Photovoltaic output information generation method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a photovoltaic output information generation method and device, equipment and a storage medium. The method comprises the steps that meteorological factors are screened based on correlation analysis, a comprehensive meteorological factor curve is generated through dimensionality reduction, and a photovoltaic output curve is synchronously generated. Five types of results of cloudy days, cloudy days, sunny days, sunny-to-cloudy days and multi-cloud-to-sunny days are extracted through three-stage clustering, an error distribution estimation model is constructed based on historical data of each scene, a classification historical data set is used for training a Bayesian neural network, an integrated model is constructed, a path is matched according to input data, and an error distribution estimation model is constructed; and inputting the generated comprehensive curve into an integration and error model, and outputting a photovoltaic daily output curve with a confidence interval as final information. Refining input with high reference value is provided for time sequence production simulation, and the refinement degree and the uncertainty quantification capability of output prediction are improved.

Description

Technical Field

[0001] This disclosure relates to the field of new energy power output prediction technology, and in particular to a method, apparatus, equipment and storage medium for generating photovoltaic power output information. Background Technology

[0002] The installed capacity of photovoltaic power generation in the power system has been increasing year by year. While bringing green electricity, the intermittency and volatility of its output have also had a significant impact on the dynamic and static characteristics of the power system. To cope with the randomness of new energy power generation, time-series production simulation technology is usually used to optimize the allocation of various flexible resources in system planning and scheduling. The accuracy of new energy power generation directly affects the accuracy of time-series production simulation.

[0003] Existing photovoltaic power generation models utilize a limited range of data types and lack sufficient data mining capabilities. Commonly used data-driven generation methods suffer from issues such as coarse characterization of features extracted from typical scenarios, mismatch between regression generation models and scenario features, and inability to efficiently quantify uncertainties, resulting in unreliable generation results. Summary of the Invention

[0004] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, device, and storage medium for generating photovoltaic power output information.

[0005] This disclosure provides a method for generating photovoltaic power output information, the method comprising: Acquire meteorological data and photovoltaic power output data, analyze the correlation between the meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data; The comprehensive meteorological factor curve and photovoltaic output curve are clustered in the first stage using statistical distance indices to obtain the first clustering results, which include cloudy, partly cloudy, and mixed types. The mixed type results are clustered in the second stage using linear distance indices to obtain sunny and changing weather types. The changing weather type results are clustered in the third stage using angular distance indices to obtain weather transition type results, which include sunny to partly cloudy and partly cloudy to sunny types. Based on the results of weather conversion type, cloudy type, partly cloudy type, and sunny type, the error distribution is estimated, and an error distribution estimation model is constructed. Obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; train several preset models based on the historical datasets; and combine all trained preset models to construct an ensemble model. Acquire the data for the period to be generated, match the preset intermediate process based on the data for the period to be generated, and obtain the daily comprehensive meteorological factor curve of the period to be generated. Input the daily comprehensive meteorological factor curve of the period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic power generation curve of the period to be generated, as photovoltaic power generation information.

[0006] The method provided in this disclosure, comprising acquiring meteorological data and photovoltaic power output data, analyzing the correlation between the meteorological data and photovoltaic power output data, selecting a target meteorological factor, generating a comprehensive meteorological factor curve based on the target meteorological factor, and generating a photovoltaic power output curve based on the photovoltaic power output data, includes: Acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological factors and photovoltaic power output parameters based on the meteorological data and photovoltaic power output data, and determine the meteorological factors with a correlation greater than a preset correlation threshold as target meteorological factors; Principal component analysis is used to reduce the dimensionality of the target meteorological factors, extract the principal component factors, and generate a comprehensive meteorological factor curve based on the principal component factors. A photovoltaic output curve is generated based on the photovoltaic output data.

[0007] The method provided in this disclosure uses a statistical distance index to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve to obtain a first clustering result, including: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated based on the statistical distance index, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated based on the statistical distance index. The comprehensive meteorological factor curve and the photovoltaic output curve are clustered in the first stage based on the mean of the first curve, the maximum value of the first curve, the mean of the second curve, and the maximum value of the second curve, to obtain the first clustering result. The first clustering result includes cloudy type result, partly cloudy type result, and mixed type result.

[0008] The method provided in this disclosure performs a second-stage clustering of the mixed-type results using a linear distance index to obtain sunny weather type results and changing weather type results, and performs a third-stage clustering of the changing weather type results using an angular distance index to obtain weather transition type results. The mixed-type results include a mixed-type comprehensive meteorological factor curve and a mixed-type photovoltaic output curve, including: The mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve are determined by using the line distance index. Based on the mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve, the overall situation of daily irradiance fluctuation is obtained. Based on the overall fluctuation of daily irradiance, a second-stage clustering was performed on the mixed-type results to obtain the sunny-day type results and the changing-weather type results. The results of the changing weather type include the changing weather type-comprehensive meteorological factor curve and the changing weather type-photovoltaic output curve; The time-period offset characteristics between the weather change type-comprehensive meteorological factor curve and the weather change type-photovoltaic output curve are determined using the angular distance index. The third-stage clustering of the weather type results is performed based on the time period offset characteristics to obtain the weather transition type results.

[0009] The method provided in this disclosure estimates the error distribution based on weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, and constructs an error distribution estimation model, including: Extract the daily comprehensive meteorological factor curves and comprehensive meteorological factor cluster center curves for all historical days in the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; The point-by-point error is calculated based on the daily comprehensive meteorological factor curve and the comprehensive meteorological factor cluster center curve. Nonparametric kernel density estimation is performed on the point-by-point errors to obtain the error distribution estimate of the daily comprehensive meteorological factor curve corresponding to each type of result, and an error distribution estimation model is constructed based on the error distribution estimates of all types.

[0010] The method provided in this disclosure embodiment obtains historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; trains several preset models based on the historical datasets; and constructs an ensemble model by combining all trained preset models, including: Retrieve historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; Several preset models are initialized, and the corresponding types of results are used to train the several preset models respectively, and the output values ​​of the corresponding types of results are obtained for each preset model after training; An ensemble model is constructed by combining all pre-trained models. The output of the ensemble model is the average of the output values ​​of all pre-trained models for the corresponding type of result.

