New energy wind and solar sample data processing method and device based on regularity constraints

Abstract

The present application relates to the field of energy data processing technology, and in particular to a method and device for processing new energy wind and solar sample data based on regularity constraints, wherein the method comprises: establishing a regular interpolation model to interpolate and supplement data missing that are prone to occur in new energy wind power and photovoltaic systems in complex environments; establishing a data quality assessment model based on a topologically aware local reversible attention mechanism to automatically assess the quality of data generated by different interpolation methods; establishing a dynamic weight fusion model to automatically adjust the weights of each data source in the data completion process, thereby optimizing the data repair effect; and establishing an abnormal data removal model to improve the integrity and reliability of new energy data. Thus, the present application solves the problems that the existing technology has poor generalization in dealing with complex and changeable severe weather conditions, is difficult to effectively balance data integrity and accuracy, and cannot meet the needs of new energy power generation systems for high-quality data.

Description

Technical Field

[0001] The present application relates to the technical field of energy data processing, and in particular to a method and device for processing new energy wind and solar sample data based on regularity constraints. Background Art

[0002] As renewable energy generation continues to expand, severe weather events are increasingly impacting data collected from wind and solar power systems. Furthermore, data collection equipment failures can lead to missing data or the generation of outliers, posing significant challenges to system stability and data quality. Accurately processing and completing missing data, as well as accurately identifying outliers, are crucial to ensuring the reliability and optimal operation of renewable energy power generation systems. Renewable energy generation data is significantly affected by external environmental factors such as light intensity, wind speed, and temperature. This is particularly true in severe weather, which can exacerbate power generation fluctuations.

[0003] In existing technology systems, the processing of new energy data usually relies on simple interpolation or correction methods based on physical models. Although traditional physical model methods can accurately model the physical characteristics and performance parameters of power generation equipment, their generalization capabilities are limited when dealing with complex and changeable severe weather conditions. Especially in severe weather conditions, the correction accuracy of the physical model will drop significantly; simple interpolation methods often find it difficult to effectively balance data integrity and accuracy when processing time series data. In addition, traditional methods have difficulty accurately identifying abnormal data that is significantly affected by the external environment, and cannot meet the needs of new energy power generation systems for high-quality data.

[0004] In summary, existing technologies have poor generalization capabilities in dealing with complex and changeable severe weather conditions, making it difficult to effectively balance data integrity and accuracy, and unable to meet the needs of new energy power generation systems for high-quality data, which urgently needs to be addressed. Summary of the Invention

[0005] The present application provides a method and device for processing new energy wind and solar sample data based on regular constraints to solve the problems that the existing technology has poor generalization in dealing with complex and changeable severe weather conditions, is difficult to effectively balance data integrity and accuracy, and cannot meet the needs of new energy power generation systems for high-quality data.

[0006] The first aspect of the present application provides a method for processing new energy wind and solar sample data based on regularity constraints, comprising the following steps: obtaining new energy wind and solar sample data corresponding to target wind power and photovoltaic systems, and performing regularity completion operations on the new energy wind and solar sample data based on a preset regular interpolation strategy to obtain corresponding wind and solar sample completion data; extracting multi-scale features of the wind and solar sample completion data based on a pre-built local reversible attention mechanism evaluator, and optimizing the wind and solar sample completion data through the multi-scale features to obtain corresponding first wind and solar sample optimization data; determining at least one data missing category in the wind and solar sample optimization data, and adjusting the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data; obtaining abnormal data that meets preset abnormality requirements in the second wind and solar sample optimization data, and eliminating the abnormal data using a pre-built outlier data elimination model to obtain target new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems.

[0007] Optionally, in one embodiment of the present application, the new energy wind and solar sample data corresponding to the target wind power and photovoltaic system is obtained, and based on a preset rule interpolation strategy, a regular completion operation is performed on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data, including: determining the generalized missing window corresponding to the new energy wind and solar sample data based on a preset sampling time length, sensor effective data range, environmental parameter vector and severe weather threshold; judging the data missing type of the new energy wind and solar sample data based on the generalized missing window; if the data missing type is a single point missing type, calculating the adjacent point average value and time series prediction value corresponding to the missing point, and determining the corresponding single point completion value based on the adjacent point average value and the time series prediction value, so as to use the single point completion value to complete the new energy wind and solar sample data. The single-point missing completion operation is performed on the source wind and solar sample data; if the data missing type is an odd-numbered point missing type, the K-nearest neighbor prediction function, wind power environmental parameter set and photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data are determined, and the odd-numbered point missing completion operation is performed on the new energy wind and solar sample data according to the K-nearest neighbor prediction function, the wind power environmental parameter set and the photovoltaic historical data feature matrix; if the data missing type is an even-numbered point missing type, the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature corresponding to the new energy wind and solar sample data are determined, and the even-numbered point missing completion operation is performed on the new energy wind and solar sample data according to the photovoltaic environmental parameter set, the wind power weight smoothing coefficient, the temperature coefficient and the power curve curvature.

[0008] Optionally, in one embodiment of the present application, the determining of at least one data missing category in the wind and light sample optimization data, and adjusting the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain the corresponding second wind and light sample optimization data, includes: determining the parity compensation item parameters corresponding to the odd point missing type and the even point missing type; calculating the maximum allowable missing length corresponding to the wind and light sample optimization data, and constructing a target interpolation formula based on the maximum allowable missing length and the parity compensation item parameters, so as to use the target interpolation formula to adjust the weight corresponding to each data missing category to obtain the second wind and light sample optimization data.

[0009] Optionally, in one embodiment of the present application, the abnormal data that meets the preset abnormal requirements in the second wind-solar sample optimization data is obtained, and the abnormal data is eliminated by using a pre-built abnormal value data elimination model to obtain the target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system, including: determining the sliding window corresponding to each data point in the second wind-solar sample optimization data, and obtaining the real-time wind speed and historical average wind speed corresponding to the target wind power and photovoltaic system, so as to calculate the dynamic threshold corresponding to each data point through the real-time wind speed and the historical average wind speed; based on the preset wind power data abnormal elimination rule, judging whether each data point exceeds the dynamic threshold, if the current If a data point exceeds the dynamic threshold, the current data point is set to 1, otherwise the current data point is set to 0; the photovoltaic panel conversion efficiency, horizontal total radiation and temperature attenuation coefficient corresponding to the target wind power and photovoltaic system are obtained, and a theoretical output benchmark is established based on the photovoltaic panel conversion efficiency, the horizontal total radiation and the temperature attenuation coefficient; the daytime statistics are calculated according to the theoretical output benchmark, and the corresponding dual-threshold constraints are constructed using the daytime statistics; the photovoltaic data anomaly elimination rule is determined through the dual-threshold constraint and the theoretical output benchmark, so as to eliminate the abnormal data in the second wind-solar sample optimization data according to the photovoltaic data anomaly elimination rule, so as to obtain the target new energy wind-solar sample data.

