Multi-source spatio-temporal data fusion signal processing system based on high-elasticity transmission conductor

Abstract

A multi-source spatio-temporal data fusion signal processing system based on a high-elasticity transmission conductor comprises a high-elasticity intelligent transmission conductor body and a low-noise light amplification system, and a multi-source data real-time fusion and intelligent processing method is used in the multi-source spatio-temporal data fusion signal processing system. The high-elasticity intelligent power transmission conductor body is a heat-resistant aluminum alloy stranded wire; the high-elasticity intelligent power transmission conductor body is combined with an optical fiber sensing technology and a meteorological technology to form a point, line and surface combined low-cost wide-area multi-element sensing technology, and a signal processing technology for mining wind dance characteristic parameters by virtue of a deep network and realizing multi-source spatio-temporal data fusion by virtue of big data association analysis is adopted, so that the understanding degree of field monitoring data is improved; the value of the multi-element sensing big data is excavated, and objective scientific basis and theoretical basis are provided for safe operation, efficient management and intelligent decision making of the ultra / extra-high voltage transmission line. The device is good in stability, large in bearing current, high in safety, low in cost and capable of reducing carbon emission, and has remarkable economic benefits and social benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of on-line monitoring, evaluation and early warning of transmission cables, and particularly relates to a signal processing system for multi-source spatio-temporal data fusion based on a highly elastic transmission wire. Background Art

[0002] With the construction of China's "clean, low-carbon, safe and efficient" energy system, the installed capacity and consumption proportion of domestic new energy have been increasing year by year. The volatility, randomness and intermittency of new energy pose major challenges to the safe and efficient transmission capacity of grid overhead transmission lines. Moreover, the distribution of China's energy resources and power loads is seriously unbalanced. The central, eastern and coastal regions with relatively developed economies and large populations have strong power demands, while energy resources are mainly distributed in the western regions. China's economic development stage and the characteristics of concentrated energy distribution determine that large-capacity and low-loss long-distance power transmission must be carried out, and power transmission projects are developing in the direction of increasing the operating voltage level. The terrain and meteorological conditions are complex, and power construction needs to overcome difficulties such as harsh environments such as high temperature and high humidity, icing, and long-span power transmission. Therefore, the volatility and randomness brought by the access of new energy to the power grid, as well as the complex environment faced by long-distance power transmission, also put forward higher requirements for the electrical conductivity, heat resistance, elastic margin of transmission capacity, and high-temperature sag of electrical engineering aluminum alloys. A large amount of human and material costs are required for operation and maintenance. To ensure the safety of the line, it is necessary to further improve the monitoring and early warning capabilities and the intelligent and digital levels of transmission lines.

[0003] The current situation of overhead transmission line conductors in China is mainly traditional steel-core aluminum stranded wires. If the method of increasing the conductor cross-sectional area to manufacture large-section steel-core aluminum stranded wires is used to improve the transmission capacity, higher-strength towers must be built to achieve this, which not only increases the corridor area but also increases the construction cost, and will not conform to the guiding principles of "resource-saving, environment-friendly, new technologies, new materials, and new processes" (i.e., two types and three news) transmission lines. In addition, it is currently difficult for common on-line ultra-long line monitoring systems for transmission cables to accurately evaluate and give early warnings. It is difficult to grasp the icing state of transmission lines in real time and globally, which poses an unprecedented challenge to the measurement distance, accuracy, positioning accuracy, etc. of sensing means. Moreover, with high voltage levels, large currents, and strong electric fields, the electromagnetic, acoustic, and vibration cross-interference between devices is large. Coupled with the changing climate conditions such as strong winds, ice and snow, and lightning strikes, the operating environment is complex. How to accurately predict and reliably evaluate the risk coefficient of icing faults at an early stage is challenging.

[0004] The technical development history closely related to the present invention shows that:

[0005] For overhead conductors, although steel stranded wires have high strength, they have high resistance and are prone to corrosion. Although copper wires have low resistance, their costs are very high.

[0006] The research on heat-resistant aluminum alloys in China started relatively late. Domestically produced heat-resistant aluminum conductors were initially mainly used for substation busbars and later were widely applied in line capacity expansion and renovation projects, which not only saved a large amount of project investment but also increased the transmission capacity by 40% - 60%, resulting in obvious economic benefits. At the beginning of this century, heat-resistant conductors gradually began to be used in newly built and expanded lines.

[0007] At present, China has been able to stably produce heat-resistant aluminum alloy conductors with a conductivity of 60% IACS and has been well applied in many important domestic lines. With the continuous in-depth research on heat-resistant aluminum alloy conductors, the conductivity of heat-resistant conductors at the 150°C level in China has now reached or exceeded 61% IACS.

[0008] In summary, currently, the heat resistance of overhead aluminum conductors used for power transmission is poor, the current-carrying capacity is low, and the renewal cost of facilities such as poles caused by replacing large-section conductors is too high. Upgrading the voltage level requires a complete replacement of the entire system. Therefore, using heat-resistant aluminum alloy conductors to replace conductors with the same cross-sectional area for capacity increase is one of the effective technologies to increase the transmission capacity. In addition, aiming at the problems of frequent faults such as line icing and overloading caused by the complex service scenarios of transmission lines, harsh meteorological conditions, and many heavy-load users, which seriously affect the safe and reliable power supply of the lines, it is necessary to continuously improve the real-time perception level of equipment, highlight dynamic prediction, autonomous early warning, and intelligent decision-making construction, effectively improve the self-healing ability of equipment, and reduce the burden on grass-roots operation and maintenance personnel. It is currently difficult to accurately evaluate and early warn common on-line monitoring systems for transmission cables, mainly reflected in the difficulties of ultra-long-distance monitoring and maintenance, inability to grasp the icing state of transmission lines in real time and globally; difficult field power supply, the timeliness of power supply and maintenance of equipment, the efficiency of fault recovery, and the fault location technology need to be improved; the environment is complex, the interference is strong, and the risk assessment and early warning are difficult. How to accurately predict and reliably evaluate the risk coefficient of icing faults at an early stage is challenging.

[0009] Therefore, in the future, more and more clean energies with characteristics such as volatility, randomness, and intermittency of transmission capacity will be converted into electrical energy through advanced power generation technologies and connected to the main network lines. To ensure the stability, controllability of the new energy grid connection output, and reliability issues brought by complex climate environments, it is necessary to develop new high-conductivity, heat-resistant, and high-elastic intelligent power transmission conductors. Summary of the Invention

[0010] The present invention provides a signal processing system for multi-source spatio-temporal data fusion based on a high-elastic power transmission conductor. The technical principle is described as follows:

[0011] When a metal is subjected to an external electric field, the carriers move directionally in the lattice field (also known as the lattice Coulomb potential field) formed by the periodically arranged ion cores, thus forming an electric current. Alloying elements, impurities, crystal defects, etc. can all cause periodic damage to the lattice field. Those abnormal ion cores or lattice atoms that disrupt the periodicity of the lattice field collide with or impede the directionally moving carriers, resulting in resistance. Therefore, the electrical conductivity of an alloy highly depends on the alloy composition and microstructure. Solute atoms, vacancies, precipitates, dislocations, etc. can all cause lattice distortion, leading to scattering of carriers during movement. With the addition of alloying elements and the increase in the addition amount, the resistivity of aluminum alloys will increase significantly. Alloying elements and their existing states have different effects on the electrical conductivity of aluminum alloys. When impurities or alloying elements are dissolved in the aluminum matrix, the electrical conductivity of aluminum alloys will be significantly reduced. After they precipitate from the solid solution, the influence on the electrical conductivity is only one fraction to one fortieth of that in the solid solution state. Therefore, it is very necessary to perform heat treatment on conductor materials.

[0012] To increase the heat-resistant temperature of an alloy, it is necessary to add microalloying elements and make the second phase that precipitates disperse and distribute to inhibit interface migration. However, these elements cause great damage to the electrical conductivity, especially when they exist in the solid solution state. In 1949, American scholars found through research that adding a small amount of zirconium element to aluminum materials can improve the heat-resistant performance of aluminum materials. Japan took the lead in developing Al-Zr series heat-resistant aluminum alloy wires, which can improve the heat-resistant performance of materials without reducing the tensile strength and electrical conductivity of the wires. Heat-resistant aluminum alloys mostly use micro-zirconium aluminum alloys. Zr atoms interact with dislocations and grain boundaries, segregate at grain boundaries, and hinder the migration of large-angle grain boundaries, thus hindering the recovery and recrystallization of aluminum alloys when the temperature rises, and thus increasing the heat-resistant temperature of aluminum alloys. However, due to the relatively large difference in atomic radii between Zr and Al, large lattice distortion will occur when Zr is dissolved in aluminum, and the diffusion rate of Zr is relatively low. Supersaturated Zr is difficult to precipitate from the aluminum matrix. Therefore, with the increase in the zirconium content in aluminum, the heat resistance is greatly improved, but the electrical conductivity is significantly reduced. To balance the electrical conductivity, the zirconium content of traditional heat-resistant aluminum alloys is generally not more than 0.1%, and the heat-resistant grade is relatively low.

[0013] To solve the contradiction among electrical conductivity, heat-resistant temperature, and strength, it is necessary to form second-phase particles that are highly dispersed, thermodynamically stable, coherent or semi-coherent with the matrix through methods such as reasonable composition design, appropriate preparation processes, and precise heat treatment control, so as to inhibit interface migration, and only then can the goal of synergistically improving electrical conductivity, heat resistance, and strength be achieved.

[0014] Intelligent power transmission sensing and monitoring mainly uses fiber Bragg grating sensing technology and convolutional neural network methods. Among them, the fiber Bragg grating sensor is an important sensing component in the fiber optic sensing system, which is made based on the sensing mechanism of fiber grating elements. The sensor encodes the wavelength, overcoming the weaknesses of intensity modulation sensors that must compensate for the losses of fiber connectors and couplers and the fluctuations in the output power of light sources. The fiber grating is an excellent sensitive element, and its peak wavelength changes with the changes of physical quantities such as temperature and stress. By designing sensitive structures for the conversion of non-optical physical quantities, non-optical quantity optical measurements can also be realized. The external parameters such as strain, pressure, temperature, wind direction, and temperature micro-vibration can be sensed directly by demodulating the wavelength signal of the fiber grating. Since the fiber grating uses the wavelength encoding method, multiplexing technologies such as wavelength division, time division, and space division can be used to form a sensing network by connecting fiber Bragg gratings in series or parallel, and quasi-distributed networked measurement of physical quantities can be realized. This is a major advantage of the fiber grating sensing system.

[0015] In recent years, machine learning algorithms have been widely used in the pattern recognition of distributed fiber optic sensing signals, and they have also been relatively well-developed in aspects such as data feature extraction, dataset construction, and pattern recognition algorithms. However, in the construction of the dataset, it is necessary to manually extract the features of the data, and with the change of the environment, the features of the data will also change accordingly. This often requires a large amount of human resources and will greatly reduce the efficiency of machine learning. Convolutional neural networks (CNNs) perfectly avoid the disadvantages of manual data feature extraction. It can automatically extract the features of the data and perform classification and recognition. Two-dimensional convolutional neural networks have played a huge role in the field of image recognition. Images can be directly used as input signals without any manual feature extraction, and the neural network can automatically extract the features of the images. One-dimensional convolutional neural networks can achieve the recognition of one-dimensional sequence signal types such as vibration and sound without setting specific feature engineering. The entire neural network system has good robustness and high operation efficiency.

[0016] Adopt the technology of combining fiber random laser technology and hybrid amplification of remotely pumped erbium-doped fiber to develop an ultra-long-distance and low-noise fiber optic sensing and monitoring system, integrate deep learning algorithms, develop an algorithm for monitoring the icing of transmission lines, realize synchronous monitoring and data sharing among the source, grid, and load, improve the level of autonomous early warning and intelligent interconnection and interaction, improve the level of operation and maintenance intelligence, and support the large-scale safe grid connection and full consumption of new energy.

[0017] In this technical solution, for the signal processing system based on the multi-source spatio-temporal data fusion of high-elastic transmission lines, the key technologies are as follows: It includes the high-elastic intelligent transmission line body and the low-noise optical amplification system, and uses the multi-source data real-time fusion and intelligent processing method; they jointly constitute the necessary auxiliary components to support the high-elastic intelligent transmission line; among them:

[0018] The high-elastic intelligent transmission line body is a domestic steel-core 58% conductivity heat-resistant aluminum alloy stranded wire NRLH58GJ-240 / 30;

[0019] In the signal processing system based on the multi-source spatio-temporal data fusion of high-elastic intelligent transmission lines, the high-elastic intelligent transmission line body combines fiber optic sensing technology and meteorological technology to form a low-cost wide-area multi-source sensing technology that combines points, lines, and surfaces. By using deep network to mine the wind dancing characteristic parameters and using big data correlation analysis to realize the signal processing technology of multi-source spatio-temporal data fusion, it improves the understanding of on-site monitoring data, explores the self-value of multi-source sensing big data, and provides an objective scientific basis and theoretical foundation for the safe operation, efficient management, and intelligent decision-making of ultra / extra-high voltage transmission lines;

[0020] I. The low-noise optical amplification system uses the following new fiber optic sensing technology:

[0021] (I) New distributed amplification technology

[0022] First, a theoretical model of signal optical power distribution based on high-order low-noise fiber random laser amplification is established

[0023] For the new low-noise optical amplification system, the power steady-state equation model shown in Equation (1) is used to conduct theoretical analysis on the high-order fiber random laser amplification part in this amplification system. The power distribution of each frequency component along the fiber length is calculated using the following model:

[0024]

[0025] Among them, the subscripts '0', '1', '2' correspond to the pump, the first-order and second-order Stokes lights respectively; the superscripts '+' and '-' represent the forward and backward waves; P0,1,2 represent the optical power; z represents the optical transmission direction coordinate; f0,1,2 represent the optical frequency; Γ1,2 represent the number of photons; where Δf1,2 = 0.25 THz represents the radiation bandwidth; T = 298 K represents the absolute temperature, KB represents the Boltzmann constant, h represents the Planck constant, α0,1,2 represent the fiber transmission loss, g1,2 represent the Raman gain coefficient, and ε0,1,2 represent the Rayleigh backscattering coefficient;

[0026] When performing numerical analysis, consider the boundary conditions: P+(0) = P, 0in P+(0) = R P-(0), P-(L) = RP+(L) to represent different cavity structures (fully open cavity structure, semi-open 1,2L1,21,21,2F1,21,2 cavity structure, and the influence of different reflectivities), where Pin represents the pump power. By constraining these boundary conditions, use the iterative method to find the numerical solution of the above equation;

[0027] Second, establish a theoretical model based on the L-ROPA technology

[0028] The working principle of the new L-ROPA technology is an erbium-doped fiber amplifier. The internal population of particles is mainly concentrated in the ground state and metastable state. The Giles simplified model can be used to analyze the gain coefficient and amplification characteristics of L-ROPA:

[0029]

[0030] Among them, the superscripts ‘+’ and ‘-’ represent the forward and backward waves, the subscript ‘k’ represents the light beam of the k-th wavelength, and N t is the total average number of particles in the two-level system, N 2 is the number of particles in the upper energy level, α k represents the absorption coefficient of wavelength k, g k represents the gain coefficient of wavelength k, Δυ k is the effective noise bandwidth, m represents the number of optical wave polarization modes, υ k represents the frequency of the light of the k-th wavelength, l is the background loss, and ζ is the saturation coefficient of the erbium-doped fiber;

[0031] Third, establish a RIN transfer theoretical model based on the low-noise fiber random laser amplification system

[0032] To realize the new low-noise fiber random laser amplification system, it is necessary to study the RIN transfer from the pump light to the signal light in the amplification system. It can be derived by the perturbation method that the RIN transfer amount H(f) under the condition of co-pumping is:

[0033]

[0034] Combined with the above theoretical model, the RIN transfer characteristics and optimization methods in the fiber optic sensing system based on the new low-noise fiber random laser amplification technology can be analyzed; the combination of the high-order low-noise fiber random laser amplification technology and the L-ROPA technology can effectively overcome the power loss of long-distance optical transmission and realize a long-distance low-noise optical amplification system, as Figure 1 shown;

