Equipment for monitoring and recording waveforms and fault diagnosis applicable to new energy access scenarios

Abstract

The present invention provides a device for monitoring and recording waves and fault diagnosis applicable to new energy access scenarios, which is used for: collecting electrical signals in new energy access scenarios in real time to obtain original electrical signals; determining the decomposition level and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals; wherein the adaptive threshold is dynamically adjusted according to the energy distribution of the signals; performing fault diagnosis on the decomposed signals based on an extreme learning machine to obtain a fault diagnosis result; wherein the parameters of the extreme learning machine are optimized through a sparrow search algorithm; marking the time of fault occurrence in the fault diagnosis result, and recording the electrical signal waveforms in a preset time period before and after the time of fault occurrence to obtain a fault diagnosis result with time marking and recording wave data. In the present invention, key information of electrical signals can be effectively extracted and the diagnosis accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a device for monitoring and recording waves and fault diagnosis applicable to new energy access scenarios. Background Art

[0002] New energy access scenarios mainly include cases where solar photovoltaic power generation, wind power generation, biomass power generation, etc. are connected to the power system. In the process of new energy accessing the power grid, it is crucial to ensure the safe and stable operation of the power grid. However, new energy power generation has characteristics such as intermittency, volatility, and uncertainty. For example, solar photovoltaic power generation is affected by light intensity, weather changes, and day-night alternation, and wind power generation depends on changes in wind speed and direction. These factors result in complex and variable electrical signals output by new energy power generation.

[0003] Traditional power system monitoring and fault diagnosis devices are mainly designed for traditional energy power generation and are difficult to meet the requirements of new energy access scenarios.

[0004] For the analysis and processing of electrical signals, some existing decomposition methods lack consideration of the energy change characteristics of new energy electrical signals, thus being unable to effectively extract key information in the electrical signals. In terms of fault diagnosis, traditional diagnosis methods have a low diagnostic accuracy when dealing with complex electrical signal faults in new energy access scenarios, resulting in unsatisfactory diagnostic effects. Summary of the Invention

[0005] The main object of the present invention is to provide a device for monitoring and recording waves and fault diagnosis applicable to new energy access scenarios, aiming to overcome the defects of being unable to effectively extract key information of electrical signals and having a low fault diagnosis accuracy at present.

[0006] To achieve the above object, the present invention provides a device for monitoring and recording waves and fault diagnosis applicable to new energy access scenarios, and the device is used for:

[0007] Real-time collect electrical signals in new energy access scenarios to obtain original electrical signals;

[0008] Determine the decomposition layer number and wavelet basis through an adaptive threshold, and decompose the original electrical signal to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signal;

[0009] Based on an extreme learning machine, perform fault diagnosis on the decomposed signals to obtain a fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized through a sparrow search algorithm;

[0010] Time-mark the fault occurrence moment in the fault diagnosis result, and record the electrical signal waveforms in a preset time period before and after the fault occurrence moment to obtain a time-marked fault diagnosis result and recording wave data.

[0011] Further, determining the decomposition level and wavelet basis through an adaptive threshold, and decomposing the original electrical signal to obtain a decomposed signal, includes:

[0012] Calculating the energy distribution of the original electrical signal to obtain energy distribution data;

[0013] Setting an initial threshold range according to the energy distribution data, and equally dividing the initial threshold range into multiple sub-threshold intervals;

[0014] Performing wavelet packet simulation decomposition on each sub-threshold interval based on the original electrical signal to obtain a decomposition effect evaluation index corresponding to each sub-threshold interval; wherein, the decomposition effect evaluation index includes information entropy and root mean square error;

[0015] Selecting the threshold value in the sub-threshold interval corresponding to the optimal decomposition effect evaluation index as the adaptive threshold;

[0016] Determining the decomposition level and wavelet basis according to the adaptive threshold, and performing wavelet packet decomposition on the original electrical signal to obtain a decomposed signal.

[0017] Further, determining the decomposition level and wavelet basis according to the adaptive threshold, and performing wavelet packet decomposition on the original electrical signal to obtain a decomposed signal, includes:

[0018] Performing discretization processing on the adaptive threshold to obtain multiple discrete threshold points;

[0019] For each discrete threshold point, performing wavelet packet decomposition on the original electrical signal with different levels, and successively increasing the level until the preset maximum level is reached;

[0020] At each level, selecting different wavelet bases according to the discrete threshold point for wavelet packet decomposition to obtain multiple decomposition results;

[0021] Calculating the energy proportion set for each decomposition result;

[0022] Selecting the level and wavelet basis corresponding to the highest energy proportion set as the final decomposition level and wavelet basis;

[0023] Using the final decomposition level and wavelet basis to perform wavelet packet decomposition on the original electrical signal to obtain the decomposed signal.