[0011] The method provided in this disclosure includes acquiring data for a time period to be generated, matching a preset intermediate process based on the data for the time period to be generated to obtain a daily comprehensive meteorological factor curve for the time period to be generated, and inputting the daily comprehensive meteorological factor curve for the time period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic daily power output curve for the time period to be generated, as photovoltaic power output information. Acquire the data for the time period to be generated, which includes daily meteorological data and daily weather type, and match a preset intermediate process based on the data for the time period to be generated; If the data to be generated for a time period is the daily meteorological data of the time period to be generated, the comprehensive meteorological factor curve of the time period to be generated is obtained based on the daily meteorological data of the time period to be generated, and the comprehensive meteorological factor curve of the time period to be generated is compared with the comprehensive meteorological factor cluster center curve in each type of result to determine the daily weather type of the time period to be generated. Input the daily weather type of the period to be generated and the daily comprehensive meteorological factor curve of the period to be generated into the integrated model, and output the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve; If the data to be generated for a period of time is the daily weather type of the period to be generated, the daily comprehensive meteorological factor error curve of the period to be generated is determined according to the error distribution estimation model and the daily weather type of the period to be generated. The daily comprehensive meteorological factor error curve and the comprehensive meteorological factor cluster center curve corresponding to the daily weather type are superimposed to obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated and the daily weather type of the period to be generated are input into the integrated model to output the photovoltaic power generation curve of the period to be generated. The photovoltaic power generation curve is the output of the integrated model under a given confidence level.

[0012] This disclosure also provides a photovoltaic power output information generation device, the device comprising: The acquisition module is used to acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data. The clustering module is used to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve using statistical distance indicators to obtain a first clustering result. The first clustering result includes cloudy type results, partly cloudy type results, and mixed type results. The mixed type results are then clustered in a second stage using linear distance indicators to obtain sunny type results and changing weather type results. Finally, the changing weather type results are clustered in a third stage using angular distance indicators to obtain weather transition type results. The weather transition type results include sunny to partly cloudy type results and partly cloudy to sunny type results. The first construction module is used to estimate the error distribution based on the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, and to construct an error distribution estimation model. The second construction module is used to obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, train several preset models based on the historical datasets, and combine all trained preset models to construct an integrated model. The output module is used to acquire the data of the period to be generated, match the preset intermediate process according to the data of the period to be generated, and obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated is input into the integrated model and the error distribution estimation model to determine the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve, which is used as photovoltaic power generation information.

[0013] The apparatus provided in this disclosure, wherein the acquisition module is specifically used for: Acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological factors and photovoltaic power output parameters based on the meteorological data and photovoltaic power output data, and determine the meteorological factors with a correlation greater than a preset correlation threshold as target meteorological factors; Principal component analysis is used to reduce the dimensionality of the target meteorological factors, extract the principal component factors, and generate a comprehensive meteorological factor curve based on the principal component factors. A photovoltaic output curve is generated based on the photovoltaic output data.

[0014] The clustering module in the apparatus provided in this disclosure is specifically used for: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated based on the statistical distance index, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated based on the statistical distance index. The comprehensive meteorological factor curve and the photovoltaic output curve are clustered in the first stage based on the mean of the first curve, the maximum value of the first curve, the mean of the second curve, and the maximum value of the second curve, to obtain the first clustering result. The first clustering result includes cloudy type result, partly cloudy type result, and mixed type result.

[0015] The clustering module in the apparatus provided in this disclosure is specifically used for: The mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve are determined by using the line distance index. Based on the mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve, the overall situation of daily irradiance fluctuation is obtained. Based on the overall fluctuation of daily irradiance, a second-stage clustering was performed on the mixed-type results to obtain the sunny-day type results and the changing-weather type results. The results of the changing weather type include the changing weather type-comprehensive meteorological factor curve and the changing weather type-photovoltaic output curve; The time-period offset characteristics between the weather change type-comprehensive meteorological factor curve and the weather change type-photovoltaic output curve are determined using the angular distance index. The third-stage clustering of the weather type results is performed based on the time period offset characteristics to obtain the weather transition type results.

[0016] The apparatus provided in this disclosure embodiment, wherein the first construction module is specifically used for: Extract the daily comprehensive meteorological factor curves and comprehensive meteorological factor cluster center curves for all historical days in the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; The point-by-point error is calculated based on the daily comprehensive meteorological factor curve and the comprehensive meteorological factor cluster center curve. Nonparametric kernel density estimation is performed on the point-by-point errors to obtain the error distribution estimate of the daily comprehensive meteorological factor curve corresponding to each type of result, and an error distribution estimation model is constructed based on the error distribution estimates of all types.

[0017] The apparatus provided in this disclosure embodiment, wherein the second construction module is specifically used for: Retrieve historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; Several preset models are initialized, and the corresponding types of results are used to train the several preset models respectively, and the output values ​​of the corresponding types of results are obtained for each preset model after training; An ensemble model is constructed by combining all pre-trained models. The output of the ensemble model is the average of the output values ​​of all pre-trained models for the corresponding type of result.

[0018] The output module of the apparatus provided in this disclosure embodiment is specifically used for: Acquire the data for the time period to be generated, which includes daily meteorological data and daily weather type, and match a preset intermediate process based on the data for the time period to be generated; If the data to be generated for a time period is the daily meteorological data of the time period to be generated, the comprehensive meteorological factor curve of the time period to be generated is obtained based on the daily meteorological data of the time period to be generated, and the comprehensive meteorological factor curve of the time period to be generated is compared with the comprehensive meteorological factor cluster center curve in each type of result to determine the daily weather type of the time period to be generated. Input the daily weather type of the period to be generated and the daily comprehensive meteorological factor curve of the period to be generated into the integrated model, and output the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve; If the data to be generated for a period of time is the daily weather type of the period to be generated, the daily comprehensive meteorological factor error curve of the period to be generated is determined according to the error distribution estimation model and the daily weather type of the period to be generated. The daily comprehensive meteorological factor error curve and the comprehensive meteorological factor cluster center curve corresponding to the daily weather type are superimposed to obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated and the daily weather type of the period to be generated are input into the integrated model to output the photovoltaic power generation curve of the period to be generated. The photovoltaic power generation curve is the output of the integrated model under a given confidence level.

[0019] This disclosure also provides an electronic device, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the photovoltaic power output information generation method provided in this disclosure.

[0020] This disclosure also provides a computer-readable storage medium storing a computer program for executing the photovoltaic power output information generation method provided in this disclosure.

[0021] The technical solution provided in this disclosure has the following advantages compared with the prior art: The photovoltaic output information generation method provided in this embodiment employs a multi-stage clustering method to refine the scene extraction of the comprehensive meteorological factor curve and photovoltaic output curve derived from meteorological data and photovoltaic output data. This yields results for cloudy, partly cloudy, sunny, partly cloudy turning sunny, and sunny turning partly cloudy types, fully mining data information and improving the matching accuracy between typical scenes and meteorological characteristics. This provides clearly categorized input for ensemble model training. Based on several types of results, an error distribution estimation model is constructed, and Bayesian ensemble models for the corresponding types of results are trained. The method outputs the photovoltaic daily output curve, providing a highly valuable refined input for time-series production simulation. Attached Figure Description

[0022] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0023] Figure 1 A schematic flowchart illustrating the photovoltaic power output information generation method provided in this embodiment of the disclosure; Figure 2 This is a schematic diagram of the structure of the photovoltaic power output information generation device provided in the embodiments of this disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0024] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0025] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0026] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0027] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0028] It should be noted that the terms "a" and "several" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "a or several".