[0035] Figure 2 is the simulation curve of the Rayleigh scattered signal light power distribution of the new low-noise optical amplification system under a sensing distance of 30 km, that is, the power distribution of light along the optical fiber;Figure 3 It is the simulation curve of the Rayleigh scattering signal optical power distribution, i.e., the optical power distribution along the optical fiber, of a new low-noise optical amplification system at a sensing distance of 50 km. Among them, the red curve corresponds to the Rayleigh scattering signal optical power in the high-order fiber random laser amplification system, and the blue curve corresponds to the Rayleigh scattering signal optical power in the new low-noise optical amplification system. The transmission loss of the 1550 nm signal light is 0.25 dB / km, the fiber random laser pump power is 1.7 W, and the length of the erbium-doped fiber is 10 m. To overcome the optical power loss caused by the transmission and splitting of the sensing signal light in the PON, the new L-ROPA technology is combined with the distributed fiber vibration sensing technology, making the signal-to-noise ratio at the end of the sensing optical fiber of the new LL-DVS system based on the optical fiber communication network higher, and enabling a high signal-to-noise ratio and high-sensitivity LL-DVS system based on the OPGW optical cable;

[0036] The signal processing system based on the multi-source spatio-temporal data fusion of high-elastic transmission wires preferably requires the protected technical content as follows: In the automatic discrimination of the icing characteristics and icing levels of high-elastic wires and the specific content requirements of constructing an icing assessment and early warning model for transmission lines based on multi-source data fusion by applying the following multi-source data real-time fusion and intelligent processing method:

[0037] Line galloping characteristic analysis and galloping level discrimination method based on deep learning:

[0038] Through the deep learning network, a non-linear mapping relationship is established by combining the icing state levels of high-elastic wires for the analysis and feature expression of line icing characteristics. During this period, methods for extracting micro-meteorological and stress time-frequency characteristics and galloping characteristics based on deep learning are applied, such as Figure 4 、 Figure 5 as shown. Figure 4 It is for the extraction of the time-series structure characteristics of line icing based on 1-D CNN, Figure 5 and for the extraction of the spatial distribution characteristics of line icing based on Bi-LSTM. Then, a mapping relationship between the spatio-temporal characteristics of line icing and the galloping level is constructed based on the 1D-CNN-Bi-LSTM network, and the automatic discrimination of the icing level is realized according to the non-linear mapping relationship between the icing characteristics and the icing level constructed by this network;

[0039] The requirements for the multi-source data fusion method based on association analysis are:

[0040] First, a data mining method based on association analysis: Association analysis is a very commonly used data mining method, and its main purpose is to help us discover the possible correlations between things to achieve a better understanding of the data. It has the ability and advantage of directly analyzing the essential laws of the data. In the leakage detection method of this paper, the Apriori algorithm, an association analysis algorithm, will be used to discover certain sets of data items that often appear simultaneously, especially the set of data items in which the characteristic items of the acoustic wave sensing signal and the target event items frequently appear simultaneously, so as to find out some important rules of things happening hidden inside these sets of data items; in association analysis, the item sets that meet the condition that their support degree is greater than the minimum support degree threshold in these sets of data items are defined as frequent item sets, and the rules of things happening in the item sets are defined as association rules, and the meaning represented by the support degree of an item set refers to the proportion of the number of data records containing this item set in all records of the data set; the main process of the above-described association rule mining is as Figure 6 shown. The Apriori algorithm is one of the most commonly used frequent item set mining algorithms in association analysis. Its main idea is to scan the data set multiple times for multi-item set statistics. Each time, all candidate frequent item sets are counted, and then the candidate sets are pruned based on the support degree threshold to generate new frequent item sets.

[0041] Second, the frequent item set mining method: In the association rule mining method, when performing frequent item set mining, in addition to the support degree must meet the threshold condition, there is an additional pruning condition: that is, the finally extracted item set must contain and only contain one thing representing the target event, and the rest of the items all represent any possible characteristic items; as Figure 7 shown is a detailed example of the frequent item set mining process related to the pipeline leakage detection method of this paper. The frequent item sets generated after pruning are shown in the 2nd, 3rd, and 4th layers in the figure. They are all candidate sets that meet the pruning conditions of the minimum support degree threshold and having one and only one target event item pruning condition. For example, Figure 7 the frequent 4-item set obtained in the 4th layer in the example: {4, 6, 7, C}, where 4, 6, and 7 are signal characteristic numbers, and C represents the target event category. The meaning represented by this frequent item set is that the characteristics {4, 6, 7} all belong to Class1 and always appear simultaneously with the occurrence of the target event C.

[0042] Third, the generation and pruning of association rules

[0043] By setting a reasonable confidence threshold, the association rules that meet the minimum confidence threshold condition are generated to complete the basic content of association analysis. However, in actual use, when the dimension of the feature set matches the number of target events, a large number of association rules that meet the requirements are generally generated. However, there is information redundancy in these rules, and even some of them have some imperceptible negative effects. Simply increasing the confidence threshold cannot achieve the purpose of compressing the redundant data volume and extracting more effective association rules. In this case, two rule evaluation indicators, the Kulc coefficient k(X→Y) and the imbalance factor r(X→Y), can be introduced to prune the existing association rules:

[0044]

[0045] The Kulc coefficient calculates the forward and backward confidence levels c(X→Y) and c(Y→X) for the same association rule and then performs an averaging process. Like the confidence level, the higher its value represents a stronger conditional probability for the rule that a feature causes an event to occur. It has the advantage of being unaffected by null values in the dataset compared to a single confidence level. The imbalance factor is often used together with the Kulc coefficient. It mainly has the advantage of being able to eliminate negative effect rules. The smaller its value represents better balance and stronger positive correlation for the association rule;

[0046] Fourth, the design and implementation of the association rule classifier: For the results of data preprocessing - two boolean feature matrices A and B that are opposite to each other, add K-dimensional binary sequence labels of K types of target events as the input for the association rule mining method based on the Apriori algorithm in Subsection 4.1. After performing two association analyses on the two boolean matrices respectively, two sets of strong association rules between the feature item sets and the target event items that meet the confidence level, Kulc coefficient, and imbalance factor threshold conditions are obtained. Among them, the association rule set mined based on the boolean matrix A is defined as R A , and the association rule set mined based on the boolean matrix B is defined as R B , both are stored in the association rule classifier as the most core reference criteria for classification and discrimination; the association rules stored in the rule sets R A , R B are each classified according to the category numbers of the K types of target events, such as: {1, 2, 3,...}, that is, the antecedents of the association rules with the same consequent are combined into small sets respectively, such as: R A1 , R A2 , R A3 , R B1 etc., for convenient use in classification and discrimination; the classifier also includes the feature selection results saved during the training period and the feature clustering center set obtained using the FCM algorithm during the binarization process. So far, the core content of the association rule classifier has been constructed;

[0047] The process of using the association rule classifier to discriminate target events is as follows:

[0048] As Figure 8 shown, the main steps implemented by the association rule classifier include:

[0049] 1) Extract the MFCC features and AR model coefficient features of the test signal, select the optimal feature columns with reference to the feature selection results, calculate the membership degrees of the feature values to the two categories of Class0 and Class1 with reference to the FCM clustering centers. Each frame of the time-domain test signal is converted to obtain two groups of positive and negative Boolean feature sequences. Specific test signal preprocessing methods;

[0050] 2) Compare the Boolean feature sequences with the rule sets of various target events in the association rule classifier, and respectively count the proportion of the number of satisfied conditions of the test signal in the rule sets of various events. Among them, the proportion of the number of the same type of event is the result of adding the proportions in the positive and negative rule sets and taking the sum value;

[0051] 3) Compare the proportion values of various events, calculate the maximum value Max, set a threshold thr for the recognized category determination. Otherwise, the discrimination completed under the condition that Max is very small is not persuasive; if Max meets the minimum discrimination threshold thr, then determine the current test signal as the target event category corresponding to the maximum value Max; if the threshold condition is not met, that is, the test signal is quite different from the features of the mined rule set, then determine the current signal as the default event or unknown event;

[0052] Based on the above association relationships between multiple parameters such as meteorology, terrain, line parameters, and icing characteristics and the icing level, realize the estimation and automatic evaluation of the line icing risk coefficient, as the data basis for safety warning.

[0053] (3) The requirements for model construction and system software development are:

[0054] Build a transmission line icing assessment and warning model based on multi-source data fusion; build a transmission line icing comprehensive monitoring, assessment and warning interaction platform and software based on the research algorithm, realize a multi-parameter fusion perception system, and on this basis realize functions such as highly reliable icing state assessment and safety warning based on multi-parameter fusion;

[0055] The system platform software is planned to be designed with a B / S architecture, as Figure 9As shown in the figure, it runs on the Linux platform and is mainly divided into a data platform (backend) and a user platform (frontend); among them: the data platform is the backend, and the user platform is the frontend; the backend is written in C# and is mainly responsible for communicating with systems such as fiber optic sensing monitoring, parsing the data sent to the software system, and processing and storing it. The storage method can choose local database storage and remote server storage; the frontend mainly uses VUE, etc. for web page writing, and is mainly responsible for interacting with users, providing functions such as user login and management, data waveform display, alarm display, electronic map display, and multi-parameter visualization display. Users obtain web page content by accessing the Apache server; a database is used as a data interface between the front and backend for data transmission and caching of the three device systems, and a TCP stream is used as a control interface for coordinated control and information interaction.

[0056] Select transmission lines of 66 kV and above for multi-source data fusion perception, comprehensive testing of system hardware and software, and actual icing experiments to complete the demonstration application of comprehensive monitoring, evaluation, and safety warning of the icing state of ultra-long-distance transmission lines.

[0057] The signal processing system based on multi-source spatio-temporal data fusion of high-elastic transmission wires uses the following line readout characteristic analysis and wind dance level discrimination method: The content of line readout characteristic analysis is as follows:

[0058] (1) Signal transmission characteristics mainly involve: high-speed signal transmission ability, signal attenuation and anti-interference, multi-channel transmission and parallel processing;

[0059] High-speed signal transmission ability: When the line reads doctoral data (which can be understood as a high-speed and large-capacity data transmission scenario), it can support high-frequency signal transmission; this means that it can process a large number of data bits in a short time to meet the need for rapid data acquisition during the readout process. For example, in modern scientific research experiments, from complex physical simulation calculation results to the reading of massive biological gene data, the line needs to have an efficient high-speed signal transmission ability to ensure the real-time and accuracy of data. The transmission rate specifically depends on factors such as the material of the line, manufacturing process, and the transmission protocol adopted; adopting advanced coding techniques and signal modulation methods can effectively reduce signal distortion and interference during transmission, improving signal integrity and reliability;

[0060] Signal attenuation and anti-interference: As the transmission distance increases, the signal in the line will inevitably attenuate. In long-distance readout data transmission, this may lead to a decline in signal quality, affecting the accuracy and integrity of data. Therefore, the line needs to have good signal attenuation control characteristics;

[0061] High-quality insulation materials and shielding technologies are adopted to reduce signal attenuation and external interference. For example, using special optical fiber materials can reduce the loss of optical signals during transmission. For electrical signal lines, a multi-layer shielding structure can effectively block external electromagnetic interference and ensure the stability of signals during transmission. At the same time, the line is also equipped with devices such as signal amplifiers to compensate for and enhance attenuated signals to ensure that the signals can still be accurately read after long-distance transmission.

[0062] Multi-channel transmission and parallel processing: To improve the transmission efficiency of read data, the line often supports multi-channel transmission. This is like multiple lanes on a highway, which can transmit different data blocks or data streams simultaneously. For example, in a large-scale data storage system, the line can read data from a disk array in parallel through multiple channels, greatly shortening the data reading time. Multi-channel transmission requires the line to have good parallel processing capabilities to coordinate data transmission between channels and avoid data conflicts and chaos. This involves complex timing control and data scheduling algorithms to ensure that data from different channels can reach the destination accurately and orderly. At the same time, the line also needs to have the function of monitoring and managing multi-channel transmission to promptly detect and solve possible channel failures or data transmission anomalies.

[0063] (2) The data processing characteristics meet the following requirements: data caching and prefetching, data error correction and verification, protocol adaptation and compatibility.

[0064] Data caching and prefetching: During the read process, the access to data has a certain degree of randomness and suddenness. To improve the response speed of data reading, the line has a data caching function. It can temporarily store data that may be frequently accessed in the near future in the cache. When these data are needed again, they can be quickly obtained directly from the cache without having to read from the data source (such as a hard disk, server, etc.) again, greatly reducing the data reading latency. In addition, the line also adopts prefetching technology. By analyzing the historical patterns and rules of data access, it predicts the data that needs to be read next in advance and performs prefetching operations in the background to preload these data into the cache. When the user really needs these data, a zero-latency or near-zero-latency reading experience can be achieved, greatly improving the efficiency and smoothness of the read operation.

[0065] Data error correction and verification: Since data may be corrupted due to various factors during line transmission, the data error correction and verification function is crucial. The line adopts various error correction coding technologies, such as parity check codes, cyclic redundancy check codes (CRC), etc., to monitor and verify the transmitted data in real time.

[0066] When a data error is detected, the circuit can automatically correct the error according to the principle of error correction coding; for some errors that cannot be automatically corrected, it will promptly issue an error report to notify the relevant system or user to take corresponding measures, such as retransmitting the data, etc. This data error correction and verification mechanism ensures the accuracy and reliability of the data read from the doctoral degree, preventing scientific research analysis errors or experimental result deviations caused by data errors.

[0067] Protocol adaptation and compatibility: When the circuit reads data from the doctoral degree, it needs to communicate and interact with various different devices and systems. Therefore, it must have good protocol adaptation and compatibility. Different scientific research devices, storage systems, and computing platforms may adopt different data transmission protocols and interface standards. The circuit needs to be able to support multiple common protocols, such as USB, PCIe, Ethernet, etc., and be able to seamlessly convert and adapt between different protocols. For example, when the circuit is connected to an experimental instrument using the USB protocol and a server using the Ethernet protocol, it can automatically recognize and adapt to these two different protocols to ensure that data can be correctly transmitted and interacted between the instrument and the server. This protocol adaptation and compatibility enables the circuit to be flexibly applied to various complex scientific research environments and scenarios of reading the doctoral degree, providing strong support for data sharing and collaborative work between different devices.

[0068] (3) The reliability and stability characteristics have the following requirements: fault tolerance design, environmental adaptability, long-term operation stability;

[0069] Fault tolerance design: The circuit needs to have a high degree of reliability during the process of reading the doctoral degree to cope with various possible failure situations. Fault tolerance design is one of the important means to improve the reliability of the circuit. For example, the circuit may adopt a redundant link design. When the main link fails, the standby link can automatically switch to ensure that data transmission is not interrupted; in addition, for key components in the circuit, such as chips, connectors, etc., redundant backups may be used. If a component fails, the backup component can immediately take over the work to ensure the normal operation of the circuit. This fault tolerance design greatly reduces the risk of the entire system for reading the doctoral degree being paralyzed due to a single point of failure, improving the availability and reliability of the system;

[0070] Environmental adaptability: The environment for pursuing a Ph.D. can be diverse, including different places such as laboratories and field monitoring stations. The circuit needs to have good environmental adaptability. In terms of temperature, the circuit may need to be able to operate normally within a wide temperature range, from low-temperature laboratory environments to high-temperature outdoor equipment boxes. It may adopt special heat dissipation designs or high-temperature resistant materials to ensure the stable performance of the circuit under different temperature conditions. In terms of humidity, the circuit needs to have a certain moisture-proof ability to prevent problems such as short circuits or corrosion caused by moisture intrusion. For environments where there may be electromagnetic interference, the circuit also needs to have good electromagnetic interference resistance to ensure the accuracy and stability of data transmission. For example, in scientific research experiments in strong electromagnetic field environments, the circuit can effectively resist external electromagnetic interference through technologies such as shielding and filtering to ensure the reliable transmission of Ph.D. data.

