New energy generating capacity real-time prediction method and system based on clustering and online learning

Abstract

The invention provides a clustering and online learning-based new energy generating capacity real-time prediction method and system. The method comprises the steps of obtaining data information of a target area; preprocessing the data information of the target area, extracting characteristic variables, performing clustering analysis, constructing an online learning prediction model according to the characteristic variables after clustering analysis, and performing real-time prediction on the new energy power generation amount of the target area to obtain a real-time power generation amount prediction result; performing real-time monitoring on the real-time generating capacity prediction result, and sending alarm information; and adjusting and optimizing the online learning prediction model according to the alarm information, and replacing the online learning prediction model. The technical effects that the data with the similar power generation characteristics are classified into one class based on the clustering algorithm, the model prediction precision is improved, meanwhile, the model can be dynamically updated in real time after receiving the new data by using the online learning method, and the real-time prediction accuracy is improved are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of renewable energy power generation prediction, and in particular to a real-time prediction method and system for renewable energy power generation based on clustering and online learning. Background Art

[0002] Real-time prediction of renewable energy power generation uses advanced prediction algorithms and models to make instant and dynamic estimates and predictions of the amount of electricity generated by renewable energy (such as solar energy, wind energy, hydropower, geothermal energy, etc.) equipment at a specific point in time or time period in the future.

[0003] However, in the process of implementing the technical solutions in the embodiments of the present application, the inventors of the present application discovered that the above-mentioned prior art has at least the following technical problems:

[0004] The existing new energy power generation prediction methods have the problems of not fully utilizing the internal similarity structure of the data and insufficient real-time prediction capabilities, resulting in the inability to identify the impact of different feature data groups on power generation, which in turn affects the power generation prediction accuracy and the low real-time prediction accuracy. Summary of the Invention

[0005] Embodiments of the present invention provide a real-time prediction method and system for renewable energy power generation based on clustering and online learning. These methods address the existing problems of renewable energy power generation prediction methods, which lack the ability to fully utilize the internal similarity structure of data and inadequate real-time prediction capabilities. This leads to an inability to identify the impact of different feature data groups on power generation, which in turn affects the accuracy of power generation predictions and the low accuracy of real-time predictions. The system achieves the technical effect of grouping data with similar power generation characteristics based on a clustering algorithm, improving model prediction accuracy. Furthermore, online learning methods are used to dynamically update the model in real time upon receiving new data, thereby improving real-time prediction accuracy.

[0006] In view of the above problems, an embodiment of the present application is proposed to provide a method and system for real-time prediction of renewable energy power generation based on clustering and online learning.

[0007] In a first aspect, the present invention provides a real-time prediction method for renewable energy power generation based on clustering and online learning, the method comprising: obtaining data information of a target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographic location information, and power generation equipment status information; preprocessing the data information of the target area to obtain preprocessed data information; extracting characteristic variables from the preprocessed historical data information, wherein the characteristic variables are characteristic variables that have an important impact on the prediction of renewable energy power generation; performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis; constructing an online learning prediction model based on the characteristic variables after cluster analysis and an online gradient descent algorithm; using the online learning prediction model to perform real-time prediction of renewable energy power generation in the target area to obtain a real-time power generation prediction result; monitoring the real-time power generation prediction result in real time and sending an alarm message; and tuning the online learning prediction model based on the alarm message and replacing the online learning prediction model.

[0008] Preferably, the method further comprises: synchronously displaying the real-time power generation prediction result through a result display module.

[0009] Preferably, the data information of the target area obtained includes: the field description of the meteorological information includes temperature, humidity, air pressure, irradiance, wind speed and collection time; the field description of the power generation data includes power generation and collection time; the field description of the geographical location information includes longitude, latitude, altitude, and collection time; the field description of the power generation equipment status information includes operating status, installed capacity, adjustable output, cumulative operating time and collection time.

[0010] Preferably, the preprocessing of the data information of the target area to obtain the preprocessed data information includes: performing missing value filling processing on the data information of the target area to obtain first data information; performing outlier processing on the first data information to obtain second data information; wherein, performing outlier processing on the first data information to obtain the second data information includes: calculating the absolute value of the difference between any two adjacent data; obtaining a data change threshold; judging whether the absolute value of the difference between any two adjacent data is greater than the data change threshold; if the absolute value of the difference between any two adjacent data is greater than the data change threshold, determining that the current data is an outlier; calculating the mean of the ten data adjacent to the outlier; replacing the outlier with the mean of the ten data adjacent to the outlier; selecting the Min-Max normalization method to perform normalization processing on the second data information to obtain the preprocessed data information, and the calculation formula is:

[0011]

[0012] Among them, D * is the normalized result,

[0013] D is the original data,

[0014] D min is the minimum value of the data,

[0015] D max The maximum value of the data.

