5G-based offshore wind power data joint compression and encryption method and early warning system

Abstract

The present application relates to a 5G-based offshore wind power data joint compression and encryption method and early warning system, the method specifically comprises: wind turbine source data feature extraction: feature extraction is carried out on the source data collected by the wind turbine side PLC; wind turbine data joint compression and encrypted transmission: the above-mentioned features are quantitatively compressed based on the threshold difference compression algorithm under the 5G network transmission mechanism, and after compression, hierarchical encryption is carried out based on the data sensitivity based on the national encryption and decryption mechanism; block chain distributed storage: the compressed and encrypted features are transmitted to the onshore switch station under the 5G network, and the onshore switch station stores the data features based on the block chain distributed storage technology; large model data feature early warning enhancement feedback: the onshore switch station carries out early warning enhancement on the feedback data features based on the RAG large model; the present application has the advantages of strong applicability, good effect, high accuracy, and improvement of the level of autonomous controllable operation.

Description

Technical Field

[0001] This invention belongs to the fields of offshore wind power fault early warning and 5G secure communication technology, specifically involving a 5G-based method for joint compression and encryption of offshore wind power data and an early warning system. Background Technology

[0002] During offshore wind power data transmission, the harsh marine environment can cause delays and interruptions in wind turbine communication, impacting the autonomous and controllable operation and maintenance of wind turbines. Furthermore, data security is not effectively guaranteed during transmission, making critical data susceptible to loss or theft. This fundamental lack of data security severely affects the safe and stable operation of wind turbines. Simultaneously, traditional application systems lack effective early warning mechanisms for wind turbine status assessment during operation, significantly impacting turbine lifespan and posing safety hazards. Therefore, it is essential to provide a 5G-based method and early warning system for joint compression and encryption of offshore wind power data, offering strong applicability, high effectiveness, high accuracy, and improved autonomous and controllable operation and maintenance. Summary of the Invention

[0003] (a) Technical issues

[0004] In view of the above-mentioned existing technology, this application mainly addresses the following technical problems:

[0005] 1. How to improve the communication capabilities of wind turbines and the security of data transmission during wind turbine data transmission;

[0006] 2. To provide an effective early warning mechanism for wind turbine status.

[0007] (II) Technical Solution

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide a 5G-based method and early warning system for joint compression and encryption of offshore wind power data, which is highly applicable, effective, accurate, and improves the level of independent and controllable operation and maintenance.

[0009] The objective of this invention is achieved as follows: Firstly, a method for joint compression and encryption of offshore wind power data based on 5G, the method comprising the following steps:

[0010] Step 1: Feature extraction of wind turbine source data: Feature extraction is performed on the source data collected by the PLC on the wind turbine side to obtain a feature dataset;

[0011] Step 2: Joint compression and encrypted transmission of wind turbine data: The above feature dataset is quantized and compressed using a threshold differential compression algorithm under the 5G network transmission mechanism, and layered encryption is performed based on the national cryptographic encryption and decryption mechanism according to the data sensitivity to obtain a compressed and encrypted feature dataset.

[0012] Step 3: Blockchain Distributed Storage: The compressed and encrypted feature dataset is transmitted to the land-based switch station with the support of the 5G network. The land-based switch station stores the compressed and encrypted feature dataset based on blockchain distributed storage technology.

[0013] Step 4: Large model enhances early warning feedback based on data features: The onshore switching station will enhance early warning feedback based on the large model of RAG.

[0014] Furthermore, the feature extraction in step 1 includes four main categories of features: SCADA features, CMS vibration features, environmental coupling features, and image features.

[0015] Furthermore, the SCADA features include gearbox temperature difference data features, bearing temperature rise gradient features, wind turbine wind energy utilization coefficient features, and wind turbine power curve deviation features; the image features include wind turbine blade crack ratio features and wind turbine edge average curvature features.

[0016] In this invention, the gearbox temperature difference data characteristic represents the relationship between the gearbox temperature and the fan condition; the bearing temperature rise gradient characteristic represents the relationship between the bearing temperature and the fan condition; the fan energy utilization coefficient characteristic represents the relationship between the fan energy utilization coefficient and the fan health status; and the fan power curve deviation characteristic represents the relationship between the fan power, the fan power generation, and the in-service status.

[0017] Among them, the characteristic of the proportion of cracks in the wind turbine blades is calculated from the number of pixels in the current wind turbine crack image and the total number of pixels in the wind turbine crack image; the characteristic of the average curvature of the wind turbine edge represents the characteristic of the wind turbine blade structure curvature describing the change of the average curvature of the wind turbine edge.

[0018] Furthermore, the CMS vibration characteristics include wind turbine vibration kurtosis characteristics, bearing fault amplitude ratio characteristics, wind turbine waveform factor characteristics, wind turbine gear sideband energy characteristics, and wavelet energy entropy characteristics; the environmental coupling characteristics include wave-tower transfer function characteristics, salt spray corrosion characteristics, and tower tilt angle RMS characteristics.