[0071] Long-term operation stability: The work of pursuing a Ph.D. requires the circuit to operate continuously for a long time. Therefore, the long-term operation stability of the circuit is crucial. During the design and manufacturing process of the circuit, strict reliability tests and aging tests will be carried out to ensure that its performance does not decline and the failure rate is low during long-term operation;

[0072] At the same time, the circuit may adopt intelligent monitoring and maintenance technologies to monitor the operating status of the circuit in real time, such as parameters like voltage, current, and temperature. When abnormal situations are detected, it can issue early warnings in a timely manner and perform self-diagnosis and repair. For example, some advanced circuit systems can analyze the operating data of the circuit through built-in software algorithms, predict possible failures in advance, and take corresponding preventive measures, such as adjusting working parameters and performing equipment maintenance, so as to ensure the stable operation of the circuit during long-term Ph.D. tasks;

[0073] The wind dance level discrimination method meets the following requirements:

[0074] (1) Discrimination method based on wind speed:

[0075] Measurement principle and equipment: Wind speed is one of the important parameters for judging the wind dance level. Commonly used wind speed measurement equipment is a mechanical anemometer or an ultrasonic anemometer. The working principle of the mechanical anemometer is based on various technologies. It rotates through devices such as wind cups under the action of wind force, and the wind speed is deduced according to the rotation speed. The ultrasonic anemometer measures the wind speed by using the principle that the propagation speed of ultrasonic waves in the air is affected by the wind speed. The anemometer is installed in an open and unobstructed position to ensure that the true wind speed can be accurately measured. When installing, attention needs to be paid to the height and direction of the instrument. Generally, it should be installed in accordance with relevant standards and specifications to ensure the accuracy and comparability of the measurement data;

[0076] The relevant technical specifications of the anemometer are:

[0077] 1) JJG 1194-2023 "Digital Vane Anemometer": It stipulates the metrological performance requirements, general technical requirements, verification conditions, verification items, verification methods, handling of verification results, etc. of digital vane anemometers;

[0078] 2) JJG(Building) 01-1992 "Verification Regulation of Thermal Ball Anemometer" and JJG(Building) 0001-1992 "Metrological Verification Regulation of Thermal Ball Anemometer": They provide specific regulation requirements for the verification of thermal ball anemometers.

[0079] 3) GJB 6550-2008 "General Specification for Military Digital Wind Direction and Speed Anemometer": It is the general specification standard for military digital wind direction and speed anemometers.

[0080] 4) GJB / J 3828-1999 "Verification Regulation of Hot-Wire / Thermal-Film Anemometer": It is applicable to the verification of hot-wire / thermal-film anemometers.

[0081] 5) ESDU 16002-2018 "Practical Guide to Turbulence and Flow Measurement Using Thermal Anemometers (CTA), Laser Doppler Anemometers (LDA) and Particle Image Velocimetry (PIV) - Part 2: Constant Temperature Anemometers" and ESDU 17008-2018 "Practical Guide to Turbulence and Flow Measurement Using Thermal Anemometers (CTA), Laser Doppler Anemometers (LDA) and Particle Image Velocimetry (PIV) - Part 3: Laser Doppler Anemometers": They provide practical guides for thermal anemometers, laser Doppler anemometers, etc. in turbulence and flow measurement.

[0082] 6) KSE 4086-2004(2009) "Portable Mine Thermal Anemometer": It is the standard for portable mine thermal anemometers.

[0083] Wind speed grade classification standard: According to the internationally common wind force grade classification standard, the wind speed range is divided into different levels; for example, level 0 wind means calm wind, with a wind speed less than 0.2 m / s; level 1 wind is a light breeze, with a wind speed between 0.3 - 1.5 m / s; level 2 wind is a gentle breeze, with a wind speed of 1.6 - 3.3 m / s, etc. As the wind speed increases, the wind grade also increases accordingly. For different application scenarios, the wind speed grade will be further subdivided or adjusted according to actual needs; for example, in the assessment of the siting of certain specific wind farms, different wind speed intervals may be more detailedly divided to more accurately evaluate the wind energy resources and the applicability of wind turbines. In the discrimination of wind grades, strictly follow the established wind speed grade classification standard and combine the actually measured wind speed data to determine the current wind grade;

[0084] Data acquisition and processing: An anemometer continuously collects wind speed data and transmits it to a data processing system. The data processing system filters and averages the collected raw data to remove noise and outliers, obtaining a more accurate wind speed value. For example, using a moving average algorithm to average the wind speed data within a certain time window can reduce the impact of instantaneous wind speed fluctuations on the determination of the wind dance level. At the same time, the data processing system also monitors and analyzes the wind speed data in real time. When the wind speed changes and reaches different level thresholds, it timely updates the determination result of the wind dance level;

[0085] (2) Discrimination method based on wind direction change

[0086] Wind direction measurement technology: The measurement of wind direction is also of great significance for the determination of the wind dance level. Wind direction sensors or / and wind vanes are devices for measuring wind direction, and common ones include wind vanes, etc. The wind vane determines the wind direction through the pointing of its pointer under the action of wind force. Modern wind direction sensors usually adopt electronic technology and can convert wind direction information into electrical signals for output, facilitating data acquisition and processing. Attention should also be paid to the installation position of the wind direction sensor. It should be ensured that it can rotate freely without being interfered by surrounding obstacles and accurately measure the wind direction.

[0087] Relationship between wind direction change characteristics and wind dance level: Different wind dance levels are accompanied by different wind direction change characteristics. In lower-level wind dances, the wind direction may be relatively stable with a small change range. As the wind dance level increases, the wind direction change may become more frequent and intense. For example, in the case of strong winds (higher wind dance levels), due to the complexity of the atmospheric circulation and the influence of factors such as terrain, the wind direction may change by a large angle in a short time. By monitoring and analyzing the wind direction change, it is possible to assist in determining the wind dance level. Specifically, by calculating the change frequency and angle range of the wind direction, combined with historical data and empirical models, the threshold of wind direction change characteristics corresponding to different wind dance levels is determined. When the actually measured wind direction change parameters exceed or meet the threshold corresponding to a certain wind dance level, it can be determined that the current situation is at that wind dance level;

[0088] Discrimination method integrating wind speed and wind direction: To more accurately determine the wind dance level, it is necessary to comprehensively consider two factors: wind speed and wind direction. Establish a wind speed-wind direction joint discrimination model and input the measured data of wind speed and wind direction into the model for analysis and processing. In the model, according to the meteorological conditions and actual wind dance phenomena corresponding to different wind speed and wind direction combinations, corresponding discrimination rules and thresholds are formulated. For example, when the wind speed reaches a certain level and the wind direction changes frequently and by a large amplitude, it may correspond to a higher wind dance level. When the wind speed is low but the wind direction is relatively stable, it may correspond to a lower wind dance level. Through this comprehensive discrimination method, the wind dance level can be judged more comprehensively and accurately, providing a more reliable basis for relevant weather forecasts, wind energy utilization, outdoor activity safety, etc.;

[0089] (3) Discrimination method based on the influence of wind on objects

[0090] Observation objects and indicators: It is a relatively intuitive method to judge the wind dance level by observing the influence of wind on different objects. Some common objects can be selected as observation objects, such as trees, flags, cooking smoke, etc. The observation indicators include the swing amplitude and morphological changes of the objects. For example, in the case of gentle wind (low wind dance level), the branches of the trees may sway gently and the flags may flutter slightly. While in strong wind (high wind dance level), the trees may sway greatly, and even the branches may break, the flags will be blown fully open and flutter violently, and the cooking smoke will be dispersed and the direction is unstable. Establish the corresponding relationship between the degree of influence and the wind dance level: Based on the observation and analysis of different objects under different wind forces, establish the corresponding relationship between the degree of influence of the objects and the wind dance level. This requires a large amount of field observations and data accumulation, and summary and induction in combination with meteorological principles and experience. For example, the swing amplitude of the trees can be divided into several levels, such as slight swing, moderate swing, large swing, etc., corresponding to different wind dance levels respectively. At the same time, corresponding level divisions are also made for the fluttering state of the flags, the diffusion form of the cooking smoke, etc., and the corresponding relationship with the wind dance level is established. When actually judging the wind dance level, by observing the actual performance of these objects and comparing with the pre-established corresponding relationship, the current wind dance level can be quickly judged;

[0091] Limitations and supplementary methods: Although this discrimination method based on the influence of wind on objects is intuitive, it also has certain limitations. For example, the degree of influence of the objects may be affected by the characteristics of the objects themselves (such as the type and height of the trees, the material of the flags, etc.) and the surrounding environment (such as the obstruction of buildings, the undulation of the terrain, etc.), resulting in certain errors in the discrimination results.

[0092] To make up for this limitation, discrimination methods such as wind speed measurement and wind direction change analysis can be combined for comprehensive judgment. At the same time, continuously improve and optimize the corresponding relationship between the observation objects and the wind dance level, consider more factors, and improve the accuracy and reliability of the discrimination method. For example, for common objects in different regions and different environments, establish more targeted corresponding relationships between the degree of influence and the wind dance level to adapt to different actual application scenarios.

[0093] The signal processing system for multi-source spatio-temporal data fusion based on the highly elastic transmission wire meets the following requirements:

[0094] (1) Theoretical and practical basis of the research content:

[0095] When alloying elements are added to an alloy to form a solid solution, the lattice points of the matrix are distorted, strengthening the aluminum alloy. However, this disrupts the periodicity of the lattice potential field, increasing electron scattering and leading to an increase in resistance. This is the case even when a metal element with a low resistivity dissolves into a metal with a high resistivity. Different alloying elements have different degrees of influence on the conductivity of aluminum alloys. Elements such as Ca, Bi, Cd, Be, Sb, Ni, Pb, In, Ba, Co, Ce, Sr, Ga, and Zn have a relatively small impact on the conductivity of aluminum alloys. Elements such as Mg, Si, Ge, Fe, Ag, and Cu do not have a significant impact on the conductivity of aluminum alloys, while elements such as Ti, V, Cr, Mn, Zr, and Li have a greater impact on the conductivity of aluminum alloys. Boron treatment is the most effective way to improve the conductivity of aluminum alloys. B can react with impurity elements such as Ti, V, Cr, Mn, and Fe to form insoluble borides or complex compounds containing impurity elements, causing the impurity elements originally dissolved in aluminum to precipitate and deposit at the bottom of the melt, thereby improving the conductivity of aluminum conductors. The maximum solid solubility of Ti in the Al-Ti binary alloy is 0.15%, which decreases to 0.01% when B is present. B also reduces the solid solubility of Fe in aluminum alloys and has a modifying effect on the coarse lamellar FeAl 3 phase, causing it to be distributed discontinuously in the form of dots or strips within and at the grain boundaries of the aluminum matrix. Research shows that when alloying elements exist in the precipitated state rather than in the solid solution state, the resistivity can be significantly reduced. Among them, Fe and Si reduce by two orders of magnitude, while Ti, Mn, Cr, and V reduce by one order of magnitude.

[0096] In addition, the chemical interaction between alloying elements affects the solid solubility of alloying elements and impurity elements in aluminum, as well as the existence, size, and dispersion degree of the second phase, thus affecting the conductivity and heat resistance of aluminum. According to the mechanism of interaction with aluminum, alloying elements can be divided into two categories. One category can indirectly affect the evolution of other phases, change the distribution and morphology of the precipitated phases in the alloy, and thus improve the microstructure and properties of the alloy. The other category can directly react with Al to form the second phase, significantly enhancing the properties of the alloy. Keith et al. analyzed the elements in the periodic table and found that elements such as Sc, Zr, Ti, Hf, Er, Tm, Yb, and Lu have a low equilibrium solid solubility in the aluminum matrix. They can be microalloyed by conventional casting methods. The second phase has a large precipitation driving force and can react with Al to form a second phase that is finely dispersed and coherent with the matrix. Since the diffusion rate of these elements in the aluminum matrix is low, they can effectively pin dislocations, subgrain boundaries, and grain boundaries at higher temperatures, making the alloy have high strength, good heat resistance, and good conductivity at the same time.

[0097] Sc belongs to the transition elements and also belongs to the rare earth elements. It not only has the function of purifying the melt and improving the casting structure of rare earth elements, but also has the function of refining grains and inhibiting recrystallization of transition elements. It is the most effective microalloying element for aluminum alloys; the limiting solid solubility of Sc in aluminum is 0.32 wt.%, and for aluminum alloys added with trace Sc, secondary Al 3 Sc particles will precipitate. Al 3 The structure and lattice constant of Sc and α-Al are almost exactly the same. Therefore, when the aluminum solid solution containing Sc decomposes, it does not go through the metastable phase stage but generates nano-scale Al with extremely high dispersion 3 Sc spherical particles, which have a significant strengthening effect on the matrix. Al 3 The Sc phase has an L1 2 type structure and has strong thermal stability, and still maintains a coherent relationship with the matrix at high temperatures. However, Sc is a strategic element with a high price and is difficult to be widely used in the industrial field.

[0098] Since Zr has a similar effect to Sc in inhibiting recrystallization, improving the heat resistance and strength of aluminum alloys, and its price is lower than that of Sc, Zr is an indispensable key element for heat-resistant aluminum alloys. The microalloying effect of Zr is closely related to its existence state. When trace Zr is added to aluminum and aluminum alloys, there are 4 different existence forms, namely, dissolved in α-Al and forming Al 3 Zr primary phase, Al 3 Zr (Ll 2 ) metastable phase, Al 3 Zr (D03) equilibrium phase. The maximum solubility of Zr element in α-Al is 0.28 wt.% (660.5 °C). As the temperature decreases, the solubility of Zr element decreases rapidly and drops to about 0.05% at 400 °C. Zr can react with Al in a peritectic reaction: Liquid + Al 3 Zr → α-Al, the peritectic point composition is 0.11% Zr, and the temperature is 660.5 °C. When the Zr content is greater than 0.11%, Al 3 Zr primary phase will be generated. If the addition amount of Zr is too high, or the control is improper during the melting and casting process, Zr is prone to segregation and form coarse primary Al 3 Zr phase, which will have an adverse effect on the alloy properties.

[0099] The metastable Al 3 Zr is an extremely effective dispersion strengthening body and recrystallization inhibition particle. Since the diffusion rate of Zr in α-Al is relatively low, only 1.2×10 -20 m 2 / s at 400 °C, therefore, the metastable Al 3The Zr phase has good thermal stability. If the cooling rate is too slow or the holding time at a relatively high temperature is too long, the Al 3 Zr equilibrium phase will be formed. Its size is relatively large, and it is semi-coherent or incoherent with the matrix, which will lead to a decrease in the precipitation driving force during the subsequent aging process of the alloy. By adopting appropriate hot deformation and heat treatment processes, the Zr dissolved in α-Al can be precipitated, forming fine and dispersedly distributed metastable Al 2 structures of Al 3 Zr phase. The particle spacing λ < 1 μm, and it maintains a good coherent relationship with the matrix. It has a strong pinning effect on dislocations and grain boundaries, can hinder the process of dislocations rearranging into sub-grain boundaries and developing into large-angle grain boundaries, and can also hinder the migration of large-angle grain boundaries and inhibit the growth of recrystallization nuclei, thereby hindering the recrystallization process. Experiments have proved that it can increase the recrystallization temperature of aluminum by more than 100 °C.

[0100] In aluminum alloys, rare earth elements have purification, modification, refinement, and microalloying effects. The solubility of rare earth elements in aluminum is extremely low, and they tend to form stable high-melting-point compounds with other elements and impurities. Adding a certain amount of rare earth to combine with impurities such as Fe and Si will reduce the solid solution amount of impurity elements in the aluminum matrix and change the shape and distribution of inclusions, thereby reducing the harmful effects of impurity elements on conductivity. Rare earth compounds exist as tiny solid particles in the aluminum alloy melt, providing a large number of non-spontaneous nuclei for the crystallization of aluminum alloys. Their segregation at the front of the crystallization interface can not only hinder grain growth but also promote the branching of dendrites, increasing the ratio of nucleation rate to grain growth rate, thereby significantly refining the as-cast grains and dendrite structure and effectively improving the alloy strength. Taking the rare earth element Er as an example, in the initial stage of solidification, Er directly forms Al 3 Er phase in the aluminum liquid. This phase is the same as Al 3 Zr and Al 3 Sc and other belong to the L1 2 type structure, similar to the structure of the aluminum matrix, and its lattice parameters are also close to Al. Al 3 Er particles can become the nucleation cores of α-Al or further grow into coarse primary Al 3 Er phase, hindering grain growth and promoting the formation of dendrite structure. Er can form a certain supersaturated solid solution in Al, and after appropriate heat treatment, it will precipitate from the matrix to form secondary Al 3 Er particles. These particles are fine and dispersed, can effectively pin dislocations and sub-grain boundaries, and improve the strength and recrystallization temperature of the alloy. However, its thermal stability is not as good as Al 3 Zr.

