Distributed photovoltaic power generation control system and method based on artificial intelligence
Through the distributed photovoltaic power generation control system based on artificial intelligence, deep learning technology is used to analyze sensor data, and real-time monitoring and adjustment of energy storage capacity, the problem of power output instability of distributed photovoltaic power generation systems is solved, and the energy utilization efficiency and system are improved stability.
Patent Information
- Application Number
- CN202410352700.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-26
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-03-26
AI Technical Summary
There is instability in the power output of distributed photovoltaic power generation systems, which makes it difficult to maximize the utilization of energy storage capacity management, affecting energy utilization efficiency and system stability.
Adopt a distributed photovoltaic power generation control system based on artificial intelligence, collect voltage signals, environmental irradiance and power values through sensors, and use deep learning technology to extract feature vectors and correlation analysis to determine whether energy storage capacity needs to be adjusted, real-time monitoring and adjustment .
It improves energy utilization efficiency, saves costs, enhances system stability, and optimizes energy planning to ensure the maximum utilization of energy storage capacity.
Smart Images

Figure CN118300519B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of photovoltaic power generation control, and more specifically, to an artificial intelligence-based distributed photovoltaic power generation control system and method. Background Art
[0002] Distributed photovoltaic power generation is a new type of energy system that uses photovoltaic modules to directly convert solar energy into electrical energy. It is characterized by being distributed in multiple different locations, including buildings, public facilities, and industrial sites. This power generation method advocates the principle of generating electricity nearby and using it nearby. The operation mode is mainly based on the user's self-generation and self-use, and the excess electricity is connected to the Internet, and it can be balanced and adjusted in the distribution system. Distributed photovoltaic power generation is clean, efficient, and decentralized. It is conducive to energy conservation and emission reduction, improves energy utilization, and has broad development prospects. It is an important means to achieve energy structure transformation and sustainable development.
[0003] The output of photovoltaic power generation is affected by many variables, including weather conditions, seasonal changes, and daily time cycles, which leads to a certain instability in its power output. Therefore, it is necessary to manage and adjust the energy storage capacity. When the power generated by the photovoltaic system exceeds the immediate demand, the excess power can be stored. When power production decreases or cannot meet demand, the stored power can be released to maintain the balance of energy supply and demand. Therefore, judging the energy storage capacity of each distributed photovoltaic device and maximizing the use of photovoltaic energy are important issues in the application of new energy.
[0004] Therefore, a distributed photovoltaic power generation control system and method based on artificial intelligence is desired. Summary of the invention
[0005] In order to solve the above technical problems, this application is proposed. The embodiment of the present application provides a distributed photovoltaic power generation control system and method based on artificial intelligence, which first obtains the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally obtains the classification result through the classifier to determine whether the energy storage capacity of the distributed photovoltaic equipment needs to be adjusted, thereby realizing real-time monitoring and adjustment of the energy storage capacity, thereby improving energy utilization efficiency, saving costs, improving system stability, and optimizing energy planning.
[0006] According to one aspect of the present application, there is provided a distributed photovoltaic power generation control system based on artificial intelligence, which includes:
[0007] A distributed photovoltaic data acquisition module, used to acquire voltage signals collected by sensors from distributed photovoltaic power generation equipment at a preset sampling frequency, environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by sensors, and energy storage capacity of distributed photovoltaic power generation equipment at multiple predetermined time points;
[0008] A distributed photovoltaic data extraction module is used to extract a distributed electric energy-related feature vector and an energy storage capacity-related feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points;
[0009] The photovoltaic energy storage capacity adjustment result generation module is used to determine whether the distributed photovoltaic device energy storage capacity needs to be adjusted based on the distributed electric energy associated feature vector and the energy storage capacity associated feature vector.
[0010] According to another aspect of the present application, there is provided a distributed photovoltaic power generation control method based on artificial intelligence, which includes:
[0011] Acquire voltage signals collected by sensors from distributed photovoltaic power generation equipment at a preset sampling frequency, ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by sensors, and energy storage capacity of distributed photovoltaic power generation equipment at multiple predetermined time points;
[0012] Extracting a distributed electric energy associated feature vector and an energy storage capacity associated feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points;
[0013] Based on the distributed electric energy associated characteristic vector and the energy storage capacity associated characteristic vector, it is determined whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted.
[0014] Compared with the prior art, the present application provides an artificial intelligence-based distributed photovoltaic power generation control system and method, which first obtains the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally obtains the classification result through the classifier to determine whether the energy storage capacity of the distributed photovoltaic equipment needs to be adjusted, thereby realizing real-time monitoring and adjustment of the energy storage capacity, thereby improving energy utilization efficiency, saving costs, improving system stability, and optimizing energy planning. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0016] Figure 1 It is a block diagram of a distributed photovoltaic power generation control system based on artificial intelligence according to an embodiment of the present application.
[0017] Figure 2 It is a block diagram of a distributed photovoltaic data extraction module in an artificial intelligence-based distributed photovoltaic power generation control system according to an embodiment of the present application.
