Photovoltaic data transmission encryption method based on cloud computing

Abstract

The invention discloses a photovoltaic data transmission encryption method based on cloud computing. The method comprises the following steps: acquiring photovoltaic data through a sensor and preprocessing the photovoltaic data; feature extraction is carried out on the preprocessed photovoltaic data, a random forest model is used for judging the sensitivity degree of different photovoltaic data, and the target is to predict the encryption requirements of the different photovoltaic data at the current moment; the method comprises the steps of encrypting power generation data by using an AES advanced encryption standard, encrypting environmental data by using a TLS transport layer security protocol, encrypting equipment state data by using RSA asymmetric encryption, and dynamically adjusting corresponding encryption levels based on encryption requirements of different photovoltaic data at the current moment; decrypting different types of photovoltaic data on the cloud platform, and storing the photovoltaic data in corresponding databases; generating a photovoltaic data correlation topological graph according to correlation strength among different statistical characteristic values in the photovoltaic data; based on the photovoltaic data correlation topological graph, analyzing the power generation data and the environmental data by using a time sequence prediction model, and predicting a future power generation amount change trend; a photovoltaic system operation optimization strategy is automatically generated according to a future power generation amount change trend and equipment state data; the method provides a scientific basis for optimizing the operation strategy of the photovoltaic system.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud computing, and in particular to a photovoltaic data transmission encryption method based on cloud computing. Background Art

[0002] Photovoltaic systems convert solar energy into electricity and are widely used in homes, businesses, and industries. To ensure the efficient operation of these systems, real-time monitoring and data transmission are crucial. However, these systems face numerous challenges in data transmission, particularly regarding data security and privacy.

[0003] Many existing PV data transmission solutions lack adequate encryption, making them vulnerable to cyberattacks and the leakage of sensitive information. For example, unencrypted data can be easily stolen or tampered with during transmission. Existing technologies often use fixed encryption methods and fail to dynamically adjust encryption levels based on data sensitivity and real-time needs, resulting in over- or under-encryption. Current data analysis methods often rely on static models and fail to fully leverage real-time data changes, resulting in inaccurate and inefficient predictions. Summary of the Invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a photovoltaic data transmission encryption method based on cloud computing to solve the problem that current data analysis methods often rely on static models and fail to fully utilize changes in real-time data, resulting in low prediction accuracy and efficiency.

[0005] Technical solution: The photovoltaic data transmission encryption method based on cloud computing described in the present invention includes the following steps:

[0006] S1. Acquire photovoltaic data through sensors and preprocess it. Feature extraction is performed on the preprocessed photovoltaic data. A random forest model is used to determine the sensitivity of different photovoltaic data. The goal is to predict the encryption requirements of different photovoltaic data at the current moment. Photovoltaic data includes power generation data, environmental data, and equipment status data.

[0007] S2. Use the AES advanced encryption standard to encrypt power generation data, the TLS transport layer security protocol to encrypt environmental data, and RSA asymmetric encryption to encrypt device status data. Dynamically adjust the encryption level based on the encryption requirements of different photovoltaic data at the current moment. After encrypting the photovoltaic data, send it to the cloud platform via the HTTPS protocol.

[0008] S3. Decrypt different categories of photovoltaic data on the cloud platform and store them in the corresponding database. The cloud platform groups the power generation data, environmental data, and equipment status data obtained from the photovoltaic system by time period. In each time period, the statistical characteristic values ​​of each type of data are sorted out and standardized and normalized. The Pearson correlation coefficient is used to quantify the linear relationship between different statistical characteristic values ​​in the photovoltaic data after standardization and normalization. Based on the correlation coefficient value, a correlation threshold is set. A value above the threshold indicates a strong correlation, and a value below the threshold indicates a weak correlation.

[0009] S4. Generate a photovoltaic data correlation topology map based on the correlation strength between different statistical characteristic values ​​in the photovoltaic data; based on the photovoltaic data correlation topology map, use a time series prediction model to analyze the power generation data and environmental data to predict future power generation trends; and automatically generate a photovoltaic system operation optimization strategy based on future power generation trends and equipment status data.

[0010] Furthermore, power generation data includes real-time power data, cumulative power generation data and power generation efficiency data; environmental data includes light intensity data, temperature data and wind speed data; equipment status data includes photovoltaic module status data, inverter status data and combiner box status data; photovoltaic data preprocessing includes missing value processing, outlier detection, unit consistency processing, timestamp formatting and data noise reduction.

[0011] Furthermore, feature extraction is performed on the preprocessed photovoltaic data to form feature data, and encryption requirement labels are set according to the sensitivity of the feature data and historical encryption requirements, where encryption requirement labels include high encryption requirement, medium encryption requirement and low encryption requirement; the feature data and encryption requirement labels are integrated to form a photovoltaic data set for training and testing; the photovoltaic data set is divided into training set and test set in proportion; the training set data is used to construct a random forest model, and the model will learn how to predict the encryption requirement label based on the feature data, and each tree of the random forest is trained using a random subset of features; the trained model is evaluated using the test set to evaluate the accuracy and recall of the model in predicting encryption requirements, thereby ensuring the effectiveness of the model on different photovoltaic data; the feature importance analysis of the random forest model is used to identify the features that have the greatest impact on encryption requirement prediction; based on the feature importance score, it is determined which data is more sensitive to encryption requirements in different scenarios.