[0101] With the continuous development of technology, the comprehensive performance requirements for heat-resistant aluminum alloy conductor materials are gradually increasing. By forming some second-phase particles with complex structures through the interaction of trace alloying elements, the microstructure of aluminum alloy can be effectively improved, thereby greatly enhancing the comprehensive performance of aluminum alloy, especially the electrical conductivity and heat resistance. Usually, a small amount of rare earth elements is added to the Al-Zr alloy. The trace RE combines with impurities such as Fe and Si in the aluminum matrix to form stable intermetallic compounds, which can effectively inhibit the harmful effects of Fe and Si elements on the electrical conductivity. Through the combined microalloying of B, Zr, and RE and supplemented with appropriate preparation processes, State Grid Smart Grid Research Institute Co., Ltd. has mastered several composition systems of high-conductivity and heat-resistant aluminum alloy conductor materials. Taking the Al-Zr-Er alloy system as an example, Zr and Er have similar physical and chemical properties. When they are compounded and added to the aluminum alloy, they will replace each other to form L1 2 structure of Al 3 (Zr x Er 1-x ) composite particles. After holding at an appropriate temperature for a period of time, Er with a faster diffusion rate precipitates first, and Al3Er particles are precipitated, promoting the enrichment of Zr with a slower diffusion rate on the outer layer of the Al 3 Er particles, forming a large number of dispersed, coherent with the matrix, L1 2 structured, nanoscale spherical Al 3 (Er,Zr) composite particles, while inhibiting the coarsening of Al 3 Er particles. On the one hand, it reduces the solid solution degree of Zr and Er in the aluminum matrix and improves the electrical conductivity of the aluminum wire. On the other hand, the dispersed, coherent with the crystal, nanoscale Al3(Er,Zr) composite particles have high thermal stability, pinning effects on dislocations, subgrains, and grain boundaries, with obvious strengthening effects and inhibition of recrystallization effects, which can effectively improve the strength and heat resistance of the aluminum conductor and can significantly improve the heat resistance of the aluminum alloy while maintaining a high electrical conductivity.

[0102] The LSW model is a classic model describing the diffusion-controlled growth of precipitation phases in binary alloys. Kuehmanm and Voorhees (KV) extended this model to obtain the KV model, enabling it to be better applied to ternary alloy systems. Figure 10 It shows the relationship between the precipitation phase size and time when Al-0.15Er and Al-0.15Zr-0.2Er alloys are annealed at 400 °C.

[0103] Although Zr can significantly improve the heat resistance of aluminum wires, Zr also has a great impact on the conductivity of aluminum alloys, and with the increase of Zr content, more serious dendritic segregation will occur. Therefore, the addition amount of Zr needs to be carefully controlled. When the Zr content does not exceed 0.1 wt.%, the supersaturation degree of Zr in the matrix decreases, and the precipitation driving force is insufficient. Moreover, due to the low diffusion rate of Zr, it needs to be kept at a high temperature for a long time to effectively precipitate, and prolonging the production cycle is not conducive to energy conservation and consumption reduction. When Zr is compounded with other alloy elements with faster diffusion rates, such as Er or Sc, two-stage aging treatment can be carried out, first pre-aging at a lower temperature for a short time, which can not only make the elements with fast diffusion rates precipitate and disperse phases at a lower temperature, but also avoid the growth of precipitation phases, and then carry out the second-stage aging at a higher temperature. Since the fine dispersed phases precipitated during the first-stage aging can provide nucleation sites for the precipitation of Zr, promoting the precipitation of Zr and forming core-shell structure particles with good high-temperature stability (such as Figure 11 as shown), it can not only improve the conductivity but also have good heat resistance and strength.

[0104] Fiber Bragg grating sensors utilize the periodic refractive index changes formed in the fiber core. When the period satisfies the Bragg condition, the light reflected by each periodic reflection surface accumulates step by step, and finally a reflection peak will be formed in the reverse direction, and the central wavelength is determined by the grating parameters. This process can be expressed by Equation (7):

[0105] λ B = 2n eff Λ (7)

[0106] where λ B is the Fiber Bragg wavelength, n eff is the refractive index of the fiber core for the central wavelength in free space, and Λ is the grating period;

[0107] The resonant peak of the FBG, that is, the central wavelength reflected by the fiber grating, depends on the grating period and the size of the effective refractive index of the core. Various physical quantities such as external vibration, strain, and temperature will change the effective refractive index and grating period of the fiber grating. The influence of physical quantities such as strain and temperature on the Bragg wavelength of the fiber grating can be expressed by the following Equation (8):

[0108]

[0109] The change of the external physical quantity acting on the fiber grating will cause the offset of the grating wavelength. The backend signal demodulation device can obtain the change magnitude of the external physical quantity by detecting the offset of the fiber grating wavelength, so as to achieve the purpose of detecting the external physical quantity; its principle is as Figure 12 shown;

[0110] The convolutional neural network CNN mainly consists of three parts, namely the convolutional layer, the pooling layer, and the fully connected layer. Each layer can generate multiple feature maps. The convolutional layer is the core of the neural network. The convolutional layer is composed of multiple groups of convolutional kernels. The convolutional kernels perform convolutional operations on the imported data, and the output result undergoes a non-linear operation through an activation function to obtain the feature array of this operation; The specific recognition process is as Figure 13 shown;

[0111] The pooling layer divides the feature array obtained by the convolutional layer into multiple independent pooling blocks, and performs pooling operations within the pooling blocks. There are two commonly used pooling methods. One is average pooling, which calculates the average value within the pooling block. The other is max pooling, which extracts the maximum value within the pooling block. Both pooling methods have their advantages and disadvantages, and can be appropriately selected according to the type of the array. As Figure 14 shown are the calculation results of different pooling methods on the same feature array.

[0112] The pooling layer can significantly reduce the size of the feature array and can also eliminate the offset of the signal. The pooling layer is also called the downsampling layer. After the input data undergoes multiple convolutions and poolings, the output is a multi-dimensional array, which contains the feature information of the original data. Finally, classification is achieved through the activation of the softmax function in the fully connected layer.

[0113] The convolutional neural network realizes the feature extraction of data through convolution and pooling, and realizes the classification of data through the fully connected layer. Moreover, compared with traditional neural networks, the convolutional neural network has the advantages of fewer training parameters, shorter training time, and no need to manually extract feature data, and is very suitable as an artificial intelligence algorithm for disturbance type recognition.

[0114] The research team has years of research foundation in transmission wires and fiber optic sensing technology, overcome a series of major core technical problems in transmission line monitoring technology, and has the ability to independently develop transmission wire and fiber optic sensing monitoring systems. Undertake research projects such as "Research on Aluminum Stress Test and Simulation of Long-Span Conductors in Power Grid Engineering Branch", "Analysis and Test Research on Aluminum Stress of the Second Long-Span Conductor in Jiangyin (Conductor Manufacturing and Parameter Determination Based on Optical Sensing Technology)", "Research on New Type Intelligent Cable Stranding Technology", "Type Test and Indoor Simulation Test of Intelligent Cable", "Fabrication and Key Performance Testing of Optical Path Amplifier, Integrated Signal Processing and Sampling System Based on EDFA Amplification and Fiber Optic Fingerprint Prototype System", etc. on research and development of transmission wire, fiber optic sensing monitoring and evaluation technology, and has good research foundation and practical experience in overhead transmission wires, line condition monitoring, fault analysis, etc. Provide hardware guarantee and reliable technical support for the specific research of the project. Based on the online monitoring status of actual operating transmission lines as data, combined with historical data of each unit in the operation and maintenance and experimental processes for comparison, integrate the research results with the operation and maintenance work to ensure the effective implementation of the research results. This project focuses on the preparation and monitoring of high-elasticity transmission wires. The research team has several successful application cases closely related to this project, accumulating a certain foundation for the research of this project.

[0115] 1. Key Points and Difficulties in Project Research

[0116] (1) Key points in project research: Solve the basic problems in the composition design and preparation process of high-conductivity heat-resistant aluminum alloy wires, reveal the influence laws of different element composite microalloying on the microstructure, electrical conductivity, and heat resistance of aluminum alloy wires, explore the action mechanism of microalloying and heat treatment on the precipitation behavior of aluminum alloys, and master the mutual influence mechanism and comprehensive optimization method among the electrical conductivity, heat resistance, and strength of aluminum alloy wires. Systematically study the microstructure evolution laws during melting, casting, rolling, drawing, and heat treatment processes, calculate, simulate, and optimize the production process based on experimental and production big data, and comprehensively utilize multi-element composite microalloying and production process control for fine microstructure regulation to obtain conductor materials and key preparation processes that take into account good heat resistance, electrical conductivity, and strength, design a reasonable industrialization procedure, and achieve low-cost mass production. On the basis of batch preparation of high-conductivity heat-resistant wires, supporting corresponding sensors and monitoring equipment, overcome signal interference caused by complex service conditions such as environment and climate, improve the signal collection and processing ability of sensors, break through the transmission wire fault diagnosis technology based on neural network convolution, and achieve the goal of online real-time intelligent monitoring.

[0117] (2) Difficulties in project research: ① Difficulty: Since conductivity, strength, and heat resistance are mutually restrictive, it is one of the difficulties of this project to increase conductivity while ensuring strength and heat resistance. Proposed solution: Through the study of the influence mechanism of the types and contents of composite microalloying elements on the microstructure and properties of materials, and by using the interaction of multiple alloy elements and multi-stage heat treatment control processes, the heat-resistant elements are transformed from the solid solution phase to the precipitation phase, obtaining a relatively stable microstructure of the material, so as to synergistically optimize strength, heat resistance, and conductivity. ② Difficulty: The acquisition and restoration of signals such as ice coating and sag of transmission lines are complex processes. Due to the large interference of the background noise of the transmission line on the effective signal, and the influence of different temperatures and humidities on the joint between the optical fiber and the environment, the signal transmission will be distorted, and it is impossible to effectively and accurately identify the operation faults of transmission lines and predict and analyze them under traditional artificial intelligence algorithms. Proposed solution: With the help of deep networks, mine the fault types and corresponding characteristic parameters of transmission lines, and use big data correlation analysis to realize signal processing technologies such as multi-source data fusion, improve the understanding of on-site monitoring data, and explore the inherent value of multi-source sensing big data, providing an objective scientific basis for the safe operation, efficient management, and intelligent decision-making of transmission lines.

[0118] (3) Innovation points of project research: Aiming at the requirements of high conductivity of 62% IACS, heat resistance (long-term heat resistance of 150 °C), and strength, based on thermodynamic and diffusion kinetics calculations and analysis of the microstructure evolution during the wire preparation process, the matching design of alloy composition, microstructure, and preparation process is carried out, broadening the composition range and production process window. By using the synergistic effect of multiple mechanisms and multiple microstructures, a microstructure with good thermal stability is obtained, realizing the synergistic improvement of high conductivity, strength, and heat resistance, and meeting the requirements of low cost, energy conservation, and environmental protection. Break through the key technologies for the preparation of high-conductivity heat-resistant aluminum alloy single wires and high-conductivity heat-resistant aluminum alloy wires, develop high-conductivity heat-resistant high-elastic intelligent transmission wires, with the ability to monitor multiple parameters such as fiber-optic micro-meteorology, ice coating, sag, and current-carrying capacity in real time, improve the monitoring accuracy of indicators such as ice coating thickness, sag distance, and current-carrying capacity, effectively reduce the operation and maintenance costs of ultra-long-distance transmission lines, realize synchronous monitoring and data sharing among the source, grid, and load, improve the level of independent early warning and intelligent interconnection and interaction, enhance the level of operation and maintenance intelligence, and ensure the safe operation of transmission lines.

[0119] (1) Research on the mechanism of microalloying of high-performance aluminum alloys, characteristic microstructural configurations, and performance regulation theory: Based on microalloying theory, thermodynamic and kinetic calculations, and analysis of the microstructure evolution during the preparation process, explore the interaction mechanism of the combined addition of multiple microalloying elements such as Zr, B, and RE. Study the corresponding relationship between the heat treatment process of aluminum alloys and the microstructure evolution. Study the influence laws of the second-phase configuration and the synergistic effect of multiphase structures on the electrical conductivity, heat resistance, and strength of electrical aluminum alloys. Master the influence of microalloying and preparation process regulation on the microstructure and properties, so as to synergistically optimize the strength, electrical conductivity, and heat resistance temperature.

[0120] (2) Research on the materials and structure technology of distributed optical fibers embedded in wires: Facing the demand for icing monitoring of long-distance transmission lines, study new low-noise optical amplification technologies, and combine with the high-order low-noise fiber random laser remote pumping erbium-doped fiber hybrid amplification (Low-noise ROPA, abbreviated as L-ROPA) scheme to realize a new long-distance low-noise fiber sensing monitoring system based on high-elastic wires. Combining the first-invented L-ROPA technology with FBG is expected to overcome the optical power loss in long-distance transmission and lay the foundation for realizing an ultra-long-distance, low-noise fiber sensing system. This project will establish a theoretical model for the relative intensity noise (RIN) transfer from the pump light to the signal light in low-noise fiber random laser amplification, study the RIN transfer mechanism and characteristics from high-order fiber random laser pumping to the signal light in the sensing system, complete the optimal design of the L-ROPA new optical amplification system, and realize a fiber sensing system with high signal-to-noise ratio and high sensitivity.

[0121] (3) Development and application of high-elastic intelligent transmission wires: Through the combination of the microstructure evolution of heat-resistant aluminum alloy materials, material property determination, and experimental verification, optimize the design of alloy compositions, study the evolution of the as-cast structure, deformed structure, and heat-treated structure of alloys, and carry out the matching design of aluminum alloy properties and ingredient formulations to obtain ingredient formulations and batch preparation processes that meet the target properties. Carry out research on the pre-forming + post-forming control system with automatic control to replace the manual adjustment of the stranding machine, eliminate the single-wire stress to the greatest extent, make the high-elastic transmission wire very compliant after stranding, and at the same time effectively solve the problems of the finished stranded wire being loose and snake-shaped, avoid strand separation after truncation, and reduce the influence of human operation and the accuracy of production equipment itself on the product performance during the stranding process of the wire.

[0122] (4) Research and Development of Multi-parameter Intelligent Monitoring System: Aiming at the problems that various sensing data of transmission lines are independent of each other, and it is difficult to predict icing and assess risks under the influence of complex geographical and changing climate environments, research on intelligent signal processing algorithms for the fusion of multi-source heterogeneous data such as line structure parameters, micro-meteorology, and micro-topography is carried out to achieve high-reliability intelligent assessment and safety warning of icing states based on real-time fusion of multi-source data. Through the deep learning network studied by this project team, establish the mapping relationship of conductor icing state levels, and conduct analysis and feature expression of line icing characteristics; based on the big data correlation analysis method, mine the correlation rules and frequent items between multi-source data and icing state characteristics; finally, realize the high-reliability real-time intelligent assessment and safety warning algorithm and system software for icing states based on multi-source data fusion.

[0123] Theoretical research steps, locations of on-site tests and test plans: Based on simulation calculations, in this project, alloy composition design is carried out by optimizing alloying elements, and through the precise control of heat treatment processes with the compound addition of alloying elements, a microstructure that meets the performance indicators is obtained. In the stage of element optimization, rare earth elements are selected according to the calculation results, and the minimum content of rare earth elements is determined. After the alloy is borided, rare earth elements are added. On the one hand, rare earth elements react with impurity elements in the aluminum liquid to form a second phase, which can reduce the lattice distortion caused by impurity atoms and improve the conductivity. At the same time, the formed second phase is generally a high-temperature resistant phase, which can improve heat resistance and strength; on the other hand, the compound addition of rare earth elements and other elements (such as Zr element) can form precipitation phases with good high-temperature stability through the process control of aging treatment, so as to achieve the coordinated improvement of conductivity, strength and heat resistance. The specific technical route of the research is as follows Figure 16 :

[0124] The development of 62% IACS high-conductivity heat-resistant aluminum alloy conductor materials and the trial production of conductors in this invention are mainly carried out by State Grid Smart Grid Research Institute Co., Ltd., and State Grid Liaoning Electric Power Co., Ltd. conducts collaborative research. The mass production of conductors will be entrusted to conductor manufacturers; after obtaining the performance test reports of the conductors and the supporting sensing intelligent system, according to the construction and application requirements of transmission lines, carry out research on engineering application technologies, and conduct on-line operation on the transmission lines of State Grid Liaoning Electric Power Co., Ltd. to analyze its operation effects and technical economy.