[0016] Preferably, the data information of the target area is filled with missing values to obtain the first data information, including: filling the missing values of the historical data in the temperature, humidity, air pressure, irradiance, wind speed, power generation, longitude, latitude, altitude, installed capacity, adjustable output and cumulative operating time columns, including: calculating the mean of the week where the historical data is missing to obtain a first mean, and filling the missing values of the historical data with the first mean; calculating the mean of the day where the historical data is missing to obtain a second mean, and filling the missing values of the historical data with the second mean; calculating the average of the first mean and the second mean to obtain a third mean, and filling the missing values of the historical data with the third mean; and filling the missing values of the historical data in the operating status column with a forward filling method.

[0017] Preferably, the real-time monitoring of the real-time power generation prediction result and the sending of alarm information include: obtaining the actual power generation of the target area; calculating the difference between the real-time power generation prediction result and the actual power generation; obtaining the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; judging whether the difference between the real-time power generation prediction result and the actual power generation exceeds the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; if the difference between the real-time power generation prediction result and the actual power generation exceeds the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation, sending an alarm information.

[0018] Preferably, the online learning prediction model is tuned and replaced according to the alarm information, including: according to the alarm information, obtaining the first moment when the difference between the real-time power generation prediction result and the actual power generation exceeds the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; obtaining the target area data information at the first moment; according to the target area data information at the first moment, optimizing and updating the online learning prediction model to obtain an optimized online learning prediction model; and replacing the online learning prediction model with the optimized online learning prediction model through dynamic loading.

[0019] In a second aspect, the present invention provides a real-time prediction system for renewable energy power generation based on clustering and online learning, the system comprising:

[0020] A data acquisition module, the data acquisition module is used to collect data information of the target area, wherein the data acquisition module includes a database collector and an API collector, the database collector is used to collect historical meteorological information, power generation data, geographical location information and power generation equipment status information of the target area, and the API collector is used to collect real-time meteorological information, power generation data, geographical location information and power generation equipment status information of the target area;

[0021] A preprocessing module, configured to perform missing value filling, outlier detection, and normalization preprocessing on the data information of the target area;

[0022] A feature extraction module, the feature extraction module is used to extract characteristic variables that have an important impact on the prediction of new energy power generation from the pre-processed historical data information;

[0023] A cluster analysis module, wherein the cluster analysis module is used to perform cluster analysis on the characteristic variables;

[0024] A model building module, wherein the model building module is used to build an online learning prediction model;

[0025] A real-time prediction module, configured to perform real-time prediction of the renewable energy power generation in the target area;

[0026] A monitoring and early warning module, which is used to monitor the real-time power generation forecast result in real time and trigger an alarm message;

[0027] A model tuning module, configured to tune the online learning prediction model according to the alarm information;

[0028] A result display module is used to display the real-time power generation prediction results output by the real-time prediction module.

[0029] In the third aspect, the present invention also provides a real-time prediction device for new energy power generation based on clustering and online learning, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor, characterized in that when the computer program is executed by the processor, the processor executes the steps of the aforementioned real-time prediction method for new energy power generation based on clustering and online learning.

[0030] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, it is used to implement the steps of the aforementioned method for real-time prediction of new energy power generation based on clustering and online learning.

[0031] The above one or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:

[0032] An embodiment of the present invention provides a real-time prediction method and system for renewable energy power generation based on clustering and online learning. The method includes: obtaining data information of a target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographic location information, and power generation equipment status information; preprocessing the data information of the target area to obtain preprocessed data information; extracting characteristic variables from the preprocessed historical data information, wherein the characteristic variables are characteristic variables that have an important impact on the prediction of renewable energy power generation; performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis; constructing an online learning prediction model based on the characteristic variables after cluster analysis and an online gradient descent algorithm; using the online learning prediction model to perform real-time prediction of renewable energy power generation in the target area to obtain a real-time power generation prediction result; monitoring the real-time power generation prediction result in real time and sending an alarm message; and tuning the online learning prediction model based on the alarm message and replacing the online learning prediction model. This technology addresses the existing problems in renewable energy power generation forecasting methods, which include underutilization of the internal similarity structure of data and insufficient real-time forecasting capabilities. This leads to an inability to identify the impact of different characteristic data groups on power generation, which in turn affects the accuracy of power generation forecasts and low real-time forecast accuracy. The technology achieves the technical effect of grouping data with similar power generation characteristics into one category based on a clustering algorithm, improving model prediction accuracy. At the same time, it utilizes online learning methods to enable the model to be dynamically updated in real time upon receiving new data, thereby improving real-time forecast accuracy.

[0033] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Schematic diagram of a process for real-time prediction of renewable energy power generation based on clustering and online learning in an embodiment of the present invention;

[0035] Figure 2Schematic diagram of a process of a real-time prediction system for renewable energy power generation based on clustering and online learning in an embodiment of the present invention;

[0036] Figure 3 2 is another structural schematic diagram of a real-time prediction device for renewable energy power generation based on clustering and online learning in an embodiment of the present invention.

[0037] Description of reference numerals: bus 300 , receiver 301 , processor 302 , transmitter 303 , memory 304 , bus interface 306 . DETAILED DESCRIPTION

[0038] The embodiments of the present invention provide a real-time prediction method and system for renewable energy power generation based on clustering and online learning, which solves the problems in the existing renewable energy power generation prediction methods, such as insufficient utilization of the internal similarity structure of data and insufficient real-time prediction capabilities, which lead to the inability to identify the impact of different feature data groups on power generation, thereby affecting the power generation prediction accuracy and the low accuracy of real-time prediction.

[0039] The overall idea of the technical solution provided by the present invention is as follows: obtaining data information of the target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographical location information and power generation equipment status information; preprocessing the data information of the target area to obtain preprocessed data information; extracting characteristic variables from the preprocessed historical data information, wherein the characteristic variables are characteristic variables that have an important impact on the prediction of new energy power generation; performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis; constructing an online learning prediction model based on the characteristic variables after cluster analysis and an online gradient descent algorithm; using the online learning prediction model, performing real-time prediction of the new energy power generation in the target area to obtain a real-time power generation prediction result; monitoring the real-time power generation prediction result in real time and sending an alarm message; tuning the online learning prediction model based on the alarm message and replacing the online learning prediction model. The technical effect of grouping data with similar power generation characteristics into one category based on clustering algorithms to improve the model prediction accuracy is achieved. At the same time, the online learning method is used to enable the model to be dynamically updated in real time after receiving new data, thereby improving the accuracy of real-time predictions.

[0040] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Unless there is a conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0041] The term "and / or" in this document simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. Additionally, the character " / " in this document generally indicates that the related objects are in an "or" relationship.

[0042] Example 1

[0043] Figure 1 FIG. 1 is a flow chart of a method for real-time prediction of renewable energy power generation based on clustering and online learning in an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a real-time prediction method for renewable energy power generation based on clustering and online learning, the method comprising:

[0044] Step s1: obtaining data information of the target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographical location information, and power generation equipment status information;

[0045] Specifically, the data acquisition module is the input end of the system. The embodiment of the present application collects the necessary data information of the target area through the data acquisition module, including historical meteorological information, power generation data, geographic location information and power generation equipment status information of the target area, and real-time meteorological information, power generation data, geographic location information and power generation equipment status information.

[0046] In step s1, the field description of the meteorological information includes temperature, humidity, air pressure, irradiance, wind speed and collection time; the field description of the power generation data includes power generation and collection time; the field description of the geographic location information includes longitude, latitude, altitude, and collection time; the field description of the power generation equipment status information includes operating status, installed capacity, adjustable output, cumulative operating time and collection time.

[0047] Step s2: preprocessing the data information of the target area to obtain preprocessed data information;

[0048] Specifically, the collected historical meteorological information, power generation data, geographic location information and power generation equipment status information are pre-processed by filling missing values, detecting outliers and normalization, and the real-time meteorological information, power generation data, geographic location information and power generation equipment status information are normalized, so as to further eliminate the problems of missing values, outliers and inconsistent dimensions in the data, ensure the quality of the collected data, and facilitate the technical effect of subsequent processing.