[0012] Furthermore, power generation data is obtained from photovoltaic data, and a random number generator is used to generate a 128-bit, 192-bit or 256-bit symmetric key for AES encryption; the power generation data is encrypted using the selected key and AES algorithm, the encrypted data is output as ciphertext, and the generated symmetric key is stored using the key management system KMS; environmental data is obtained from photovoltaic data, and before sending the environmental data, a TLS connection is established, a handshake is performed through the TLS protocol, the identity of the server is verified, and a session key is generated; wherein the session key is used to encrypt data transmission in the current connection; the environmental data is encrypted through the TLS protocol using the established session key; during the transmission process, the environmental data will be segmented and encryption and integrity verification will be added; the device status data is obtained from the photovoltaic data, and a pair of RSA keys is generated, including a public key and a private key, wherein the public key is used to encrypt data and the private key is used for decryption; the device status data is encrypted using the recipient's public key to generate ciphertext; RSA is an asymmetric encryption, and the recipient holding the private key can decrypt the device status data.

[0013] Furthermore, the encryption requirements of different photovoltaic data at the current moment are obtained; for power generation data with high encryption requirements, use 256-bit key for AES encryption and select Galois / CounterMode mode; for power generation data with medium encryption requirements, use 192-bit key and use CipherBlock Chaining mode; for power generation data with low encryption requirements, select the electronic codebook mode and perform security restrictions; for environmental data with high encryption requirements, use TLS1.3 and the key exchange algorithm ECDHE for encryption; for environmental data with medium encryption requirements, use TLS1.2 and the DHE algorithm for encryption; for environmental data with low encryption requirements, use TLS1.2 and RSA key exchange for encryption; for device status data with high encryption requirements, use 2048-bit or higher RSA keys and ensure that each transmitted data is the latest; for device status data with medium encryption requirements, use 1536-bit RSA keys; for device status data with low encryption requirements, use 1024-bit RSA keys; after encrypting the photovoltaic data, send it to the cloud platform via the HTTPS protocol.

[0014] Furthermore, the cloud platform receives encrypted photovoltaic data through HTTPS. For power generation data, the AES algorithm is used to decrypt it, using a pre-shared key; for environmental data, the TLS protocol automatically handles decryption; for device status data, the RSA algorithm is used to decrypt it, using a private key for operation; the decrypted power generation data, environmental data, and device status data are stored in the cloud platform's database respectively, using a table partitioning strategy for storage; the time granularity is set, and the decrypted data is grouped according to the set time period to form a time series data set; statistical characteristic value calculations are performed to obtain statistical characteristics of power generation data, environmental data, and device status data; when calculating statistical characteristics of power generation data, for real-time power data, the average value, maximum value, minimum value, and standard deviation are calculated; for cumulative power generation data, the average value, maximum value, minimum value, and standard deviation are calculated. Calculate the total amount and change rate; for power generation efficiency data, calculate the average daily efficiency and peak efficiency; when calculating the statistical characteristics of environmental data, for light intensity, calculate the average light and peak light; for temperature data, calculate the average daily temperature and extreme temperature; for wind speed data, calculate the average wind speed and wind speed change rate; when calculating the statistical characteristics of equipment status data, for PV module status, calculate the failure rate and normal working time; for inverter status, calculate the frequency of abnormal occurrence and operating time; for junction box status, calculate the frequency of state change; perform Z-score normalization on each type of statistical eigenvalue, convert each statistical eigenvalue to a form with a mean of 0 and a standard deviation of 1; perform Min-Max normalization, convert each statistical eigenvalue to be in the range of [0,1] to eliminate the dimensional influence between different statistical eigenvalues.

[0015] Furthermore, the standardized and normalized statistical feature values ​​are organized into a data frame, where each column represents a statistical feature and each row represents data for a time period. The Pearson correlation coefficient is used to measure the linear relationship between two statistical feature values, with a value range of [-1, 1]. Its formula is:

[0016]

[0017] Among them, r xy is the correlation coefficient between the first statistical eigenvalue x and the second statistical eigenvalue y, n is the number of samples, x i and y i are the i-th sample values ​​of the first statistical eigenvalue x and the second statistical eigenvalue y, respectively. and are the means of the first statistical eigenvalue x and the second statistical eigenvalue y respectively;

[0018] Set the correlation threshold T, if |r xy |>T, then there is a strong correlation between the two statistical eigenvalues. If |r xy|≤T, then there is a weak correlation between the two statistical eigenvalues.

[0019] Furthermore, Python's NetworkX is used to draw a topological graph. Each statistical eigenvalue is used as a node in the graph, and edges are added for each pair of strongly correlated features according to the correlation matrix. xy |>T, then add an edge between the two nodes, and the weight of the edge is set to the absolute value of the correlation coefficient |r xy |; The thickness and color of the edges are adjusted according to the edge weights, and the node positions are positioned using force-directed layout. Based on the photovoltaic data correlation topology, environmental features related to power generation are selected as input features. An input feature set containing power generation data and environmental data for the previous m moments is constructed for time series model training, where m is a positive integer.