[0024] Further, the fault diagnosis based on the extreme learning machine for the decomposed signal to obtain a fault diagnosis result includes:

[0025] Feature extraction is performed on the decomposed signal based on the extreme learning machine to obtain a feature vector set; the feature vector set includes amplitude features, frequency features, phase features, and time-frequency domain features extracted by wavelet transform;

[0026] Fault diagnosis is performed on the feature vector set based on the extreme learning machine to obtain a fault diagnosis result.

[0027] Furthermore, the parameters of the extreme learning machine are optimized by the sparrow search algorithm, specifically:

[0028] Individuals of the sparrow population are randomly generated in the sparrow search algorithm; each individual of the sparrow population represents a set of extreme learning machine parameters;

[0029] The diagnostic accuracy rate of the extreme learning machine corresponding to each individual of the sparrow population on the training set is calculated as the fitness value;

[0030] According to the discoverer-joiner-vigilant rule of the sparrow search algorithm, the individuals of the sparrow population are updated and iterated to the maximum number of iterations to obtain the optimal extreme learning machine parameters.

[0031] Furthermore, the new energy access scenario is an application scenario where a new energy power system is connected to the power grid.

[0032] Furthermore, after obtaining the fault diagnosis result and the oscillogram data with time stamps, it includes:

[0033] The oscillogram data is discretized to obtain discrete data points; curve fitting is performed according to the discrete data points to generate a signal curve;

[0034] Hash calculation is performed on the fault diagnosis result and the oscillogram data with time stamps to obtain the corresponding hash value;

[0035] Based on the hash value and the signal curve, a communication key is generated;

[0036] Based on the communication key, the fault diagnosis result and the oscillogram data are encrypted and transmitted to the management terminal.

[0037] The present invention also provides a method for monitoring, recording and fault diagnosis applicable to new energy access scenarios, including the following steps: collecting electrical signals in new energy access scenarios in real time to obtain original electrical signals; determining the decomposition level and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signals; performing fault diagnosis on the decomposed signals based on an extreme learning machine to obtain a fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized through a sparrow search algorithm; marking the time of fault occurrence in the fault diagnosis result, and recording the electrical signal waveforms in a preset time period before and after the time of fault occurrence to obtain a fault diagnosis result with time marks and recording data.

[0038] The device for monitoring, recording and fault diagnosis applicable to new energy access scenarios provided by the present invention is used for: collecting electrical signals in new energy access scenarios in real time to obtain original electrical signals; determining the decomposition level and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signals; performing fault diagnosis on the decomposed signals based on an extreme learning machine to obtain a fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized through a sparrow search algorithm; marking the time of fault occurrence in the fault diagnosis result, and recording the electrical signal waveforms in a preset time period before and after the time of fault occurrence to obtain a fault diagnosis result with time marks and recording data. In the present invention, by determining the decomposition level and wavelet basis through an adaptive threshold and decomposing the original electrical signals, the key information of the electrical signals can be effectively extracted; by performing fault diagnosis on the decomposed signals through an extreme learning machine, the obtained fault diagnosis result can improve the accuracy rate. Description of the Drawings

[0039] Figure 1 It is a schematic flowchart of monitoring, recording and fault diagnosis applicable to new energy access scenarios in an embodiment of the present invention;

[0040] Figure 2 It is a schematic diagram of the method steps of monitoring, recording and fault diagnosis applicable to new energy access scenarios in an embodiment of the present invention;

[0041] Figure 3 It is a schematic block diagram of the structure of a computer device in an embodiment of the present invention.