[0019] In this invention, the kurtosis feature of the wind turbine vibration represents the relationship between the wind turbine vibration factors and the wind turbine's operational stability; the bearing failure amplitude ratio feature, the wind turbine waveform factor feature, the wind turbine gear sideband energy feature, and the wavelet energy entropy feature are all used as sub-features of the wind turbine's CMS vibration; the wind turbine gear sideband energy feature represents the relationship between the energy at the edge of the wind turbine gear and the wind turbine's operating state.

[0020] Among them, the wave-tower transfer function characteristic, as a sub-characteristic of environmental coupling, is calculated from the stress response density and wave spectral density; the salt spray corrosion characteristic is calculated from the chloride ion concentration; and the tower tilt angle RMS characteristic is calculated from the tower tilt angle per unit time.

[0021] Furthermore, the joint compression and encryption method in step 2 specifically involves: after compressing the 14 types of sub-features, performing layered encryption on the data based on data sensitivity characteristics and the advantages of the national cryptographic algorithm. The specific steps are as follows:

[0022] Step 2.1: For the two types of sensitive features, SCADA features and image features, the SM2 algorithm is used for compression based on the national cryptographic algorithm mechanism;

[0023] Step 2.2: For the two types of high-volume time-series features, namely CMS vibration features and environmental coupling features, the SM4 algorithm is used for compression based on the national cryptographic algorithm mechanism;

[0024] Step 2.3: Perform binary conversion on the two types of features encrypted using the SM2 and SM4 algorithms. First, perform binary conversion on the two types of features encrypted by different mechanisms in order, and then combine them into a binary string in order. At this time, record the position sequence table of each feature in the binary string.

[0025] Step 2.4: Hash the currently calculated binary string using the SM3 hash algorithm;

[0026] Step 2.5: Concatenate the binary string with the hash value to form a new binary string, and encrypt and transmit it based on the SM2 and SM4 national cryptographic algorithms according to data sensitivity, and transmit the position sequence list of each feature in the binary string together;

[0027] Step 2.6: The received data at the booster station is decrypted using SM2 and SM4 mechanisms to obtain a binary string and a hash value;

[0028] Step 2.7: Read the binary string according to the position sequence table of the binary string, remove the hash value, and restore the binary string of the real data;

[0029] Step 2.8: Perform a decimal conversion on the actual binary string to obtain the differentially compressed data.

[0030] Furthermore, the blockchain distributed storage in step 3 specifically involves: vertical management of cold and hot data based on blockchain technology, wherein cold data is stored in SM4-XTS mode and hot data is stored in SM4-CTR mode, and the storage results are returned to the chain record in the form of CID.

[0031] Furthermore, in step 4, the large model performs data feature early warning enhancement feedback, specifically as follows: after obtaining the compressed decimal data features at the offshore booster station, the large model analysis results are fed back to the onshore switch station based on the established RAG technology large language model enhancement model; wherein the RAG technology large language model enhancement model includes: using RAG+LLM to realize the establishment of wind turbine fault feature database, compressed feature comparison and early warning enhancement technology; wherein the database consists of various types of structured, unstructured and semi-structured data.

[0032] Secondly, a 5G-based offshore wind power data joint compression and encryption early warning system includes a wind turbine equipment layer, a compression layer, an encryption / decryption layer, a large model calculation layer, and a storage layer. The 5G-based offshore wind power data joint compression and encryption early warning system is used to execute the 5G-based offshore wind power data joint compression and encryption method described above. The wind turbine equipment layer is connected to the compression layer, the compression layer is connected to both the wind turbine equipment layer and the encryption / decryption layer, the encryption / decryption layer is connected to both the compression layer and the large model calculation layer, the large model calculation layer is connected to both the encryption / decryption layer and the storage layer, and the storage layer is connected to the large model calculation layer.

[0033] Furthermore, the wind turbine equipment layer includes a wind turbine 5G communication module and a feature extraction classifier; the compression layer includes a threshold differential compression module; the encryption / decryption layer includes a national cryptographic encryption / decryption module; the large model calculation layer includes a large model early warning enhancement module based on RAG technology; and the storage layer includes a designed blockchain storage module.

[0034] The wind turbine 5G communication module is responsible for establishing a 5G network communication channel with the 5G base station of the offshore substation; the feature extraction classifier is used to extract features from the source data collected by the wind turbine-side PLC to obtain a feature dataset, including SCADA features, CMS vibration features, environmental coupling features and image features.

[0035] The threshold differential compression module is used to quantize and compress the above feature dataset based on the threshold differential compression algorithm under the 5G network transmission mechanism; at the same time, an SM3 hash value is introduced into the compressed binary string.

[0036] The national cryptographic encryption and decryption module is used to perform layered encryption on the compressed data based on the national cryptographic encryption and decryption mechanism according to data sensitivity, so as to obtain a compressed encrypted feature dataset.