[0020] Furthermore, LSTM is selected as the time series prediction model. The selected model is trained using historical data, hyperparameters are adjusted to optimize model performance, and cross-validation is used to evaluate the model effect. The trained time series model is used to predict future power generation, and the power generation trend at the next m moments is obtained. The accuracy of the prediction is evaluated by comparing it with the actual data. Based on the future power generation trend and equipment status data, the photovoltaic system operation optimization strategy is automatically generated.

[0021] The photovoltaic data transmission encryption system based on cloud computing described in the present invention includes:

[0022] Photovoltaic data acquisition and preprocessing module: includes a data acquisition unit, a data preprocessing unit, and a sensitivity analysis unit. The data acquisition unit acquires photovoltaic data through sensors, including power generation data, environmental data, and equipment status data. The data preprocessing unit preprocesses the photovoltaic data. The sensitivity analysis unit extracts features from the preprocessed photovoltaic data and uses a random forest model to determine the sensitivity of different photovoltaic data. The goal is to predict the encryption requirements of different photovoltaic data at the current moment.

[0023] Data encryption and transmission module: includes AES encryption unit, TLS encryption unit, RSA encryption unit, encryption policy configuration unit and data transmission unit; among them, the AES encryption unit uses the AES advanced encryption standard to encrypt power generation data, the TLS encryption unit uses the TLS transport layer security protocol to encrypt environmental data, and the RSA encryption unit uses RSA asymmetric encryption to encrypt device status data. The encryption policy configuration unit dynamically adjusts the corresponding encryption level based on the encryption requirements of different photovoltaic data at the current moment; after encrypting the photovoltaic data, the data transmission unit sends it to the cloud platform via the HTTPS protocol;

[0024] Data decryption and storage module: includes a data decryption unit, a time series construction unit, and a feature statistics unit. The data decryption unit decrypts different categories of photovoltaic data on the cloud platform and stores them in the corresponding database. In the time series construction unit, the cloud platform groups the power generation data, environmental data, and equipment status data obtained from the photovoltaic system by time period. In each time period, the statistical feature values ​​of each type of data are sorted out and standardized and normalized. The feature statistics unit uses the Pearson correlation coefficient to quantify the linear relationship between different statistical feature values ​​in the photovoltaic data after standardization and normalization. Based on the correlation coefficient value, a correlation threshold is set. A value above the threshold indicates a strong correlation, and a value below the threshold indicates a weak correlation.

[0025] Data analysis and optimization module: includes a correlation topology map generation unit, a time series prediction unit and an optimization strategy generation unit; among them, the correlation topology map generation unit generates a photovoltaic data correlation topology map based on the correlation strength between different statistical characteristic values ​​in the photovoltaic data; the time series prediction unit uses a time series prediction model to analyze the power generation data and environmental data based on the photovoltaic data correlation topology map to predict future power generation trends; the optimization strategy generation unit automatically generates a photovoltaic system operation optimization strategy based on future power generation trends and equipment status data.

[0026] Beneficial effects: Compared with the existing technology, the present invention has the following significant advantages: The present invention adopts a combination of multiple encryption technologies to provide a hierarchical encryption strategy based on the sensitivity of the data type, thereby effectively protecting the security of photovoltaic data during transmission. The sensitivity of the data is judged by the random forest model, which realizes the real-time prediction of encryption requirements. The encryption level can be dynamically adjusted according to the sensitivity of the photovoltaic data at the current moment, thereby improving the encryption efficiency and flexibility. The use of time series prediction models to analyze photovoltaic data can accurately predict the future trend of power generation changes, thereby providing a scientific basis for optimizing the operation strategy of the photovoltaic system and improving the operational efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of the method of the present invention;

[0028] Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0029] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0030] like Figure 1 As shown, an embodiment of the present invention provides a photovoltaic data transmission encryption method based on cloud computing, comprising the following steps:

[0031] S100: Acquire photovoltaic data through sensors, the photovoltaic data including power generation data, environmental data, and equipment status data, and perform photovoltaic data preprocessing; extract features from the preprocessed photovoltaic data, and use a random forest model to determine the sensitivity of different photovoltaic data, with the goal of predicting the encryption requirements of different photovoltaic data at the current moment;

[0032] S200 uses the AES advanced encryption standard to encrypt power generation data, the TLS transport layer security protocol to encrypt environmental data, and RSA asymmetric encryption to encrypt device status data. The encryption level is dynamically adjusted based on the encryption requirements of different photovoltaic data at the current moment. After encrypting the photovoltaic data, it is sent to the cloud platform via the HTTPS protocol.

[0033] S300. Decrypting different categories of photovoltaic data on the cloud platform and storing them in corresponding databases. The cloud platform groups power generation data, environmental data, and equipment status data obtained from the photovoltaic system by time period. In each time period, statistical characteristic values ​​of each type of data are sorted out and standardized and normalized. The Pearson correlation coefficient is used to quantify the linear relationship between different statistical characteristic values ​​in the photovoltaic data after standardization and normalization. A correlation threshold is set based on the correlation coefficient value. A value above the threshold indicates a strong correlation, and a value below the threshold indicates a weak correlation.