[0042] The implementation, functional features and advantages of the present invention will be further described in conjunction with the embodiments with reference to the drawings. Detailed Embodiments

[0043] In order to make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0044] Combined with Figure 1 the flow schematic diagram of monitoring, recording, and fault diagnosis applicable to the new energy access scenario in

[0045] collect the electrical signals in the new energy access scenario in real time to obtain the original electrical signals;

[0046] determine the decomposition level and wavelet basis through an adaptive threshold, and decompose the original electrical signal to obtain the decomposed signal; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signal;

[0047] perform fault diagnosis on the decomposed signal based on the extreme learning machine to obtain the fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized through the sparrow search algorithm;

[0048] mark the time of fault occurrence in the fault diagnosis result, and record the electrical signal waveforms in the preset time period before and after the time of fault occurrence to obtain the fault diagnosis result with time mark and the recording data.

[0049] In this embodiment, a monitoring, recording, and fault diagnosis device specifically for the new energy access scenario is proposed. The device aims to effectively address the complex electrical signal monitoring and fault diagnosis problems brought about by the access of new energy power generation to the power grid, ensure the safe and stable operation of the power grid, and ensure the security of relevant data transmission.

[0050] First of all, in the new energy access scenario, whether it is solar energy, wind energy, or other new energy power generation forms, the electrical signals generated have unique characteristics. First, through the signal acquisition module, these electrical signals (such as voltage, current, etc.) can be obtained in real time and accurately. The acquired electrical signals contain rich information, such as amplitude, frequency, phase, etc. These original electrical signals are the basis for subsequent processing and analysis. The design of the acquisition module takes into account factors such as the high frequency, wide dynamic range, and possible noise interference of new energy electrical signals to ensure that the quality of the acquired original data meets the requirements of subsequent processing.

[0051] Furthermore, determine the decomposition level and wavelet basis through an adaptive threshold, and decompose the original electrical signal to obtain the decomposed signal; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signal.

[0052] Due to the complexity and variability of the electrical signals in the above new energy access scenarios, traditional fixed decomposition methods cannot effectively extract the key information therein. The present invention adopts an adaptive threshold method to determine the number of decomposition layers and wavelet bases of wavelet packet decomposition. Dynamically adjusting the threshold according to the energy distribution of the signal is the core innovation of this step. By analyzing the energy distribution of the original electrical signal, the energy concentration degree of the signal in different frequency bands and at different times can be understood. Dynamically determining the threshold according to the energy distribution can make the selection of the decomposition layer number and wavelet basis more conform to the actual characteristics of the signal. For example, when the energy of the signal changes violently in a certain frequency band, the number of decomposition layers can be increased to analyze the signal components in this frequency band more finely; while for the frequency band with relatively stable energy, the number of decomposition layers can be appropriately reduced to improve the processing efficiency. Selecting the appropriate wavelet basis is also equally important. Different wavelet bases have different characteristics in time-frequency analysis. Dynamically selecting the wavelet basis according to the energy distribution can better highlight the local characteristics and detailed information of the signal, so as to obtain a more accurate decomposed signal. The above decomposed signal will contain the information of the original electrical signal in different frequency bands and scales, providing a richer data basis for subsequent fault diagnosis.

[0053] Furthermore, fault diagnosis is carried out on the decomposed signal based on the extreme learning machine to obtain a fault diagnosis result; the above extreme learning machine (ELM) is an efficient machine learning algorithm, which is suitable for quickly and accurately classifying and diagnosing complex data. In this solution, the decomposed signal is input into the extreme learning machine for fault diagnosis. The performance of the extreme learning machine largely depends on its parameter settings. Traditional parameter selection methods often have poor effects. The present invention introduces the sparrow search algorithm (SSA) to optimize the parameters of the extreme learning machine. The sparrow search algorithm simulates the foraging and anti-predation behaviors of a sparrow group, and has strong global search ability and convergence speed. Through the sparrow search algorithm, the parameter combination that makes the diagnosis accuracy of the extreme learning machine the highest when processing the decomposed signal can be found. For example, optimize the parameters such as the number of hidden layer neurons, input weights, and biases of the extreme learning machine, so as to improve the ability to identify the fault modes of new energy electrical signals, and more accurately diagnose information such as the fault type, location, and severity, and obtain a reliable fault diagnosis result.