[0010] Optionally, in one embodiment of the present application, the target interpolation formula is:

[0011]

[0012] in, represents the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that based on phys interpolation method t Estimated value of the moment; 、n Indicates the amount of missing data; Represents the angle parameter; Represents a linear interpolation function.

[0013] A second aspect of the present application provides a new energy wind and solar sample data processing device based on regularity constraints, including: an interpolation and completion module, configured to obtain new energy wind and solar sample data corresponding to target wind power and photovoltaic systems, and perform regularity completion operations on the new energy wind and solar sample data based on a preset regular interpolation strategy to obtain corresponding wind and solar sample completion data; a feature optimization module, configured to extract multi-scale features of the wind and solar sample completion data based on a pre-built local reversible attention mechanism evaluator, and optimize the wind and solar sample completion data through the multi-scale features to obtain corresponding first wind and solar sample optimization data; a weight adjustment module, configured to determine at least one data missing category in the wind and solar sample optimization data, and adjust the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data; a data elimination module, configured to obtain abnormal data that meets preset abnormality requirements in the second wind and solar sample optimization data, and eliminate the abnormal data using a pre-built outlier data elimination model to obtain target new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems.

[0014] Optionally, in one embodiment of the present application, the interpolation completion module includes: a first determination unit, for determining the generalized missing window corresponding to the new energy wind and solar sample data based on a preset sampling time length, a sensor valid data range, an environmental parameter vector and a severe weather threshold; a first judgment unit, for judging the data missing type of the new energy wind and solar sample data based on the generalized missing window; a single point missing unit, for calculating the average value and time series prediction value of the adjacent points corresponding to the missing point if the data missing type is a single point missing type, and determining the corresponding single point completion value based on the average value and time series prediction value of the adjacent points, so as to perform a single point missing completion operation on the new energy wind and solar sample data using the single point completion value; an odd point missing unit, for determining the data missing type if the data missing type is a single point missing type If the data missing type is an odd point missing type, the K nearest neighbor prediction function, wind power environment parameter set and photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data are determined, and the odd point missing completion operation is performed on the new energy wind and solar sample data according to the K nearest neighbor prediction function, the wind power environment parameter set and the photovoltaic historical data feature matrix; an even point missing unit is used to determine the photovoltaic environment parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature corresponding to the new energy wind and solar sample data if the data missing type is an even point missing type, and perform an even point missing completion operation on the new energy wind and solar sample data according to the photovoltaic environment parameter set, the wind power weight smoothing coefficient, the temperature coefficient and the power curve curvature.

[0015] Optionally, in one embodiment of the present application, the weight adjustment module includes: a second determination unit, used to determine the parity compensation item parameters corresponding to the odd point missing type and the even point missing type; a construction unit, used to calculate the maximum allowable missing length corresponding to the wind and light sample optimization data, and construct a target interpolation formula based on the maximum allowable missing length and the parity compensation item parameters, so as to use the target interpolation formula to adjust the weight corresponding to each data missing category to obtain the second wind and light sample optimization data.

[0016] Optionally, in one embodiment of the present application, the data elimination module includes: a first calculation unit, used to determine the sliding window corresponding to each data point in the second wind-solar sample optimization data, and obtain the real-time wind speed and historical average wind speed corresponding to the target wind power and photovoltaic system, so as to calculate the dynamic threshold corresponding to each data point through the real-time wind speed and the historical average wind speed; a second judgment unit, used to judge whether each data point exceeds the dynamic threshold based on a preset wind power data anomaly elimination rule, if the current data point exceeds the dynamic threshold, the current data point is set to 1, otherwise the current data point is set to 0; an acquisition unit , used to obtain the photovoltaic panel conversion efficiency, horizontal total radiation and temperature attenuation coefficient corresponding to the target wind power and photovoltaic system, and establish a theoretical output benchmark based on the photovoltaic panel conversion efficiency, the horizontal total radiation and the temperature attenuation coefficient; the second calculation unit is used to calculate the daytime statistics according to the theoretical output benchmark, and use the daytime statistics to construct the corresponding dual-threshold constraint; the third determination unit is used to determine the photovoltaic data anomaly elimination rule through the dual-threshold constraint and the theoretical output benchmark, so as to eliminate the abnormal data in the second wind-solar sample optimization data according to the photovoltaic data anomaly elimination rule, so as to obtain the target new energy wind-solar sample data.

[0017] Optionally, in one embodiment of the present application, the target interpolation formula is:

[0018]

[0019] in, represents the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that based on phys interpolation method t Estimated value of the moment; 、 n Indicates the amount of missing data; Represents the angle parameter; Represents a linear interpolation function.

[0020] The third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the new energy wind and solar sample data processing method based on regularity constraints as described in the above embodiment.

[0021] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned new energy wind and solar sample data processing method based on regularity constraints.

[0022] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned new energy wind and solar sample data processing method based on regularity constraints.

[0023] Therefore, the embodiments of the present application have the following beneficial effects:

[0024] The embodiments of the present application can obtain new energy wind and solar sample data corresponding to target wind power and photovoltaic systems, and perform regular complementation operations on the new energy wind and solar sample data based on a preset regular interpolation strategy to obtain corresponding wind and solar sample complement data; extract multi-scale features of the wind and solar sample complement data based on a pre-built local reversible attention mechanism evaluator, and optimize the wind and solar sample complement data using the multi-scale features to obtain corresponding first wind and solar sample optimization data; determine at least one data missing category in the wind and solar sample optimization data, and adjust the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data; obtain abnormal data that meets preset abnormality requirements in the second wind and solar sample optimization data, and use a pre-built outlier data elimination model to eliminate abnormal data to obtain target new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems, thereby improving the integrity and reliability of the new energy data and improving the operating efficiency of the new energy system. Thus, the present application solves the problems that the existing technology has poor generalization in dealing with complex and changeable severe weather conditions, is difficult to effectively balance data integrity and accuracy, and cannot meet the needs of new energy power generation systems for high-quality data.