[0037] The blockchain distributed storage module is used to transmit the compressed and encrypted feature dataset to the land-based switch station with the support of the 5G network. The land-based switch station stores the compressed and encrypted feature dataset based on blockchain distributed storage technology.

[0038] The large model early warning enhancement module based on RAG technology is used in onshore switching stations to enhance the early warning of feedback data features based on the large model of RAG.

[0039] Thirdly, a computer system includes: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the 5G-based offshore wind power data joint compression and encryption method as described above when running the computer program.

[0040] (III) Beneficial Effects

[0041] 1. This invention relies on 5G communication technology and large model technology, and uses a combination of compression algorithm and national cryptographic encryption and decryption algorithm to extract features, transmit features and provide fault warning for offshore wind power communication data. 5G technology is widely used in the industrial field and has excellent characteristics such as low transmission latency and stable communication security. It has a very good robust effect on the reliable transmission of offshore wind power.

[0042] 2. The large-scale model of this invention has unique advantages in logical thinking and computing power in intelligent decision-making, and is capable of deploying early warning systems for offshore wind power faults;

[0043] 3. The classifier designed in this invention classifies offshore wind power data features according to data sensitivity and data characteristics, including classifying data based on data characteristics and classifying and encrypting data based on national cryptographic algorithms based on data sensitivity. This can effectively classify, statistically analyze, and utilize offshore wind power data.

[0044] 4. The threshold differential compression algorithm designed in this invention can effectively compress the features of offshore wind power data. Compared with the current mainstream compression algorithms, it can improve compression efficiency while ensuring data quality, and effectively overcome the shortcomings of traditional compression algorithms in terms of compression quality and compression efficiency.

[0045] 5. This invention uses large-scale model technology and blockchain technology to classify and store data according to the frequency of occurrence or use, and to provide fault early warning. This can meet the requirements of enhanced fault early warning, efficient data storage and security of large-scale model reading, assist technicians in accurately judging the status of wind turbines, and further improve the available hours of offshore wind turbine power generation. Attached Figure Description

[0046] Figure 1 This is a system overall structure framework diagram of the present invention.

[0047] Figure 2 This is a system communication structure diagram of the present invention.

[0048] Figure 3 This is a flowchart of the early warning enhancement mechanism of the present invention.

[0049] Figure 4 This is a flowchart of the distributed storage process for data features in this invention. Detailed Implementation

[0050] The purpose of this invention is to overcome the problems of insufficient security of offshore wind power communication data, high cost of communication infrastructure, and insufficient accuracy of existing application systems in fault early warning. It proposes a highly applicable, effective, and accurate method and early warning system for joint compression and encryption of offshore wind power data based on 5G and large models. This method addresses the shortcomings of existing systems in terms of security and reliability, and optimizes and improves data security protection, reliable data communication, enhanced fault early warning, and efficient data storage, thereby further improving the safe and reliable operation and autonomous and controllable operation and maintenance level of offshore wind power.

[0051] The present invention will be further described below with reference to the embodiments and / or accompanying drawings.

[0052] Example 1

[0053] like Figure 1-4 As shown, a joint compression and encryption method for offshore wind power data based on 5G is described, the method comprising the following steps:

[0054] Step 1: Key Feature Extraction from Multi-Data Fusion of Offshore Wind Turbines: Feature extraction is performed on the source data collected by the PLC on the wind turbine side to obtain a feature dataset;

[0055] In this embodiment, offshore wind turbines generate different types and frequencies of data during operation. These data are closely related to the health status of the turbines. Data features are extracted from offshore wind turbines from four dimensions: SCADA data, CMS vibration data, environmental coupling data, and image data, as detailed below:

[0056] ① SCADA data characteristics

[0057] 1) During data transmission in the fan, the temperature of the fan gearbox is closely related to the fan's condition. Here, we extract the input and output temperature data from the PLC sensor of the fan gearbox, and the characteristics of the gearbox temperature difference data. for: In the formula: Indicates the gearbox input temperature; Indicates the gearbox output temperature; when The gearbox of the blower may experience oil circuit blockage. Here, 8℃ is used as the threshold limit for the temperature difference characteristic of the gearbox.

[0058] 2) During the operation of the fan, the bearing temperature is closely related to the fan's condition. Here, the temperature of the fan bearing is extracted from the PLC sensor. and with Calculate the bearing temperature rise gradient characteristics for the sampling interval time. Specifically: ,when At this time, the fan bearing may experience a failure of the indicated lubricating oil. Here, 1℃ / min is used as the threshold limit for the bearing temperature rise gradient characteristic.

[0059] 3) During wind turbine operation, the wind energy utilization coefficient is closely related to the wind turbine's health status. Here, we extract the air density collected by the wind turbine's PLC. Wind speed and pressure Calculate the characteristics of wind turbine wind energy utilization coefficient Specifically: In the formula, This indicates the radius of the wind turbine; here The threshold range is defined in .