[0018] Figure 3 It is a block diagram of an electric energy data feature extraction unit in an artificial intelligence-based distributed photovoltaic power generation control system according to an embodiment of the present application.
[0019] Figure 4 This is a flow chart of a distributed photovoltaic power generation control method based on artificial intelligence according to an embodiment of the present application.
[0020] Figure 5 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION
[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.
[0022] Exemplary Systems
[0023] Figure 1FIG. 1 is a block diagram of a distributed photovoltaic power generation control system based on artificial intelligence according to an embodiment of the present application. Figure 1 As shown, according to the embodiment of the present application, the distributed photovoltaic power generation control system 100 based on artificial intelligence includes: a distributed photovoltaic data acquisition module 110, which is used to obtain the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points; a distributed photovoltaic data extraction module 120, which is used to extract the distributed electric energy associated feature vector and the energy storage capacity associated feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points; a photovoltaic energy storage capacity adjustment result generation module 130, which is used to determine whether the energy storage capacity of the distributed photovoltaic equipment needs to be adjusted based on the distributed electric energy associated feature vector and the energy storage capacity associated feature vector.
[0024] In the above-mentioned distributed photovoltaic power generation control system 100 based on artificial intelligence, the distributed photovoltaic data acquisition module 110 is used to obtain the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power value at multiple predetermined time points in a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points. It should be understood that by collecting and analyzing the voltage signal, the stability, power quality and power output of the system can be understood. Environmental irradiance refers to the energy flow density of solar radiation per unit area, and the power value refers to the actual output power of the photovoltaic power generation system. Environmental irradiance and power value are both important factors affecting the power generation efficiency of the photovoltaic power generation system. By collecting environmental irradiance and power values at multiple predetermined time points, a correlation model between the two can be established, so that the impact of environmental irradiance on power output can be better understood. The energy storage capacity of the distributed photovoltaic power generation system is an important parameter for adjusting the power supply capacity and power balance of the system. By collecting the energy storage capacity of distributed photovoltaic power generation equipment at multiple predetermined time points, the energy storage status and change trend of the system can be understood. This helps to determine whether the energy storage capacity needs to be adjusted to meet different load requirements and optimize the operation of the system. This information is of great significance for performance evaluation, power optimization and energy storage capacity adjustment of distributed photovoltaic power generation systems.
[0025] In the above-mentioned distributed photovoltaic power generation control system 100 based on artificial intelligence, the distributed photovoltaic data extraction module 120 is used to extract the distributed electric energy associated feature vector and the energy storage capacity associated feature vector from the voltage signal collected by the sensor at a preset sampling frequency from the distributed photovoltaic power generation equipment, the environmental irradiance and power value at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points. It should be understood that the distributed electric energy associated feature vector and the energy storage capacity associated feature vector provide a comprehensive description of the system operation status, which can help system managers better understand and manage the power generation system. Specifically, the extraction of the distributed electric energy associated feature vector is mainly through feature extraction and encoding of data such as voltage signals, environmental irradiance and power values. These features may include frequency domain features, time domain features, statistical features, etc., which are used to describe the correlation and characteristic patterns between electric energy. The extraction of the energy storage capacity associated feature vector is mainly to extract features from the energy storage capacity data of the distributed photovoltaic power generation equipment. These features may include the changing trend of the energy storage capacity, the degree of fluctuation of the energy storage capacity, the utilization rate of the energy storage capacity, etc., which are used to describe the state and change of the energy storage capacity. By extracting the energy storage capacity-related feature vectors, the energy storage status of each power generation equipment can be evaluated, providing a basis for the adjustment and optimization of the energy storage capacity.
[0026] Figure 2 FIG. 1 is a block diagram of a distributed photovoltaic data extraction module in a distributed photovoltaic power generation control system based on artificial intelligence according to an embodiment of the present application. Figure 2As shown, in a specific embodiment of the present application, the distributed photovoltaic data extraction module 120 includes: an electric energy data feature extraction unit 121, which is used to extract features from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency and the environmental irradiance and power value at multiple predetermined time points within a predetermined time period collected by the sensor to obtain the distributed electric energy associated feature vector; an energy storage capacity data feature extraction unit 122, which is used to extract features from the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points to obtain the energy storage capacity associated feature vector. It should be understood that through feature extraction, representative features can be extracted from data such as voltage signals, environmental irradiance and power values, and these features can reflect the correlation and mutual influence between different electric energies. For example, frequency domain features can be extracted to describe the frequency relationship between electric energies, and time domain features can be extracted to describe the time series relationship between electric energies, thereby revealing the interaction and synergistic power generation capabilities between different electric energies in the photovoltaic power generation system. Further, feature extraction is performed on the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points, and features representing the state of the energy storage capacity can be extracted from the energy storage capacity data. These characteristics may include the changing trend of energy storage capacity, the degree of fluctuation of energy storage capacity, the utilization rate of energy storage capacity, etc. By analyzing these characteristics, the energy storage status of each power generation equipment can be evaluated, the remaining degree and availability of energy storage capacity can be understood, and a basis can be provided for the adjustment and optimization of energy storage capacity.