[0125] The beneficial effects of this invention:

[0126] Based on the research on the conductivity mechanism and heat resistance mechanism of microalloyed aluminum alloys, through the correlation research of microalloying, process, microstructure, and properties, appropriate microalloying and optimized casting-rolling and heat treatment processes are adopted to obtain a microstructure with good high-temperature stability. The conductivity of the heat-resistant aluminum alloy is increased to 62% IACS, and the long-term operating temperature is ≥150 °C, thereby reducing the power loss of transmission lines. On this basis, it is planned to use advanced optical fiber sensing technology to achieve all-weather real-time online monitoring of the icing state of ultra-long-distance and low-cost transmission lines. On the other hand, by integrating rich multi-source information such as meteorological data, geographic information, and line parameters, the correlation between icing levels and multi-source data is mined based on big data analysis methods, so as to construct an accurate icing index evaluation system and assessment system, and realize a highly reliable line galloping disaster warning system. In short, this project organically combines high-conductivity heat-resistant wires with advanced optical fiber sensing and traditional meteorological technologies to form a low-cost wide-area multi-sensor technology that combines points, lines, and surfaces. By using deep networks to mine galloping characteristic parameters and using big data correlation analysis to achieve advanced signal processing technologies such as multi-source spatio-temporal data fusion, the understanding of on-site monitoring data is improved, and the self-value of multi-sensor big data is mined, providing an objective scientific basis for the safe operation, efficient management, and intelligent decision-making of ultra / extra-high voltage transmission lines.

[0127] The implementation of the project will break through the key technologies for the preparation of high-conductivity heat-resistant aluminum alloy single wires and high-conductivity heat-resistant aluminum alloy wires, develop high-conductivity heat-resistant and high-elastic intelligent transmission wires, effectively reduce the operation and maintenance costs of ultra-long-distance transmission lines, achieve synchronous monitoring and data sharing among the power source, grid, and load, improve the level of independent early warning and intelligent interconnection and interaction, enhance the level of intelligent operation and maintenance, and support the large-scale safe grid connection and full consumption of new energy. The results can be transformed in cable manufacturing-related manufacturers inside and outside the system, which is of great significance for promoting the development of overhead conductors for transmission lines in China, improving the manufacturing level of the cable industry, enhancing the technical level of transmission lines of the State Grid Corporation, and reducing the power loss of transmission lines.

[0128] In order to better achieve the capacity expansion and transformation of urban power grids at low cost, aiming at problems such as long-distance power transmission and large fluctuations in new energy access, effectively utilize the original tower bases and line corridors, conduct intelligent monitoring, and grasp the status of transmission lines in real time. Developing high-conductivity, heat-resistant, and highly elastic smart aluminum alloy conductors to achieve high-efficiency intelligent power transmission has become one of the trends in conductor research and development. However, the existing heat-resistant aluminum alloy conductors have poor comprehensive performance (the highest conductivity is 61% IACS (20 °C), and the long-term heat-resistant temperature is 150 °C), which increases line losses, causes a large amount of electrical energy waste, and the reliability of the smart monitoring system is poor. Developing heat-resistant aluminum alloy conductor materials with a conductivity of 62% IACS and their conductors, and even aluminum alloy conductor materials with higher conductivity and heat resistance, and supporting corresponding smart real-time monitoring systems are the current trends in the development of heat-resistant aluminum alloy conductor technologies. Therefore, this invention focuses on overcoming the key technical problems of high-conductivity, heat-resistant, and highly elastic smart power transmission aluminum alloy conductor materials and their conductors with a long-term operating temperature of 150 °C and a conductivity of 62% IACS. The successful development, popularization, and application of the conductor materials and conductors will greatly promote the technological development of heat-resistant aluminum alloy conductors in China, the development of power transmission technology level, and the improvement of the intelligent operation and maintenance level, supporting the large-scale and safe grid connection and full consumption of new energy; at the same time, the successful development and grid connection operation of 62% IACS high-conductivity, heat-resistant, and highly elastic smart power transmission aluminum alloy conductors will surely play a demonstrative role in the development of new products by Chinese cable enterprises and lead the technological progress of cable enterprises.

[0129] The 62% IACS high-conductivity, heat-resistant, and highly elastic smart power transmission aluminum alloy conductors developed in this invention will be applied in engineering on the transmission lines under the jurisdiction of the State Grid Corporation. After systematically summarizing the key technologies such as construction, application, and maintenance of the 62% IACS high-conductivity, heat-resistant, and highly elastic smart power transmission aluminum alloy conductors through grid connection operation, corresponding technical specifications will be formulated, and then it will be promoted and applied to other transmission lines. The achievements of this invention will lay a foundation for global energy interconnection, large-scale renewable energy access, and the construction of smart grids. It is necessary to make full use of the two markets and two resources at home and abroad, promote cross-border grid interconnection, and the promotion channels are as follows:

[0130] (1) Newly built long-distance and large-capacity power transmission lines;

[0131] (2) Long-span and grounding electrode lines;

[0132] (2) Smart power transmission, intelligent monitoring of lines;

[0133] (3) Capacity expansion and transformation lines under the condition of limited original iron towers and line corridors;

[0134] (4) Transmission projects for new energy sending, AC in and out lines of converter stations, and short-term overloaded transmission lines.

[0135] The 62% IACS high-conductivity heat-resistant aluminum alloy wire can operate stably at a higher operating temperature, allowing for a greater current-carrying capacity. It can not only improve the safety and stability of line operation, but also reduce construction and maintenance costs, extend the lifespan of transmission lines, save line corridor resources, and reduce carbon emissions, presenting significant economic and social benefits.

[0136] (1) The implementation of the present invention can effectively enhance the line's ultimate transmission capacity, extend the line's lifespan, and reduce line construction and maintenance costs. Replacing ordinary electrical aluminum wire with heat-resistant wire of 62% IACS, calculated based on 50,000 kilometers, can reduce transmission line losses by approximately 1.07×10 9 kWh per year. Calculated at an electricity price of 0.5 yuan / kWh, it can save 535 million yuan in electricity bills, generating significant economic benefits. During the operation period, it can quickly compensate for the incremental one-time investment. In the urban network line expansion and renovation project, especially in areas with narrow line corridors, replacing the in-service wire with a high-conductivity heat-resistant wire of similar cross-sectional specifications can increase the transmission capacity by 40% - 60%. It can meet the requirements of strength and the safety of the wire's ground clearance without replacing the iron tower, saving a large amount of project investment.

[0137] (2) The present invention significantly improves the power grid's resource allocation ability, safety and stability level, and the interactivity between the power grid, power sources, and users. It can improve energy utilization efficiency, effectively solve the contradiction between China's energy resources and the uneven distribution of electricity loads, indirectly reduce the consumption of other energy sources and carbon emissions, meet the overall requirements of building a resource-conserving and environment-friendly society in China, and conform to the trend of sustainable development.

[0138] (3) Effectively reduce the operation and maintenance costs of ultra-long-distance transmission lines, achieve synchronous monitoring and data sharing among the source, grid, and load, improve the level of independent early warning and intelligent interconnection and interaction, enhance the intelligent level of operation and maintenance, and support the large-scale and safe grid connection and full consumption of new energy.

[0139] (4) Relying on the industrialization of the State Grid Corporation and related enterprises, the present invention can enhance the scientific and technological achievement transformation ability of the aluminum wire manufacturing industry, promote the improvement of China's wire and cable manufacturing technology level and product competitiveness, promote the formation of a high-performance aluminum alloy wire industrial cluster, and at the same time drive the transformation and upgrading of the aluminum industry. Description of the Drawings

[0140] Figure 1 It is a schematic diagram of the composition scheme of a new type of low-noise optical amplification system;

[0141] Figure 2 It is the simulation curve of the Rayleigh scattering signal optical power distribution of the new type of low-noise optical amplification system at a sensing distance of 30 km, that is, the optical power distribution along the optical fiber;

[0142] Figure 3It is the simulation curve of the Rayleigh scattering signal optical power distribution, i.e., the optical power distribution along the optical fiber, of the new low-noise optical amplification system at a sensing distance of 50 km.

[0143] Figure 4 It is the extraction of the time-series structural features of line icing based on 1-D CNN.

[0144] Figure 5 It is the extraction of the spatial distribution features of line icing based on Bi-LSTM.

[0145] Figure 6 It is the schematic diagram of the method for mining the association rules between icing levels and multi-sensor data based on association analysis.

[0146] Figure 7 It is an example of the frequent item set mining process containing only one target event item.

[0147] Figure 8 It is the main process of the association rule classifier discrimination.

[0148] Figure 9 It is the B / S architecture design of the system software.

[0149] Figure 10 It is the relationship between the size and time of the precipitated phase during heat preservation at 400 °C.

[0150] Figure 11 It is the formation mechanism of the second-phase particles by the composite addition of Zr with Er and Sc.

[0151] Figure 12 It is the modulation principle of the fiber grating sensor by external physical quantities.

[0152] Figure 13 It is the schematic diagram of the convolutional neural network (CNN) structure.

[0153] Figure 14 It is the schematic diagram of the pooling process (average pooling).

[0154] Figure 15 It is the schematic diagram of the pooling process (max pooling).

[0155] Figure 16 It is the key material development and application technology roadmap for high-elasticity intelligent power transmission conductors. Specific implementation manners

[0156] The present invention will be further described below in conjunction with the embodiments and the accompanying drawings of the specification, but is not limited thereto.

[0157] Embodiment 1

[0158] The present invention provides a signal processing system for multi-source spatio-temporal data fusion based on a highly elastic transmission wire. The related technical principles are described as follows: When a metal is subjected to an external electric field, the carriers move directionally in the lattice field (also known as the lattice Coulomb potential field) formed by the periodically arranged ion cores to form an electric current. Alloying elements, impurities, crystal defects, etc. will all cause periodic damage to the lattice field. Those abnormal ion cores or lattice atoms that have damaged the periodicity of the lattice field will collide with or impede the directionally moving carriers to generate resistance. Therefore, the conductivity of an alloy highly depends on the alloy composition and microstructure. Solute atoms, vacancies, precipitates, dislocations, etc. will all cause lattice distortion, resulting in scattering of the carriers during movement. With the addition of alloying elements and the increase in the addition amount, the resistivity of aluminum alloy will increase significantly. Alloying elements and their existing states have different effects on the conductivity of aluminum alloy. When impurities or alloying elements are dissolved in the aluminum matrix, the conductivity of aluminum alloy will be significantly reduced. After they precipitate from the solid solution, the effect on conductivity is only one fraction to one fortieth of that in the solid solution state. Therefore, it is very necessary to perform heat treatment on the conductor material.

[0159] To increase the heat resistance temperature of an alloy, it is necessary to add micro-alloying elements and make the precipitated second phase disperse to inhibit interface migration. However, these elements cause great damage to the conductivity, especially when they exist in the solid solution state. In 1949, American scholars found through research that adding a small amount of zirconium element to aluminum can improve the heat resistance of aluminum. Japan took the lead in developing Al-Zr series heat-resistant aluminum alloy wires, which can improve the heat resistance of the material without reducing the tensile strength and conductivity of the wire. Heat-resistant aluminum alloys mostly use micro-zirconium aluminum alloys. Zr atoms interact with dislocations and grain boundaries, segregate at the grain boundaries, and impede the migration of large-angle grain boundaries, thus impeding the recovery and recrystallization of aluminum alloy when the temperature rises, and thus increasing the heat resistance temperature of aluminum alloy. However, due to the relatively large difference in atomic radius between Zr and Al, large lattice distortion will occur when Zr is dissolved in aluminum, and the diffusion rate of Zr is relatively low. Supersaturated Zr is difficult to precipitate from the aluminum matrix. Therefore, as the zirconium content in aluminum increases, the heat resistance is greatly improved, but the conductivity is significantly reduced. To balance the conductivity, the zirconium content of traditional heat-resistant aluminum alloys is generally not more than 0.1%, and the heat resistance level is relatively low.

[0160] To solve the contradiction among conductivity, heat resistance temperature, and strength, it is necessary to form second-phase particles with highly dispersed distribution, thermodynamic stability, and coherent or semi-coherent with the matrix through methods such as reasonable composition design, appropriate preparation process, and precise heat treatment control, so as to inhibit interface migration and achieve the goal of synergistically improving electrical conductivity, heat resistance, and strength.

[0161] Intelligent power transmission sensing and monitoring mainly uses fiber Bragg grating sensing technology and convolutional neural network methods. Among them, the fiber Bragg grating sensor is an important sensing component in the fiber optic sensing system, which is made based on the sensing mechanism of fiber grating elements. The sensor encodes the wavelength, overcoming the weakness of intensity modulation sensors that must compensate for the losses of fiber connectors and couplers as well as the fluctuations in the output power of light sources. Fiber gratings are sensitive elements with excellent performance. Their peak wavelengths change with physical quantities such as temperature and stress. By designing sensitive structures for the conversion of non-optical physical quantities, optical measurement of non-optical quantities can also be achieved. External parameters such as strain, pressure, temperature, wind direction and speed, and temperature-induced micro-vibrations can be sensed directly by demodulating the wavelength signals of fiber gratings. Since fiber gratings use wavelength encoding methods, multiplexing technologies such as wavelength division, time division, and space division can be used. Fiber Bragg gratings are combined in series or parallel to form a sensing network, enabling quasi-distributed networked measurement of physical quantities. This is a major advantage of fiber grating sensing systems.

[0162] In recent years, machine learning algorithms have been widely used in the pattern recognition of distributed fiber optic sensing signals, and they have also developed relatively well in aspects such as data feature extraction, dataset construction, and pattern recognition algorithms. However, in the construction of datasets, it is necessary to manually extract data features, and with the change of the environment, the data features will also change accordingly, which often requires a large amount of human resources and will greatly reduce the efficiency of machine learning. Convolutional neural networks (CNNs) perfectly avoid the disadvantages of manual data feature extraction. They can automatically extract data features and perform classification and recognition. Two-dimensional convolutional neural networks have played a huge role in the field of image recognition. Images can be directly used as input signals without any manual feature extraction. The neural network can automatically extract the features of the images. One-dimensional convolutional neural networks can achieve the recognition of one-dimensional sequence signal types such as vibration and sound without setting specific feature engineering. The entire neural network system has good robustness and high computing efficiency.

[0163] Adopt the technology of combining fiber random laser technology and remote-pumped erbium-doped fiber hybrid amplification to develop an ultra-long-distance, low-noise fiber optic sensing and monitoring system. Integrate deep learning algorithms, develop transmission line icing monitoring algorithms, realize synchronous monitoring and data sharing among the source, grid, and load, improve the level of autonomous early warning and intelligent interconnection and interaction, enhance the level of operation and maintenance intelligence, and support the large-scale safe grid connection and full consumption of new energy.