[0025] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0027] Figure 1 Flowchart of a method for processing new energy wind and solar sample data based on regularity constraints according to an embodiment of the present application;

[0028] Figure 2 A schematic diagram of the execution logic of a method for processing new energy wind and solar sample data based on regularity constraints provided in one embodiment of the present application;

[0029] Figure 3 This is an example diagram of a new energy wind and solar sample data processing device based on regularity constraints according to an embodiment of the present application;

[0030] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0031] Among them, 10-new energy wind and solar sample data processing device based on regularity constraints; 100-interpolation completion module, 200-feature optimization module, 300-weight adjustment module, 400-data elimination module; 401-memory, 402-processor, 403-communication interface. DETAILED DESCRIPTION

[0032] The following describes in detail embodiments of the present application, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0033] The following describes a method and device for processing new energy wind and solar sample data based on regularity constraints according to an embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides a new energy wind and solar sample data processing method based on regularity constraints. In this method, by obtaining new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, and based on a preset regular interpolation strategy, performing regularity completion operations on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data; based on a pre-built local reversible attention mechanism evaluator, extracting multi-scale features of the wind and solar sample completion data, and optimizing the wind and solar sample completion data through multi-scale features to obtain corresponding first wind and solar sample optimization data; determining at least one data missing category in the wind and solar sample optimization data, and based on a preset dynamic weight fusion strategy, adjusting the weight corresponding to each data missing category in at least one data missing category to obtain corresponding second wind and solar sample optimization data; obtaining abnormal data that meets the preset abnormality requirements in the second wind and solar sample optimization data, and using a pre-built outlier data elimination model to eliminate the abnormal data to obtain the target new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, thereby improving the integrity and reliability of the new energy data and improving the operating efficiency of the new energy system. This solves the problems of existing technologies such as poor generalization in dealing with complex and changeable severe weather conditions, difficulty in effectively balancing data integrity and accuracy, and inability to meet the needs of new energy power generation systems for high-quality data.

[0034] Specifically, Figure 1 This is a flowchart of a method for processing new energy wind and solar sample data based on regularity constraints provided in an embodiment of the present application.

[0035] like Figure 1As shown, the new energy wind and solar sample data processing method based on regularity constraints includes the following steps:

[0036] In step S101, new energy wind and solar sample data corresponding to the target wind power and photovoltaic system is obtained, and based on a preset rule interpolation strategy, a regular completion operation is performed on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data.

[0037] The embodiment of the present application can first obtain new energy wind and solar sample data corresponding to wind power and photovoltaic systems under severe weather conditions, and based on the rule interpolation method, and according to historical data and data patterns under similar weather conditions, regularly fill in the missing data in the wind power and photovoltaic systems, thereby ensuring the continuity and consistency of the data.

[0038] Optionally, in one embodiment of the present application, new energy wind and solar sample data corresponding to the target wind power and photovoltaic system is obtained, and based on a preset rule interpolation strategy, a regular completion operation is performed on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data, including: determining the generalized missing window corresponding to the new energy wind and solar sample data based on a preset sampling time length, sensor effective data range, environmental parameter vector and severe weather threshold; judging the data missing type of the new energy wind and solar sample data based on the generalized missing window; if the data missing type is a single point missing type, calculating the adjacent point average value and time series prediction value corresponding to the missing point, and determining the corresponding single point completion value based on the adjacent point average value and time series prediction value, so as to utilize the single point completion value Perform single-point missing completion operations on the new energy wind and solar sample data; if the data missing type is an odd-numbered point missing type, determine the K-nearest neighbor prediction function, wind power environmental parameter set, and photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data, and perform odd-numbered point missing completion operations on the new energy wind and solar sample data based on the K-nearest neighbor prediction function, wind power environmental parameter set, and photovoltaic historical data feature matrix; if the data missing type is an even-numbered point missing type, determine the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient, and power curve curvature corresponding to the new energy wind and solar sample data, and perform even-numbered point missing completion operations on the new energy wind and solar sample data based on the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient, and power curve curvature.

[0039] In the actual implementation process, the embodiment of the present application can first determine the mathematical expressions corresponding to the wind turbine and photovoltaic in the wind power and photovoltaic system, wherein the mathematical expression of the wind turbine is:

[0040] (1)

[0041] The mathematical expression for photovoltaics is:

[0042] (2)

[0043] in, Indicates the electric power output by the fan; Indicates the air density; Indicates the conversion efficiency of photovoltaic cells; Indicates the power coefficient of the wind turbine; It represents the area of ​​the circle covered by the rotation of the fan blades; Indicates the effective area of ​​the photovoltaic panel that receives sunlight; Indicates the electrical power output by the photovoltaic cell; Indicates the solar radiation power received per unit area; Represents the cube of wind speed (the wind turbine output power increases in a cubic relationship with wind speed); Represents the cosine of the sun's incidence angle.

[0044] Secondly, to address the data loss problem caused by severe weather, the present embodiment can use multiple regular interpolation methods to complete the data and evaluate the effects of different interpolation methods through a convolutional neural network (local reversible attention mechanism). The specific steps are as follows:

[0045] For isolated point missing window , the embodiment of the present application may introduce a dynamic hybrid interpolation strategy, as shown in the following formula:

[0046]

[0047] Among them, the parameter adaptation mechanism can be defined as:

[0048] (4)

[0049] Where, is the Sigmoid function; represents a T-dimensional real vector space; is the ARIMA model parameter matrix; Represents the adaptive weight coefficient, which is used to dynamically adjust the contribution ratio of simple interpolation and ARIMA model prediction; It represents the variance within the local window; Represents historical data; Represents a time series at time t and the local data window vector composed of the observation values ​​at adjacent moments; Represents a time series at time t The predicted value of Represents the time series at time t -1 observations; Represents the time series at time t +1 observations; Indicates function operation on the ARIMA model parameter matrix.