[0049] In step s2, the data information of the target area is preprocessed to obtain preprocessed data information, specifically including:

[0050] Performing missing value filling processing on the data information of the target area to obtain first data information;

[0051] performing outlier processing on the first data information to obtain second data information;

[0052] The performing outlier processing on the first data information to obtain the second data information includes:

[0053] Calculate the absolute value of the difference between any two adjacent data, that is, max{|D i+1 -D i |,|D i -D i-1 |}, where D i+1 is the i+1th data, D i is the i-th data, D i-1 is the i-1th data;

[0054] Get the data change threshold ε i ;

[0055] Determine whether the absolute value of the difference between any two adjacent data is greater than the data change threshold, that is, max{|D i+1 -D i |,|D i -D i-1 |}>ε i ;

[0056] If the absolute value of the difference between any two adjacent data is greater than the data change threshold, the current data is determined to be an abnormal value, that is, if max{|D i+1 -D i |,|D i -D i-1 |}>ε i , then D i For abnormal data;

[0057] Calculate the mean of the ten data adjacent to the abnormal value, where the calculation formula is:

[0058] Replace the outlier with the mean of the ten data points adjacent to the outlier;

[0059] The Min-Max normalization method is selected to normalize the second data information to obtain preprocessed data information. The calculation formula is:

[0060]

[0061] Among them, D * is the normalized result,

[0062] D is the original data,

[0063] D min is the minimum value of the data,

[0064] D max The maximum value of the data.

[0065] In step s2, the data information of the target area is filled with missing values to obtain first data information, which specifically includes:

[0066] Fill in missing values in historical data for temperature, humidity, air pressure, irradiance, wind speed, power generation, longitude, latitude, altitude, installed capacity, adjustable output, and cumulative operating time, including:

[0067] Calculating the mean of the week in which the historical data has missing values to obtain a first mean, and filling the missing values in the historical data with the first mean;

[0068] Calculating the mean of the days on which the historical data has missing values to obtain a second mean, and filling the missing values of the historical data with the second mean;

[0069] Calculating an average of the first mean and the second mean to obtain a third mean, and filling in the missing values of the historical data by using the third mean;

[0070] The missing values of the historical data of the running status column are filled using the forward filling method.

[0071] Specifically, missing values in the historical data for the temperature, humidity, air pressure, irradiance, wind speed, power generation, longitude, latitude, altitude, installed capacity, adjustable output, and cumulative operating time columns are filled using the mean of the week in which the missing value occurred, the mean of the day in which the missing value occurred, or both methods to calculate the missing value and take the average of the two as the final missing value. Missing values in the historical data for the operating status column are filled using the forward filling method.

[0072] Step s3: extracting characteristic variables from the pre-processed historical data information, wherein the characteristic variables are characteristic variables that have a significant impact on the prediction of renewable energy power generation;

[0073] Specifically, since new energy power generation data has strong time characteristics, the embodiment of the present application selects the feature extraction algorithm as the time series analysis method ARIMA, and uses the selected ARIMA algorithm to perform feature extraction on the preprocessed historical meteorological information, power generation data, geographic location information, and power generation equipment status information, and extract characteristic variables that have an important impact on the prediction of new energy power generation.

[0074] Step s4: performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis;

[0075] Specifically, to address the complex and variable patterns in renewable energy power generation data and the unknown number of clusters, this embodiment of the present invention uses DBSCAN to cluster feature-extracted historical meteorological information, power generation data, geographic location information, and power generation equipment status information. This clustering process groups renewable energy power generation equipment with similar power generation characteristics, environmental conditions, and equipment status into a single category. The clustering results are then verified and evaluated to ensure they meet expectations. This further achieves the technical effect of fully exploring the similarity structure in renewable energy power generation data and improving the model's prediction accuracy.

[0076] Step s5: constructing an online learning prediction model based on the characteristic variables after the cluster analysis and the online gradient descent algorithm;

[0077] Specifically, based on the characteristic variables extracted after the cluster analysis, the online learning algorithm is configured as an online gradient descent algorithm, the loss function in the online gradient descent algorithm is selected as the MSE, and the optimizer is selected as the Adam optimizer. The clustering results and the online gradient descent algorithm are used to construct a real-time prediction model, namely the online learning prediction model. This achieves the technical effect of enabling the model to be dynamically updated in real time upon receiving new data, thereby improving the accuracy of real-time predictions.