[0029] The names of messages or information exchanged between several devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0030] To address the aforementioned issues, this disclosure provides a method for generating photovoltaic power output information, which will be described below with reference to specific embodiments.

[0031] Figure 1 This is a flowchart illustrating a photovoltaic power output information generation method provided in an embodiment of the present disclosure. The method can be executed by a photovoltaic power output information generation device, which can be implemented using software and / or hardware and is generally integrated into an electronic device.

[0032] Example 1: This embodiment of the present disclosure provides a method for generating photovoltaic power output information, the method comprising: Acquire meteorological data and photovoltaic power output data, analyze the correlation between the meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data; The comprehensive meteorological factor curve and photovoltaic output curve are clustered in the first stage using statistical distance indices to obtain the first clustering results, which include cloudy, partly cloudy, and mixed types. The mixed type results are clustered in the second stage using linear distance indices to obtain sunny and changing weather types. The changing weather type results are clustered in the third stage using angular distance indices to obtain weather transition type results, which include sunny to partly cloudy and partly cloudy to sunny types. Based on the results of weather conversion type, cloudy type, partly cloudy type, and sunny type, the error distribution is estimated, and an error distribution estimation model is constructed. Obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; train several preset models based on the historical datasets; and combine all trained preset models to construct an ensemble model. Acquire the data for the period to be generated, match the preset intermediate process based on the data for the period to be generated, and obtain the daily comprehensive meteorological factor curve of the period to be generated. Input the daily comprehensive meteorological factor curve of the period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic power generation curve of the period to be generated, as photovoltaic power generation information.

[0033] In this embodiment, meteorological data refers to observational or forecast data describing atmospheric conditions. For example, hourly data recorded daily by a weather station includes total irradiance, direct irradiance, diffuse irradiance, ambient temperature, ambient humidity, wind speed, wind direction, and air pressure.

[0034] In this embodiment, photovoltaic output data refers to the actual power generation data output by the photovoltaic power station. For example, the active power sequence recorded by the power station monitoring system corresponding to the timestamps of meteorological data.

[0035] In this embodiment, the target meteorological factors are key meteorological factors that are strongly correlated with photovoltaic output and screened out through correlation analysis. For example, total irradiance and ambient temperature are screened out based on thresholds.

[0036] In this embodiment, the comprehensive meteorological factor curve is a single-dimensional time series curve obtained by projecting the original meteorological data onto the first principal component, and it is a comprehensive representation of the core meteorological driving force.

[0037] In this embodiment, the photovoltaic output curve is a curve showing the change in the power generation of a photovoltaic power station over time within a day. For example, it is a time-series curve showing the power output climbing from 0 kW to a peak of 5500 kW and then decreasing back to 0 kW from sunrise to sunset on a certain day.

[0038] In this embodiment, the correlation between meteorological factors and photovoltaic output parameters is analyzed, and target meteorological factors with a correlation greater than a preset correlation threshold are screened. Principal component analysis is performed on the target meteorological factors to reduce dimensionality, extract principal component factors, generate a comprehensive meteorological factor curve, and generate a photovoltaic output curve based on the photovoltaic output data.

[0039] In this embodiment, the statistical distance index is a metric used to quantify the difference between two curves. Specifically, the index is defined and calculated by two statistical measures: the mean of the curve (representing the average level of the whole day) and the maximum value of the curve (representing the peak value of the whole day).

[0040] In this embodiment, the first-stage clustering is a process of using the aforementioned statistical distance index as a similarity measure to perform the first clustering of all comprehensive meteorological factor curves and photovoltaic output curves for historical dates. The purpose is to perform an initial classification based on macroscopic overall intensity. For example, the K-means clustering algorithm (K=3) is used to cluster all historical daily curves into 3 major categories.

[0041] In this embodiment, the first clustering result is the grouping result obtained after the first stage of clustering, which divides all historical days into three initial categories. For example, 365 historical day curves are assigned to three clusters: cluster 1 contains 120 curves, cluster 2 contains 100 curves, and cluster 3 contains 145 curves.

[0042] In this embodiment, the cloudy day type result is a category in the first clustering result, and its curve characteristics are characterized by low mean and low maximum value. For example, for dates belonging to this category, the value of the comprehensive meteorological factor curve fluctuates slightly between 0.1 and 0.3 throughout the day, and the peak power of the corresponding photovoltaic output curve is only about 20% of the installed capacity.

[0043] In this embodiment, the cloudy type result is one category in the first clustering result, and its curve characteristics show moderate mean and moderate maximum values. For example, the comprehensive meteorological factor curve for this type of date fluctuates between 0.3 and 0.7, and the photovoltaic output curve has obvious peaks and valleys, with the peak power being approximately 50%-70% of the installed capacity.

[0044] In this embodiment, the mixed-type result is a category within the first clustering result, characterized by high mean and high maximum value. The mixed-type result includes sunny-day type results and change-type results (partly cloudy to sunny and sunny to partly cloudy).

[0045] In this embodiment, the mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated. Based on these four statistical characteristics, a first-stage clustering is performed on the two curves, resulting in three clustering results: cloudy, partly cloudy, and mixed types.

[0046] In this embodiment, the line distance index is a distance metric that quantifies the overall similarity of the shapes of two curves by comparing them point by point. It is used to characterize the overall differences in the mean and fluctuation patterns of the curves. For example, Euclidean distance or dynamic time warping distance is calculated by taking the square root of the sum of the squares of the differences in power values ​​between the two output curves at each corresponding time point.

[0047] In this embodiment, the second-stage clustering uses a line distance metric to group all curves under the mixed-type results a second time, aiming to subdivide them into purer categories based on fluctuation patterns.

[0048] In this embodiment, the sunny day type result is a category obtained from the second-stage clustering, and its curve characteristics are high mean, high maximum value and gentle fluctuation. For example, the comprehensive meteorological factor curve has a smooth single-peak shape, and the photovoltaic output curve is a full arch shape with no sharp drop throughout the day.

[0049] In this embodiment, the changing weather type result is another category obtained from the second-stage clustering. Its curve characteristics are high mean, high maximum value, but drastic fluctuations. For example, the comprehensive meteorological factor curve shows multiple sharp rises and falls within a day, and the photovoltaic output curve correspondingly shows multiple peaks and valleys.

[0050] In this embodiment, the angular distance metric is a measure that quantifies the similarity of curve trends by comparing vector directions (angles), and is particularly suitable for identifying the relative strength relationship between the first and second halves of a curve. For example, the cosine distance treats the first and second halves of the curve as two vectors and calculates the cosine value of the angle between these two vectors.