[0164] In this technical solution, for the signal processing system based on the multi-source spatio-temporal data fusion of high-elastic transmission wires, the key technology lies in: it includes the high-elastic intelligent transmission wire body and the low-noise optical amplification system, which uses the multi-source data real-time fusion and intelligent processing method; together they constitute the necessary auxiliary components to support the high-elastic intelligent transmission wire; among them:

[0165] The high-elastic intelligent transmission wire body is the domestic steel-core 58% conductivity heat-resistant aluminum alloy stranded wire NRLH58GJ-240 / 30;

[0166] In the signal processing system based on the multi-source spatio-temporal data fusion of high-elastic intelligent transmission wires, the high-elastic intelligent transmission wire body combines fiber optic sensing technology and meteorological technology to form a low-cost wide-area multi-source sensing technology that combines points, lines, and surfaces. By using deep network to mine the wind dance characteristic parameters and using big data correlation analysis to realize the signal processing technology of multi-source spatio-temporal data fusion, it improves the understanding of on-site monitoring data, excavates the self-value of multi-source sensing big data, and provides an objective scientific basis and theoretical foundation for the safe operation, efficient management, and intelligent decision-making of ultra / extra-high voltage transmission lines;

[0167] I. The low-noise optical amplification system uses the following new fiber optic sensing technology:

[0168] (I) New distributed amplification technology

[0169] First, a theoretical model of signal optical power distribution based on high-order low-noise fiber random laser amplification is established

[0170] For the new low-noise optical amplification system, the power steady-state equation model shown in Equation (1) is used to conduct theoretical analysis on the high-order fiber random laser amplification part in this amplification system, and the power distribution of each frequency component along the fiber length is calculated using the following model:

[0171]

[0172] Among them, the subscripts '0', '1', '2' correspond to the pump, first-order, and second-order Stokes lights respectively; the superscripts '+' and '-' represent the forward and backward waves; P0,1,2 represent the optical power; z represents the optical transmission direction coordinate; f0,1,2 represent the optical frequencies; Γ1,2 represent the photon numbers; where Δf1,2 = 0.25 THz represents the radiation bandwidth; T = 298 K represents the absolute temperature, KB represents the Boltzmann constant, h represents the Planck constant, α0,1,2 represent the fiber transmission losses, g1,2 represent the Raman gain coefficients, and ε0,1,2 represent the Rayleigh backscattering coefficients;

[0173] When performing numerical analysis, consider the boundary conditions: P+(0) = P, 0in P+(0) = R P-(0), P-(L) = RP+(L) to represent different cavity structures (fully open cavity structure, semi-open 1,2L1,21,21,2F1,21,2 cavity structure, and the influence of different reflectivities), where Pin represents the pump power. By constraining these boundary conditions, the numerical solution of the above equation is obtained using the iterative method;

[0174] Second, establish a theoretical model based on the L-ROPA technology

[0175] The working principle of the new L-ROPA technology is an erbium-doped fiber amplifier. The internal population of particles is mainly concentrated in the ground state and the metastable state. The Giles simplified model can be used to analyze the gain coefficient and amplification characteristics of L-ROPA:

[0176]

[0177] Among them, the superscripts ‘+’ and ‘-’ represent the forward and backward waves, the subscript ‘k’ represents the light beam of the kth wavelength, and N t is the total average number of particles in the two-level system, N 2 is the number of particles in the upper energy level, α k represents the absorption coefficient of wavelength k, and g k represents the gain coefficient of wavelength k, Δυ k is the effective noise bandwidth, m represents the number of optical wave polarization modes, υ k represents the frequency of the light of the kth wavelength, l is the background loss, and ζ is the saturation coefficient of the erbium-doped fiber;

[0178] Third, establish a RIN transfer theoretical model based on the low-noise fiber random laser amplification system

[0179] To realize the new low-noise fiber random laser amplification system, it is necessary to study the RIN transfer from the pump light to the signal light in the amplification system. It can be derived by the perturbation method that the RIN transfer amount H(f) under the condition of co-pumping is:

[0180]

[0181] Combined with the above theoretical model, the RIN transfer characteristics and optimization methods in the fiber optic sensing system based on the new low-noise fiber random laser amplification technology can be analyzed; the combination of the high-order low-noise fiber random laser amplification technology and the L-ROPA technology can effectively overcome the power loss in long-distance optical transmission and realize a long-distance low-noise optical amplification system, as Figure 1 shown;

[0182] Figure 2 is the simulation curve of the Rayleigh scattered signal light power distribution of the new low-noise optical amplification system under the sensing distance of 30 km, that is, the power distribution of light along the optical fiber;Figure 3 It is the simulation curve of the Rayleigh scattering signal optical power distribution, i.e., the optical power distribution along the optical fiber, of a new type of low-noise optical amplification system at a sensing distance of 50 km. Among them, the red curve corresponds to the Rayleigh scattering signal optical power in the high-order fiber random laser amplification system, and the blue curve corresponds to the Rayleigh scattering signal optical power in the new type of low-noise optical amplification system. The transmission loss of the 1550 nm signal light is 0.25 dB / km, the fiber random laser pump power is 1.7 W, and the length of the erbium-doped fiber is 10 m. To overcome the optical power loss caused by the transmission and splitting of the sensing signal light in the PON, the new type of L-ROPA technology is combined with the distributed fiber vibration sensing technology, making the signal-to-noise ratio at the end of the sensing optical fiber of the new type of LL-DVS system based on the optical fiber communication network higher, and enabling the realization of a high signal-to-noise ratio and high-sensitivity LL-DVS system based on the OPGW optical cable;

[0183] The signal processing system based on the multi-source spatio-temporal data fusion of high-elastic transmission wires preferably requires the protected technical content to be: applying the following multi-source data real-time fusion and intelligent processing method for the automatic discrimination of the icing characteristics and icing levels of high-elastic wires, and the specific content requirements for constructing the transmission line icing assessment and early warning model based on multi-source data fusion:

[0184] Line reading characteristics analysis and wind dance level discrimination method based on deep learning:

[0185] Through the deep learning network, a non-linear mapping relationship is established by combining the icing state levels of high-elastic wires for the analysis and feature expression of line icing characteristics. During this period, methods for extracting micro-meteorological, stress time-frequency characteristics and wind dance characteristics based on deep learning are applied, such as Figure 4 、 Figure 5 as shown. Figure 4 It is the extraction of the time-series structure characteristics of line icing based on 1-D CNN, Figure 5 and the extraction of the spatial distribution characteristics of line icing based on Bi-LSTM. Then, a mapping relationship between the spatio-temporal characteristics of line icing and the wind dance level is constructed based on the 1D-CNN-Bi-LSTM network, and the automatic discrimination of the icing level is realized according to the non-linear mapping relationship between the icing characteristics and the icing level constructed by this network;

[0186] The requirements for the multi-source data fusion method based on association analysis are:

[0187] First, a data mining method based on association analysis: The Apriori algorithm, an association analysis algorithm, is used to discover sets of data items that often appear simultaneously, especially sets of data items where the characteristic items of acoustic wave sensing signals and target event items frequently appear simultaneously, thereby finding certain important rules of event occurrence hidden within these sets of data items; in association analysis, item sets that satisfy the condition that their support is greater than the minimum support threshold are defined as frequent item sets, and the rules of event occurrence in these item sets are defined as association rules, and the meaning represented by the support of an item set refers to the proportion of the number of data records containing this item set in all records of the data set;

[0188] Second, a frequent item set mining method: In the association rule mining method, when performing frequent item set mining, in addition to the support having to meet the threshold condition, there is an additional pruning condition: that is, the finally extracted item set must contain and only contain one thing representing the target event, and the rest of the items all represent any possible characteristic items; as Figure 7 shown is a detailed example of the frequent item set mining process related to the pipeline leakage detection method in this article. In the figure, the frequent item sets generated after pruning are shown in the 2nd, 3rd, and 4th layers. They are all candidate sets that meet the pruning conditions of the minimum support threshold and have one and only one target event item. For example Figure 7 the frequent 4-item set obtained in the 4th layer in the example: {4, 6, 7, C}, where 4, 6, and 7 are signal characteristic numbers, and C represents the target event category. The meaning represented by this frequent item set is that the characteristics {4, 6, 7} all belong to Class1 and always appear simultaneously with the occurrence of the target event C.

[0189] Third, the generation and pruning of association rules

[0190] By setting a reasonable confidence threshold, association rules that meet the minimum confidence threshold condition are generated to complete the basic content of association analysis; however, in actual use, when the dimension of the feature set and the number of target events are relatively matched, generally a large number of association rules that meet the requirements will be generated, but there is information redundancy in these rules, and even some of them will have some imperceptible negative effects. Simply increasing the confidence threshold cannot achieve the purpose of compressing the redundant data volume and then extracting more effective association rules. In this case, two rule evaluation indicators, the Kulc coefficient k(X→Y) and the imbalance factor r(X→Y), can be introduced to prune the existing association rules:

[0191]

[0192] The Kulc coefficient is to calculate the positive and negative confidences c(X→Y) and c(Y→X) for the same association rule, and then do an average. Like the confidence, the higher the value, the stronger the conditional probability that the rule of this feature leads to the occurrence of the event. Compared with a single confidence, it has the advantage of not being affected by null values ​​in the data set.

[0193] Fourth, the design and implementation of association rule classifier: The result of data preprocessing - two positive and negative Boolean feature matrices A and B, add K-dimensional binary sequence labels of K-type target events as the input of the association rule mining method based on the Apriori algorithm. After performing two association analyses on the two Boolean matrices respectively, two sets of strong association rules between feature item sets and target event items that meet the confidence, Kulc coefficient and imbalance factor threshold conditions are obtained; the association rule set mined based on the Boolean matrix A is defined as R A , the association rule set mined based on the Boolean matrix B is defined as R B , are stored in the association rule classifier as the core reference standard for classification and judgment; stored in the rule set R A , R B The association rules in the example are divided into categories according to the category numbers of the K target events, such as {1, 2, 3, ...}, that is, the association rule antecedents with the same rule consequents are combined into small sets, such as R A1 , R A2 , R A3 , R B1 etc., which is convenient for classification and discrimination; the classifier also includes the feature selection results saved during the training period and the feature cluster center set obtained by using the FCM algorithm in the binarization process; so far, the core content of the association rule classifier has been constructed;

[0194] The process of using association rule classifier to identify target events is as follows:

[0195] like Figure 8 As shown in the figure, the main steps of implementing the association rule classifier include:

[0196] 1) Extract the MFCC features and AR model coefficient features of the test signal, select the optimal feature column according to the feature selection results, calculate the membership of the feature value to Class0 and Class1 according to the FCM clustering center, and convert each frame of the time domain test signal into two sets of positive and negative Boolean feature sequences. The specific test signal preprocessing method;

[0197] 2) Compare the Boolean feature sequence with the rule sets of various target events in the association rule classifier, and separately count the proportion of the number of test signals satisfied in the rule sets of various events. The proportion of the number of the same type of events is the sum of the proportions in the positive and negative rule sets.

[0198] 3) Compare the proportion values of various events, calculate the maximum value Max, and set a threshold thr for the recognized category determination. Otherwise, the discrimination completed under the condition that Max is very small is not persuasive. If Max meets the minimum discrimination threshold thr, the current test signal is determined to be the target event category corresponding to the maximum value Max. If the threshold condition is not met, that is, the test signal is quite different from the features of the mined rule set, the current signal is determined to be the default event or unknown event.

[0199] Based on the above correlation relationships between multiple parameters such as meteorology, terrain, line parameters, and icing characteristics and the icing level, realize the estimation and automatic evaluation of the line icing risk coefficient, which serves as the data basis for safety warning.

[0200] (III) The requirements for model construction and system software development are as follows:

[0201] Build a transmission line icing evaluation and warning model based on multi-source data fusion; based on the research algorithm, build an interactive platform and software for comprehensive monitoring, evaluation, and warning of transmission line icing, realize a multi-parameter fusion perception system, and on this basis, realize functions such as highly reliable icing state evaluation and safety warning based on multi-parameter fusion;

[0202] The system platform software is planned to be designed with a B / S architecture, as Figure 9 shown, running on the Linux platform, mainly divided into a data platform (background) and a user platform (foreground); among them: the data platform is the background, and the user platform is the foreground; the background is written in C#, mainly responsible for communicating with systems such as fiber optic sensing monitoring, parsing the data sent to the software system and processing and storing it. The storage method can choose local database storage and remote server storage; the foreground is mainly written in web pages using VUE, etc., mainly responsible for interacting with users, providing functions such as user login and management, data waveform display, alarm display, electronic map display, and multi-parameter visualization display. Users obtain web page content by accessing the Apache server; a database is used as the data interface between the foreground and background for data transmission and caching of the three device systems, and a TCP stream is used as the control interface for coordination control and information interaction;

[0203] Select transmission lines of 66 kV and above for multi-source data fusion perception, comprehensive testing of system hardware and software, and actual icing experiments to complete the demonstration application of comprehensive monitoring, evaluation, and safety warning of the icing state of extra-long-distance transmission lines.

[0204] The signal processing system for multi-source spatio-temporal data fusion based on highly elastic power transmission lines uses the following line reading characteristics analysis and wind dance level discrimination method: The content of line reading characteristics analysis is as follows:

[0205] (1) Signal transmission characteristics mainly involve: high-speed signal transmission ability, signal attenuation and anti-interference, multi-channel transmission and parallel processing;

[0206] High-speed signal transmission ability: When the line reads doctoral data, it can support high-frequency signal transmission; the transmission rate specifically depends on the material of the line, manufacturing process, and the transmission protocol adopted; the line needs to have good signal attenuation control characteristics;

[0207] Insulating materials and shielding technologies are used to reduce signal attenuation and external interference; using a multi-layer shielding structure can effectively block external electromagnetic interference and ensure the stability of the signal during transmission; the line is also equipped with signal amplifiers to compensate and enhance the attenuated signal to ensure that the signal can still be accurately read after long-distance transmission;

[0208] Multi-channel transmission and parallel processing: The line often supports multi-channel transmission; multi-channel transmission requires the line to have good parallel processing ability to coordinate data transmission between channels and avoid data conflicts and chaos; this involves complex timing control and data scheduling algorithms to ensure that data from different channels can reach the destination accurately and orderly; at the same time, the line also needs to have the function of monitoring and managing multi-channel transmission to promptly detect and solve possible channel failures or data transmission anomalies;

[0209] (3) Data processing characteristics meet the following requirements: data caching and pre-reading, data error correction and verification, protocol adaptation and compatibility;

[0210] Data caching and pre-reading: During the doctoral reading process, data access has a certain degree of randomness and suddenness; to improve the response speed of data reading, the line has a data caching function; the line also uses pre-reading technology, which analyzes the historical patterns and rules of data access, predicts the data that needs to be read next in advance, and performs pre-reading operations in the background to pre-load this data into the cache; when the user really needs this data, zero-latency or near-zero-latency reading can be achieved, greatly improving the efficiency and smoothness of the doctoral reading operation;

[0211] Data error correction and verification: The line uses a variety of error correction coding technologies to monitor and verify the transmitted data in real time;

[0212] When a data error is detected, the line can automatically correct the error according to the principle of error correction coding;

[0213] Protocol Adaptation and Compatibility: When the circuit reads the doctor data, it needs to communicate and interact with various different devices and systems.

[0214] (3) The reliability and stability features have the following requirements: fault tolerance design, environmental adaptability, long-term operation stability;

[0215] Fault Tolerance Design: The circuit needs to have a high degree of reliability during the process of reading doctor data to cope with various possible fault situations. Fault tolerance design is one of the important means to improve the reliability of the circuit. For example, the circuit may adopt a redundant link design. When the main link fails, the backup link can automatically switch to ensure that data transmission is not interrupted. In addition, for key components in the circuit, such as chips and connectors, redundant backups may be used. If a component fails, the backup component can immediately take over the work to ensure the normal operation of the circuit. This fault tolerance design greatly reduces the risk of the entire doctor reading system crashing due to a single-point failure, improving the availability and reliability of the system;

[0216] Environmental Adaptability: The doctor reading environment may vary widely, including different places such as laboratories and field monitoring stations. The circuit needs to have good environmental adaptability. In terms of temperature, the circuit may need to be able to work properly within a wide temperature range, from a low-temperature laboratory environment to a high-temperature outdoor equipment box. It may adopt special heat dissipation designs or high-temperature resistant materials to ensure the stable performance of the circuit under different temperature conditions. In terms of humidity, the circuit needs to have a certain moisture-proof ability to prevent problems such as short circuits or corrosion caused by moisture intrusion. For some environments where there may be electromagnetic interference, the circuit also needs to have good anti-electromagnetic interference performance to ensure the accuracy and stability of data transmission. For example, in scientific research experiments in a strong electromagnetic field environment, the circuit can effectively resist external electromagnetic interference through shielding and filtering technologies to ensure the reliable transmission of doctor reading data.