[0050] 1. Missing window definition

[0051] In this embodiment of the application, the new energy time series (i.e., new energy wind and solar sample data) can be set as ,in For the The data of the sensors, Denote the sensor number and define the generalized missing window:

[0052] (5)

[0053] Where, is the sampling time length; For the Valid data range of each sensor; is the environmental parameter vector; is the severe weather threshold; Indicates the k The valid lower limit of sensor data; Indicates the k The effective upper limit of sensor data; Indicates the k Sensors at time t wind speed; Indicates the k Sensors at time t irradiance; Indicates the maximum wind speed threshold, used to determine whether the wind speed is too high; Indicates the minimum irradiance threshold, used to determine whether the irradiance is too low.

[0054] 2. Single point missing completion (i.e. cardinality missing):

[0055] Applicable scenario: isolated point missing

[0056] The interpolation method is shown below:

[0057] (6)

[0058] in, Wind power data in time The completion value of t -1) is the wind power data at the previous moment; t +1) is the wind power data at the next moment; Represents the adaptive weight coefficient, which is used to dynamically adjust the contribution ratio of simple interpolation and ARIMA model prediction; Indicates that the ARIMA model is t The wind power data forecast value.

[0059] 3. Continuous missing completion:

[0060] (1) Odd number of points missing: ( is an odd number, and m and n are used to locate the double index of the data point);

[0061] Applicable scenario: 3 points missing in a row ( )

[0062] The interpolation method is as follows:

[0063] (7)

[0064] in, represents the K nearest neighbor prediction function based on photovoltaic data; Represents a set of photovoltaic environmental parameters (such as irradiance, conversion efficiency, etc.); Represents the photovoltaic historical data feature matrix; Represents the photovoltaic prediction results; Represents photovoltaic power generation power.

[0065] (2) Even-numbered points missing: ( is an even number)

[0066] Applicable scenario: 2 points missing continuously ( )

[0067] The interpolation strategy is as follows:

[0068] (8)

[0069] The weight distribution rule is as follows:

[0070] (9)

[0071] in, is the second-order derivative of the wind power interpolation function with respect to wind speed, representing the curvature of the power curve; Represents wind power physical parameters (wind speed, air density, wind turbine swept area); is the second-order derivative of the wind power interpolation function with respect to wind speed; is the wind power weight smoothing coefficient, which is related to the dynamic range of wind speed; represents the second-order derivative of the photovoltaic interpolation function with respect to the irradiation intensity; is a set of photovoltaic environmental parameters.

[0072] In step S102, based on a pre-built local reversible attention mechanism evaluator, multi-scale features of the scenery sample completion data are extracted, and the scenery sample completion data is optimized through the multi-scale features to obtain corresponding first scenery sample optimized data.

[0073] Furthermore, in view of the advantages and disadvantages of different interpolation methods, the embodiment of the present application can construct a topological structure-based local reversible attention mechanism evaluator to extract multi-scale features from the scenery sample completion data, and optimize the scenery sample completion data through multi-scale features to obtain the corresponding first scenery sample optimization data, thereby improving the reliability of the interpolation result. The specific steps are as follows:

[0074] 1. Input feature construction:

[0075] First, construct the wind power input tensor and photovoltaic input tensor, where the wind power input tensor is shown as follows:

[0076]

[0077] Where, the channel dimension Include: , ; Indicates the length of wind power data; represents the wind power generation at time T; express t Wind power generation at time; Represents the total number of time units in a year.

[0078] The photovoltaic input tensor is shown below:

[0079] ,

[0080] Where, the same dimension Include: , ; express t PV power generation at time; Represents the photovoltaic power generation at time T.

[0081] 2. Multi-scale local reversible attention mechanism structure:

[0082] The channel convolution formula is shown below:

[0083] (10)

[0084] in, i and j Represents the height and width (i.e. spatial position) of the output feature map; m andn Represents the spatial size of the convolution kernel (different from the interpretation in Equation (13) and Equation (14)); p The padding size is used to control the size of the output feature map; is the convolution kernel size; is the input channel index; Represents the channel-by-channel convolution bias;

[0085] The point-by-point convolution formula is determined by the above channel convolution formula:

[0086] (11)

[0087] in, Indicates the number of input channels; Represents the point-by-point convolution bias; Represents the weight of point-by-point convolution; c and k Indicates the channel index.

[0088] Integrating the channel convolution formula and the point-by-point convolution formula, we can get the expression of the multi-scale local reversible attention mechanism:

[0089] (12)

[0090] in, represents the attention mask matrix; represents the weight matrix of the attention mechanism, a and b Indicates the size of the convolution kernel; c Indicates the number of channels; p Indicates the padding size.

[0091] Therefore, the topologically based local reversible attention mechanism evaluator of the embodiment of the present application combines the spatial and temporal characteristics of wind power and photovoltaic system data, which can deeply explore the potential patterns in the data and further improve the accuracy and effect of data completion.

[0092] In step S103, at least one data missing category in the wind and solar sample optimization data is determined, and based on a preset dynamic weight fusion strategy, the weight corresponding to each data missing category in the at least one data missing category is adjusted to obtain corresponding second wind and solar sample optimization data.

[0093] Afterwards, the embodiment of the present application can adjust the weight corresponding to each data missing category based on the preset dynamic weight fusion strategy to obtain the corresponding second wind and light sample optimization data, thereby ensuring the rationality of the interpolation data.

[0094] Optionally, in one embodiment of the present application, at least one data missing category in the wind and solar sample optimization data is determined, and based on a preset dynamic weight fusion strategy, the weight corresponding to each data missing category in at least one data missing category is adjusted to obtain the corresponding second wind and solar sample optimization data, including: determining the parity compensation item parameters corresponding to the odd point missing type and the even point missing type; calculating the maximum allowable missing length corresponding to the wind and solar sample optimization data, and constructing a target interpolation formula based on the maximum allowable missing length and the parity compensation item parameters, so as to use the target interpolation formula to adjust the weight corresponding to each data missing category to obtain the second wind and solar sample optimization data.