[0051] In this embodiment, the third-stage clustering uses the angular distance index to group all curves under the changing weather type results a third time, aiming to perform a final subdivision based on their time-shifting patterns. For example, using K-means (K=2), all curves with drastic fluctuations are divided into two categories: those with a high initial value followed by a low initial value and those with a low initial value followed by a high initial value.

[0052] In this embodiment, the weather transition type result is the final category obtained from the third-stage clustering, which specifically describes the direction of the weather transition. For example, sunny to cloudy (the vector intensity in the morning is significantly greater than that in the afternoon) and cloudy to sunny (the vector intensity in the afternoon is significantly greater than that in the morning).

[0053] In this embodiment, the historical dataset is a collection of data prepared for model training after preliminary steps (data preprocessing, clustering and classification) for each final determined weather type (such as cloudy, sunny, and partly cloudy to sunny). Each dataset contains paired data for all historical days under that type. For example, the historical dataset for sunny weather results contains 100 data samples, where the input for each sample is a comprehensive meteorological factor curve, and the output is the corresponding photovoltaic power output curve.

[0054] In this embodiment, the preset model is a machine learning model with a pre-selected and predefined network structure. This refers to three types of Bayesian neural networks: Bayesian-BP, Bayesian-CNN, and Bayesian-LSTM. The input is a comprehensive meteorological factor curve for a specific weather type. The output is a specific photovoltaic power output curve sampled from the above probability distribution during a specific Monte Carlo forward propagation. For example, inputting a sunny day curve into a trained sunny-day-Bayesian-LSTM model, the model performs one sampling and outputs a predicted power output curve. Repeating this process multiple times will yield several slightly different predicted curves.

[0055] In this embodiment, the ensemble model organizes multiple pre-trained models for a specific weather type, allowing them to work collaboratively according to specific rules (such as averaging). Based on the weather type identifier, the corresponding pre-trained model is invoked. The input curve is simultaneously fed into these three models, and each model independently performs multiple Monte Carlo samplings, generating its own set of prediction curves. The input is daily meteorological data or daily weather type, and the output is the average of the output values ​​of all pre-trained models for the corresponding type.

[0056] In this embodiment, the data to be generated for a specific time period refers to the input information for the future time period for which photovoltaic power output forecasting is needed. For example, it is necessary to forecast the three days from next Monday to Wednesday.

[0057] In this embodiment, the photovoltaic output information includes the photovoltaic daily output curve obtained by matching a preset intermediate process or the photovoltaic daily output curve based on the daily weather type.

[0058] In this embodiment, the daily comprehensive meteorological factor curve for the period to be generated is a single-dimensional curve representing the core meteorological driving force of that day, generated using daily meteorological data for the period to be generated through steps such as principal component analysis. For example, based on the forecast meteorological data for next Monday, a curve is generated where the value rises from 0.1 to 0.9 in the morning, remains high in the afternoon, and declines in the evening.

[0059] In this embodiment, the photovoltaic daily power output curve is the final output of the model, a curve showing the change in the photovoltaic power generation of a photovoltaic power station over time on a given day. It contains information about uncertainties. For example, to predict the photovoltaic output for next Monday, the model outputs a mean curve and the upper and lower boundaries around that mean curve.

[0060] The working principle and beneficial effects of this embodiment are as follows: A multi-stage clustering method is used to refine the scene extraction of the comprehensive meteorological factor curve and photovoltaic output curve obtained from meteorological data and photovoltaic output data, and successively obtain the type results of cloudy, partly cloudy, sunny, partly cloudy to sunny, and sunny to partly cloudy, so as to fully explore the data information, improve the matching accuracy of typical scenes and meteorological features, and provide clearly classified input for the training of the integrated model; an error distribution estimation model is constructed based on several types of results, and the Bayesian integrated model of the corresponding types of results is trained to output the photovoltaic daily output curve, providing a refined input with high reference value for time series production simulation.

[0061] Example 2: The method provided in this embodiment of the present disclosure acquires meteorological data and photovoltaic power output data, analyzes the correlation between the meteorological data and photovoltaic power output data, selects a target meteorological factor, generates a comprehensive meteorological factor curve based on the target meteorological factor, and generates a photovoltaic power output curve based on the photovoltaic power output data, including: Acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological factors and photovoltaic power output parameters based on the meteorological data and photovoltaic power output data, and determine the meteorological factors with a correlation greater than a preset correlation threshold as target meteorological factors; Principal component analysis is used to reduce the dimensionality of the target meteorological factors, extract the principal component factors, and generate a comprehensive meteorological factor curve based on the principal component factors. A photovoltaic output curve is generated based on the photovoltaic output data.

[0062] In this embodiment, meteorological factors refer to individual variables selected from meteorological data that may affect photovoltaic power generation. Examples include total irradiance, direct irradiance, diffuse irradiance, and ambient temperature.

[0063] In this embodiment, the photovoltaic output parameter refers to the photovoltaic output data itself, which is the object of analysis and is usually considered as the dependent variable. For example, the instantaneous power generation value collected synchronously with meteorological data.

[0064] In this embodiment, correlation is a statistical indicator used to quantify the strength and direction of the linear relationship between meteorological factors and photovoltaic output, typically using the Pearson correlation coefficient. For example, the correlation coefficient between total irradiance and photovoltaic output is calculated to be 0.92.

[0065] In this embodiment, the preset correlation threshold is a manually set critical value for the correlation coefficient used to screen important meteorological factors. For example, the threshold is set to |r|>0.7.

[0066] In this embodiment, principal component analysis is a statistical dimensionality reduction method that transforms multiple potentially correlated variables into a few uncorrelated comprehensive variables. The principal component extraction process reduces the dimensionality of the feature matrix formed by multidimensional strongly correlated meteorological condition data to form a single-dimensional comprehensive meteorological factor curve.

[0067] In this embodiment, the principal component factor is a new variable obtained through principal component analysis, where the first principal component typically contains the most important information from the original variables. For example, the first principal component PC1 = 0.96 * (normalized irradiance) + 0.28 * (normalized temperature).

[0068] The working principle and beneficial effects of this embodiment are as follows: The correlation between meteorological factors and photovoltaic output parameters is analyzed, and target meteorological factors with a correlation greater than a preset correlation threshold are selected. Principal component analysis is performed on the target meteorological factors to reduce dimensionality, extract principal component factors, and generate a comprehensive meteorological factor curve. A photovoltaic output curve is then generated based on the photovoltaic output data. Selecting strongly correlated meteorological factors and reducing dimensionality effectively removes redundancy and noise, forming a comprehensive factor characterizing the core meteorological drivers. This comprehensive curve is directly related to the photovoltaic output curve, providing high signal-to-noise ratio and dimensionally simple input features for subsequent clustering and prediction, thus improving model efficiency and accuracy.