[0060] 4) The power output of the wind turbine significantly affects its power generation and operational status. Here, the air density collected by the wind turbine's PLC is used. Sweeped area Wind speed Actual power and operating condition coefficient Calculate the deviation characteristics of the wind turbine power curve Specifically: In the formula, when If the deviation exceeds 10 minutes, it indicates that the wind turbine may have an abnormal power output. Here, a deviation of more than 15% for more than 10 minutes is used as the threshold limit for the power curve deviation characteristic.

[0061] ②CMS vibration characteristics

[0062] 1) The vibration factor of the fan is closely related to the operating stability of the fan. Here, we assume that the total amount of fan data is... The vibration kurtosis characteristics of the wind turbine It can be represented as: In the formula, Indicates the total amount of wind turbine data The mean; Indicates the first One vibration data point; This indicates the total number of data points for the fan vibration signal; Indicates the total amount of wind turbine data Standard deviation; when A value greater than 4 indicates that the bearing is at risk of impact damage. Here, 4 is taken as the threshold limit for the vibration kurtosis of the fan.

[0063] 2) Bearing failures occur frequently and with a high probability during fan vibration. This is based on the characteristic of bearing failure amplitude ratio. As a sub-characteristic of CMS vibration in wind turbines, it specifically includes: In the formula, Indicates the frequency of failures in the bearing's inner ring; Indicates the noise frequency. This indicates the total number of frequency data points for fan noise. This indicates the average amplitude of faults in the inner ring. Indicates the average amplitude of background noise; Indicates the rotational frequency of the shaft; when A value greater than 8 indicates a serious bearing damage accident. Here, 8 is used as the threshold limit for the amplitude ratio characteristic of wind turbine bearing failure.

[0064] 3) During long-term operation, the fan is prone to fatigue crack propagation. Here, the fan waveform factor is calculated. As a sub-feature of CMS vibration, specifically: In the formula, Indicates the valid values ​​of the waveform data; Represents waveform data; Indicates the total number of data points for the wind turbine waveform signal; when the waveform factor A value greater than 1.8 indicates that the wind turbine may experience fatigue crack propagation. Here, 1.8 is taken as the threshold limit for the waveform factor feature.

[0065] 4) The energy at the edge of the wind turbine gear is closely related to the wind turbine's operating state. Here, we calculate the energy characteristics of the wind turbine gear sideband. As a characteristic of the CMS vibrational oscillator, specifically: In the formula, Indicates the meshing frequency; Indicates the width of the sideband; Represents the edge energy function; when the gear sideband energy characteristic When the value is greater than three times the baseline, it indicates that the wind turbine gear may have a tooth breakage warning. Here, three times the baseline is used as the threshold limit for the energy characteristics of the gear sideband.

[0066] 5) Wavelet energy entropy can well characterize the vibration state of wind turbine CMS. Here, wavelet energy entropy is used as the vibration sub-feature of wind turbine CMS.

[0067] ③ Environmental coupling characteristics

[0068] 1) During operation, wind turbines are susceptible to resonance due to the influence of marine natural conditions. Stress response density is used here. and wave spectral density Calculate the wave-tower transfer function As a sub-feature of environmental coupling, it is specifically: In the formula, when A value greater than 0.2 indicates that the wind turbine tower may be in resonance. Here, 0.2 is used as the threshold limit for the wave-tower transfer function characteristic.

[0069] 2) Due to the corrosive effects of salt spray at sea, the foundation structure of wind turbines may face risks of fracture, tilting, or even collapse during prolonged operation at sea. Here, the chloride ion concentration is used as an example. Calculate the characteristics of salt spray corrosion Specifically: In the formula, when the salt spray corrosion characteristics When the value is greater than 0.2 mm / year, it indicates that the wind turbine needs salt spray corrosion protection. Here, 0.2 mm / year is used as the characteristic threshold limit for salt spray corrosion.

[0070] 3) During wind turbine operation, the tower is prone to tilting due to offshore winds, leading to varying degrees of foundation settlement. This calculation is performed per unit time. Calculation of tower tilt angle and RMS characteristic of tower tilt angle Specifically: In the formula, Indicates the tower's inclination angle; when When the wind turbine foundation settlement occurs, 0.3° is used as the threshold limit for the wind turbine tower tilt angle characteristic.

[0071] ④ Image features

[0072] 1) During long-term operation of the wind turbine, the wind turbine blades are prone to cracking due to factors such as wind resistance and operating frequency. Here, we use the current number of pixels in the wind turbine crack image. Total number of pixels in the image of the wind turbine crack Calculate the characteristics of crack proportion in wind turbine blades Specifically: In the formula, when When the wind turbine needs to be shut down for maintenance, 75% is used as the threshold limit for the proportion of wind turbine blade cracks.

[0073] 2) During wind turbine operation, the curvature of the wind turbine blade structure can effectively describe the change in the average curvature of the wind turbine edge. Here, we calculate the characteristic of the average curvature of the wind turbine edge. Specifically: In the formula, and These represent the first derivatives of the contour; and Let each represent the second derivative of the contour; when A value greater than 2% indicates that the wind turbine may have suffered structural damage. Here, 2% is used as the threshold limit for the edge curvature feature.