[0095] Specifically, the process of adaptively adjusting weight distribution according to the data missing class in the embodiment of the present application is as follows:

[0096] 1. Compensation coefficient for missing base:

[0097] First, the embodiment of the present application may define a parity compensation term as shown in the following formula:

[0098] (13)

[0099] in, Represents the target parameter, when (i.e., symmetry is missing), ; When it is asymmetric missing, ; k is a constant that can be used to adjust the slope of the exponential function.

[0100] 2. Construct the final interpolation formula (i.e. target interpolation formula):

[0101] Secondly, the embodiments of the present application can determine the final interpolation formula based on the cardinality loss compensation coefficient.

[0102] Optionally, in one embodiment of the present application, the target interpolation formula is:

[0103]

[0104] in, Indicates the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that the interpolation method is based on phys (i.e., interpolation based on different physical models). t Estimated value of the moment; 、 n Indicates the amount of missing data; θ Represents the angle parameter; Represents a linear interpolation function.

[0105] In an embodiment of the present application, the final interpolation formula determined based on the cardinality loss compensation coefficient is as follows:

[0106] (14)

[0107] in, Indicates the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that based on phys interpolation method t Estimated value of the moment; 、 n Indicates the amount of missing data; , which represents the angle parameter, fixed at 90°, used for orthogonal fusion and ; Represents a linear interpolation function.

[0108] 3. Adaptive threshold control:

[0109] Calculation of the maximum allowable missing length:

[0110] (15)

[0111] Among them, for new energy data, the embodiment of this application can take , so .

[0112] Therefore, the embodiment of the present application adopts a dynamic weight fusion strategy to automatically adjust the weight of the data source according to different environmental conditions to obtain the corresponding second wind and light sample optimization data, so as to be able to flexibly respond to complex weather changes and system fluctuations, and ensure that the data repair process is more accurate and effective.

[0113] In step S104, abnormal data that meets preset abnormal requirements is obtained from the second wind-solar sample optimization data, and the abnormal data is eliminated using a pre-built outlier data elimination model to obtain target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system.

[0114] Furthermore, the embodiments of the present application can establish an abnormal data elimination model to eliminate abnormal data in the second wind-solar sample optimization data through the outlier data elimination model to obtain new energy wind-solar sample data, thereby improving the integrity and reliability of the new energy data.

[0115] Therefore, the embodiment of the present application removes abnormal data through an improved filter, thereby ensuring that the data conforms to actual physical laws, avoiding the negative impact of data anomalies on system performance, and effectively improving the prediction accuracy and stability of the new energy system.

[0116] Optionally, in one embodiment of the present application, abnormal data that meets preset abnormal requirements is obtained from the second wind-solar sample optimization data, and the abnormal data is eliminated using a pre-built outlier data elimination model to obtain target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system, including: determining a sliding window corresponding to each data point in the second wind-solar sample optimization data, and obtaining the real-time wind speed and historical average wind speed corresponding to the target wind power and photovoltaic system, so as to calculate the dynamic threshold corresponding to each data point through the real-time wind speed and the historical average wind speed; based on the preset wind power data abnormal elimination rule, judging whether each data point exceeds the dynamic threshold, such as If the current data point exceeds the dynamic threshold, the current data point is set to 1, otherwise the current data point is set to 0; the photovoltaic panel conversion efficiency, horizontal total radiation and temperature attenuation coefficient corresponding to the target wind power and photovoltaic system are obtained, and based on the photovoltaic panel conversion efficiency, horizontal total radiation and temperature attenuation coefficient, a theoretical output benchmark is established; the daytime statistics are calculated according to the theoretical output benchmark, and the corresponding dual-threshold constraints are constructed using the daytime statistics; the photovoltaic data anomaly elimination rule is determined through the dual-threshold constraint and the theoretical output benchmark, so as to eliminate the abnormal data in the second wind and solar sample optimization data according to the photovoltaic data anomaly elimination rule, so as to obtain the target new energy wind and solar sample data.

[0117] In the actual implementation process, the embodiment of the present application can adopt corresponding abnormal data elimination strategies for wind power and photovoltaic power, and build an outlier data elimination model. For example, wind power adopts the improved Hampel filter method, and photovoltaic power adopts the dynamic power constraint method to eliminate abnormal data, as described below:

[0118] 1. Abnormal elimination of wind power data based on improved Hampel filter:

[0119] First, wind power data is affected by multiple physical parameters. In this embodiment of the application, a wind power calculation formula can be defined as follows:

[0120] (16)

[0121] in, is the fan power coefficient; is the air density; is the swept area of ​​the wind wheel, is the real-time wind speed; is the fan efficiency.

[0122] Secondly, the embodiment of the present application can set the wind power historical data as , and use the improved Hampel filter to detect anomalies in wind power historical data, as shown in the following formula:

[0123] (17)

[0124] in, is the historical mean; is the standard deviation; is the empirical threshold; Indicates the threshold of the Hampel filter.

[0125] It should be noted that if is the historical wind power data at time t, then Abnormal data are removed.

[0126] 1) Sliding window definition:

[0127] In the embodiment of the present application, the time series can be set as , and define the sliding window as:

[0128] (18)

[0129] Among them, the window length is 5% of the total data volume; is the maximum length of the time series.

[0130] 2) Dynamic threshold calculation:

[0131] In the embodiment of the present application, the calculation expression of the dynamic threshold is as follows:

[0132] (19)

[0133] in, Indicates time t the median; Indicates time t dataset; Indicates time t The mean absolute deviation of Indicates time t Dynamic wind speed threshold.

[0134] It should be noted that the embodiment of the present application may introduce a wind speed correlation coefficient adjustment threshold, as shown in the following formula:

[0135] (20)

[0136] in, is the real-time wind speed; is the historical average wind speed; Indicates the maximum wind speed; Indicates time t The mean absolute deviation of Indicates the adjusted threshold.

[0137] 3) Elimination rules:

[0138] In the embodiment of the present application, the mathematical expression of the rejection rule based on the improved Hampel filter (i.e., the wind power data abnormality rejection rule) is:

[0139]

[0140] In the actual implementation process, the embodiment of the present application can construct a sliding window for each data point in the second wind and light sample optimization data to calculate the robust statistics and , and determine whether the data point exceeds the dynamic threshold range.