[0027] Figure 3 FIG. 1 is a block diagram of an electric energy data feature extraction unit in an artificial intelligence-based distributed photovoltaic power generation control system according to an embodiment of the present application. Figure 3As shown, in a specific embodiment of the present application, the electric energy data feature extraction unit 121 includes: an electric energy data preprocessing subunit 1211, which is used to perform data preprocessing on the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency and the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor to obtain multiple voltage frequency domain information, ambient irradiance input vectors and power input vectors, wherein the multiple voltage frequency domain information includes the voltage frequency, voltage amplitude and voltage phase of each sampling point; a voltage signal feature encoding subunit 1212, which is used to feature encode the multiple voltage frequency domain information to obtain a voltage frequency domain correlation feature vector; an electric energy data feature encoding subunit 1213, which is used to feature encode the ambient irradiance input vector and the power input vector to obtain an ambient irradiance-power correlation feature vector; an electric energy correlation subunit 1214, which is used to associate the voltage frequency domain correlation feature vector with the ambient irradiance-power correlation feature vector to obtain the distributed electric energy correlation feature vector. It should be understood that the analysis of the voltage signal in the frequency domain can provide more detailed and comprehensive voltage characteristics. By performing Fourier transform or other frequency domain analysis methods on the voltage signal, the voltage signal in the time domain can be converted into a representation in the frequency domain to obtain information such as voltage frequency, voltage amplitude and voltage phase. These frequency domain information can reveal the periodicity, harmonic components and phase relationship of the voltage signal, which is helpful for analyzing the stability and quality of the voltage waveform. Ambient irradiance is one of the main factors affecting the output power of a photovoltaic power generation system. By collecting the ambient irradiance values at multiple predetermined time points and forming them into an input vector, information about the time series change of ambient irradiance can be provided. This helps to analyze the trend, periodicity and potential impact of ambient irradiance. The power value is the output of the photovoltaic power generation system, and its change is closely related to the ambient irradiance and the system state. By collecting the power values at multiple predetermined time points and forming them into an input vector, information about the time series change of power can be provided. This helps to analyze the volatility, trend and response characteristics of power. In this way, the voltage characteristics, environmental impact and power changes of the photovoltaic power generation system can be analyzed and described from different angles and dimensions.
[0028] Furthermore, considering that the voltage frequency domain information usually contains a large number of data points, directly using the original data for analysis and modeling may have problems of high dimensionality and data redundancy. Therefore, feature encoding is used to convert the complex voltage frequency domain information into a more compact feature vector, reduce the dimensionality and redundancy of the data, and improve the data processing efficiency and the interpretability of the model. Furthermore, by feature encoding the ambient irradiance input vector and the power input vector, the correlation features between them can be extracted. These correlation features can include the time lag relationship between ambient irradiance and power, the degree of influence of ambient irradiance on power, etc. By analyzing these features, we can deeply understand the relationship between ambient irradiance and power, and provide a basis for the evaluation and optimization of the power generation capacity of photovoltaic power generation systems.
[0029] In particular, the voltage frequency domain correlation eigenvector reflects the frequency domain characteristics and correlation of the voltage signal in the photovoltaic power generation system, while the ambient irradiance-power correlation eigenvector reflects the correlation between ambient irradiance and system power. By correlating these two eigenvectors, the relationship between voltage and ambient irradiance-power can be revealed, and the operating status and performance of the photovoltaic power generation system can be further understood to provide more comprehensive information, including the frequency domain characteristics of the voltage signal, changes in environmental conditions, and the power generation capacity of the system. Specifically, the voltage frequency domain correlation eigenvector and the ambient irradiance-power correlation eigenvector can be spliced or combined at corresponding positions to obtain a new eigenvector, namely the distributed power correlation eigenvector.
[0030] In a specific embodiment of the present application, the electric energy data preprocessing subunit includes: performing a discrete fast Fourier transform on the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency to extract the multiple voltage frequency domain information; arranging the environmental irradiance and power values at multiple predetermined time points in the predetermined time period collected by the sensor according to the time dimension to obtain the environmental irradiance input vector and the power input vector. It should be understood that the voltage signal collected by the sensor is filtered out of high-frequency disturbances through a Kalman filter, and then the voltage signal collected by the sensor is processed based on a discrete fast Fourier transform to extract the frequency domain information of the voltage signal, including voltage frequency, voltage amplitude, and voltage phase. The voltage signal changes in the time domain, and the signal can be converted to the frequency domain through the Fourier transform. Discrete fast Fourier transform is an efficient algorithm that can discretize continuous time signals and extract the frequency components of the signal in the frequency domain. By performing a discrete fast Fourier transform on the voltage signal, the voltage frequency domain information can be obtained, revealing the intensity and phase information of each frequency component contained in the voltage signal. In addition, the voltage signal in the photovoltaic power generation system usually contains multiple frequency components, including the fundamental frequency (usually the grid frequency) and its harmonic components. By extracting the frequency domain information of the voltage signal, the amplitude and phase of each frequency component can be analyzed to understand the contribution and characteristics of different frequency components in the system.