[0217] Long-Term Operation Stability: The doctor reading work requires the circuit to run continuously for a long time. During the design and manufacturing process of the circuit, strict reliability tests and aging tests will be carried out to ensure that its performance does not decline and the failure rate is low during long-term operation;

[0218] At the same time, the circuit may adopt intelligent monitoring and maintenance technologies to monitor the operating status of the circuit in real time; when an abnormal situation is detected, it can issue an early warning in a timely manner and perform self-diagnosis and repair; thus ensuring the stable operation of the circuit during long-term doctor reading tasks;

[0219] The wind dance level discrimination method meets the following requirements:

[0220] (1) Discrimination Method Based on Wind Speed:

[0221] Measurement Principle and Equipment: The wind speed measurement equipment is a mechanical anemometer or an ultrasonic anemometer; the anemometer is installed in an open and unobstructed position to ensure accurate measurement of the true wind speed; during installation, attention should be paid to the height and direction of the instrument, and generally it should be installed in accordance with relevant standards and specifications to ensure the accuracy and comparability of measurement data;

[0222] The relevant technical specifications for anemometers are as follows: 1) JJG 1194-2023 "Digital Vane Anemometer": It stipulates the metrological performance requirements, general technical requirements, verification conditions, verification items, verification methods, handling of verification results, etc. for digital vane anemometers; 2) JJG(Building) 01-1992 "Verification Regulation for Thermal Ball Anemometer" and JJG(Building) 0001-1992 "Metrological Verification Regulation for Thermal Ball Anemometer": Provide specific regulation requirements for the verification of thermal ball anemometers. 3) GJB 6550-2008 "General Specification for Military Digital Wind Direction and Speed Anemometer": It is the general specification standard for military digital wind direction and speed anemometers. 4) GJB / J 3828-1999 "Verification Regulation for Hot Wire / Thermal Film Anemometer": Applicable to the verification of hot wire / thermal film anemometers. 5) ESDU 16002-2018 "Practical Guide for Turbulence and Flow Measurement Using Thermal Anemometers (CTA), Laser Doppler Anemometers (LDA) and Particle Image Velocimetry (PIV) - Part 2: Constant Temperature Anemometers" and ESDU 17008-2018 "Practical Guide for Turbulence and Flow Measurement Using Thermal Anemometers (CTA), Laser Doppler Anemometers (LDA) and Particle Image Velocimetry (PIV) - Part 3: Laser Doppler Anemometers": Provide practical guides for thermal anemometers, laser Doppler anemometers, etc. in turbulence and flow measurement. 6) KSE4086-2004(2009) "Portable Mine Thermal Anemometer": It is the standard for portable mine thermal anemometers.

[0223] Wind Speed Grade Classification Standard: According to the internationally common wind force grade classification standard, the wind speed range is divided into different levels; for example, a wind speed of 0 indicates calm wind, with a wind speed less than 0.2 m / s; a wind speed of 1 is a light breeze, with a wind speed between 0.3 - 1.5 m / s; a wind speed of 2 is a gentle breeze, with a wind speed of 1.6 - 3.3 m / s, etc. As the wind speed increases, the wind grade also increases accordingly. For different application scenarios, the wind speed grade may be further subdivided or adjusted according to actual needs; for example, in the siting assessment of certain specific wind farms, different wind speed intervals may be more detailedly divided to more accurately evaluate the wind energy resources and the applicability of wind turbines. In wind grade determination, strictly follow the established wind speed grade classification standard and combine the actually measured wind speed data to determine the current wind grade;

[0224] Data collection and processing: The anemometer continuously collects wind speed data and transmits it to the data processing system; the data processing system filters and averages the collected raw data to remove noise and outliers, obtaining a more accurate wind speed value; for example, using a moving average algorithm to average the wind speed data within a certain time window can reduce the impact of instantaneous wind speed fluctuations on the determination of wind dance levels. At the same time, the data processing system also monitors and analyzes the wind speed data in real time. When the wind speed changes and reaches different level thresholds, it timely updates the determination result of the wind dance level;

[0225] (2) Discrimination method based on wind direction changes

[0226] Wind direction measurement technology: The measurement of wind direction is also of great significance for the determination of wind dance levels. Wind direction sensors or / and wind vanes are devices for measuring wind direction

[0227] Relationship between wind direction change characteristics and wind dance levels: Different wind dance levels are accompanied by different wind direction change characteristics; by monitoring and analyzing the wind direction changes, it is possible to assist in judging the wind dance levels; specifically, by calculating the change frequency and angular range of the wind direction, combined with historical data and empirical models, to determine the threshold values of wind direction change characteristics corresponding to different wind dance levels; when the actually measured wind direction change parameters exceed or meet the threshold value corresponding to a certain wind dance level, it can be judged that the current is at this wind dance level;

[0228] Discrimination method integrating wind speed and wind direction: In order to more accurately determine the wind dance levels, it is necessary to comprehensively consider two factors, wind speed and wind direction; establish a wind speed-wind direction joint discrimination model, and input the measured data of wind speed and wind direction into the model for analysis and processing at the same time; in the model, according to the meteorological conditions and actual wind dance phenomena corresponding to different wind speed and wind direction combinations, formulate corresponding discrimination rules and thresholds;

[0229] (3) Discrimination method based on the impact of wind on objects

[0230] Observation objects and indicators: Determine the wind dance levels by observing the impact of wind on different objects; the observed indicators include the swing amplitude and morphological changes of the objects; establish the corresponding relationship between the degree of influence and the wind dance levels: Based on the observation and analysis of different objects under different wind forces, establish the corresponding relationship between the degree of influence of the objects and the wind dance levels; when actually determining the wind dance levels, by observing the actual performance of these objects and comparing with the pre-established corresponding relationship, it is possible to quickly judge the current wind dance level;

[0231] Limitations and supplementary methods:

[0232] Comprehensive judgment is carried out in combination with methods such as wind speed measurement and wind direction change analysis and discrimination; meanwhile, the corresponding relationship between the observed object and the wind level is continuously improved and optimized, considering more factors to improve the accuracy and reliability of the discrimination method.

[0233] The signal processing system based on multi-source spatio-temporal data fusion of high-elastic transmission wires meets the following requirements:

[0234] (I) Theoretical and practical basis of the research content:

[0235] The fiber Bragg grating sensor utilizes the periodic refractive index change formed in the fiber core. When the period satisfies the Bragg condition, the light reflected back by each periodic reflection surface accumulates step by step, and finally a reflection peak will be formed in the reverse direction. The central wavelength is determined by the grating parameters. This process can be expressed by Equation (7):

[0236] λ B =2n eff Λ (7)

[0237] Among them, λ B is the fiber Bragg wavelength, n eff is the refractive index of the fiber core with respect to the central wavelength in free space, and Λ is the grating period;

[0238] The resonant peak of the FBG, that is, the central wavelength reflected by the fiber grating, depends on the grating period and the size of the effective refractive index of the core. Various physical quantities such as external vibration, strain, and temperature will change the effective refractive index and grating period of the fiber grating. The influence of physical quantities such as strain and temperature on the Bragg wavelength of the fiber grating can be expressed by the following Equation (8):

[0239]

[0240] The change in the external physical quantity acting on the fiber grating will cause the offset of the grating wavelength. The backend signal demodulation device can obtain the change magnitude of the external physical quantity by detecting the wavelength offset of the fiber grating, so as to achieve the purpose of detecting the external physical quantity; its principle is as Figure 12 shown;

[0241] The convolutional neural network CNN mainly consists of three parts, namely the convolutional layer, the pooling layer, and the fully connected layer. Each layer can generate multiple feature maps. The convolutional layer is the core of the neural network. The convolutional layer is composed of multiple groups of convolutional kernels. The convolutional kernels perform convolutional operations with the input data, and the output result undergoes a non-linear operation through an activation function to obtain the feature array of this operation; the specific recognition process is as Figure 13 shown;

[0242] The pooling layer divides the feature array obtained by the convolutional layer into multiple independent pooling blocks, and performs pooling operations within the pooling blocks. There are two common pooling methods. One is average pooling, which calculates the average value within the pooling block. The other is max pooling, which extracts the maximum value within the pooling block. Each of the two pooling methods has its own advantages and disadvantages, and can be appropriately selected according to the type of the array. As Figure 14 shown are the calculation results of different pooling methods for the same feature array.

[0243] The pooling layer can significantly reduce the size of the feature array and eliminate signal offsets. The pooling layer is also called the downsampling layer. After the input data passes through multiple convolutional and pooling layers, the output is a multi-dimensional array, which contains the feature information of the original data. Finally, classification is achieved through the activation of the softmax function in the fully connected layer.

[0244] The convolutional neural network realizes the feature extraction of data through convolution and pooling, and realizes the classification of data through the fully connected layer. Moreover, compared with the traditional neural network, the convolutional neural network has the advantages of fewer training parameters, shorter training time, and no need to manually extract feature data, and is very suitable as an artificial intelligence algorithm for disturbance type recognition.

[0245] The specific technical route of the research is as Figure 16 :

[0246] In this embodiment, the development of the 62% IACS high-conductivity heat-resistant aluminum alloy conductor material and the trial production of the wire are mainly carried out by the State Grid Smart Grid Research Institute Co., Ltd., and the State Grid Liaoning Electric Power Co., Ltd. conducts cooperative research. The mass production of the wire will be entrusted to the wire manufacturer; after obtaining the performance test report of the wire and the supporting sensing intelligent system, according to the requirements of transmission line construction and application, engineering application technology research will be carried out, and it will be put into operation on the transmission line of the State Grid Liaoning Electric Power Co., Ltd., and its operation effect and technical economy will be analyzed.

[0247] The beneficial effects of this embodiment:

[0248] This embodiment is based on the research of the electrical conductivity mechanism and heat resistance mechanism of microalloyed aluminum alloys. Through the correlation research of microalloying, process, microstructure, and properties, appropriate microalloying and optimized casting-rolling and heat treatment processes are adopted to obtain a microstructure with good high-temperature stability. The electrical conductivity of the heat-resistant aluminum alloy is increased to 62% IACS, and the long-term operating temperature is ≥150 °C, thereby reducing the power loss of transmission lines. On this basis, it is planned to use advanced optical fiber sensing technology to achieve all-weather real-time online monitoring of the icing state of ultra-long-distance and low-cost transmission lines. On the other hand, by integrating rich multi-source information such as meteorological data, geographical information, and line parameters, the correlation between icing levels and multi-source data is mined based on big data analysis methods, and thus a precise icing index evaluation system and assessment system are constructed to realize a highly reliable line galloping disaster warning system. In short, this project organically combines high-conductivity heat-resistant wires with advanced optical fiber sensing and traditional meteorological technologies to form a low-cost wide-area multi-sensor technology that combines points, lines, and surfaces. With the help of advanced signal processing technologies such as using deep networks to mine galloping characteristic parameters and using big data correlation analysis to achieve multi-source spatio-temporal data fusion, the understanding of on-site monitoring data is improved, and the self-value of multi-sensor big data is mined, providing an objective scientific basis for the safe operation, efficient management, and intelligent decision-making of ultra / extra-high voltage transmission lines.

[0249] The implementation of this embodiment will break through the key technologies for the preparation of high-conductivity heat-resistant aluminum alloy single wires and high-conductivity heat-resistant aluminum alloy wires, develop high-conductivity heat-resistant and high-elastic intelligent transmission wires, effectively reduce the operation and maintenance costs of ultra-long-distance transmission lines, realize synchronous monitoring and data sharing among the source, grid, and load, improve the level of independent early warning and intelligent interconnection and interaction, enhance the level of intelligent operation and maintenance, and support the large-scale safe grid connection and full consumption of new energy. The results can be transformed in cable manufacturing-related manufacturers inside and outside the system, which is of great significance for promoting the development of overhead conductors for transmission lines in China, improving the manufacturing level of the cable industry, enhancing the technical level of transmission lines of the State Grid Corporation, and reducing the power loss of transmission lines.

[0250] In order to better achieve the capacity expansion and transformation of urban power grids at low cost, aiming at problems such as long-distance power transmission and large fluctuations in new energy access, effectively utilize the original tower bases and line corridors, conduct intelligent monitoring, and grasp the status of transmission lines in real time. Developing high-conductivity, heat-resistant, and highly elastic intelligent aluminum alloy conductors to achieve high-efficiency intelligent power transmission has become one of the trends in conductor research and development. However, the existing heat-resistant aluminum alloy conductors have poor comprehensive performance (the highest conductivity is 61% IACS (20 °C), and the long-term heat-resistant temperature is 150 °C), which increases line losses, causes a large amount of electrical energy waste, and the reliability of the intelligent monitoring system is poor. Developing heat-resistant aluminum alloy conductor materials with a conductivity of 62% IACS and their wires, and even aluminum alloy conductor materials with higher conductivity and heat resistance, and supporting corresponding intelligent real-time monitoring systems are the current trends in the development of heat-resistant aluminum alloy wire technology. Therefore, the present invention focuses on overcoming the key technical problems of high-conductivity, heat-resistant, and highly elastic intelligent power transmission aluminum alloy conductor materials and their wires with a long-term operating temperature of 150 °C and a conductivity of 62% IACS. The successful development, popularization, and application of the conductor materials and wires will greatly promote the technological development of heat-resistant aluminum alloy wires in China, the development of power transmission technology level, and the improvement of the intelligent operation and maintenance level, supporting the large-scale safe grid connection and full consumption of new energy; at the same time, the successful development and grid connection operation of 62% IACS high-conductivity, heat-resistant, and highly elastic intelligent power transmission aluminum alloy wires will surely play a demonstrative role in the development of new products by Chinese cable enterprises and lead the technological progress of cable enterprises.

Claims

1. A signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines, characterized by: It includes a highly flexible smart transmission conductor body and a low-noise optical amplification system, which uses real-time fusion and intelligent processing methods of multi-source data; Together they form the necessary auxiliary components to support highly flexible smart transmission lines; among them: The main body of the highly elastic intelligent transmission conductor is domestically produced steel core 58% conductivity heat-resistant aluminum alloy stranded wire NRLH58GJ-240 / 30; In the signal processing system based on multi-source spatiotemporal data fusion of highly elastic intelligent transmission lines, the highly elastic intelligent transmission line itself combines optical fiber sensing technology and meteorological technology to form a low-cost wide-area multi-sensing technology combining points, lines and surfaces. The signal processing technology of multi-source spatiotemporal data fusion is realized by mining wind dance characteristic parameters with the help of deep networks and using big data correlation analysis, which improves the understanding of on-site monitoring data, mines the inherent value of multi-sensor big data, and provides an objective scientific basis and theoretical foundation for the safe operation, efficient management and intelligent decision-making of ultra-high / ultra-high voltage transmission lines. The low-noise optical amplification system utilizes the following novel fiber optic sensing technologies:

1. New distributed amplification technology First, a theoretical model of signal light power distribution based on high-order low-noise fiber random laser amplification was established. A novel low-noise optical amplifier system is theoretically analyzed using the power steady-state equation model shown in equation (1) for the high-order fiber random laser amplification part of the amplifier system, where the power distribution of each frequency component along the fiber length is calculated using the following model: Wherein, the subscripts '0', '1', '2' correspond to pump, first-order and second-order Stokes light respectively; the superscripts '+' and '-' represent forward and backward waves; P0,1,2 represents optical power; z represents the coordinate of the optical transmission direction; f0,1,2 represents the optical frequency; Γ1,2 represents the number of photons; where Δf1,2=0.25THz represents the radiation bandwidth; T=298K represents the absolute temperature, KB represents the Boltzmann constant, h represents the Planck constant, α0,1,2 represents the optical fiber transmission loss, g1,2 represents the Raman gain coefficient, and ε0,1,2 represents the Rayleigh backscattering coefficient; In numerical analysis, the boundary conditions are considered: P+(0)=P, 0in P+(0)=R P-(0), P-(L)=R P+(L) to represent different cavity structures, where Pin represents the pump power. By constraining these boundary conditions, the above equations are numerically solved using an iterative method; Second, establish a theoretical model based on L-ROPA technology The working principle of the new L-ROPA technology is an erbium-doped fiber amplifier, whose internal population is mainly concentrated in the ground state and metastable state. The Giles simplified model can be used to analyze the L-ROPA gain coefficient and amplification characteristics: Wherein, the superscripts '+' and '-' represent the forward and backward waves, the subscript 'k' represents the kth wavelength beam, and N t is the total average number of particles in the two-level system, N2 is the number of particles in the upper energy level, and α k represents the absorption coefficient at wavelength k, g k represents the gain coefficient for wavelength k, Δυ k is the effective noise bandwidth, m represents the number of light polarization modes, υ k represents the frequency of the kth wavelength light, l is the background loss, and ζ is the saturation coefficient of the erbium-doped fiber; Third, establish a theoretical model of RIN transfer based on low-noise fiber random laser amplification system In order to realize a new type of low-noise fiber random laser amplification system, it is necessary to study the RIN transfer from pump light to signal light in the amplification system. The perturbation method can be used to derive the RIN transfer amount H(f) under the same-direction pumping condition: Combined with the above theoretical model, the RIN transfer characteristics and optimization methods in the fiber optic sensing system based on the new low-noise fiber random laser amplification technology can be analyzed; the combination of high-order low-noise fiber random laser amplification technology and L-ROPA technology can effectively overcome the power loss of long-distance optical transmission and realize the long-distance low-noise optical amplification system to meet the requirements: the broadband power supply is connected to the wavelength division multiplexing WDM through the acousto-optic modulator and the erbium-doped fiber amplifier EDFA, and the host computer is connected to the wavelength division multiplexing WDM through the spectrum demodulation module and the erbium-doped fiber amplifier EDFA; the signal generator is connected to the acousto-optic modulator and the spectrum demodulation module respectively; the pump light source is connected to the spectrum demodulation module, and is connected to multiple fiber Bragg gratings FBG through the erbium-doped fiber EDF; The Rayleigh scattering signal optical power distribution of the new low-noise optical amplifier system at sensing distances of 30km and 50km, that is, the power distribution simulation curve of the light along the optical fiber; among them, the red curve corresponds to the Rayleigh scattering signal optical power in the high-order optical fiber random laser amplifier system, and the blue curve corresponds to the Rayleigh scattering signal optical power in the new low-noise optical amplifier system. The 1550nm signal light transmission loss is 0.25dB / km, the fiber random laser pump power is 1.7W, and the erbium-doped fiber length is 10m. In order to overcome the optical power loss caused by the transmission and beam splitting of the sensing signal light in PON, the new L-ROPA technology is combined with the distributed optical fiber vibration sensing technology, so that the signal-to-noise ratio of the sensing optical fiber tail end of the new LL-DVS system based on the optical fiber communication network is higher, and a high signal-to-noise ratio and high-sensitivity LL-DVS system based on OPGW optical cable can be realized.