[0069] Example 3: The method provided in this embodiment of the present disclosure uses a statistical distance index to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve, and obtains the first clustering result, including: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated based on the statistical distance index, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated based on the statistical distance index. The comprehensive meteorological factor curve and the photovoltaic output curve are clustered in the first stage based on the mean of the first curve, the maximum value of the first curve, the mean of the second curve, and the maximum value of the second curve, to obtain the first clustering result. The first clustering result includes cloudy type result, partly cloudy type result, and mixed type result.

[0070] The working principle and beneficial effects of this embodiment are as follows: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated, and the mean and maximum values ​​of the second curve of the photovoltaic power output curve are calculated. Based on these four statistical characteristics, a first-stage clustering is performed on the two curves, yielding three clustering results: cloudy, partly cloudy, and mixed types. By integrating the core statistical characteristics of meteorological driving forces and power generation response for clustering, the classification results are ensured to simultaneously possess meteorological and physical significance and consistency in power output patterns, achieving a rapid and robust macroscopic initial classification of historical power output patterns.

[0071] Example 4: The method provided in this embodiment uses a linear distance index to perform a second-stage clustering of the mixed-type results to obtain sunny weather type results and changing weather type results, and uses an angular distance index to perform a third-stage clustering of the changing weather type results to obtain weather transition type results. The mixed-type results include a mixed-type comprehensive meteorological factor curve and a mixed-type photovoltaic output curve, including: The mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve are determined by using the line distance index. Based on the mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve, the overall situation of daily irradiance fluctuation is obtained. Based on the overall fluctuation of daily irradiance, a second-stage clustering was performed on the mixed-type results to obtain the sunny-day type results and the changing-weather type results. The results of the changing weather type include the changing weather type-comprehensive meteorological factor curve and the changing weather type-photovoltaic output curve; The time-period offset characteristics between the weather change type-comprehensive meteorological factor curve and the weather change type-photovoltaic output curve are determined using the angular distance index. The third-stage clustering of the weather type results is performed based on the time period offset characteristics to obtain the weather transition type results.

[0072] In this embodiment, the mixed-type-integrated meteorological factor curve is the daily integrated meteorological factor curve included in the mixed-type results obtained from the first-stage clustering. For example, there is a batch of curves with high values ​​but different shapes, some smooth and some fluctuating sharply.

[0073] In this embodiment, the hybrid type photovoltaic output curve is the daily photovoltaic output curve corresponding to the aforementioned meteorological curve. For example, a batch of output curves with very high peak power but different shapes, some are single-peaked arches, and some are multi-peaked sawtooth shapes.

[0074] In this embodiment, the overall situation of daily irradiance fluctuation is a description of the above-mentioned curve morphology characteristics, referring to the severity and pattern of fluctuations in irradiance and photovoltaic power output throughout the day.

[0075] In this embodiment, the weather change type-comprehensive meteorological factor curve is the daily comprehensive meteorological factor curve belonging to the weather change type. For example, a curve shows a value that drops sharply from 0.8 to 0.3 in the morning, rebounds to 0.9 in the afternoon, and then drops rapidly again in the evening.

[0076] In this embodiment, the changing weather type-photovoltaic output curve is the daily photovoltaic output curve corresponding to the curve mentioned above. For example, the power suddenly drops from 4.5 MW to 1.0 MW at noon due to cloud cover, and then quickly recovers to 4.0 MW after the clouds clear.

[0077] In this embodiment, the time-period offset characteristic refers to the difference in the distribution of curve energy or intensity at different times of the day. For example, the energy is stronger in the morning than in the afternoon, or vice versa.

[0078] In this embodiment, a three-stage clustering algorithm is proposed. Based on indicators such as the average daily power, the maximum daily power, the weighted fluctuation range, and the custom skewness, three comprehensive indicators—statistical distance, linear distance, and angular distance—are constructed as clustering features for the first, second, and third stages of photovoltaic daily power generation. These indicators accurately characterize the impact of meteorological factors such as the magnitude of daily irradiance, the degree of change in irradiance throughout the day, and the changes in irradiance at different times of the day on photovoltaic power generation, thereby achieving refined extraction of typical weather scenarios for photovoltaic output.

[0079] The working principle and beneficial effects of this embodiment are as follows: By using the linear distance index to measure the mean and fluctuation characteristics of the mixed-type results, clustering is performed to obtain clear weather type and changing weather type. The angular distance index is used to measure the time-period offset characteristics of the changing weather type results, and clustering is performed to obtain weather transition type results. Through progressive clustering using linear and angular distances, refined separation of complex high-irradiance scenes is achieved, identifying stable and changing patterns respectively, and further analyzing the temporal offset patterns within the changing patterns to accurately depict the weather transition process.

[0080] Example 5: The method provided in this embodiment of the present disclosure estimates the error distribution based on the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, and constructs an error distribution estimation model, including: Extract the daily comprehensive meteorological factor curves and comprehensive meteorological factor cluster center curves for all historical days in the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; The point-by-point error is calculated based on the daily comprehensive meteorological factor curve and the comprehensive meteorological factor cluster center curve. Nonparametric kernel density estimation is performed on the point-by-point errors to obtain the error distribution estimate of the daily comprehensive meteorological factor curve corresponding to each type of result, and an error distribution estimation model is constructed based on the error distribution estimates of all types.

[0081] In this embodiment, the daily comprehensive meteorological factor curve refers to the comprehensive meteorological factor curve for each specific historical day belonging to a certain weather type.

[0082] In this embodiment, the integrated meteorological factor cluster center curve refers to the integrated meteorological factor curve obtained through clustering, which represents the average shape or typical pattern of a certain weather type. It is the average template of all historical daily curves under this type.

[0083] In this embodiment, point-by-point error refers to the difference between the observed value of the daily curve and the typical value of the cluster center curve at the same time point.

[0084] In this embodiment, nonparametric kernel density estimation is a probability density function estimation method that does not presuppose that the data follows a specific distribution. By placing a smooth kernel at each data point and then superimposing all kernel functions, a continuous curve describing the shape of the data distribution is obtained.

[0085] In this embodiment, the error distribution estimation is a mathematical model obtained by nonparametric kernel density estimation, which describes the probability of a certain type of error value occurring at a specific time point.

[0086] In this embodiment, the error distribution estimation model stores the error probability density function for each weather type learned through KDE. The input is the weather type and time information; the output is that, given a weather type, the model can generate a random error value that conforms to the historical error statistics of that type. For example, when a request is made to generate an error for the sunny weather type, the model will randomly select a value, such as +0.02, from its fitted sunny weather error distribution.

[0087] The working principle and beneficial effects of this embodiment are as follows: Daily comprehensive meteorological factor curves and corresponding cluster center curves for each type of historical day are extracted, and point-by-point errors are calculated. Non-parametric kernel density estimation is performed on various errors to obtain error distribution estimates for each type of comprehensive meteorological factor curve, and an error distribution estimation model is constructed. Kernel density estimation quantifies the random fluctuation characteristics of comprehensive meteorological factors for each weather type around their center curves, establishing a probabilistic model describing the sources of uncertainty, and providing a random error sampling basis that conforms to historical statistical patterns for subsequent scene generation.