[0031] Furthermore, by arranging the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor according to the time dimension, an ambient irradiance input vector and a power input vector can be obtained. In this way, the time series information can be retained. It should be understood that the performance and power generation capacity of the photovoltaic power generation system usually change with time, so arranging the ambient irradiance and power values according to time can capture this time series relationship.
[0032] In a specific embodiment of the present application, the voltage signal feature encoding subunit includes: passing the multiple voltage frequency domain information through a voltage frequency domain information semantic encoder including an embedding layer to obtain multiple voltage frequency domain information feature vectors; arranging the multiple voltage frequency domain information feature vectors in two dimensions to obtain a voltage frequency domain information feature matrix; passing the voltage frequency domain information feature matrix through a voltage frequency domain feature extractor based on a convolutional neural network to obtain the voltage frequency domain associated feature vector. It should be understood that the voltage frequency domain information contains the characteristics of the voltage signal in the frequency domain, including frequency, amplitude, phase, etc. By using the voltage frequency domain information semantic encoder of the embedding layer, the voltage frequency domain information can be converted into a feature vector with semantic meaning. Such a feature vector is more expressive and helps to improve the performance and generalization ability of the model. Specifically, in the encoding process of the voltage frequency domain information semantic encoder including the embedding layer, the multiple voltage frequency domain information is first mapped to the embedding vector using the embedding layer to obtain a sequence of embedded vectors, wherein the embedding layer can be constructed using a knowledge graph so that the prior information of the voltage feature is introduced in the process of converting the frequency, amplitude and phase of each sampling point into the embedded vector. Next, the voltage frequency domain information semantic encoder including the embedding layer adopts the converter-based Bert model to encode the sequence of embedded vectors, wherein the converter-based Bert model performs global semantic encoding on the sequence of embedded vectors based on the global mask structure through the intrinsic mask structure of the converter to obtain the multiple voltage frequency domain information feature vectors.
[0033] Furthermore, considering that there is a dynamic correlation between the voltage spectrum features of each sampling point, multiple voltage frequency domain information feature vectors are arranged in two dimensions to obtain a voltage frequency domain information feature matrix. In this way, the information between different feature vectors can be integrated to form a more comprehensive data representation. Specifically, the rows of the matrix can represent different time points or sampling points, and the columns of the matrix can represent different features. Furthermore, after the two-dimensional arrangement, the obtained voltage frequency domain information feature matrix is passed through a voltage frequency domain feature extractor based on a convolutional neural network to extract high-dimensional local correlation features in the feature matrix. It should be understood that the feature extractor based on the convolutional neural network has a strong generalization ability and can adapt to different voltage frequency domain information data. By using a convolutional neural network to extract features, the reliance on manual feature engineering can be reduced, making the model more versatile and adaptable, thereby better capturing the correlation and important features in the voltage frequency domain information. Specifically, each layer of the voltage frequency domain feature extractor based on the convolutional neural network is used to perform convolution processing, pooling processing along the channel dimension, activation processing and dimension reduction processing on the energy storage capacity input matrix in the forward transfer of the layer to generate the voltage frequency domain associated feature vector by the last layer of the voltage frequency domain feature extractor based on the convolutional neural network, wherein the input of the first layer of the voltage frequency domain feature extractor based on the convolutional neural network is the voltage frequency domain information feature matrix.
[0034] In a specific embodiment of the present application, the electric energy data feature encoding subunit includes: calculating the association matrix of the ambient irradiance input vector and the power input vector to obtain the ambient irradiance-power association input matrix; passing the ambient irradiance-power association input matrix through an ambient irradiance-power association feature filter based on a convolutional neural network to obtain the ambient irradiance-power association feature vector. It should be understood that calculating the association matrix of the ambient irradiance input vector and the power input vector can reveal the association between ambient irradiance and power, as well as their impact on the performance of the photovoltaic power generation system. Specifically, the association matrix can be regarded as the feature extraction result between ambient irradiance and power, and then reflects the linear relationship between the two vectors, thereby integrating and converting the information in the original input vector into higher-level features. More specifically, the product between the transposed vector of the ambient irradiance input vector and the power input vector is calculated to obtain the ambient irradiance-power association input matrix.
[0035] Furthermore, by passing the ambient irradiance-power correlation input matrix through an ambient irradiance-power correlation feature filter based on a convolutional neural network, an ambient irradiance-power correlation feature vector can be obtained. The purpose of this is to extract important features of the ambient irradiance-power correlation from the correlation matrix in order to better understand and utilize this correlation. Specifically, the convolutional neural network can gradually abstract and extract higher-level features through a combination of convolutional layers and pooling layers, thereby better capturing the correlation between ambient irradiance and power. More specifically, the ambient irradiance-power correlation feature filter based on a convolutional neural network is used to process the ambient irradiance-power correlation input matrix with the following convolution formula to obtain the ambient irradiance-power correlation feature vector;
[0036] Wherein, the convolution formula is:
[0037] f i =GP{Sigmoid(N i ×f i-1 +B i )}
[0038] Among them, f i-1 is the input of the i-th layer convolutional neural network model, f i is the output of the i-th layer convolutional neural network model, N i is the filter of the i-th layer convolutional neural network model, and B i is the bias matrix of the i-th layer convolutional neural network model, Sigmoid represents the nonlinear activation function, and GP represents the pooling operation.