[0074] Step 2: Joint compression and encrypted transmission of wind turbine data: The above feature dataset is quantized and compressed using a threshold differential compression algorithm under the 5G network transmission mechanism, and layered encryption is performed based on the national cryptographic encryption and decryption mechanism according to the data sensitivity to obtain a compressed and encrypted feature dataset.

[0075] In this embodiment, ① threshold differential compression method

[0076] After extracting 14 sub-features, including wind turbine gearbox temperature difference, bearing temperature gradient, wind energy utilization coefficient, power curve deviation, vibration kurtosis, waveform factor, bearing fault amplitude ratio, gear sideband energy, wavelet packet energy entropy, wave-tower transfer function, salt spray corrosion, tower tilt angle, blade crack ratio, and average edge curvature, these 14 features are compressed. This compression is based on a threshold-based approach, classifying data between normal and abnormal data to improve transmission efficiency and save bandwidth costs while ensuring data quality. The specific compression rules are as follows: In the formula, This indicates that the current feature is within the threshold range; This indicates that the current feature is not within the threshold range; This indicates that the current feature belongs to a single threshold; This indicates that the current feature belongs to the double threshold category; This represents a single feature from 14 sub-features, including gearbox temperature difference characteristics, bearing temperature gradient characteristics, wind energy utilization coefficient characteristics, power curve deviation characteristics, vibration kurtosis characteristics, waveform factor characteristics, bearing fault amplitude ratio characteristics, gear sideband energy characteristics, wavelet packet energy entropy characteristics, wave-tower transfer function characteristics, salt spray corrosion characteristics, tower tilt angle characteristics, blade crack ratio characteristics, and edge average curvature characteristics. The threshold representing a single threshold feature; This represents the upper limit of the double-threshold feature; This represents the lower limit of the double threshold feature. Threshold differential compression of feature data based on the threshold limit can further improve data transmission efficiency and save data transmission bandwidth.

[0077] ② Combined compression and encryption methods

[0078] After compressing 14 sub-features, including gearbox temperature difference, bearing temperature gradient, wind energy utilization coefficient, power curve deviation, vibration kurtosis, waveform factor, bearing fault amplitude ratio, gear sideband energy, wavelet packet energy entropy, wave-tower transfer function, salt spray corrosion, tower tilt angle, blade crack ratio, and edge average curvature, data transmission security is further considered during data transmission. Based on data sensitivity characteristics and the advantages of national cryptographic algorithms, layered encryption is applied to the data. The specific steps are as follows:

[0079] 1) For the two types of features, SCADA features and image features, including six sub-features such as bearing temperature gradient features, wind energy utilization coefficient features, gear sideband energy features, wavelet packet energy entropy features, tower tilt angle features, and edge average curvature features, these features are sensitive features. Here, based on the national cryptographic algorithm mechanism, the SM2 algorithm is used for compression.

[0080] 2) For the two major features of CMS vibration characteristics and environmental coupling characteristics, including the eight minor features of wind turbine gearbox temperature difference characteristics, power curve deviation characteristics, vibration kurtosis characteristics, waveform factor characteristics, bearing fault amplitude ratio characteristics, wave-tower transfer function characteristics, salt spray corrosion characteristics, and blade crack ratio characteristics, which belong to high-flow-rate time-series features, the SM4 algorithm is used for compression based on the national cryptographic algorithm mechanism.

[0081] 3) Perform binary conversion on the two types of features encrypted using the SM2 and SM4 algorithms. First, convert the features encrypted using the two different mechanisms into binary sequences, and then combine them into a single binary string in that order. At this point, record the position sequence table of each feature within the binary string. ;

[0082] 4) Hash the currently calculated binary string using the SM3 hash algorithm, specifically: In the formula, This represents the value after hashing using the SM3 algorithm; Representation of the number of bits in a binary string; This represents 2B data representing the conversion of a binary string to an integer. and Represents two elements of the elliptic curve equation obtained using the Chinese cryptographic algorithm; and Represents the coordinates of the base point of a known elliptic curve; and The coordinates representing the public key;

[0083] 5) Combine binary string with The hash values ​​are concatenated into a new binary string and encrypted using the SM2 and SM4 national cryptographic algorithms based on data sensitivity. The purpose of concatenating the hash values ​​here is to improve the data's resistance to decryption, reduce the risk of data leakage, and establish a sequence table of the positions of each feature in the binary string. Transmitted together;

[0084] 6) The received data at the booster station is decrypted using SM2 and SM4 mechanisms to obtain the binary string and hash value. ;

[0085] 7) Based on the position sequence table of the binary string Read the binary string and remove the hash value. To restore the binary string of the actual data;

[0086] 8) Perform a decimal conversion on the real binary string to obtain the differentially compressed data.