[0034] S400. Generate a photovoltaic data correlation topology map based on the strength of correlation between different statistical characteristic values ​​in the photovoltaic data; based on the photovoltaic data correlation topology map, use a time series prediction model to analyze the power generation data and environmental data to predict future power generation trends; and automatically generate a photovoltaic system operation optimization strategy based on the future power generation trends and equipment status data.

[0035] According to step S100, the power generation data includes real-time power data, cumulative power generation data, and power generation efficiency data; the environmental data includes light intensity data, temperature data, and wind speed data; and the device status data includes photovoltaic module status data, inverter status data, and combiner box status data;

[0036] The photovoltaic data preprocessing includes missing value processing, outlier detection, unit consistency processing, timestamp formatting and data noise reduction.

[0037] According to step S100, feature extraction is performed on the preprocessed photovoltaic data to form feature data. Encryption requirement labels are set based on the sensitivity of the feature data and historical encryption requirements, wherein the encryption requirement labels include high encryption requirement, medium encryption requirement, and low encryption requirement. The feature data and the encryption requirement labels are integrated to form a photovoltaic dataset for training and testing. The photovoltaic dataset is divided into a training set and a test set in proportion, with 70% used for training and 30% used for testing.

[0038] Using the training set data, a random forest model was constructed. The model will learn how to predict the encryption demand label based on the feature data. Each tree of the random forest is trained using a random subset of features, and the number of trees is set to 100. The trained model is evaluated using the test set to assess the accuracy and recall of the model in predicting encryption demand, ensuring the effectiveness of the model on different photovoltaic data, with an accuracy of 95%, a recall of 92%, and an F1-score of 93.5%. Feature importance analysis using the random forest model identifies the features that have the greatest impact on encryption demand prediction. Based on the feature importance scores, it is determined which data is more sensitive to encryption demand in different scenarios.

[0039] Through feature importance analysis of the random forest model, the features that have the greatest impact on encryption demand prediction are identified. The feature importance scores are as follows: the importance score of light intensity is 0.30, the importance score of real-time power is 0.25, the importance score of temperature is 0.20, the importance score of cumulative power generation is 0.15, the importance score of component status is 0.05, and the importance score of inverter status is 0.05. Based on the importance scores of the features, the sensitivity to encryption demand in different scenarios is judged: Highly sensitive features: Light intensity: has the greatest impact on encryption demand and needs to be given priority. Real-time power: follows closely behind and has a greater impact. Medium-sensitivity features: Temperature and cumulative power generation: affect encryption demand within a certain range. Low-sensitivity features: Component status and inverter status: have little impact on encryption demand.

[0040] According to step S200, power generation data is obtained from the photovoltaic data, and a 128-bit, 192-bit, or 256-bit symmetric key is generated using a random number generator for AES encryption; the power generation data is encrypted using the selected key and the AES algorithm, the encrypted data is output as ciphertext, and the generated symmetric key is stored using a key management system (KMS);

[0041] Environmental data is obtained from photovoltaic data. Before sending the environmental data, a TLS connection is established. A handshake is performed through the TLS protocol to verify the server's identity and generate a session key. The session key is used to encrypt data transmission in the current connection. The environmental data is encrypted through the TLS protocol using the established session key. During transmission, the environmental data will be segmented and encrypted and integrity checked.

[0042] The device status data is obtained from the photovoltaic data and a pair of RSA keys is generated, including a public key and a private key. The public key is used to encrypt the data and the private key is used to decrypt it. The device status data is encrypted using the recipient's public key to generate ciphertext. RSA is an asymmetric encryption method, and the recipient holding the private key can decrypt the device status data.

[0043] According to step S200, the encryption requirements of different photovoltaic data at the current moment are obtained;

[0044] For power generation data with high encryption requirements, use a 256-bit key for AES encryption and select the Galois / Counter Mode mode. For power generation data with medium encryption requirements, use a 192-bit key and the Cipher Block Chaining mode. For power generation data with low encryption requirements, select the Electronic Codebook mode and implement security restrictions.

[0045] For data in environments with high encryption requirements, TLS 1.3 and the ECDHE key exchange algorithm are used for encryption. For data in environments with medium encryption requirements, TLS 1.2 and the DHE algorithm are used for encryption. For data in environments with low encryption requirements, TLS 1.2 and RSA key exchange are used for encryption.

[0046] For device status data with high encryption requirements, use a 2048-bit or higher RSA key to ensure that each transmitted data is the latest. For device status data with medium encryption requirements, use a 1536-bit RSA key. For device status data with low encryption requirements, use a 1024-bit RSA key.

[0047] After encrypting the photovoltaic data, it is sent to the cloud platform via the HTTPS protocol.

[0048] In this example, at a certain moment, we collected the following photovoltaic data: Power generation data: real-time power data: 150kW, cumulative power generation data: 3000kWh, power generation efficiency data: 85%. Environmental data: light intensity: 800W / m 2, Temperature: 25°C, Wind Speed: 5m / s. Device Status Data: PV Module Status: Normal, Inverter Status: Normal, Combiner Box Status: Normal.

[0049] Based on historical data and feature sensitivity analysis, the current encryption requirements are judged as follows: power generation data: high encryption requirements, environmental data: medium encryption requirements, equipment status data: high encryption requirements.