[0054] Finally, time-stamp the fault occurrence moment in the fault diagnosis result, record the electrical signal waveforms in a preset time period before and after the fault occurrence moment, and obtain the fault diagnosis result with time stamp and the oscillogram data. In the operation of the power grid, accurately recording the time of fault occurrence is crucial for fault analysis and power grid restoration. In this step, after obtaining the fault diagnosis result, the fault occurrence moment is accurately marked. At the same time, considering that the electrical signal waveforms before and after the fault occurrence contain rich fault information, the device will record the electrical signal waveforms in a preset time period before and after the fault occurrence moment. The selection of this preset time period is fully considered, which should include sufficient fault initiation and development information and not be too long to cause excessive data volume. In this way, the fault diagnosis result with time stamp and the oscillogram data obtained can provide comprehensive and accurate data support for subsequent detailed fault analysis, power grid operation state evaluation, and fault cause tracing, etc.

[0055] In this embodiment, through adaptive signal decomposition and optimized fault diagnosis algorithms, the fault characteristics in the new energy electrical signals can be captured more accurately, the ability to detect and identify faults can be improved, and the false alarm and missed alarm situations can be reduced.

[0056] In one embodiment, determining the decomposition layer number and wavelet basis through an adaptive threshold, and decomposing the original electrical signal to obtain decomposed signals includes:

[0057] Calculate the energy distribution of the original electrical signal to obtain energy distribution data;

[0058] Set an initial threshold range according to the energy distribution data, and equally divide the initial threshold range into multiple sub-threshold intervals;

[0059] Perform wavelet packet simulation decomposition on each sub-threshold interval based on the original electrical signal to obtain the decomposition effect evaluation indexes corresponding to each sub-threshold interval; wherein, the decomposition effect evaluation indexes include information entropy and root mean square error;

[0060] Select the threshold in the sub-threshold interval corresponding to the optimal decomposition effect evaluation index as the adaptive threshold;

[0061] Determine the decomposition layer number and wavelet basis according to the adaptive threshold, and perform wavelet packet decomposition on the original electrical signal to obtain decomposed signals.

[0062] In this embodiment, the original electrical signal contains rich information, and its energy distribution is uneven in different frequency bands and over time. Calculating the energy distribution data is a key step in understanding the characteristics of the electrical signal. Through specific mathematical algorithms (such as relevant energy calculation methods based on Fourier transform or wavelet transform, etc.), the energy values of the original electrical signal in each frequency band and the variation of energy over time can be obtained. These energy distribution data will provide a basis for determining the threshold range in the subsequent steps. For example, if the energy is concentrated and changes frequently in a certain frequency band, this will affect the corresponding setting of the threshold range in that frequency band.

[0063] Setting the initial threshold range based on the calculated energy distribution data is an adaptive strategy based on the characteristics of the signal itself. Since the energy distribution of electrical signals varies greatly under different new energy access scenarios, this method can ensure the matching degree between the threshold range and the current signal. Dividing the initial threshold range into multiple sub-threshold intervals can more finely explore the influence of different threshold levels on the decomposition effect. Each sub-threshold interval represents a possible threshold selection range. By performing subsequent operations within these sub-intervals, the decomposition situation under different thresholds can be comprehensively evaluated.

[0064] The above-mentioned wavelet packet simulation decomposition for each sub-threshold interval is an effective method to comprehensively explore different threshold possibilities. Wavelet packet decomposition can decompose the original electrical signal at different scales and frequency bands to obtain more detailed signal component information. And introducing information entropy and root mean square error as evaluation indicators for the decomposition effect is of great significance. Information entropy can measure the uncertainty or information content of the decomposed signal. A lower information entropy indicates that the signal is more orderly, that is, the decomposition effect is better, and different signal components can be separated more clearly. The root mean square error reflects the error magnitude between the decomposed signal and the original signal. A smaller root mean square error means that the decomposed signal can more accurately represent the characteristics of the original signal. By considering these two evaluation indicators simultaneously, the wavelet packet decomposition effect under each sub-threshold interval can be comprehensively and objectively evaluated.

[0065] Furthermore, after obtaining the evaluation indicators of the decomposition effect for each sub-threshold interval, select the threshold in the sub-threshold interval corresponding to the optimal evaluation indicator. The selection of this optimal threshold is the result of comprehensively considering information entropy and root mean square error. For example, the threshold in the sub-threshold interval that minimizes both the information entropy and the root mean square error may be selected as the adaptive threshold. This selection method ensures that the finally determined threshold can achieve the best balance between ensuring the information content and accuracy of the decomposed signal, thus making the decomposition effect optimal.