[0078] Step s6: using the online learning prediction model to perform real-time prediction of renewable energy power generation in the target area to obtain a real-time power generation prediction result;

[0079] Specifically, in the online learning prediction model, a new real-time prediction task is added, and the real-time prediction model is configured as an online gradient descent prediction model. Real-time meteorological information, geographic location information and power generation equipment status information that have been normalized and preprocessed are selected to perform real-time prediction of the new energy power generation in the target area, thereby achieving the technical effect of obtaining real-time power generation prediction results.

[0080] Step s7: monitoring the real-time power generation prediction result in real time and sending an alarm message;

[0081] Specifically, the predicted power generation results of the new energy power generation equipment in the target area are compared and analyzed with the actual power generation data of the new energy power generation equipment in the target area. When the difference between the real-time power generation prediction result and the actual power generation exceeds a threshold, the system triggers an early warning, sends an alarm message, and notifies the model tuning module to tune the online learning prediction model. This achieves the technical effect of real-time monitoring of the operation status of the online learning prediction model and regular updates.

[0082] In step s7, the real-time power generation prediction result is monitored in real time and an alarm message is sent, specifically including: obtaining the actual power generation of the target area; calculating the difference between the real-time power generation prediction result and the actual power generation; obtaining the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; judging whether the difference between the real-time power generation prediction result and the actual power generation exceeds the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; if the difference between the real-time power generation prediction result and the actual power generation exceeds the alarm threshold of the difference between the real-time power generation prediction result and the actual power generation, an alarm message is sent.

[0083] Step s8: According to the alarm information, the online learning prediction model is tuned and replaced.

[0084] Specifically, the embodiment of the present application uses the data at the moment when the difference between the power generation prediction result and the actual power generation exceeds the alarm threshold based on the alarm information to tune the online learning prediction model to adapt it to the latest power generation environment. By dynamically loading, the tuned online learning prediction model replaces the online learning prediction model, thereby achieving the technical effect of ensuring that the prediction performance does not drop sharply.

[0085] In step s8, the online learning prediction model is tuned and replaced according to the alarm information, including:

[0086] Obtaining, according to the alarm information, a first moment when a difference between the real-time power generation prediction result and the actual power generation exceeds an alarm threshold of the difference between the real-time power generation prediction result and the actual power generation;

[0087] Obtaining target area data information at the first moment;

[0088] Optimizing and updating the online learning prediction model according to the target area data information at the first moment to obtain an optimized online learning prediction model;

[0089] The online learning prediction model is replaced by the optimized online learning prediction model in a dynamic loading manner.

[0090] Furthermore, the method further includes: synchronously displaying the real-time power generation prediction result through a result display module.

[0091] Specifically, after obtaining the real-time power generation forecast result, the real-time power generation forecast result is synchronously displayed to provide intuitive data support for energy scheduling decisions.

[0092] Example 2

[0093] The embodiment of the present invention also provides a real-time prediction system for renewable energy power generation based on clustering and online learning, such as Figure 2 As shown, the system includes:

[0094] A data acquisition module, the data acquisition module is used to collect data information of the target area, wherein the data acquisition module includes a database collector and an API collector, the database collector is used to collect historical meteorological information, power generation data, geographical location information and power generation equipment status information of the target area, and the API collector is used to collect real-time meteorological information, power generation data, geographical location information and power generation equipment status information of the target area;

[0095] Specifically, the data acquisition module is the input end of the system, wherein the database collector and the API collector respectively collect historical and real-time weather information, power generation data, geographic location information, and power generation equipment status information of the target area. The database collector captures historical weather information, power generation data, equipment geographic location information, and power generation equipment status information by connecting to the business database of the new energy power station in the target area. The API collector obtains the latest weather information, power generation data, and power generation equipment status information of the new energy power station in real time through the API interface.

[0096] A preprocessing module, configured to perform missing value filling, outlier detection, and normalization preprocessing on the data information of the target area;

[0097] A feature extraction module, the feature extraction module is used to extract characteristic variables that have an important impact on the prediction of renewable energy power generation from the pre-processed historical data information;

[0098] A cluster analysis module, wherein the cluster analysis module is used to perform cluster analysis on the characteristic variables;

[0099] A model building module, wherein the model building module is used to build an online learning prediction model;

[0100] A real-time prediction module, configured to perform real-time prediction of the renewable energy power generation in the target area;

[0101] A monitoring and early warning module, which is used to monitor the real-time power generation forecast result in real time and trigger an alarm message;

[0102] A model tuning module, configured to tune the online learning prediction model according to the alarm information;

[0103] A result display module is used to display the real-time power generation prediction results output by the real-time prediction module.