[0050] Implement corresponding encryption according to encryption requirements: Power generation data (high encryption requirements):

[0051] The AES algorithm uses a 256-bit encryption key in Galois / Counter Mode (GCM). The following example shows encrypted data: Original data: [150, 3000, 85], Encrypted data: [0xA1B2C3...0xD4E5F6].

[0052] Environmental data (medium encryption requirements): Encrypted using TLS 1.2 and the DHE algorithm.

[0053] The encrypted data example is as follows: original data: [800,25,5], encrypted data: [0xC1D2E3...0xF4G5H6].

[0054] Device status data (high encryption requirements): Encrypted using a 2048-bit RSA key to ensure that each transmitted data is up to date.

[0055] The encrypted data example is as follows: Original data: [Normal, Normal, Normal], Encrypted data: [0xE1F2G3...0xH4I5J6].

[0056] All encrypted data is sent to the cloud platform via the HTTPS protocol: encrypted power generation data: [0xA1B2C3...0xD4E5F6], encrypted environmental data: [0xC1D2E3...0xF4G5H6], encrypted device status data: [0xE1F2G3...0xH4I5J6].

[0057] According to step S300, the cloud platform receives the encrypted photovoltaic data via HTTPS. The power generation data is decrypted using the AES algorithm, using a pre-shared key. The environmental data is automatically decrypted using the TLS protocol. The device status data is decrypted using the RSA algorithm, using a private key. The decrypted power generation data, environmental data, and device status data are stored separately in the cloud platform database, using a separate table strategy.

[0058] Set the time granularity and group the decrypted data according to the set time period to form a time series data set; perform statistical characteristic value calculations to obtain the statistical characteristics of power generation data, environmental data, and equipment status data;

[0059] When calculating the statistical characteristics of power generation data, for real-time power data, the average, maximum, minimum, and standard deviation are calculated; for cumulative power generation data, the total amount and change rate are calculated; for power generation efficiency data, the daily average efficiency and peak efficiency are calculated; when calculating the statistical characteristics of environmental data, for light intensity, the average light intensity and peak light intensity are calculated; for temperature data, the daily average temperature and extreme temperature are calculated; for wind speed data, the average wind speed and wind speed change rate are calculated; when calculating the statistical characteristics of equipment status data, for PV module status, the failure rate and normal operating time are calculated; for inverter status, the frequency of abnormal occurrences and operating time are calculated; for combiner box status, the frequency of state changes is calculated;

[0060] Z-score normalization is performed on each type of statistical eigenvalue to convert each statistical eigenvalue to a form with a mean of 0 and a standard deviation of 1; Min-Max normalization is performed to convert each statistical eigenvalue to be within the range of [0,1] to eliminate the dimensional influence between different statistical eigenvalues.

[0061] According to step S300, the standardized and normalized statistical feature values ​​are organized into a data frame, where each column represents a statistical feature and each row represents data for a time period. The Pearson correlation coefficient is used to measure the linear relationship between two statistical feature values, and its value range is [-1, 1]. Its formula is:

[0062]

[0063] Among them, r xy is the correlation coefficient between the first statistical eigenvalue x and the second statistical eigenvalue y, n is the number of samples, x i and y i are the i-th sample values ​​of the first statistical eigenvalue x and the second statistical eigenvalue y, respectively. and are the means of the first statistical eigenvalue x and the second statistical eigenvalue y respectively;

[0064] Set a correlation threshold T, if |r xy |>T, then there is a strong correlation between the two statistical eigenvalues, if |r xy |≤T, then there is a weak correlation between the two statistical eigenvalues.

[0065] The statistical eigenvalues ​​after standardization and normalization are as follows:

[0066] Time period T1: Real-time power (kW): 0.75, cumulative power generation (kWh): 0.60, power generation efficiency (%): 0.85, light intensity (W / m 2 ): 0.90, temperature (℃): 0.70, wind speed (m / s): 0.80.

[0067] Time period T2: Real-time power (kW): 0.80, cumulative power generation (kWh): 0.70, power generation efficiency (%): 0.88, light intensity (W / m 2 ):0.92, temperature(℃):0.75, wind speed(m / s):0.85.

[0068] Time period T3: Real-time power (kW): 0.70, cumulative power generation (kWh): 0.55, power generation efficiency (%): 0.82, light intensity (W / m 2 ): 0.85, temperature (℃): 0.68, wind speed (m / s): 0.78.

[0069] Time period T4: Real-time power (kW): 0.90, cumulative power generation (kWh): 0.80, power generation efficiency (%): 0.90, light intensity (W / m 2 ): 0.95, temperature (℃): 0.80, wind speed (m / s): 0.90.

[0070] The Pearson correlation coefficient of real-time power and light intensity is r≈0.85, and the correlation threshold T is set to 0.7, so there is a strong correlation between "real-time power" and "light intensity".

[0071] The calculation results of other features are as follows: real-time power and cumulative power generation: 0.78, strong correlation; real-time power and power generation efficiency: 0.65, weak correlation; cumulative power generation and power generation efficiency: 0.82, strong correlation; light intensity and temperature: 0.50, weak correlation.