[0087] Ultimately, the joint compression and encryption of wind turbine data features are completed, improving data transmission efficiency, reducing bandwidth costs, and ensuring data transmission security.

[0088] Step 3: Blockchain Distributed Storage: The compressed and encrypted feature dataset is transmitted to the land-based switch station with the support of the 5G network. The land-based switch station stores the compressed and encrypted feature dataset based on blockchain distributed storage technology.

[0089] Step 4: Large model enhances early warning feedback based on data features: The onshore switching station will enhance early warning feedback based on the large model of RAG.

[0090] In this embodiment, ① after obtaining the compressed decimal data features at the offshore booster station, the large-scale model analysis results are fed back to the onshore switchyard technicians based on the established RAG technology large-scale language model enhancement technology. Here, the RAG technology large-scale language model enhancement technology includes the establishment of a wind turbine fault feature database, compressed feature comparison, and early warning enhancement technology based on RAG+LLM. In terms of early warning enhancement, in addition to the application system analysis of traditional industrial control systems, the combination of early warning enhancement technology based on large models can further guide the onshore switchyard technicians to make decisions on the current status of the wind turbines and guide the technicians to issue control commands to the wind turbines. The database here consists of various types of structured, unstructured, and semi-structured data, and the database is dynamically updated.

[0091] ② Based on data transmission frequency, data is categorized into cold data and hot data. As is well known, cold data is data that is infrequently generated or used, while hot data is data that is frequently generated or used. This approach utilizes blockchain technology for vertical management of cold and hot data. Cold data is stored using the SM4-XTS mode, while hot data is stored using the SM4-CTR mode. The storage results are returned to the blockchain in CID format. When technicians access specific data characteristics, the CID is linked to the binary string's position sequence table. By performing mapping and comparison, data can be read from the database with zero trust, thereby improving the security of reading large models.

[0092] In summary, a joint compression and encryption method for offshore wind power data based on 5G can be realized.

[0093] This invention relates to a 5G-based method and early warning system for joint compression and encryption of offshore wind power data. In practice, this invention extracts features from source data collected by the wind turbine-side PLC, obtaining four main categories of features: SCADA features, CMS vibration features, environmental coupling features, and image features. These features are then quantized and compressed using a threshold differential compression algorithm. After compression, layered encryption is performed based on national cryptographic algorithms according to data sensitivity. The data features are then transmitted to an onshore switching station via a 5G network. The onshore switching station uses a large RAG model to enhance the early warning of the feedback data features and stores the data features using blockchain distributed storage technology. This 5G-based method and system for joint compression and encryption of offshore wind power data can improve the security of wind turbine communication and the reliability of wind turbine early warning in offshore wind power data communication and wind turbine operation status early warning, greatly improving the level of automated operation and maintenance of wind turbines. This invention has the advantages of being scientifically sound, highly applicable, effective, accurate, and enhancing the level of autonomous and controllable operation and maintenance.

[0094] Example 2

[0095] like Figure 1-4 As shown, a 5G-based offshore wind power data joint compression and encryption early warning system includes a wind turbine equipment layer, a compression layer, an encryption / decryption layer, a large model calculation layer, and a storage layer. The wind turbine equipment layer is connected to the compression layer, the compression layer is connected to both the wind turbine equipment layer and the encryption / decryption layer, the encryption / decryption layer is connected to both the compression layer and the large model calculation layer, the large model calculation layer is connected to both the encryption / decryption layer and the storage layer, and the storage layer is connected to the large model calculation layer.

[0096] The wind turbine equipment layer includes: a wind turbine 5G communication module and a feature extraction classifier.

[0097] Specifically, the wind turbine 5G communication module is responsible for establishing a 5G network communication channel with the 5G base station of the offshore substation;

[0098] The feature extraction classifier extracts features from offshore wind turbine source data, including SCADA features, CMS vibration features, environmental coupling features, and image features.

[0099] As a specific implementation method, the SCADA features include gearbox temperature difference features, bearing temperature gradient features, wind energy utilization coefficient features, and power curve deviation features.

[0100] CMS vibration characteristics include vibration kurtosis characteristics, waveform factor characteristics, bearing fault amplitude ratio characteristics, gear sideband energy characteristics, and wavelet packet energy entropy characteristics.

[0101] Environmental coupling characteristics include wave-to-tower transfer function characteristics, salt spray corrosion characteristics, and tower tilt angle characteristics;

[0102] Image features include blade crack ratio features and average edge curvature features; a total of 14 sub-features.

[0103] The compression layer includes a threshold differential compression module.

[0104] Specifically, the threshold differential compression module includes a designed threshold differential compression algorithm, which is responsible for quantizing and compressing SCADA features, CMS vibration features, environmental coupling features and image features according to the designed calculation rules; at the same time, it introduces SM3 hash values ​​into the compressed binary string to improve the anti-attack and anti-decryption properties of compressed data transmission.

[0105] The encryption / decryption layer includes a national standard encryption / decryption module.