[0088] Example 6: The method provided in this embodiment of the present disclosure obtains historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; trains several preset models based on the historical datasets; and constructs an ensemble model by combining all trained preset models, including: Retrieve historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; Several preset models are initialized, and the corresponding types of results are used to train the several preset models respectively, and the output values ​​of the corresponding types of results are obtained for each preset model after training; An ensemble model is constructed by combining all pre-trained models. The output of the ensemble model is the average of the output values ​​of all pre-trained models for the corresponding type of result.

[0089] In this embodiment, based on the Bagging ensemble framework, three neural networks with Bayesian layers—Bayesian-LSTM, Bayesian-BP, and Bayesian-CNN—are constructed as preset models. The basic structure of Bayesian-LSTM is the same as that of a traditional LSTM neural network. The Bayesian-BP neural network contains three hidden layers, and Bayesian-CNN consists of two one-dimensional convolutional layers, two linearization layers, and two normalization layers. The final output of the ensemble model is set as the average of the outputs of the three base learners.

[0090] In this embodiment, the training process involves inputting parameters such as the power plant name, a list of weather types, the number of daily curve data points, the maximum number of training iterations, the number of hidden neurons in the neural network, the training batch size, the training error limit, the training gradient descent learning rate, and the training weight decay coefficient. The model is then trained for each type of weather in the weather type list to obtain the parameters of the Bayesian-LSTM, Bayesian-BP, and Bayesian-CNN neural networks for each weather type.

[0091] In this embodiment, the Bagging ensemble learning method is employed, and three types of neural network base learners with Bayesian layers are constructed. Bayesian layers are embedded before the linear layers of LSTM, CNN, and BP neural networks. By probabilistically representing the weights and variances, the uncertainty of power generation is accurately quantified.

[0092] The working principle and beneficial effects of this embodiment are as follows: Historical datasets corresponding to the results of each weather type are obtained, and several preset models are initialized and trained for each type of result. An ensemble model is constructed by combining all trained models, and its output is the average of the output values ​​of all models for the corresponding type. By training dedicated model sets for different weather types and then ensembled, the models are highly matched to specific meteorological scenarios, effectively improving the prediction accuracy for each type. The ensemble averaging reduces the random error of a single model, enhancing the overall stability and robustness of the prediction.

[0093] Example 7: The method provided in this embodiment of the present disclosure obtains data of a time period to be generated, matches a preset intermediate process based on the data of the time period to be generated, obtains a daily comprehensive meteorological factor curve of the time period to be generated, and inputs the daily comprehensive meteorological factor curve of the time period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic daily power output curve of the time period to be generated, as photovoltaic power output information, including: Acquire the data for the time period to be generated, which includes daily meteorological data and daily weather type, and match a preset intermediate process based on the data for the time period to be generated; If the data to be generated for a time period is the daily meteorological data of the time period to be generated, the comprehensive meteorological factor curve of the time period to be generated is obtained based on the daily meteorological data of the time period to be generated, and the comprehensive meteorological factor curve of the time period to be generated is compared with the comprehensive meteorological factor cluster center curve in each type of result to determine the daily weather type of the time period to be generated. Input the daily weather type of the period to be generated and the daily comprehensive meteorological factor curve of the period to be generated into the integrated model, and output the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve; If the data to be generated for a period of time is the daily weather type of the period to be generated, the daily comprehensive meteorological factor error curve of the period to be generated is determined according to the error distribution estimation model and the daily weather type of the period to be generated. The daily comprehensive meteorological factor error curve and the comprehensive meteorological factor cluster center curve corresponding to the daily weather type are superimposed to obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated and the daily weather type of the period to be generated are input into the integrated model to output the photovoltaic power generation curve of the period to be generated. The photovoltaic power generation curve is the output of the integrated model under a given confidence level.

[0094] In this embodiment, the daily meteorological data refers to the specific meteorological observations or forecasts for each day within the time period to be generated. For example, a 24-hour sequence of indicators such as temperature, irradiance, and humidity for next Monday.

[0095] In this embodiment, the daily weather type is a pre-defined macro-weather category label for each day within the time period to be generated. For example, the weather forecast indicates that next Monday will be sunny and next Tuesday will be cloudy turning sunny.

[0096] In this embodiment, the preset intermediate process refers to two different internal processing paths automatically selected based on the type of input data. For example, path one inputs meteorological data, generates a composite curve, matches the weather type, calls the model, and outputs the result; path two inputs the weather type, generates an error curve, synthesizes the composite curve, calls the model, and outputs the result.

[0097] In this embodiment, comparison refers to calculating the similarity between the newly generated daily integrated meteorological factor curve for the period to be generated and the cluster center curves of various integrated meteorological factors obtained from historical clustering, in order to determine the most likely weather type. For example, if the new curve is the closest to the sunny weather center curve, then the day is determined to be a sunny day.

[0098] In this embodiment, the daily integrated meteorological factor error curve is an error sequence randomly sampled from the error distribution model corresponding to the weather type when the input is a weather type. It simulates the reasonable random fluctuations of meteorological conditions around the cluster center curve of the integrated meteorological factors corresponding to that type.

[0099] In this embodiment, the given confidence level refers to the probability requirement for the range of uncertainty in the prediction result, which is a parameter preset by the user. For example, the confidence level is set to 90%. The upper and lower boundaries of the photovoltaic daily power output curve output by the model mean that there is a 90% probability that the actual power output will fall within this range. For example, at noon, the upper limit of power output is 5.5 MW, and the lower limit is 4.9 MW.

[0100] The working principle and beneficial effects of this embodiment are as follows: different paths are matched according to the data type of the time period to be generated. When meteorological data is input, a comprehensive curve is generated and matched with the weather type. When the weather type is input, an error curve is generated using an error distribution model. The comprehensive curve is obtained by superimposing the center curve. Both types of paths ultimately input the comprehensive curve and the weather type into the integrated model, outputting a photovoltaic power output curve with a given confidence level. This achieves a unified prediction process under two input modes: meteorological data and weather type. Through the synergy of the integrated model and the error distribution model, regardless of the input data, a photovoltaic power output result that combines scenario adaptability and probabilistic uncertainty information can be generated, improving the practicality and flexibility of the method.

[0101] To achieve the above embodiments, this disclosure also proposes a photovoltaic power output information generation device.