[0141] 2. Eliminate abnormal photovoltaic data based on the dynamic power constraint method:

[0142] In the embodiment of the present application, the calculation expression of photovoltaic power is:

[0143] (twenty two)

[0144] in, is the photovoltaic panel conversion efficiency; is the solar radiation intensity; is the photovoltaic panel area; is the angle of incidence.

[0145] The mathematical expression for abnormal data removal is:

[0146] (twenty three)

[0147] 1) Theoretical power modeling: The present embodiment can establish a theoretical output benchmark based on photovoltaic physical characteristics, as shown in the following formula:

[0148] (twenty four)

[0149] in, is the photovoltaic panel conversion efficiency (calibrated value); Indicates total horizontal radiation; meteorological data represents the solar incidence angle; , which represents the temperature attenuation coefficient; Indicates the reference temperature; Indicates the current temperature.

[0150] 2) Abnormality determination criteria:

[0151] In this embodiment of the application, a dual threshold constraint may be defined:

[0152] (25)

[0153] The calculation expression of daily statistics is:

[0154] (26)

[0155] in, Indicates the theoretical photovoltaic output power; represents the daily statistics; Indicates standard deviation; Indicates the threshold value of photovoltaic power generation; Indicates the length of the time series used to calculate the average, that is, how many time points are used to calculate the average; The index variable for summation, used to traverse each time point during the summation process.

[0156] 3) Elimination rules:

[0157] In the embodiment of the present application, the mathematical expression of the elimination rule based on the dynamic power constraint method (ie, photovoltaic data abnormality elimination rule) is:

[0158] (27)

[0159] in, Represents an outlier judgment function, used to judge whether The actual measured value Whether it is an outlier.

[0160] Therefore, the embodiments of the present application can effectively handle the missing and anomaly problems of small sample data of new energy wind and solar power under severe weather conditions by integrating multiple technical means such as regularity completion, topology-aware local reversible attention mechanism evaluator, dynamic weight fusion and physical constraint anomaly elimination, thereby significantly improving the accuracy and robustness of data repair.

[0161] The following describes the execution logic of the new energy wind and solar sample data processing method based on regularity constraints of the present application in conjunction with the accompanying drawings.

[0162] Figure 2 This is a schematic diagram of the execution logic of the new energy wind and solar sample data processing method based on regularity constraints of this application. Figure 2 As shown, the execution steps of the new energy wind and solar sample data processing method based on regularity constraints of this application are as follows:

[0163] S201: Establish a regular method to complete missing data under severe weather conditions and interpolate to complete missing small sample data;

[0164] S202: Establish a topology-based local reversible attention mechanism evaluator to improve the reliability of interpolation results through multi-scale feature extraction;

[0165] S203: Establish a dynamic weight fusion strategy to adaptively adjust weight distribution according to the type of data missing;

[0166] S204: Establish an outlier data elimination model to eliminate outliers by classifying wind power and photovoltaic data respectively.

[0167] According to the new energy wind and solar sample data processing method based on regularity constraints proposed in the embodiment of the present application, by obtaining the new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, and based on a preset regular interpolation strategy, performing a regularity completion operation on the new energy wind and solar sample data to obtain the corresponding wind and solar sample completion data; based on a pre-built local reversible attention mechanism evaluator, extracting the multi-scale features of the wind and solar sample completion data, and optimizing the wind and solar sample completion data through the multi-scale features to obtain the corresponding first wind and solar sample optimization data; determining at least one data missing category in the wind and solar sample optimization data, and based on a preset dynamic weight fusion strategy, adjusting the weight corresponding to each data missing category in at least one data missing category to obtain the corresponding second wind and solar sample optimization data; obtaining abnormal data that meets the preset abnormality requirements in the second wind and solar sample optimization data, and using a pre-built outlier data elimination model to eliminate the abnormal data to obtain the target new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, thereby improving the integrity and reliability of the new energy data and improving the operating efficiency of the new energy system.

[0168] Next, a new energy wind and solar sample data processing device based on regularity constraints proposed in an embodiment of the present application will be described with reference to the accompanying drawings.

[0169] Figure 3 It is a block diagram of a new energy wind and solar sample data processing device based on regularity constraints in an embodiment of the present application.

[0170] like Figure 3 As shown, the new energy wind and solar sample data processing device 10 based on regularity constraints includes: an interpolation and completion module 100, a feature optimization module 200, a weight adjustment module 300 and a data elimination module 400.

[0171] Among them, the interpolation and completion module 100 is used to obtain the new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, and based on the preset regular interpolation strategy, perform regular completion operations on the new energy wind and solar sample data to obtain the corresponding wind and solar sample completion data.

[0172] The feature optimization module 200 is used to extract multi-scale features of the scenery sample completion data based on a pre-built local reversible attention mechanism evaluator, and optimize the scenery sample completion data through the multi-scale features to obtain corresponding first scenery sample optimized data.

[0173] The weight adjustment module 300 is used to determine at least one data missing category in the wind and solar sample optimization data, and adjust the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data.

[0174] The data elimination module 400 is used to obtain abnormal data that meets the preset abnormal requirements in the second wind-solar sample optimization data, and eliminate the abnormal data using a pre-built outlier data elimination model to obtain the target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system.

[0175] Optionally, in one embodiment of the present application, the interpolation and completion module 100 includes: a first determination unit, a first judgment unit, a single point missing unit, an odd point missing unit, and an even point missing unit.

[0176] Among them, the first determination unit is used to determine the generalized missing window corresponding to the new energy wind and solar sample data based on the preset sampling time length, sensor valid data range, environmental parameter vector and severe weather threshold.

[0177] The first judgment unit is used to judge the data missing type of the new energy wind and solar sample data based on the generalized missing window.

[0178] The single point missing unit is used to calculate the adjacent point average value and time series prediction value corresponding to the missing point if the data missing type is a single point missing type, and determine the corresponding single point completion value based on the adjacent point average value and time series prediction value, so as to use the single point completion value to perform single point missing completion operation on the new energy wind and solar sample data.

[0179] The odd point missing unit is used to determine the K nearest neighbor prediction function, wind power environment parameter set and photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data if the data missing type is an odd point missing type, and perform odd point missing completion operations on the new energy wind and solar sample data based on the K nearest neighbor prediction function, wind power environment parameter set and photovoltaic historical data feature matrix.