[0106] Specifically, the national cryptographic encryption and decryption module will perform layered encryption based on SCADA features, CMS vibration features, environmental coupling features, and image features, taking into account data sensitivity characteristics and data timing characteristics.

[0107] As a specific implementation method, SCADA features and image features are sensitive data and will be encrypted and decrypted using SM2 based on the national cryptographic encryption and decryption mechanism; CMS vibration features and environmental coupling features are high-volume time-series data and will be encrypted and decrypted using SM4 based on the national cryptographic encryption and decryption mechanism.

[0108] The large model computation layer includes a large model early warning enhancement module based on RAG technology.

[0109] Specifically, the large model early warning enhancement module based on RAG technology is responsible for mapping and comparing the data features transmitted from the offshore booster station to the onshore switch station, and feeding back, transmitting, and controlling the next action of the offshore wind turbine based on the results.

[0110] The storage layer includes a designed blockchain storage module.

[0111] Specifically, the blockchain distributed storage module is responsible for classifying and storing the SCADA features, CMS vibration features, environmental coupling features, and image features transmitted by the offshore substation according to the frequency of data occurrence or use.

[0112] As one feasible implementation, data that appears or is used frequently is stored in a distributed manner using the SM4-CTR mode, while data that appears or is used in a distributed manner is stored in a distributed manner using the SM4-XTS mode.

[0113] This invention relates to a 5G-based method and early warning system for joint compression and encryption of offshore wind power data. In practice, this invention extracts features from source data collected by the wind turbine-side PLC, obtaining four main categories of features: SCADA features, CMS vibration features, environmental coupling features, and image features. These features are then quantized and compressed using a threshold differential compression algorithm. After compression, layered encryption is performed based on national cryptographic algorithms according to data sensitivity. The data features are then transmitted to an onshore switching station via a 5G network. The onshore switching station uses a large RAG model to enhance the early warning of the feedback data features and stores the data features using blockchain distributed storage technology. This 5G-based method and system for joint compression and encryption of offshore wind power data can improve the security of wind turbine communication and the reliability of wind turbine early warning in offshore wind power data communication and wind turbine operation status early warning, greatly improving the level of automated operation and maintenance of wind turbines. This invention has the advantages of being scientifically sound, highly applicable, effective, accurate, and enhancing the level of autonomous and controllable operation and maintenance.

[0114] Example 3

[0115] A computer system includes: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the 5G-based offshore wind power data joint compression and encryption method as described above when running the computer program.

[0116] This invention, based on its objective, utilizes an organic combination of existing technologies to provide an innovative method for enhancing secure communication and fault early warning in the offshore wind power sector. It solves the problems of existing technologies and provides a safety guarantee for the reliable operation of offshore wind turbines. The hardware of this invention consists of commercially available products based on existing technologies, and the software was developed by those skilled in the art based on communication mechanisms, automation technology, and machine learning algorithms, making it easy to implement.

[0117] The above description is merely a specific embodiment of the present invention, but the scope of the present invention is not limited thereto. Any other embodiments derived by those skilled in the art based on the technical solution of the present invention also fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A 5G-based method for joint compression and encryption of offshore wind power data, characterized in that: The method includes the following steps: Step 1: Feature extraction of wind turbine source data: Feature extraction is performed on the source data collected by the PLC on the wind turbine side to obtain a feature dataset; Step 2: Joint compression and encrypted transmission of wind turbine data: The above feature dataset is quantized and compressed using a threshold differential compression algorithm under the 5G network transmission mechanism, and layered encryption is performed based on the national cryptographic encryption and decryption mechanism according to the data sensitivity to obtain a compressed and encrypted feature dataset. Step 3: Blockchain Distributed Storage: The compressed and encrypted feature dataset is transmitted to the land-based switch station with the support of the 5G network. The land-based switch station stores the compressed and encrypted feature dataset based on blockchain distributed storage technology. Step 4: Large model enhances early warning feedback based on data features: The onshore switching station will enhance early warning feedback based on the large model of RAG.

2. The 5G-based joint compression and encryption method for offshore wind power data as described in claim 1, characterized in that: The feature extraction in step 1 includes four main categories: SCADA features, CMS vibration features, environmental coupling features, and image features.

3. The 5G-based joint compression and encryption method for offshore wind power data as described in claim 2, characterized in that: The SCADA features include gearbox temperature difference data features, bearing temperature rise gradient features, wind turbine wind energy utilization coefficient features, and wind turbine power curve deviation features; the image features include wind turbine blade crack ratio features and wind turbine edge average curvature features.

4. The 5G-based joint compression and encryption method for offshore wind power data as described in claim 2, characterized in that: The CMS vibration characteristics include wind turbine vibration kurtosis characteristics, bearing fault amplitude ratio characteristics, wind turbine waveform factor characteristics, wind turbine gear sideband energy characteristics, and wavelet energy entropy characteristics. The environmental coupling characteristics include wave-tower transfer function characteristics, salt spray corrosion characteristics, and tower tilt angle RMS characteristics.