[0104] Example 3

[0105] Based on the same inventive concept as the method for real-time prediction of new energy power generation based on clustering and online learning in the aforementioned embodiment, the present invention also provides a method device for real-time prediction of new energy power generation based on clustering and online learning, which stores a computer program. When the program is executed by a processor, it implements the steps of any method of the aforementioned method for real-time prediction of new energy power generation based on clustering and online learning.

[0106] Among them, Figure 3 In the embodiment of the present invention, a bus architecture (represented by bus 300) is shown. Bus 300 may include any number of interconnected buses and bridges. Bus 300 links various circuits together, including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also link various other circuits together, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are not described further herein. Bus interface 306 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same component, namely a transceiver, which provides a means for communicating with various other devices over a transmission medium.

[0107] The processor 302 is responsible for managing the bus 300 and general processing, while the memory 304 may be used to store data used by the processor 302 when performing operations.

[0108] Example 4

[0109] Based on the same inventive concept as the real-time prediction method for new energy power generation based on clustering and online learning in the aforementioned embodiment, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, it is used to implement the steps of the real-time prediction method for new energy power generation based on clustering and online learning described above.

[0110] The above one or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:

[0111] An embodiment of the present invention provides a real-time prediction method and system for renewable energy power generation based on clustering and online learning. The method includes: obtaining data information of a target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographic location information, and power generation equipment status information; preprocessing the data information of the target area to obtain preprocessed data information; extracting characteristic variables from the preprocessed historical data information, wherein the characteristic variables are characteristic variables that have an important impact on the prediction of renewable energy power generation; performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis; constructing an online learning prediction model based on the characteristic variables after cluster analysis and an online gradient descent algorithm; using the online learning prediction model to perform real-time prediction of renewable energy power generation in the target area to obtain a real-time power generation prediction result; monitoring the real-time power generation prediction result in real time and sending an alarm message; and tuning the online learning prediction model based on the alarm message and replacing the online learning prediction model. This technology addresses the existing problems in renewable energy power generation forecasting methods, which include underutilization of the internal similarity structure of data and insufficient real-time forecasting capabilities. This leads to an inability to identify the impact of different characteristic data groups on power generation, which in turn affects the accuracy of power generation forecasts and low real-time forecast accuracy. The technology achieves the technical effect of grouping data with similar power generation characteristics into one category based on a clustering algorithm, improving model prediction accuracy. At the same time, it utilizes online learning methods to enable the model to be dynamically updated in real time upon receiving new data, thereby improving real-time forecast accuracy.

[0112] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0116] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A real-time prediction method for renewable energy power generation based on clustering and online learning, characterized in that: The method comprises: Obtaining data information of the target area, wherein the data information includes historical and real-time meteorological information, power generation data, geographic location information, and power generation equipment status information; Preprocessing the data information of the target area to obtain preprocessed data information; Extracting characteristic variables from the pre-processed historical data information, wherein the characteristic variables are characteristic variables that have a significant impact on the prediction of renewable energy power generation; Performing cluster analysis on the characteristic variables, dividing dates with similar power generation characteristics into the same category, and obtaining characteristic variables after cluster analysis; Constructing an online learning prediction model based on the characteristic variables after the cluster analysis and the online gradient descent algorithm; Using the online learning prediction model, the renewable energy power generation in the target area is predicted in real time to obtain a real-time power generation prediction result; Monitor the real-time power generation forecast result in real time and send an alarm message; According to the alarm information, the online learning prediction model is tuned and replaced.

2. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 1, characterized in that: The method further comprises: The real-time power generation forecast result is synchronously displayed through the result display module.

3. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 1, characterized in that: The obtaining of data information of the target area includes: The field description of the meteorological information includes temperature, humidity, air pressure, irradiance, wind speed and collection time; The field description of the power generation data includes power generation and collection time; The field description of the geographic location information includes longitude, latitude, altitude, and collection time; The field description of the power generation equipment status information includes operating status, installed capacity, adjustable output, cumulative operating time and collection time.

4. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 1, characterized in that: The preprocessing of the data information of the target area to obtain the preprocessed data information includes: Performing missing value filling processing on the data information of the target area to obtain first data information; Performing outlier processing on the first data information to obtain second data information; The performing outlier processing on the first data information to obtain the second data information includes: Calculate the absolute value of the difference between any two adjacent data; Get data change threshold; Determine whether the absolute value of the difference between any two adjacent data is greater than the data change threshold; If the absolute value of the difference between any two adjacent data is greater than the data change threshold, the current data is determined to be an abnormal value; Calculate the mean of the ten data adjacent to the abnormal value; Replace the outlier with the mean of the ten data points adjacent to the outlier; The Min-Max normalization method is selected to normalize the second data information to obtain preprocessed data information. The calculation formula is: Among them, D * is the normalized result, D is the original data, D min is the minimum value of the data, D max The maximum value of the data.

5. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 4, characterized in that: The performing missing value filling processing on the data information of the target area to obtain the first data information includes: Fill in missing values in historical data for temperature, humidity, air pressure, irradiance, wind speed, power generation, longitude, latitude, altitude, installed capacity, adjustable output, and cumulative operating time, including: Calculating the mean of the week in which the historical data has missing values to obtain a first mean, and filling the missing values in the historical data with the first mean; Calculating the mean of the days on which the historical data has missing values to obtain a second mean, and filling the missing values of the historical data with the second mean; Calculating an average of the first mean and the second mean to obtain a third mean, and filling in the missing values of the historical data by using the third mean; The missing values of the historical data of the running status column are filled using the forward filling method.

6. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 1, characterized in that: The real-time monitoring of the real-time power generation prediction result and sending of alarm information include: Obtain the actual power generation in the target area; Calculating the difference between the real-time power generation prediction result and the actual power generation; Obtaining an alarm threshold value of a difference between the real-time power generation prediction result and the actual power generation; Determining whether a difference between the real-time power generation prediction result and the actual power generation exceeds an alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; If the difference between the real-time power generation prediction result and the actual power generation exceeds an alarm threshold of the difference between the real-time power generation prediction result and the actual power generation, an alarm message is sent.

7. The method for real-time prediction of renewable energy power generation based on clustering and online learning according to claim 6, characterized in that: The step of optimizing the online learning prediction model and replacing the online learning prediction model according to the alarm information includes: Obtaining, according to the alarm information, a first moment when a difference between the real-time power generation prediction result and the actual power generation exceeds an alarm threshold of the difference between the real-time power generation prediction result and the actual power generation; Obtaining target area data information at the first moment; Optimizing and updating the online learning prediction model according to the target area data information at the first moment to obtain an optimized online learning prediction model; The online learning prediction model is replaced by the optimized online learning prediction model in a dynamic loading manner.

8. A real-time prediction system for renewable energy power generation based on clustering and online learning, characterized in that: The system comprises: A data acquisition module, the data acquisition module is used to collect data information of the target area, wherein the data acquisition module includes a database collector and an API collector, the database collector is used to collect historical meteorological information, power generation data, geographical location information and power generation equipment status information of the target area, and the API collector is used to collect real-time meteorological information, power generation data, geographical location information and power generation equipment status information of the target area; A preprocessing module, configured to perform missing value filling, outlier detection, and normalization preprocessing on the data information of the target area; A feature extraction module, the feature extraction module is used to extract characteristic variables that have an important impact on the prediction of renewable energy power generation from the pre-processed historical data information; A cluster analysis module, wherein the cluster analysis module is used to perform cluster analysis on the characteristic variables; A model building module, wherein the model building module is used to build an online learning prediction model; A real-time prediction module, configured to perform real-time prediction of the renewable energy power generation in the target area; A monitoring and early warning module, which is used to monitor the real-time power generation forecast result in real time and trigger an alarm message; A model tuning module, configured to tune the online learning prediction model according to the alarm information; A result display module is used to display the real-time power generation prediction results output by the real-time prediction module.

9. A real-time prediction device for renewable energy power generation based on clustering and online learning, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by the processor, the processor executes the steps of the method for real-time prediction of new energy power generation based on clustering and online learning as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it is used to implement the steps of the real-time prediction method of new energy power generation based on clustering and online learning as described in any one of claims 1 to 7 above.

Images