[0180] The even point missing unit is used to determine the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature corresponding to the new energy wind and solar sample data if the data missing type is the even point missing type, and perform the even point missing completion operation on the new energy wind and solar sample data according to the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature.

[0181] Optionally, in one embodiment of the present application, the weight adjustment module 300 includes: a second determining unit and a constructing unit.

[0182] The second determining unit is configured to determine parity compensation item parameters corresponding to the odd point loss type and the even point loss type.

[0183] A construction unit is used to calculate the maximum allowable missing length corresponding to the wind and solar sample optimization data, and to construct a target interpolation formula based on the maximum allowable missing length and the parity compensation item parameters, so as to use the target interpolation formula to adjust the weight corresponding to each data missing category to obtain the second wind and solar sample optimization data.

[0184] Optionally, in one embodiment of the present application, the data elimination module 400 includes: a first calculation unit, a second judgment unit, an acquisition unit, a second calculation unit and a third determination unit.

[0185] Among them, the first calculation unit is used to determine the sliding window corresponding to each data point in the second wind-solar sample optimization data, and obtain the real-time wind speed and historical average wind speed corresponding to the target wind power and photovoltaic system, so as to calculate the dynamic threshold corresponding to each data point through the real-time wind speed and historical average wind speed.

[0186] The second judgment unit is used to judge whether each data point exceeds the dynamic threshold based on the preset wind power data abnormality elimination rule, and if the current data point exceeds the dynamic threshold, the current data point is set to 1, otherwise the current data point is set to 0.

[0187] The acquisition unit is used to obtain the photovoltaic panel conversion efficiency, horizontal global radiation and temperature attenuation coefficient corresponding to the target wind power and photovoltaic system, and establish a theoretical output benchmark based on the photovoltaic panel conversion efficiency, horizontal global radiation and temperature attenuation coefficient.

[0188] The second calculation unit is used to calculate the daily statistics according to the theoretical output benchmark and construct the corresponding double threshold constraints using the daily statistics.

[0189] The third determination unit is used to determine the photovoltaic data anomaly elimination rule through dual threshold constraints and theoretical output benchmarks, so as to eliminate the abnormal data in the second wind and solar sample optimization data according to the photovoltaic data anomaly elimination rule to obtain the target new energy wind and solar sample data.

[0190] Optionally, in one embodiment of the present application, the target interpolation formula is:

[0191]

[0192] in, Indicates the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that based on phys interpolation method t Estimated value of the moment; 、n Indicates the amount of missing data; Represents the angle parameter; Represents a linear interpolation function.

[0193] It should be noted that the above explanation of the embodiment of the method for processing new energy wind and solar sample data based on regularity constraints is also applicable to the new energy wind and solar sample data processing device based on regularity constraints in this embodiment, and will not be repeated here.

[0194] According to the new energy wind and solar sample data processing device based on regularity constraints proposed in the embodiment of the present application, it includes an interpolation and completion module 100, which is used to obtain the new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, and based on a preset regular interpolation strategy, perform regularity completion operations on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data; a feature optimization module 200, which is used to extract multi-scale features of the wind and solar sample completion data based on a pre-built local reversible attention mechanism evaluator, and optimize the wind and solar sample completion data through the multi-scale features to obtain corresponding first wind and solar sample optimization data; a weight adjustment module 300, used to determine at least one data missing category in the wind and solar sample optimization data, and based on a preset dynamic weight fusion strategy, adjust the weight corresponding to each data missing category in at least one data missing category to obtain the corresponding second wind and solar sample optimization data; data elimination module 400, used to obtain abnormal data that meets the preset abnormal requirements in the second wind and solar sample optimization data, and use a pre-built outlier data elimination model to eliminate the abnormal data to obtain the target new energy wind and solar sample data corresponding to the target wind power and photovoltaic system, thereby improving the integrity and reliability of the new energy data and improving the operating efficiency of the new energy system.

[0195] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0196] Memory 401 , processor 402 , and computer programs stored in the memory 401 and executable on the processor 402 .

[0197] When the processor 402 executes the program, the new energy wind and solar sample data processing method based on regularity constraints provided in the above embodiment is implemented.

[0198] Furthermore, the electronic device further includes:

[0199] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0200] The memory 401 is used to store computer programs that can be run on the processor 402 .

[0201] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0202] If the memory 401, processor 402, and communication interface 403 are implemented independently, the communication interface 403, memory 401, and processor 402 can be connected to each other via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0203] Optionally, in a specific implementation, if the memory 401 , the processor 402 and the communication interface 403 are integrated on a chip, the memory 401 , the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0204] The processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0205] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the above-mentioned method for processing new energy wind and solar sample data based on regularity constraints is implemented.

[0206] An embodiment of the present application further provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned new energy wind and solar sample data processing method based on regularity constraints.

[0207] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" 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 application. 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 can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0208] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0209] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing a custom logical function or process step, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed in a different order than shown or discussed, including performing functions in a substantially simultaneous manner or in a reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application pertain.

[0210] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" is any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (not exhaustive) of computer-readable media include: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program can be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0211] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, any of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0212] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0213] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0214] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A method for processing new energy wind and solar sample data based on regularity constraints, characterized in that: The following steps are involved: Obtaining new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems, and performing a regular completion operation on the new energy wind and solar sample data based on a preset regular interpolation strategy to obtain corresponding wind and solar sample completion data; Based on a pre-built local reversible attention mechanism evaluator, extract multi-scale features of the scenery sample completion data, and optimize the scenery sample completion data by using the multi-scale features to obtain corresponding first scenery sample optimized data; Determining at least one data missing category in the wind and solar sample optimization data, and adjusting a weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data; Abnormal data that meets preset abnormal requirements in the second wind-solar sample optimization data is obtained, and the abnormal data is eliminated using a pre-built outlier data elimination model to obtain the target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system.