[0102] Figure 2 This is a schematic diagram of the structure of a photovoltaic output information generation device provided in an embodiment of this disclosure. This device can be implemented by software and / or hardware, and is generally integrated into an electronic device. For example... Figure 2 As shown, the device 200 includes: an acquisition module 201, a clustering module 202, a first construction module 203, a second construction module 204, and an output module 205, wherein, The acquisition module 201 is used to acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data. Clustering module 202 is used to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve using statistical distance indicators to obtain a first clustering result. The first clustering result includes cloudy type results, partly cloudy type results, and mixed type results. The mixed type results are then clustered in a second stage using linear distance indicators to obtain sunny type results and changing weather type results. Finally, the changing weather type results are clustered in a third stage using angular distance indicators to obtain weather transition type results. The weather transition type results include sunny to partly cloudy type results and partly cloudy to sunny type results. The first construction module 203 is used to estimate the error distribution based on the weather conversion type results, cloudy type results, partly cloudy type results and sunny type results, and to construct an error distribution estimation model. The second construction module 204 is used to obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results and sunny type results, train several preset models based on the historical datasets, and combine all trained preset models to construct an integrated model. The output module 205 is used to acquire the data of the period to be generated, match the preset intermediate process according to the data of the period to be generated, obtain the daily comprehensive meteorological factor curve of the period to be generated, and input the daily comprehensive meteorological factor curve of the period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve, as photovoltaic power generation information.

[0103] The apparatus provided in this disclosure, wherein the acquisition module 201 is specifically used for: Acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological factors and photovoltaic power output parameters based on the meteorological data and photovoltaic power output data, and determine the meteorological factors with a correlation greater than a preset correlation threshold as target meteorological factors; Principal component analysis is used to reduce the dimensionality of the target meteorological factors, extract the principal component factors, and generate a comprehensive meteorological factor curve based on the principal component factors. A photovoltaic output curve is generated based on the photovoltaic output data.

[0104] The clustering module 202 in the apparatus provided in this disclosure is specifically used for: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated based on the statistical distance index, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated based on the statistical distance index. The comprehensive meteorological factor curve and the photovoltaic output curve are clustered in the first stage based on the mean of the first curve, the maximum value of the first curve, the mean of the second curve, and the maximum value of the second curve, to obtain the first clustering result. The first clustering result includes cloudy type result, partly cloudy type result, and mixed type result.

[0105] The clustering module 202 in the apparatus provided in this disclosure is specifically used for: The mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve are determined by using the line distance index. Based on the mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve, the overall situation of daily irradiance fluctuation is obtained. Based on the overall fluctuation of daily irradiance, a second-stage clustering was performed on the mixed-type results to obtain the sunny-day type results and the changing-weather type results. The results of the changing weather type include the changing weather type-comprehensive meteorological factor curve and the changing weather type-photovoltaic output curve; The time-period offset characteristics between the weather change type-comprehensive meteorological factor curve and the weather change type-photovoltaic output curve are determined using the angular distance index. The third-stage clustering of the weather type results is performed based on the time period offset characteristics to obtain the weather transition type results.

[0106] The apparatus provided in this disclosure embodiment, wherein the first construction module 203 is specifically used for: Extract the daily comprehensive meteorological factor curves and comprehensive meteorological factor cluster center curves for all historical days in the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; The point-by-point error is calculated based on the daily comprehensive meteorological factor curve and the comprehensive meteorological factor cluster center curve. Nonparametric kernel density estimation is performed on the point-by-point errors to obtain the error distribution estimate of the daily comprehensive meteorological factor curve corresponding to each type of result, and an error distribution estimation model is constructed based on the error distribution estimates of all types.

[0107] The apparatus provided in this disclosure embodiment, wherein the second construction module 204 is specifically used for: Retrieve historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; Several preset models are initialized, and the corresponding types of results are used to train the several preset models respectively, and the output values ​​of the corresponding types of results are obtained for each preset model after training; An ensemble model is constructed by combining all pre-trained models. The output of the ensemble model is the average of the output values ​​of all pre-trained models for the corresponding type of result.

[0108] The output module 205 of the apparatus provided in this embodiment is specifically used for: Acquire the data for the time period to be generated, which includes daily meteorological data and daily weather type, and match a preset intermediate process based on the data for the time period to be generated; If the data to be generated for a time period is the daily meteorological data of the time period to be generated, the comprehensive meteorological factor curve of the time period to be generated is obtained based on the daily meteorological data of the time period to be generated, and the comprehensive meteorological factor curve of the time period to be generated is compared with the comprehensive meteorological factor cluster center curve in each type of result to determine the daily weather type of the time period to be generated. Input the daily weather type of the period to be generated and the daily comprehensive meteorological factor curve of the period to be generated into the integrated model, and output the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve; If the data to be generated for a period of time is the daily weather type of the period to be generated, the daily comprehensive meteorological factor error curve of the period to be generated is determined according to the error distribution estimation model and the daily weather type of the period to be generated. The daily comprehensive meteorological factor error curve and the comprehensive meteorological factor cluster center curve corresponding to the daily weather type are superimposed to obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated and the daily weather type of the period to be generated are input into the integrated model to output the photovoltaic power generation curve of the period to be generated. The photovoltaic power generation curve is the output of the integrated model under a given confidence level.

[0109] The photovoltaic power output information generation device provided in this disclosure can execute the photovoltaic power output information generation method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects of executing the method.

[0110] To implement the above embodiments, this disclosure also proposes a computer program product, including a computer program / instructions, which, when executed by a processor, implements the photovoltaic power output information generation method in the above embodiments.

[0111] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure.

[0112] The following is a detailed reference. Figure 3 The diagram illustrates a structural schematic suitable for implementing the electronic device 300 in the embodiments of this disclosure. The electronic device 300 in the embodiments of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0113] like Figure 3 As shown, the electronic device 300 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a memory 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processor 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0114] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0115] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 309, or installed from a memory 308, or installed from a ROM 302. When the computer program is executed by the processor 301, it performs the functions defined in the photovoltaic power output information generation method of embodiments of this disclosure.

[0116] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0117] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0118] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0119] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the aforementioned photovoltaic power output information generation method.

[0120] Electronic devices can be programmed with computer program code in one or more programming languages ​​or combinations thereof to perform the operations of this disclosure. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0122] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0123] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0124] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0125] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0126] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in several embodiments.

[0127] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for generating photovoltaic power output information, characterized in that, include: Acquire meteorological data and photovoltaic power output data, analyze the correlation between the meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data; The comprehensive meteorological factor curve and photovoltaic output curve are clustered in the first stage using statistical distance indices to obtain the first clustering results, which include cloudy, partly cloudy, and mixed types. The mixed type results are clustered in the second stage using linear distance indices to obtain sunny and changing weather types. The changing weather type results are clustered in the third stage using angular distance indices to obtain weather transition type results, which include sunny to partly cloudy and partly cloudy to sunny types. Based on the results of weather conversion type, cloudy type, partly cloudy type, and sunny type, the error distribution is estimated, and an error distribution estimation model is constructed. Obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; train several preset models based on the historical datasets; and combine all trained preset models to construct an ensemble model. Acquire the data for the period to be generated, match the preset intermediate process based on the data for the period to be generated, and obtain the daily comprehensive meteorological factor curve of the period to be generated. Input the daily comprehensive meteorological factor curve of the period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic power generation curve of the period to be generated, as photovoltaic power generation information.