[0066] Finally, once the adaptive threshold is determined, the decomposition level and wavelet basis can be determined based on this threshold. There is an inherent relationship between the adaptive threshold, the decomposition level, and the wavelet basis. For example, a larger threshold corresponds to a finer decomposition level to better separate the complex components in the signal; at the same time, the threshold may also affect the selection of the wavelet basis, so that the selected wavelet basis can better highlight the signal characteristics under the current threshold. By performing wavelet packet decomposition in this way, decomposition signals that better conform to the characteristics of the original electrical signal can be obtained, and these decomposition signals will be more conducive to subsequent processing steps such as fault diagnosis.

[0067] In this embodiment, it is carried out adaptively based on the energy distribution of the original electrical signal. Compared with the traditional method of fixed threshold, fixed decomposition level, and wavelet basis, this adaptive strategy can better adapt to the complex and variable electrical signals in the new energy access scenario, improving the accuracy and effectiveness of the decomposition. By introducing two evaluation indicators, information entropy and root mean square error, the decomposition effect is evaluated from different perspectives. This multi-dimensional evaluation method avoids the one-sidedness that may be brought by a single evaluation indicator and can more accurately determine the optimal threshold, decomposition level, and wavelet basis, so as to obtain high-quality decomposition signals.

[0068] In one embodiment, the decomposition level and wavelet basis are determined according to the adaptive threshold, and wavelet packet decomposition is performed on the original electrical signal to obtain decomposition signals, including:

[0069] Perform discretization processing on the adaptive threshold to obtain multiple discrete threshold points;

[0070] For each discrete threshold point, perform wavelet packet decomposition on the original electrical signal with different levels, and gradually increase the level until the preset maximum level is reached;

[0071] At each level, select different wavelet bases according to the discrete threshold point to perform wavelet packet decomposition to obtain multiple decomposition results;

[0072] Calculate the energy ratio set for each decomposition result;

[0073] Select the level and wavelet basis corresponding to the highest energy ratio set as the final decomposition level and wavelet basis;

[0074] Use the final decomposition level and wavelet basis to perform wavelet packet decomposition on the original electrical signal to obtain the decomposition signal.

[0075] In this embodiment, first, the adaptive threshold is discretized to obtain a plurality of discrete threshold points. The adaptive threshold is a continuous numerical range or a specific numerical value. However, in the actual processing, in order to more systematically explore its influence on the decomposition level and wavelet basis selection, it needs to be discretized. Through the discretization process, the adaptive threshold can be transformed into a plurality of discrete threshold points. These discrete threshold points form the basis for subsequent analysis. Each discrete threshold point represents a possible threshold selection, which helps to comprehensively evaluate the decomposition situation at different threshold levels.

[0076] For each discrete threshold point, performing wavelet packet decomposition with different levels is a method to comprehensively explore the decomposition possibilities. Starting from a lower level and gradually increasing the level until the preset maximum level can cover different degrees of decomposition fineness. This is because different signal features may be better reflected at different decomposition levels. For example, for complex new energy access electrical signals, if the decomposition level is too low, it may not be possible to separate the subtle fault features hidden in the signal; while too many decomposition levels may lead to excessive computational complexity and may introduce unnecessary noise. By this way of gradually increasing the level, the most suitable decomposition level range can be found for each discrete threshold point.

[0077] After determining each discrete threshold point and the corresponding decomposition level, different wavelet bases are further selected for decomposition according to the discrete threshold point. Different wavelet bases have different time-frequency localization characteristics and are suitable for analyzing different types of signal features. For example, some wavelet bases perform better when dealing with signals with sharp mutations, while other wavelet bases are more suitable for dealing with signals with smooth changes. By trying different wavelet bases at each level, multiple decomposition results can be obtained, so as to comprehensively explore the optimal wavelet basis selection at this discrete threshold point and level.

[0078] Calculating the energy ratio set is a key step in evaluating the quality of the decomposition result. By calculating the energy ratio set for each decomposition result, the energy distribution of the signal in different frequency bands can be understood. The energy ratio set reflects the proportion of the energy of the decomposed signal in each frequency band to the total energy. It can intuitively show whether the decomposition result can effectively concentrate the energy of the original electrical signal in a specific frequency band, thereby highlighting the main features of the signal and potential fault-related features. For example, if a decomposition result can make the energy ratio of the fault-related frequency band significantly higher than other frequency bands, then this decomposition result is more conducive to subsequent fault diagnosis.

