the wind turbine generator set in each of the time periods within the target processing cycle; S3, obtaining the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period; S4, obtaining the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation, and if the difference is greater than a preset threshold, adjusting each of the time weights, and repeating steps S3 to S4 until the difference is lower than the preset threshold; S5. Take the current processing cycle as a new target processing cycle, and repeat steps S2 to S3 until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

2. A single machine power prediction method according to claim 1, characterized in that: The sum of the time weights is 1. According to the average power generation and the time weights corresponding to the time periods, the predicted power generation of the wind turbine in the observation period is obtained, including: The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

3. A single machine power prediction method according to claim 1, characterized in that: Step S2 includes: Obtain the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle; For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval.

4. A single machine power prediction method according to claim 3, characterized in that: Get the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle, including: For each time period of the target processing cycle, the ratio of the total power generation corresponding to different wind speed intervals to the number of occurrences of the corresponding different wind speed intervals is used as the average power generation corresponding to the different wind speed intervals.

5. A single machine power prediction method according to claim 3, characterized in that: The processing cycle before the current processing cycle is used as the target processing cycle, including: Based on the differentiated time value method, the previous processing cycle of the current processing cycle is used as the target processing cycle.

6. A single machine power prediction method according to any one of claims 1 to 5, characterized in that: The duration of each time period in the same processing cycle is sorted from long to short.

7. A method for controlling full-field energy, characterized in that: According to a single-machine power prediction method as described in any one of claims 1 to 6, the predicted power generation power of the wind turbine in the observation section of the current processing cycle is obtained, combined with the target power value issued by the host computer, to perform full-field energy control.

8. A single machine power prediction system, characterized in that: include: A time weight processing module is used to set a processing cycle, wherein the processing cycle includes a plurality of time periods arranged in chronological order and an observation period located after a preset time length of the last time period; and a time weight is assigned to each time period of the processing cycle; The first execution model takes the previous processing cycle of the current processing cycle as the target processing cycle, and sends the target processing cycle to the average power generation acquisition module; An average power generation acquisition module is used to acquire the average power generation of the wind turbine generator set in each time period within the target processing cycle and send it to the prediction module; A prediction module, which obtains the predicted power generation of the wind turbine in the observation period according to the average power generation and the time weight corresponding to each time period, and sends it to the calibration module; A calibration module, which obtains the difference between the actual power generation of the wind turbine in the observation section and the predicted power generation. If the difference is greater than a preset threshold, the time weights are adjusted and the prediction module is returned until the difference is lower than the preset threshold and the second execution module is entered; The second execution model sets the current processing cycle as a new target processing cycle and returns to the first execution model until the predicted power generation of the wind turbine in the current processing cycle under the observation section is obtained.

9. A single machine power prediction system according to claim 8, characterized in that: The sum of the time weights is 1, and the prediction module includes: The product of each of the average generated power and the time weight corresponding to each of the time periods is taken as the power impact value, and the sum of the power impact values ​​is taken as the predicted generated power of the wind turbine in the observation period.

10. A single machine power prediction system according to claim 8, characterized in that: The average power generation acquisition module includes: Obtain the average power generation corresponding to different wind speed ranges in each time period of the target processing cycle; For each of the wind speed intervals, subsequent steps are performed according to the average power generation in each time period corresponding to the wind speed interval.

Images