[0039] In a specific embodiment of the present application, the energy storage capacity data feature extraction unit includes: arranging the energy storage capacity of the distributed photovoltaic power generation equipment at the plurality of predetermined time points according to the time dimension and the sample dimension to obtain an energy storage capacity input matrix; passing the energy storage capacity input matrix through an energy storage capacity timing encoder to obtain the energy storage capacity associated feature vector. It should be understood that in order to integrate the relationship between time and sample into a matrix so as to better analyze and process the energy storage capacity data, in the technical solution of the present application, the energy storage capacity of the distributed photovoltaic power generation equipment at the plurality of predetermined time points is arranged according to the time dimension and the sample dimension to obtain the energy storage capacity input matrix, so as to better organize and visualize the data.
[0040] Furthermore, the energy storage capacity is time series data that changes over time. Through the energy storage capacity time series encoder, the energy storage capacity data can be modeled in time series to capture the evolution and change trend of the energy storage capacity at different time points. It should be understood that the time series encoder can learn the long-term dependence and short-term fluctuations of the energy storage capacity, so as to better understand the dynamic characteristics of the energy storage capacity. Moreover, the energy storage capacity time series encoder can automatically learn and extract important features from the energy storage capacity data. Through the encoding process of the encoder, the energy storage capacity input matrix will be converted into an energy storage capacity associated feature vector, which contains an abstract representation of the energy storage capacity data. Specifically, each layer of the energy storage capacity time series encoder is used to perform the following operations on the energy storage capacity input matrix in the forward transfer of the layer: convolution processing is performed on the energy storage capacity input matrix to obtain a convolution feature map; mean pooling is performed on the convolution feature map based on the local channel dimension to obtain a pooling feature map; and nonlinear activation is performed on the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the energy storage capacity time series encoder is the energy storage capacity associated feature vector, and the input of the first layer of the energy storage capacity time series encoder is the energy storage capacity input matrix.
[0041] In the above-mentioned artificial intelligence-based distributed photovoltaic power generation control system 100, the photovoltaic energy storage capacity adjustment result generation module 130 is used to determine whether the distributed photovoltaic device energy storage capacity needs to be adjusted based on the distributed electric energy associated feature vector and the energy storage capacity associated feature vector.
[0042] In a specific embodiment of the present application, the photovoltaic energy storage capacity adjustment result generation module includes: a fusion unit, which is used to fuse the distributed electric energy associated feature vector and the energy storage capacity associated feature vector to obtain a distributed electric energy classification feature vector; an optimization unit, which is used to perform target dimension joint correlation migration on the distributed electric energy classification feature vector to obtain an optimized distributed electric energy classification feature vector; a result judgment unit, which is used to pass the optimized distributed electric energy classification feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted. It should be understood that the performance and status of the distributed photovoltaic system not only depends on the distributed electric energy, but also on the energy storage capacity. By fusing the distributed electric energy associated feature vector and the energy storage capacity associated feature vector, these two aspects of information can be comprehensively considered. Doing so can more comprehensively describe the electric energy classification of the distributed photovoltaic system and provide a more accurate feature vector representation.
[0043] In particular, in the technical solution of the present application, in the technical solution for power state analysis and classification of distributed photovoltaic power generation equipment, there may be outlier response values in the feature dimension of the distributed power classification feature vector, which may affect the classification effect. Outlier response values may be caused by sensor errors, equipment failures, and environmental changes. Sensors may have errors when collecting data such as voltage signals, environmental irradiance, and power values, and these outliers may be mistakenly recorded as outlier response values. In addition, the distributed photovoltaic power generation equipment itself may have faults or abnormal conditions, such as circuit component damage, connection problems, or system configuration errors, and these outliers may also be mistakenly recorded as outlier response values. In addition, changes in the working environment, such as weather conditions, shadows or pollution, may cause abnormal changes in parameters such as voltage, irradiance, and power, and then generate outlier response values. These outlier response values may have multiple effects on the classification effect. First, outlier response values may interfere with the normal operation of the feature extractor, which is usually designed and trained based on the statistical characteristics of normal data. When outlier response values exist, the extractor may not accurately capture the feature patterns of normal samples, thereby affecting the classification effect. Secondly, outlier response values may introduce imbalance and affect the weight distribution in the feature fusion process. If outlier response values occupy a large proportion in the feature fusion process, the distribution of the final feature vector may deviate from the distribution of normal samples, thereby affecting the decision boundary of the classifier. In addition, outlier response values may also interfere with the training and prediction process of the classifier. The classifier will be affected by the sample distribution during learning and prediction. When outlier response values exist, the classifier may overfit the outliers or fail to correctly distinguish them from normal samples, resulting in a decrease in classification effect. Based on this, the distributed power classification feature vector is subjected to target dimension joint correlation migration to obtain an optimized distributed power classification feature vector to eliminate or reduce the outlier response values in the feature vector, thereby improving the representation ability and classification performance of the feature vector.