[0072] According to step S400, Python's NetworkX is used to draw a topological graph, each statistical eigenvalue is used as a node in the graph, and edges are added for each pair of strongly correlated features according to the correlation matrix; if |r xy |>T, then add an edge between the two nodes, and the weight of the edge is set to the absolute value of the correlation coefficient |r xy |; The thickness and color of the edge are adjusted according to the edge weight, and the position of the node uses force-directed layout.

[0073] According to step S400, based on the photovoltaic data correlation topology, environmental features related to power generation are selected as input features, and an input feature set including power generation data and environmental data for the previous m moments is constructed for training the time series model, where m is a positive integer;

[0074] We selected LSTM as the time series prediction model, trained the model using historical data, adjusted hyperparameters to optimize model performance, and used cross-validation to evaluate the model's effectiveness. We used the trained time series model to predict future power generation, obtaining power generation trends for the next m moments. The accuracy of the predictions was evaluated by comparing them with actual data.

[0075] Automatically generate photovoltaic system operation optimization strategies based on future power generation trends and equipment status data.

[0076] like Figure 2 As shown, the present invention also provides a photovoltaic data transmission encryption system based on cloud computing, comprising:

[0077] Photovoltaic data acquisition and preprocessing module: includes: a data acquisition unit, a data preprocessing unit, and a sensitivity analysis unit. The data acquisition unit acquires photovoltaic data through sensors, including power generation data, environmental data, and equipment status data. The data preprocessing unit preprocesses the photovoltaic data. The sensitivity analysis unit extracts features from the preprocessed photovoltaic data and uses a random forest model to determine the sensitivity of different photovoltaic data. The goal is to predict the encryption requirements of different photovoltaic data at the current moment.

[0078] Data encryption and transmission module: includes: AES encryption unit, TLS encryption unit, RSA encryption unit, encryption policy configuration unit and data transmission unit; among them, the AES encryption unit uses the AES advanced encryption standard to encrypt power generation data, the TLS encryption unit uses the TLS transport layer security protocol to encrypt environmental data, and the RSA encryption unit uses RSA asymmetric encryption to encrypt device status data. The encryption policy configuration unit dynamically adjusts the corresponding encryption level based on the encryption requirements of different photovoltaic data at the current moment; after encrypting the photovoltaic data, the data transmission unit sends it to the cloud platform via the HTTPS protocol;

[0079] Data decryption and storage module: includes: data decryption unit, time series construction unit and feature statistics unit; the data decryption unit decrypts different categories of photovoltaic data on the cloud platform and stores them in the corresponding database; in the time series construction unit, the cloud platform groups the power generation data, environmental data and equipment status data obtained from the photovoltaic system by time period, sorts out the statistical feature values ​​of each type of data in each time period, and performs standardization and normalization processing; the feature statistics unit uses the Pearson correlation coefficient to quantify the linear relationship between different statistical feature values ​​in the photovoltaic data after standardization and normalization. According to the correlation coefficient value, a correlation threshold is set. A value above the threshold indicates strong correlation, and a value below the threshold indicates weak correlation.

[0080] Data analysis and optimization module: includes: correlation topology map generation unit, time series prediction unit and optimization strategy generation unit; among them, the correlation topology map generation unit generates a photovoltaic data correlation topology map based on the correlation strength between different statistical characteristic values ​​in the photovoltaic data; the time series prediction unit uses the time series prediction model to analyze the power generation data and environmental data based on the photovoltaic data correlation topology map to predict future power generation trends; the optimization strategy generation unit automatically generates photovoltaic system operation optimization strategies based on future power generation trends and equipment status data.

[0081] In this embodiment, the correlation topology map generating unit of the data analysis and optimization module inputs the statistical characteristic values ​​of the photovoltaic data, calculates the correlation between different characteristics using the Pearson correlation coefficient, and generates a correlation topology map to show the relationship between each characteristic.

[0082] The time series forecasting unit uses power generation and environmental data selected from the correlation topology map. It trains the model using records from T1 to T4, sets hyperparameters, and evaluates model performance using cross-validation. The model predicts power generation trends for the next five moments (T5 to T9) as follows: T5: 0.85, T6: 0.88, T7: 0.80, T8: 0.90, and T9: 0.92.

[0083] The optimization strategy generation unit inputs future power generation trends (forecast results from T5 to T9) and current equipment status data. If the predicted power generation increases significantly, it considers optimizing equipment loads to increase energy output. If the predicted power generation is stable or decreasing, it recommends checking the equipment status and performing maintenance.

[0084] The current equipment status is normal, and the following optimization strategies are generated: Strategy 1: During T5 to T7, optimize energy output and increase power output to 0.90kW. Strategy 2: During T8 and T9, perform equipment status checks to ensure operational efficiency.