5. The 5G-based joint compression and encryption method for offshore wind power data as described in any one of claims 2-4, characterized in that: The joint compression and encryption method in step 2 specifically involves: after compressing the 14 types of sub-features, performing layered encryption on the data based on data sensitivity characteristics and the advantages of the national cryptographic algorithm. The specific steps are as follows: Step 2.1: For the two types of sensitive features, SCADA features and image features, the SM2 algorithm is used for compression based on the national cryptographic algorithm mechanism; Step 2.2: For the two types of high-volume time-series features, namely CMS vibration features and environmental coupling features, the SM4 algorithm is used for compression based on the national cryptographic algorithm mechanism; Step 2.3: Perform binary conversion on the two types of features encrypted using the SM2 and SM4 algorithms. First, perform binary conversion on the two types of features encrypted by different mechanisms in order, and then combine them into a binary string in order. At this time, record the position sequence table of each feature in the binary string. Step 2.4: Hash the currently calculated binary string using the SM3 hash algorithm; Step 2.5: Concatenate the binary string with the hash value to form a new binary string, and encrypt and transmit it based on the SM2 and SM4 national cryptographic algorithms according to data sensitivity, and transmit the position sequence list of each feature in the binary string together; Step 2.6: The received data at the booster station is decrypted using SM2 and SM4 mechanisms to obtain a binary string and a hash value; Step 2.7: Read the binary string according to the position sequence table of the binary string, remove the hash value, and restore the binary string of the real data; Step 2.8: Perform a decimal conversion on the actual binary string to obtain the differentially compressed data.

6. The 5G-based joint compression and encryption method for offshore wind power data as described in claim 1, characterized in that: The blockchain distributed storage in step 3 specifically involves vertical management of hot and cold data based on blockchain technology. Cold data is stored in SM4-XTS mode, and hot data is stored in SM4-CTR mode. The storage results are returned to the chain record in the form of CID.

7. The 5G-based joint compression and encryption method for offshore wind power data as described in claim 1, characterized in that: The large model in step 4 performs data feature early warning enhancement feedback, specifically: after obtaining the compressed decimal data features at the offshore booster station, the large model analysis results are fed back to the onshore switch station based on the established RAG technology large language model enhancement model. Among them, the RAG technology big language model enhancement model includes: using RAG+LLM to realize the establishment of wind turbine fault feature database, compressed feature comparison and early warning enhancement technology; The database contains various types of structured, unstructured, and semi-structured data.

8. A 5G-based offshore wind power data joint compression and encryption early warning system, comprising a wind turbine equipment layer, a compression layer, an encryption / decryption layer, a large model calculation layer, and a storage layer, characterized in that: The 5G-based offshore wind power data joint compression and encryption early warning system is used to execute the 5G-based offshore wind power data joint compression and encryption method as described in any one of claims 1-7; the wind turbine equipment layer is connected to the compression layer, the compression layer is connected to both the wind turbine equipment layer and the encryption / decryption layer, the encryption / decryption layer is connected to both the compression layer and the large model calculation layer, the large model calculation layer is connected to both the encryption / decryption layer and the storage layer, and the storage layer is connected to the large model calculation layer.

9. The 5G-based offshore wind power data joint compression and encryption early warning system as described in claim 8, characterized in that: The wind turbine equipment layer includes a wind turbine 5G communication module and a feature extraction classifier; the compression layer includes a threshold differential compression module; the encryption / decryption layer includes a national cryptographic encryption / decryption module; the large model calculation layer includes a large model early warning enhancement module based on RAG technology; and the storage layer includes a designed blockchain storage module. The wind turbine 5G communication module is responsible for establishing a 5G network communication channel with the 5G base station of the offshore substation; the feature extraction classifier is used to extract features from the source data collected by the wind turbine-side PLC to obtain a feature dataset, including SCADA features, CMS vibration features, environmental coupling features and image features. The threshold differential compression module is used to quantize and compress the above feature dataset based on the threshold differential compression algorithm under the 5G network transmission mechanism; at the same time, an SM3 hash value is introduced into the compressed binary string. The national cryptographic encryption and decryption module is used to perform layered encryption on the compressed features based on the national cryptographic encryption and decryption mechanism according to the data sensitivity characteristics, so as to obtain a compressed encrypted feature dataset. The blockchain distributed storage module is used to transmit the compressed and encrypted feature dataset to the land-based switch station with the support of the 5G network. The land-based switch station stores the compressed and encrypted feature dataset based on blockchain distributed storage technology. The large model early warning enhancement module based on RAG technology is used in onshore switching stations to enhance the early warning of feedback data features based on the large model of RAG.

10. A computer system, comprising: A memory and processor connected by communication, and a computer program stored on the memory and capable of running on the processor, characterized in that: when the processor runs the computer program, it executes the 5G-based offshore wind power data joint compression and encryption method as described in any one of claims 1-7.

Images