[0044] Specifically, the distributed power classification feature vector is subjected to target dimension joint correlation migration using the following optimization formula to obtain an optimized distributed power classification feature vector;
[0045] Wherein, the optimization formula is:
[0046]
[0047] Among them, v i represents the eigenvalue of the ith position of the distributed electric energy classification feature vector, represents the mean value of all eigenvalues of the distributed power classification feature vector, L represents the length of the distributed power classification feature vector, softmax represents the classification function, log represents the logarithmic function value with base 2, and vi ′ represents the eigenvalue of the i-th position of the optimized distributed electric energy classification feature vector.
[0048] That is, considering that there are outlier response values in the feature dimension of the distributed power classification feature vector, the classification effect is affected. In order to solve this problem, this paper proposes a method based on joint correlation migration of target dimension, which can perform joint correlation migration of target dimension on the distributed power classification feature vector, that is, through a method based on target dimension, the feature value of each position can obtain a joint correlation migration of target dimension according to its correlation or dependence with the classification target dimension. Specifically, the method obtains a general probability distribution containing a special distribution by interpreting the joint correlation migration of target dimension as the negative logarithm of the univariate difference, that is, through a method based on probability theory, the feature value of each position can obtain a probability distribution according to its difference with the classification target dimension, thereby ensuring the concealment of outlier instances (instances) in the feature value samples in the distributed power classification feature vector for the overall disturbance, that is, through a method based on concealment, the feature value of each position can obtain a concealment factor according to its disturbance with the whole, in this way, the overall certainty of the feature expression of the distributed power classification feature vector is improved. In this way, by realizing the joint correlation migration of the target dimension of the feature vector, the outlier response values in the feature vector can be effectively eliminated or reduced, thereby improving the representation ability and classification performance of the feature vector.
[0049] Furthermore, the classifier can make classification judgments based on the similarity of known categories and feature vectors, and based on the classification results, it can be judged whether the energy storage capacity of each distributed photovoltaic device needs to be adjusted. For example, if the classification results show that the power situation of a certain device is abnormal or overloaded, then the energy storage capacity of the device may need to be increased to cope with the high load situation. On the contrary, if the classification results show that the power situation of a certain device is normal, then its current energy storage capacity can be maintained.
[0050] In summary, the embodiment of the present application first obtains the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the three. Finally, a classification result is obtained through a classifier to determine whether the energy storage capacity of the distributed photovoltaic equipment needs to be adjusted, thereby realizing real-time monitoring and adjustment of the energy storage capacity, thereby improving energy utilization efficiency, saving costs, improving system stability, and optimizing energy planning.
[0051] As described above, the distributed photovoltaic power generation control system 100 based on artificial intelligence according to the embodiment of the present application can be implemented in various terminal devices, such as a server deployed with a distributed photovoltaic power generation control algorithm based on artificial intelligence. In one example, the distributed photovoltaic power generation control system 100 based on artificial intelligence can be integrated into the terminal device as a software module and / or a hardware module. For example, the distributed photovoltaic power generation control system 100 based on artificial intelligence can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the distributed photovoltaic power generation control system 100 based on artificial intelligence can also be one of the many hardware modules of the terminal device.
[0052] Alternatively, in another example, the artificial intelligence-based distributed photovoltaic power generation control system 100 and the terminal device may also be separate devices, and the artificial intelligence-based distributed photovoltaic power generation control system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.
[0053] Exemplary Methods
[0054] Figure 4 FIG. 1 is a flow chart of a distributed photovoltaic power generation control method based on artificial intelligence according to an embodiment of the present application. Figure 4 As shown, according to the artificial intelligence-based distributed photovoltaic power generation control method of the embodiment of the present application, it includes: S110, obtaining the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at multiple predetermined time points; S120, extracting the distributed electric energy associated feature vector and the energy storage capacity associated feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points; S130, judging whether the energy storage capacity of the distributed photovoltaic equipment needs to be adjusted based on the distributed electric energy associated feature vector and the energy storage capacity associated feature vector.
[0055] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned distributed photovoltaic power generation control method based on artificial intelligence have been referred to above. Figures 1 to 3 The description of the artificial intelligence-based distributed photovoltaic power generation control system has been introduced in detail, and therefore, its repeated description will be omitted.
[0056] Exemplary Electronic Devices
[0057] In addition, an embodiment of the present invention further provides an electronic device, including:
[0058] A processor and a data storage system according to any one of the above embodiments of the present invention;
[0059] When the processor runs the data storage system, the units in the data storage system of any of the above embodiments of the present invention are run.
[0060] In addition, an embodiment of the present invention further provides another electronic device, comprising: one or more processors, a memory, a communication component and a communication bus, wherein the processor, the memory and the communication component communicate with each other via the communication bus;
[0061] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute an operation corresponding to the data processing method of any of the above embodiments of the present invention.