Claims

1. A photovoltaic data transmission encryption method based on cloud computing, characterized in that: The following steps are involved: S1. Acquire photovoltaic data through sensors and preprocess it. Feature extraction is performed on the preprocessed photovoltaic data. A random forest model is used to determine the sensitivity of different photovoltaic data. The goal is to predict the encryption requirements of different photovoltaic data at the current moment. Photovoltaic data includes power generation data, environmental data, and equipment status data. S2. Use the AES advanced encryption standard to encrypt power generation data, the TLS transport layer security protocol to encrypt environmental data, and RSA asymmetric encryption to encrypt device status data. Dynamically adjust the encryption level based on the encryption requirements of different photovoltaic data at the current moment. After encrypting the photovoltaic data, send it to the cloud platform via the HTTPS protocol. S3. Decrypt different categories of photovoltaic data on the cloud platform and store them in the corresponding database. The cloud platform groups the power generation data, environmental data, and equipment status data obtained from the photovoltaic system by time period. In each time period, the statistical characteristic values ​​of each type of data are sorted out and standardized and normalized. The Pearson correlation coefficient is used to quantify the linear relationship between different statistical characteristic values ​​in the photovoltaic data after standardization and normalization. Based on the correlation coefficient value, a correlation threshold is set. A value above the threshold indicates a strong correlation, and a value below the threshold indicates a weak correlation. S4. generating a photovoltaic data correlation topology map based on the correlation strength between different statistical characteristic values ​​in the photovoltaic data; Based on the photovoltaic data correlation topology map, a time series prediction model is used to analyze power generation data and environmental data to predict future power generation trends. Based on future power generation trends and equipment status data, a photovoltaic system operation optimization strategy is automatically generated.

2. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: Power generation data includes real-time power data, cumulative power generation data and power generation efficiency data; environmental data includes light intensity data, temperature data and wind speed data; equipment status data includes photovoltaic module status data, inverter status data and combiner box status data; photovoltaic data preprocessing includes missing value processing, outlier detection, unit consistency processing, timestamp formatting and data noise reduction.

3. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: Feature extraction is performed on the preprocessed photovoltaic data to form feature data. Encryption requirement labels are set based on the sensitivity of the feature data and historical encryption requirements, where encryption requirement labels include high encryption requirement, medium encryption requirement, and low encryption requirement. The feature data and encryption requirement labels are integrated to form a photovoltaic dataset for training and testing. The photovoltaic dataset is divided into training set and test set in proportion. A random forest model is constructed using the training set data. The model will learn how to predict the encryption requirement label based on the feature data. Each tree of the random forest is trained using a random subset of features. The trained model is evaluated using the test set to evaluate the accuracy and recall of the model in predicting encryption requirements, ensuring the effectiveness of the model on different photovoltaic data. Feature importance analysis of the random forest model is used to identify the features that have the greatest impact on encryption requirement prediction. Based on the feature importance scores, it is determined which data are more sensitive to encryption requirements in different scenarios.

4. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: The power generation data is obtained from the photovoltaic data, and a 128-bit, 192-bit, or 256-bit symmetric key is generated using a random number generator for AES encryption. The power generation data is encrypted using the selected key and AES algorithm, and the encrypted data is output as ciphertext. The generated symmetric key is stored using the key management system (KMS). The environmental data is obtained from the photovoltaic data. Before sending the environmental data, a TLS connection is established, a handshake is performed using the TLS protocol, the server's identity is verified, and a session key is generated. The session key is used to encrypt data transmission in the current connection. The environmental data is encrypted using the established session key using the TLS protocol. During transmission, the environmental data will be segmented and encryption and integrity checks will be added. The device status data is obtained from the photovoltaic data, and a pair of RSA keys, including a public key and a private key, are generated. The public key is used to encrypt data, and the private key is used for decryption. The device status data is encrypted using the recipient's public key to generate ciphertext. RSA is an asymmetric encryption method, and the recipient holding the private key can decrypt the device status data.

5. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: Obtain the encryption requirements of different photovoltaic data at the current moment; for power generation data with high encryption requirements, use a 256-bit key for AES encryption and select the Galois / CounterMode mode; for power generation data with medium encryption requirements, use a 192-bit key and use CipherBlock Chaining mode; for power generation data with low encryption requirements, select the electronic codebook mode and perform security restrictions; for environmental data with high encryption requirements, use TLS1.3 and the key exchange algorithm ECDHE for encryption; for environmental data with medium encryption requirements, use TLS1.2 and the DHE algorithm for encryption; for environmental data with low encryption requirements, use TLS1.2 and RSA key exchange for encryption; for device status data with high encryption requirements, use 2048-bit or higher RSA keys and ensure that each transmitted data is the latest; for device status data with medium encryption requirements, use 1536-bit RSA keys; for device status data with low encryption requirements, use 1024-bit RSA keys; after encrypting the photovoltaic data, send it to the cloud platform via the HTTPS protocol.

6. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: The cloud platform receives encrypted PV data via HTTPS. For power generation data, the AES algorithm is used to decrypt it, using a pre-shared key. For environmental data, the TLS protocol automatically handles decryption. For device status data, the RSA algorithm is used to decrypt it, using a private key. The decrypted power generation data, environmental data, and device status data are stored separately in the cloud platform's database, using a partitioned table strategy. A time granularity is set, and the decrypted data is grouped by the set time period to form a time series data set. Statistical eigenvalues ​​are calculated to obtain statistical characteristics of power generation data, environmental data, and device status data. When calculating the statistical characteristics of power generation data, for real-time power data, the average value, maximum value, minimum value, and standard deviation are calculated; for cumulative power generation data, the total amount and rate of change are calculated; for power generation efficiency data, the daily average efficiency and peak efficiency are calculated; When calculating the statistical characteristics of environmental data, for light intensity, the average light intensity and peak light intensity are calculated; for temperature data, the average daily temperature and extreme temperature are calculated; for wind speed data, the average wind speed and wind speed change rate are calculated; When calculating the statistical characteristics of device status data, for PV module status, the failure rate and normal operating time are calculated; for inverter status, the frequency of abnormal situations and operating time are calculated; for combiner box status, the frequency of state changes is calculated. Z-score normalization is performed on each type of statistical characteristic value, converting each statistical characteristic value into a form with a mean of 0 and a standard deviation of 1. Min-Max normalization is performed to convert each statistical eigenvalue into the range of [0, 1] to eliminate the dimensional influence between different statistical eigenvalues.

7. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: The standardized and normalized statistical feature values ​​are organized into a data frame, where each column represents a statistical feature and each row represents data for a time period. The Pearson correlation coefficient is used to measure the linear relationship between two statistical feature values, with a value range of [-1, 1]. Its formula is: Among them, r xy is the correlation coefficient between the first statistical eigenvalue x and the second statistical eigenvalue y, n is the number of samples, x i and y i are the i-th sample values ​​of the first statistical eigenvalue x and the second statistical eigenvalue y, respectively. and are the means of the first statistical eigenvalue x and the second statistical eigenvalue y respectively; Set the correlation threshold T, if |r xy |>T, then there is a strong correlation between the two statistical eigenvalues. If |r xy |≤T, then there is a weak correlation between the two statistical eigenvalues.

8. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: Use Python's NetworkX to draw the topology graph. Each statistical eigenvalue is a node in the graph. According to the correlation matrix, add edges for each pair of strongly correlated features. xy |>T, then add an edge between the two nodes, and the weight of the edge is set to the absolute value of the correlation coefficient |r xy |; The thickness and color of the edges are adjusted according to the edge weights, and the node positions are positioned using force-directed layout. Based on the photovoltaic data correlation topology, environmental features related to power generation are selected as input features. An input feature set containing power generation data and environmental data for the previous m moments is constructed for time series model training, where m is a positive integer.

9. The photovoltaic data transmission encryption method based on cloud computing according to claim 1, characterized in that: LSTM was selected as the time series prediction model. Historical data was used to train the selected model, hyperparameters were adjusted to optimize model performance, and cross-validation was used to evaluate the model effect. The trained time series model was used to predict future power generation, and the power generation trend at the next m moments was obtained. The accuracy of the prediction was evaluated by comparing it with the actual data. Based on the future power generation trend and equipment status data, an optimization strategy for the photovoltaic system operation was automatically generated.

10. A photovoltaic data transmission encryption system based on cloud computing, characterized in that: include: Photovoltaic data acquisition and preprocessing module: includes a data acquisition unit, a data preprocessing unit, and a sensitivity analysis unit. The data acquisition unit acquires photovoltaic data through sensors, including power generation data, environmental data, and equipment status data. The data preprocessing unit preprocesses the photovoltaic data. The sensitivity analysis unit extracts features from the preprocessed photovoltaic data and uses a random forest model to determine the sensitivity of different photovoltaic data. The goal is to predict the encryption requirements of different photovoltaic data at the current moment. Data encryption and transmission module: includes AES encryption unit, TLS encryption unit, RSA encryption unit, encryption policy configuration unit and data transmission unit; among them, the AES encryption unit uses the AES advanced encryption standard to encrypt power generation data, the TLS encryption unit uses the TLS transport layer security protocol to encrypt environmental data, and the RSA encryption unit uses RSA asymmetric encryption to encrypt device status data. The encryption policy configuration unit dynamically adjusts the corresponding encryption level based on the encryption requirements of different photovoltaic data at the current moment; after encrypting the photovoltaic data, the data transmission unit sends it to the cloud platform via the HTTPS protocol; Data decryption and storage module: includes a data decryption unit, a time series construction unit, and a feature statistics unit. The data decryption unit decrypts different categories of photovoltaic data on the cloud platform and stores them in the corresponding database. In the time series construction unit, the cloud platform groups the power generation data, environmental data, and equipment status data obtained from the photovoltaic system by time period. In each time period, the statistical feature values ​​of each type of data are sorted out and standardized and normalized. The feature statistics unit uses the Pearson correlation coefficient to quantify the linear relationship between different statistical feature values ​​in the photovoltaic data after standardization and normalization. Based on the correlation coefficient value, a correlation threshold is set. A value above the threshold indicates a strong correlation, and a value below the threshold indicates a weak correlation. Data analysis and optimization module: includes a correlation topology map generation unit, a time series prediction unit and an optimization strategy generation unit; among them, the correlation topology map generation unit generates a photovoltaic data correlation topology map based on the correlation strength between different statistical characteristic values ​​in the photovoltaic data; the time series prediction unit uses a time series prediction model to analyze the power generation data and environmental data based on the photovoltaic data correlation topology map to predict future power generation trends; the optimization strategy generation unit automatically generates a photovoltaic system operation optimization strategy based on future power generation trends and equipment status data.

Images