2. The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines according to claim 1 is characterized by: The specific requirements for applying the following multi-source data real-time fusion and intelligent processing methods to automatically identify the ice characteristics and ice levels of high-elastic conductors and to build an ice assessment and early warning model for transmission lines based on multi-source data fusion are as follows: (I) Analysis of route study characteristics and wind dance level discrimination method based on deep learning Through the deep learning network, the ice coverage status level of the highly elastic conductor is combined with the construction of a nonlinear mapping relationship to analyze and express the line ice coverage characteristics. During this period, the micro-meteorological and stress time-frequency feature extraction and wind dance characteristic analysis methods based on deep learning are applied. Then, based on the 1D-CNN-Bi-LSTM network, a mapping relationship between the spatiotemporal characteristics of line ice coverage and the wind dance level is constructed. According to the nonlinear mapping relationship between the ice coverage characteristics and the ice coverage level constructed by the network, the ice coverage level can be automatically identified.

3. The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines according to claim 2 is characterized by: The requirements of multi-source data fusion method based on association analysis are: First, data mining method based on association analysis: Apriori algorithm is used to discover certain sets of data items that often appear at the same time, especially the sets of data items where the characteristic items of acoustic sensor signals and target event items frequently appear at the same time, so as to find out some important occurrence rules implied in these sets of data items; in association analysis, the item sets in these sets of data items that meet the condition that their support is greater than the minimum support threshold are defined as frequent item sets, and the occurrence rules of these things in the item sets are defined as association rules, and the support of an item set refers to the proportion of the number of data records containing the item set in all records of the data set; Second, the frequent item set mining method: In the association rule mining method, in addition to the support must meet the threshold condition, there is an additional pruning condition when mining frequent item sets: that is, the final extracted item set must contain and only contain one thing representing the target event, and the remaining items all represent any possible feature items; Third, the generation and pruning of association rules: By setting a reasonable confidence threshold, the association rules that meet the minimum confidence threshold conditions are generated to complete the basic content of association analysis; the Kulc coefficient k(X→Y) and the imbalance factor r(X→Y) are introduced as two rule evaluation indicators to prune the existing association rules: The Kulc coefficient is to calculate the positive and negative confidences c(X→Y) and c(Y→X) for the same association rule, and then do an average. Like the confidence, the higher the value, the stronger the conditional probability that the rule of this feature leads to the occurrence of the event. Compared with a single confidence, it has the advantage of not being affected by null values ​​in the data set. Fourth, the design and implementation of association rule classifier: The result of data preprocessing - two positive and negative Boolean feature matrices A and B, add K-dimensional binary sequence labels of K-type target events as the input of the association rule mining method based on the Apriori algorithm. After performing two association analyses on the two Boolean matrices respectively, two sets of strong association rules between feature item sets and target event items that meet the confidence, Kulc coefficient and imbalance factor threshold conditions are obtained; the association rule set mined based on the Boolean matrix A is defined as R A , the association rule set mined based on the Boolean matrix B is defined as R B , are stored in the association rule classifier as the core reference standard for classification and judgment; stored in the rule set R A , R B The association rules in the classifier are divided into categories according to the category numbers of K target events, such as {1, 2, 3, ...}, that is, the antecedents of the association rules with the same consequents are combined into small sets for easy use in classification and discrimination; the classifier also includes the feature selection results saved during the training period and the feature clustering center set obtained by the FCM algorithm during the binarization process; At this point, the core content of the association rule classifier has been constructed.

4. The signal processing system based on multi-source spatiotemporal data fusion of high elasticity transmission line according to claim 3 is characterized by: The process of using association rule classifier to identify target events is as follows: 1) Extract the MFCC features and AR model coefficient features of the test signal, select the optimal feature column according to the feature selection results, calculate the membership of the feature value to Class0 and Class1 according to the FCM clustering center, and convert each frame of the time domain test signal into two sets of positive and negative Boolean feature sequences. The specific test signal preprocessing method; 2) Compare the Boolean feature sequence with the rule sets of various target events in the association rule classifier, and count the proportion of the number of test signals that meet the rule sets of various events, where the proportion of the number of events of the same type is the sum of the proportions in the positive and negative rule sets; 3) Compare the proportions of various events, calculate the maximum value Max, and set a threshold thr for the recognition category judgment. Otherwise, the judgment completed under the condition that Max is very small is not convincing; if Max meets the minimum judgment threshold thr, the current test signal is judged as the target event category corresponding to the maximum value Max; If the threshold condition is not met, that is, the test signal differs greatly from the mined rule set features, the current signal is judged as a default event or an unknown event; Based on the correlation between the above-mentioned meteorological, topographic, line parameters, icing characteristics and other multiple parameters and the icing level, the line icing risk factor estimation and automatic evaluation are achieved as the data basis for safety warning.

5. The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines according to claim 4 is characterized by: The requirements for model building and system software development are: Construct a transmission line icing assessment and early warning model based on multi-source data fusion; Based on the research algorithm, build a comprehensive monitoring, assessment and early warning interactive platform for transmission line icing, realize a multi-parameter fusion perception system, and on this basis, realize a highly reliable icing status assessment and safety early warning function based on multi-parameter fusion; The system platform software runs on the Linux platform and is mainly divided into a data platform and a user platform. The data platform is the backend, and the user platform is the frontend. The backend is mainly responsible for communicating with systems such as optical fiber sensor monitoring, parsing data sent to the software system, and processing and storing it. The storage method can be local database storage or remote server storage. The frontend is mainly responsible for interacting with users, providing user login and management, data waveform display, alarm display, electronic map display, and multi-parameter visualization display functions. Users obtain web page content by accessing the Apache server. The database is used as the data interface between the front and back ends for data transmission and caching of the three equipment systems, and the TCP stream is used as the control interface for coordinated control and information interaction. Select transmission lines of 66kV and above, conduct multi-data fusion perception, system hardware and software comprehensive testing and actual icing, and complete comprehensive monitoring, evaluation and safety warning of the icing status of ultra-long distance transmission lines.

6. The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines according to claim 5 is characterized by: The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines uses the following line reading characteristics analysis and wind dance level discrimination method: The analysis of the characteristics of the doctoral program is as follows: (i) Signal transmission characteristics mainly involve: high-speed signal transmission capability, signal attenuation and anti-interference, multi-channel transmission and parallel processing; High-speed signal transmission capability: The line can support high-frequency signal transmission when reading doctor data; the transmission rate depends on the material, manufacturing process and transmission protocol of the line; the line needs to have good signal attenuation control characteristics; Insulation materials and shielding technology are used to reduce signal attenuation and external interference; the multi-layer shielding structure can effectively block external electromagnetic interference and ensure the stability of the signal during transmission; The line is also equipped with a signal amplifier to compensate and enhance the attenuated signal to ensure that the signal can still be accurately read after long-distance transmission; Multi-channel transmission and parallel processing: Lines often support multi-channel transmission; Multi-channel transmission requires the line to have good parallel processing capabilities in order to coordinate the data transmission between channels and avoid data conflicts and confusion; this involves complex timing control and data scheduling algorithms to ensure that data from different channels can reach the destination accurately and orderly; at the same time, the line also needs to have the function of monitoring and managing multi-channel transmission in order to promptly discover and resolve possible channel failures or data transmission anomalies; (2) The data processing characteristics meet the following requirements: data caching and pre-reading, data error correction and verification, protocol adaptation and compatibility; Data caching and pre-reading: During the doctoral study process, data access is somewhat random and sudden. In order to improve the response speed of data reading, the line has a data caching function. The line also uses pre-reading technology to predict the data that needs to be read next by analyzing the historical patterns and rules of data access, and pre-reading operations are performed in the background to pre-load this data into the cache. When users really need this data, they can read it with zero or near-zero latency, greatly improving the efficiency and fluency of doctoral reading operations; Data error correction and verification: The line uses a variety of error correction coding technologies to monitor and verify the transmitted data in real time; When a data error is detected, the line can automatically perform error correction based on the principle of error correction coding; Protocol adaptation and compatibility: When reading doctoral data, the line needs to communicate and interact with various devices and systems; (III) Reliability and stability characteristics have the following requirements: fault-tolerant design, environmental adaptability, and long-term operational stability; Fault-tolerant design: The line needs to have a high degree of reliability during the doctoral study process to cope with various possible fault conditions. Fault-tolerant design is one of the important means to improve line reliability; Environmental adaptability: PhD study environments may vary, including laboratories, field monitoring stations and other different places. The line needs to have good environmental adaptability. In environments with high temperature, humidity and possible electromagnetic interference, the line also needs to have good anti-electromagnetic interference performance to ensure the accuracy and stability of data transmission. Long-term operation stability: PhD work requires the circuit to run continuously for a long time. During the design and manufacturing process, the circuit will undergo strict reliability tests and aging tests to ensure that its performance does not degrade during long-term operation and the failure rate is low; At the same time, the line may use intelligent monitoring and maintenance technology to monitor the operating status of the line in real time; When an abnormal situation is found, it can issue an early warning and perform self-diagnosis and repair in time, thus ensuring the stable operation of the line in the long-term doctoral study mission; The wind dance level determination method meets the following requirements: (I) Identification method based on wind speed Measurement principle and equipment: The wind speed measurement equipment is a mechanical anemometer or an ultrasonic anemometer. The anemometer is installed in an open, unobstructed location to ensure that the actual wind speed can be accurately measured; Wind speed classification standards: According to the internationally accepted wind force classification standards, the wind speed range is divided into different levels; for different application scenarios, the wind speed levels will be further subdivided or adjusted according to actual needs; in the determination of wind dance levels, the current wind dance level is determined in strict accordance with the established wind speed classification standards and combined with the actual measured wind speed data; Data collection and processing: The anemometer continuously collects wind speed data and transmits it to the data processing system; The data processing system filters and averages the collected raw data to remove noise and outliers to obtain more accurate wind speed values; At the same time, the data processing system also monitors and analyzes wind speed data in real time. When the wind speed changes and reaches different level thresholds, the wind dance level judgment results are updated in time; (II) Identification method based on wind direction change Wind direction measurement technology: Wind direction sensor or / and wind vane is a device that measures wind direction; Relationship between wind direction change characteristics and wind dance level: Different wind dance levels are accompanied by different wind direction change characteristics; monitoring and analysis of wind direction changes can assist in determining wind dance levels; specifically, by calculating the wind direction change frequency and angle range, combined with historical data and empirical models, the wind direction change characteristic thresholds corresponding to different wind dance levels are determined; when the actually measured wind direction change parameters exceed or meet the threshold corresponding to a certain wind dance level, it can be determined that the current wind dance level is at that level; Comprehensive wind speed and wind direction identification method: In order to more accurately identify the wind dance level, it is necessary to comprehensively consider the two factors of wind speed and wind direction; establish a wind speed-wind direction joint identification model, and input the measured data of wind speed and wind direction into the model for analysis and processing; in the model, according to the meteorological conditions and actual wind dance phenomena corresponding to different wind speed and wind direction combinations, formulate corresponding identification rules and thresholds; (III) Judgment method based on the impact of wind on objects Observation objects and indicators: The wind dance level is determined by observing the impact of wind on different objects; the indicators to be observed include the swing amplitude and shape changes of the objects; Establishing the correspondence between the degree of influence and the wind dance level: Based on the observation and analysis of different objects under different wind forces, the correspondence between the degree of influence of objects and the wind dance level is established; when actually judging the wind dance level, by observing the actual performance of these objects and comparing with the pre-established correspondence, the current wind dance level can be quickly determined; Limitations and supplementary methods: Combine wind speed measurement, wind direction change analysis and judgment methods to make a comprehensive judgment; at the same time, continuously improve and optimize the correspondence between the observed object and the wind dance level, consider more factors, and improve the accuracy and reliability of the judgment method.

7. The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines according to claim 6 is characterized by: The signal processing system based on multi-source spatiotemporal data fusion of highly elastic transmission lines meets the following requirements:

1. Theoretical and practical basis of the research content: The fiber Bragg grating sensor uses the periodic refractive index change formed in the fiber core. When the period meets the Bragg condition, the light reflected from each periodic reflection surface is gradually accumulated, and finally a reflection peak is formed in the reverse direction. The central wavelength is determined by the grating parameters. This process can be expressed by formula (7): l B =2n eff L (7) Among them, λ B is the fiber Bragg wavelength, n eff is the refractive index of the fiber core for the central wavelength in free space, and Λ is the grating period; The resonance peak of FBG, that is, the central wavelength of the fiber grating reflection, depends on the grating period and the effective refractive index of the fiber core. External vibration, strain, temperature and other physical quantities will change the effective refractive index and grating period of the fiber grating. The influence of physical quantities such as strain and temperature on the Bragg wavelength of the fiber grating is expressed by the following formula (8): The change of external physical quantity acting on the fiber Bragg grating will cause the wavelength of the grating to shift. The back-end signal demodulation equipment detects the wavelength shift of the fiber Bragg grating to obtain the change of the external physical quantity, thereby achieving the purpose of detecting the external physical quantity; The convolutional neural network CNN is mainly composed of three parts, namely the convolution layer, the pooling layer and the fully connected layer. Each layer can generate multiple feature maps. The convolution layer is the core of the neural network. The convolution layer is composed of multiple groups of convolution kernels. The convolution kernel is imported; Perform convolution operation, and the output result is processed by activation function for nonlinear operation to obtain the feature array of this operation; The pooling layer divides the feature array obtained by the convolution layer into multiple independent pooling blocks, and performs pooling operations in the pooling blocks. There are two common pooling methods: one is average pooling, which calculates the average value in the pooling block; the other is maximum pooling, which extracts the maximum value in the pooling block; The pooling layer can significantly reduce the size of the feature array and eliminate the offset of the signal. After the input data undergoes multiple layers of convolution and pooling, the output is a multidimensional array that contains the feature information of the original data. Finally, the classification is achieved through the softmax function activation of the fully connected layer. Convolutional neural networks realize data feature extraction through convolution and pooling, and data classification through fully connected layers. Compared with traditional neural networks, convolutional neural networks have the advantages of fewer training parameters, shorter training time, and no need for manual extraction of feature data. They are very suitable as artificial intelligence algorithms for disturbance type identification.

Images