[0062] Figure 5 is a block diagram of an electronic device according to an embodiment of the present application. Figure 5 , which shows a schematic diagram of the structure of an electronic device suitable for implementing a terminal device or a server in an embodiment of the present application. Figure 5 As shown, the electronic device includes one or more processors, a communication unit, etc. The one or more processors are, for example: one or more central processing units (CPUs), and / or one or more graphics processing units (GPUs), etc. The processor can perform various appropriate actions and processes according to executable instructions stored in a read-only memory (ROM) or executable instructions loaded from a storage part to a random access memory (RAM). The communication unit may include but is not limited to a network card, and the network card may include but is not limited to an IB (Infiniband) network card. The processor may communicate with a read-only memory and / or a random access memory to execute executable instructions, and is connected to the communication unit through a bus and communicates with other target devices through the communication unit, thereby completing the operation corresponding to any method provided in the embodiments of the present application, for example, the primary shard manager shards the features in the feature queue; for any shard, the primary shard manager stores the any shard in the corresponding shard feature library, and records the mapping relationship between the any shard and the corresponding shard feature library in a mapping table; in response to a failure of the primary shard manager, a shard manager is reselected from at least one slave shard manager as the new primary shard manager; the new primary shard manager obtains shard information of the shard currently being written from all shard feature libraries, and recreates the mapping relationship of the shard currently being written based on the shard information of the shard currently being written in all shard feature libraries.
[0063] In addition, various programs and data required for the operation of the device can also be stored in the RAM. The CPU, ROM and RAM are connected to each other via a bus. In the case of RAM, the ROM is an optional module. The RAM stores executable instructions, or writes executable instructions into the ROM at runtime, and the executable instructions enable the processor to perform operations corresponding to any of the above methods of the present invention. The input / output (I / O) interface is also connected to the bus. The communication unit can be integrated or can be configured to have multiple submodules (e.g., multiple IB network cards) and be linked on the bus.
[0064] The following components are connected to the I / O interface: an input section including a keyboard, a mouse, etc.; an output section including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section including a hard disk, etc.; and a communication section including a network interface card such as a LAN card, a modem, etc. The communication section performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that a computer program read therefrom is installed into the storage section as needed.
[0065] If necessary, Figure 5 The architecture shown is only an optional implementation method. In the specific practice, the above can be modified according to actual needs. Figure 5 The number and type of components can be selected, deleted, added or replaced; different functional components can be set up by separate settings or integrated settings, for example, the GPU and CPU can be set up separately or the GPU can be integrated on the CPU, the communication unit can be set up separately or integrated on the CPU or GPU, etc. These replaceable implementations all fall within the protection scope disclosed in the present invention.
[0066] In addition, an embodiment of the present invention further provides a computer program, including computer-readable code. When the computer-readable code runs on a device, a processor in the device executes instructions for implementing each step in the data processing method of any of the above embodiments of the present invention.
[0067] In addition, an embodiment of the present invention further provides a computer-readable storage medium for storing computer-readable instructions, which, when executed, implement the operations of each step in the data processing method of any of the above embodiments of the present invention.
[0068] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0069] The method and apparatus of the present invention may be implemented in many ways. For example, the method and apparatus of the present invention may be implemented by software, hardware, firmware, or any combination of software, hardware, firmware. The above order of steps for the method is only for illustration, and the steps of the method of the present invention are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present invention may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Thus, the present invention also covers a recording medium storing a program for executing the method according to the present invention.
[0070] The description of the present invention is given for the purpose of illustration and description, and is not intended to be exhaustive or to limit the invention to the disclosed form. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for particular uses.
Claims
1. A distributed photovoltaic power generation control system based on artificial intelligence, characterized in that: include: A distributed photovoltaic data acquisition module, used to acquire voltage signals collected by sensors from distributed photovoltaic power generation equipment at a preset sampling frequency, environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by sensors, and energy storage capacity of distributed photovoltaic power generation equipment at multiple predetermined time points; A distributed photovoltaic data extraction module is used to extract a distributed electric energy-related feature vector and an energy storage capacity-related feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the environmental irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points; A photovoltaic energy storage capacity adjustment result generation module is used to determine whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted based on the distributed electric energy associated feature vector and the energy storage capacity associated feature vector; Wherein, the photovoltaic energy storage capacity adjustment result generation module includes: A fusion unit, used for fusing the distributed electric energy associated feature vector and the energy storage capacity associated feature vector to obtain a distributed electric energy classification feature vector; An optimization unit, configured to perform target dimension joint correlation migration on the distributed power classification feature vector to obtain an optimized distributed power classification feature vector; A result judgment unit, used for passing the optimized distributed electric energy classification feature vector through a classifier to obtain a classification result, wherein the classification result is used to judge whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted; Wherein, the optimization unit comprises: Performing target dimension joint correlation migration on the distributed power classification feature vector using the following optimization formula to obtain an optimized distributed power classification feature vector; Wherein, the optimization formula is: in, The first The eigenvalues at the positions, represents the mean value of all eigenvalues of the distributed electric energy classification feature vector, represents the length of the distributed power classification feature vector, represents the classification function, represents the logarithmic function value with base 2, The first character vector representing the optimized distributed power classification feature vector The feature value at each position.
2. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 1 is characterized in that: The distributed photovoltaic data extraction module comprises: An electric energy data feature extraction unit, configured to extract features from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency and the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor to obtain the distributed electric energy associated feature vector; The energy storage capacity data feature extraction unit is used to extract features from the energy storage capacity of the distributed photovoltaic power generation equipment at the plurality of predetermined time points to obtain the energy storage capacity associated feature vector.
3. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 2 is characterized in that: The electric energy data feature extraction unit comprises: An electric energy data preprocessing subunit is used to perform data preprocessing on the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency and the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor to obtain multiple voltage frequency domain information, ambient irradiance input vectors and power input vectors, wherein the multiple voltage frequency domain information includes the voltage frequency, voltage amplitude and voltage phase of each sampling point; A voltage signal feature encoding subunit, used for performing feature encoding on the plurality of voltage frequency domain information to obtain a voltage frequency domain associated feature vector; An electric energy data feature encoding subunit, used for feature encoding the ambient irradiance input vector and the power input vector to obtain an ambient irradiance-power correlation feature vector; The electric energy correlation subunit is used to correlate the voltage frequency domain correlation feature vector and the ambient irradiance-power correlation feature vector to obtain the distributed electric energy correlation feature vector.
4. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 3 is characterized in that: The electric energy data preprocessing subunit comprises: Performing a discrete fast Fourier transform on the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency to extract the multiple voltage frequency domain information; The ambient irradiance and power values at a plurality of predetermined time points within a predetermined time period collected by the sensor are arranged according to a time dimension to obtain the ambient irradiance input vector and the power input vector.
5. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 4 is characterized in that: The voltage signal characteristic encoding subunit comprises: Passing the plurality of voltage frequency domain information through a voltage frequency domain information semantic encoder including an embedding layer to obtain a plurality of voltage frequency domain information feature vectors; Arranging the plurality of voltage frequency domain information feature vectors in two dimensions to obtain a voltage frequency domain information feature matrix; The voltage frequency domain information feature matrix is passed through a voltage frequency domain feature extractor based on a convolutional neural network to obtain the voltage frequency domain associated feature vector.
6. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 5 is characterized in that: The electric energy data feature encoding subunit includes: Calculating the correlation matrix of the ambient irradiance input vector and the power input vector to obtain an ambient irradiance-power correlation input matrix; The ambient irradiance-power correlation input matrix is passed through an ambient irradiance-power correlation feature filter based on a convolutional neural network to obtain the ambient irradiance-power correlation feature vector.
7. The distributed photovoltaic power generation control system based on artificial intelligence according to claim 6 is characterized in that: The energy storage capacity data feature extraction unit comprises: Arranging the energy storage capacities of the distributed photovoltaic power generation equipment at the plurality of predetermined time points according to the time dimension and the sample dimension to obtain an energy storage capacity input matrix; The energy storage capacity input matrix is passed through an energy storage capacity time series encoder to obtain the energy storage capacity associated feature vector.
8. A distributed photovoltaic power generation control method based on artificial intelligence, characterized in that: include: Acquire voltage signals collected by sensors from distributed photovoltaic power generation equipment at a preset sampling frequency, ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by sensors, and energy storage capacity of distributed photovoltaic power generation equipment at multiple predetermined time points; Extracting a distributed electric energy associated feature vector and an energy storage capacity associated feature vector from the voltage signal collected by the sensor from the distributed photovoltaic power generation equipment at a preset sampling frequency, the ambient irradiance and power values at multiple predetermined time points within a predetermined time period collected by the sensor, and the energy storage capacity of the distributed photovoltaic power generation equipment at the multiple predetermined time points; Based on the distributed electric energy associated characteristic vector and the energy storage capacity associated characteristic vector, determining whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted; Wherein, judging whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted based on the distributed electric energy associated characteristic vector and the energy storage capacity associated characteristic vector includes: The distributed electric energy associated feature vector and the energy storage capacity associated feature vector are merged to obtain a distributed electric energy classification feature vector; Performing target dimension joint correlation migration on the distributed electric energy classification feature vector to obtain an optimized distributed electric energy classification feature vector; Passing the optimized distributed electric energy classification feature vector through a classifier to obtain a classification result, wherein the classification result is used to determine whether the energy storage capacity of the distributed photovoltaic device needs to be adjusted; The distributed power classification feature vector is subjected to target dimension joint correlation migration to obtain an optimized distributed power classification feature vector, including: Performing target dimension joint correlation migration on the distributed power classification feature vector using the following optimization formula to obtain an optimized distributed power classification feature vector; Wherein, the optimization formula is: in, The first The eigenvalues at the positions, represents the mean value of all eigenvalues of the distributed electric energy classification feature vector, represents the length of the distributed power classification feature vector, represents the classification function, represents the logarithmic function value with base 2, The first character vector representing the optimized distributed power classification feature vector The feature value at each position.