2. The method for processing new energy wind and solar sample data based on regularity constraints according to claim 1 is characterized in that: The acquiring of new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems, and performing a regular complement operation on the new energy wind and solar sample data based on a preset regular interpolation strategy to obtain corresponding wind and solar sample complement data, includes: Determine the generalized missing window corresponding to the new energy wind and solar sample data based on a preset sampling time length, a sensor valid data range, an environmental parameter vector, and a severe weather threshold; Based on the generalized missing window, determining the data missing type of the new energy wind and solar sample data; If the data missing type is a single point missing type, then the adjacent point average value and the time series prediction value corresponding to the missing point are calculated, and the corresponding single point completion value is determined according to the adjacent point average value and the time series prediction value, so as to perform a single point missing completion operation on the new energy wind and solar sample data using the single point completion value; If the data missing type is an odd-numbered point missing type, determining the K nearest neighbor prediction function, the wind power environment parameter set, and the photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data, and performing an odd-numbered point missing completion operation on the new energy wind and solar sample data according to the K nearest neighbor prediction function, the wind power environment parameter set, and the photovoltaic historical data feature matrix; If the data missing type is an even point missing type, determine the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature corresponding to the new energy wind and solar sample data, and perform an even point missing completion operation on the new energy wind and solar sample data based on the photovoltaic environmental parameter set, the wind power weight smoothing coefficient, the temperature coefficient and the power curve curvature.

3. The method for processing new energy wind and solar sample data based on regularity constraints according to claim 2 is characterized in that: The determining of at least one data missing category in the wind and solar sample optimization data, and adjusting the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data, includes: Determine parity compensation item parameters corresponding to the odd point missing type and the even point missing type; Calculate the maximum allowable missing length corresponding to the wind and light sample optimization data, and construct a target interpolation formula based on the maximum allowable missing length and the parity compensation item parameters, so as to use the target interpolation formula to adjust the weight corresponding to each data missing category to obtain the second wind and light sample optimization data.

4. The method for processing new energy wind and solar sample data based on regularity constraints according to claim 3 is characterized in that: The step of obtaining abnormal data that meets preset abnormal requirements from the second wind-solar sample optimization data and eliminating the abnormal data using a pre-built outlier data elimination model to obtain target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system includes: Determine a sliding window corresponding to each data point in the second wind-solar sample optimization data, and obtain the real-time wind speed and the historical average wind speed corresponding to the target wind power and photovoltaic system, so as to calculate the dynamic threshold corresponding to each data point based on the real-time wind speed and the historical average wind speed; Based on a preset wind power data anomaly rejection rule, determining whether each data point exceeds the dynamic threshold, if the current data point exceeds the dynamic threshold, setting the current data point to 1, otherwise setting the current data point to 0; Obtaining photovoltaic panel conversion efficiency, horizontal global radiation, and temperature attenuation coefficient corresponding to the target wind power and photovoltaic system, and establishing a theoretical output benchmark based on the photovoltaic panel conversion efficiency, the horizontal global radiation, and the temperature attenuation coefficient; Calculating daily statistics based on the theoretical output benchmark, and constructing corresponding dual-threshold constraints using the daily statistics; The photovoltaic data abnormality elimination rule is determined by the dual threshold constraint and the theoretical output benchmark, so as to eliminate abnormal data in the second wind-solar sample optimization data according to the photovoltaic data abnormality elimination rule to obtain the target new energy wind-solar sample data.

5. The method for processing new energy wind and solar sample data based on regularity constraints according to claim 3 is characterized in that: The target interpolation formula is: in, represents the parity compensation item parameter; Indicates that the difference strategy based on convolutional neural network is t Estimated value of the moment; Indicates that based on phys interpolation method t Estimated value of the moment; 、 n Indicates the amount of missing data; Represents the angle parameter; Represents a linear interpolation function.

6. A new energy wind and solar sample data processing device based on regularity constraints, characterized in that: include: An interpolation and completion module is used to obtain new energy wind and solar sample data corresponding to the target wind power and photovoltaic systems, and based on a preset regular interpolation strategy, perform regular completion operations on the new energy wind and solar sample data to obtain corresponding wind and solar sample completion data; a feature optimization module, configured to extract multi-scale features of the scenery sample completion data based on a pre-built local reversible attention mechanism evaluator, and optimize the scenery sample completion data using the multi-scale features to obtain corresponding first scenery sample optimized data; a weight adjustment module, configured to determine at least one data missing category in the wind and solar sample optimization data, and adjust the weight corresponding to each data missing category in the at least one data missing category based on a preset dynamic weight fusion strategy to obtain corresponding second wind and solar sample optimization data; A data elimination module is used to obtain abnormal data that meets preset abnormal requirements in the second wind-solar sample optimization data, and eliminate the abnormal data using a pre-built outlier data elimination model to obtain the target new energy wind-solar sample data corresponding to the target wind power and photovoltaic system.

7. The new energy wind and solar sample data processing device based on regularity constraints according to claim 6 is characterized in that: The interpolation and completion module includes: A first determining unit is configured to determine a generalized missing window corresponding to the new energy wind and solar sample data based on a preset sampling time length, a sensor valid data range, an environmental parameter vector, and a severe weather threshold; A first judgment unit is configured to judge the data missing type of the new energy wind and solar sample data based on the generalized missing window; A single point missing unit is configured to calculate, if the data missing type is a single point missing type, the average value of the adjacent points and the time series prediction value corresponding to the missing point, and determine the corresponding single point completion value according to the average value of the adjacent points and the time series prediction value, so as to perform a single point missing completion operation on the new energy wind and solar sample data using the single point completion value; an odd point missing unit, configured to, if the data missing type is an odd point missing type, determine a K-nearest neighbor prediction function, a set of wind power environmental parameters, and a photovoltaic historical data feature matrix corresponding to the new energy wind and solar sample data, and perform an odd point missing completion operation on the new energy wind and solar sample data according to the K-nearest neighbor prediction function, the set of wind power environmental parameters, and the photovoltaic historical data feature matrix; An even point missing unit is used to determine the photovoltaic environmental parameter set, wind power weight smoothing coefficient, temperature coefficient and power curve curvature corresponding to the new energy wind and solar sample data if the data missing type is an even point missing type, and perform an even point missing completion operation on the new energy wind and solar sample data according to the photovoltaic environmental parameter set, the wind power weight smoothing coefficient, the temperature coefficient and the power curve curvature.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for processing new energy wind and solar sample data based on regularity constraints as described in any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the new energy wind and solar sample data processing method based on regularity constraints as described in any one of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that The computer program is executed to implement the new energy wind and solar sample data processing method based on regularity constraints as described in any one of claims 1 to 5.

Images