2. The method according to claim 1, characterized in that, Acquire meteorological data and photovoltaic (PV) output data, analyze the correlation between the meteorological data and PV output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a PV output curve based on the PV output data, including: Acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological factors and photovoltaic power output parameters based on the meteorological data and photovoltaic power output data, and determine the meteorological factors with a correlation greater than a preset correlation threshold as target meteorological factors; Principal component analysis is used to reduce the dimensionality of the target meteorological factors, extract the principal component factors, and generate a comprehensive meteorological factor curve based on the principal component factors. A photovoltaic output curve is generated based on the photovoltaic output data.

3. The method according to claim 1, characterized in that, The statistical distance index was used to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve, yielding the first clustering results, including: The mean and maximum values ​​of the first curve of the comprehensive meteorological factor curve are calculated based on the statistical distance index, and the mean and maximum values ​​of the second curve of the photovoltaic output curve are calculated based on the statistical distance index. The comprehensive meteorological factor curve and the photovoltaic output curve are clustered in the first stage based on the mean of the first curve, the maximum value of the first curve, the mean of the second curve, and the maximum value of the second curve, to obtain the first clustering result. The first clustering result includes cloudy type result, partly cloudy type result, and mixed type result.

4. The method according to claim 1, characterized in that, A second-stage clustering method is used to perform clustering on the mixed-type results using the linear distance index, yielding clear weather type results and changing weather type results. A third-stage clustering method is then used to perform clustering on the changing weather type results, yielding weather transition type results. The mixed-type results include a mixed-type comprehensive meteorological factor curve and a mixed-type photovoltaic output curve, including: The mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve are determined by using the line distance index. Based on the mean and fluctuation characteristics of the mixed type-comprehensive meteorological factor curve and the mixed type-photovoltaic power output curve, the overall situation of daily irradiance fluctuation is obtained. Based on the overall fluctuation of daily irradiance, a second-stage clustering was performed on the mixed-type results to obtain the sunny-day type results and the changing-weather type results. The results of the changing weather type include the changing weather type-comprehensive meteorological factor curve and the changing weather type-photovoltaic output curve; The time-period offset characteristics between the weather change type-comprehensive meteorological factor curve and the weather change type-photovoltaic output curve are determined using the angular distance index. The third-stage clustering of the weather type results is performed based on the time period offset characteristics to obtain the weather transition type results.

5. The method according to claim 1, characterized in that, Error distribution estimation is performed based on weather type results, including cloudy, overcast, and sunny weather results. An error distribution estimation model is then constructed, comprising: Extract the daily comprehensive meteorological factor curves and comprehensive meteorological factor cluster center curves for all historical days in the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; The point-by-point error is calculated based on the daily comprehensive meteorological factor curve and the comprehensive meteorological factor cluster center curve. Nonparametric kernel density estimation is performed on the point-by-point errors to obtain the error distribution estimate of the daily comprehensive meteorological factor curve corresponding to each type of result, and an error distribution estimation model is constructed based on the error distribution estimates of all types.

6. The method according to claim 1, characterized in that, Obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results. Train several preset models based on the historical datasets, and construct an ensemble model by combining all trained preset models, including: Retrieve historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results; Several preset models are initialized, and the corresponding types of results are used to train the several preset models respectively, and the output values ​​of the corresponding types of results are obtained for each preset model after training; An ensemble model is constructed by combining all pre-trained models. The output of the ensemble model is the average of the output values ​​of all pre-trained models for the corresponding type of result.

7. The method according to claim 1, characterized in that, Acquire the data for the period to be generated, match the data with a preset intermediate process to obtain the daily comprehensive meteorological factor curve for the period to be generated, and input the daily comprehensive meteorological factor curve for the period to be generated into the integrated model and the error distribution estimation model to determine the daily weather type and the corresponding photovoltaic power output curve for the period to be generated, as photovoltaic power output information, including: Acquire the data for the time period to be generated, which includes daily meteorological data and daily weather type, and match a preset intermediate process based on the data for the time period to be generated; If the data to be generated for a time period is the daily meteorological data of the time period to be generated, the comprehensive meteorological factor curve of the time period to be generated is obtained based on the daily meteorological data of the time period to be generated, and the comprehensive meteorological factor curve of the time period to be generated is compared with the comprehensive meteorological factor cluster center curve in each type of result to determine the daily weather type of the time period to be generated. Input the daily weather type of the period to be generated and the daily comprehensive meteorological factor curve of the period to be generated into the integrated model, and output the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve; If the data to be generated for a period of time is the daily weather type of the period to be generated, the daily comprehensive meteorological factor error curve of the period to be generated is determined according to the error distribution estimation model and the daily weather type of the period to be generated. The daily comprehensive meteorological factor error curve and the comprehensive meteorological factor cluster center curve corresponding to the daily weather type are superimposed to obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated and the daily weather type of the period to be generated are input into the integrated model to output the photovoltaic power generation curve of the period to be generated. The photovoltaic power generation curve is the output of the integrated model under a given confidence level.

8. A photovoltaic power output information generation device, the device comprising: The acquisition module is used to acquire meteorological data and photovoltaic power output data, analyze the correlation between meteorological data and photovoltaic power output data, select target meteorological factors, generate a comprehensive meteorological factor curve based on the target meteorological factors, and generate a photovoltaic power output curve based on the photovoltaic power output data. The clustering module is used to perform a first-stage clustering of the comprehensive meteorological factor curve and the photovoltaic output curve using statistical distance indicators to obtain a first clustering result. The first clustering result includes cloudy type results, partly cloudy type results, and mixed type results. The mixed type results are then clustered in a second stage using linear distance indicators to obtain sunny type results and changing weather type results. Finally, the changing weather type results are clustered in a third stage using angular distance indicators to obtain weather transition type results. The weather transition type results include sunny to partly cloudy type results and partly cloudy to sunny type results. The first construction module is used to estimate the error distribution based on the weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, and to construct an error distribution estimation model. The second construction module is used to obtain historical datasets corresponding to weather conversion type results, cloudy type results, partly cloudy type results, and sunny type results, train several preset models based on the historical datasets, and combine all trained preset models to construct an integrated model. The output module is used to acquire the data of the period to be generated, match the preset intermediate process according to the data of the period to be generated, and obtain the daily comprehensive meteorological factor curve of the period to be generated. The daily comprehensive meteorological factor curve of the period to be generated is input into the integrated model and the error distribution estimation model to determine the daily weather type of the period to be generated and the corresponding photovoltaic power generation curve, which is used as photovoltaic power generation information.

9. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the photovoltaic output information generation method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, causes the processor to implement the photovoltaic power output information generation method provided in any one of claims 1-7.

Images