[0079] After obtaining all the decomposition results and their corresponding energy ratio sets, select the case with the highest energy ratio set. This means that the selected number of layers and wavelet basis combination can make the decomposed signal most conform to the expectation in terms of energy distribution, that is, it can concentrate the energy on the key frequency bands to the greatest extent and effectively highlight the important information in the original electrical signal. This selection method ensures that the finally determined decomposition level and wavelet basis are the optimal combination selected from multiple possibilities under the current adaptive threshold, and can provide the most valuable decomposed signal for subsequent fault diagnosis and other operations.

[0080] Finally, use the determined final decomposition level and wavelet basis to perform wavelet packet decomposition on the original electrical signal, and the obtained decomposed signal will have the best feature representation. These decomposed signals can more accurately reflect the fault information in the original electrical signal during the subsequent fault diagnosis process, improving the accuracy and reliability of fault diagnosis.

[0081] In this embodiment, through the discretization of the adaptive threshold and the comprehensive exploration of the combinations of different numbers of layers and wavelet bases at each discrete threshold point, a wide range of possibilities are covered. It is ensured that no possible optimal decomposition parameter combination is missed, and the probability of finding the most suitable decomposition method for the original electrical signal is increased. Compared with traditional empirical or fixed-parameter decomposition methods, this scheme can better adapt to the complexity and variability of the electrical signal with new energy access, and provide a better decomposed signal basis for subsequent fault diagnosis and other operations.

[0082] In one embodiment, the fault diagnosis of the decomposed signal based on the extreme learning machine to obtain the fault diagnosis result includes:

[0083] Feature extraction is performed on the decomposed signal based on the extreme learning machine to obtain a set of feature vectors; the set of feature vectors includes amplitude features, frequency features, phase features, and time-frequency domain features extracted by wavelet transform;

[0084] Fault diagnosis is performed on the set of feature vectors based on the extreme learning machine to obtain the fault diagnosis result.

[0085] In this embodiment, first, feature extraction is performed on the decomposed signal based on the extreme learning machine to obtain a set of feature vectors; the set of feature vectors includes amplitude features, frequency features, phase features, and time-frequency domain features extracted by wavelet transform.

[0086] The above amplitude feature is one of the basic attributes of the electrical signal and reflects the strength of the signal. For the decomposed signal, by analyzing and extracting the amplitudes of its various components, the strength information of the signal at different frequency bands and time scales can be obtained. These amplitude features help to identify whether there are abnormal strength changes in the signal. For example, in the case of a fault, the amplitudes of certain frequency bands may increase or decrease significantly.

[0087] The above frequency feature is another key feature of the electrical signal. The decomposed signal has its own frequency components in different frequency bands. Extracting the frequency feature can help understand the frequency distribution of the signal. In the scenario of new energy access, different types of faults may cause frequency deviation, harmonic generation, or changes in specific frequency components. By accurately extracting the frequency feature, the frequency change information related to the fault can be captured.

[0088] The above phase feature reflects the relative position relationship of the electrical signal in time. In the complex electrical signal of new energy access, the change in phase may imply synchronization problems between different power generation units or signal phase deviation caused by faults. Extracting the phase feature helps discover these potential fault clues, especially in the new energy access system involving multiple power sources or complex circuit structures.

[0089] Wavelet transform is a powerful signal processing tool that can analyze the signal in both the time domain and the frequency domain simultaneously. For the decomposed signal, wavelet transform can be used to extract the time-frequency domain features at different time and frequency scales. These features can more comprehensively describe the local characteristics of the signal, such as signal mutations, transient phenomena, etc. When a fault occurs, the electrical signal often shows specific patterns in the time-frequency domain. The time-frequency domain features extracted by wavelet transform can effectively capture these fault-related patterns, providing more abundant information for subsequent fault diagnosis.

[0090] Extreme learning machine is an efficient machine learning algorithm with the ability of fast learning and accurate classification. After inputting the feature vector set containing amplitude, frequency, phase, and time-frequency domain features into the extreme learning machine, it can analyze the input feature vectors according to the pre-trained model or the patterns learned during the training process. In the training stage, the extreme learning machine learns through a large number of feature vector data of known faults and normal states, adjusts its own parameters and weights to establish the mapping relationship between the feature vectors and the fault types. In actual diagnosis, for the newly input feature vector set from the decomposed signal, the extreme learning machine uses the established mapping relationship to make a judgment and output the corresponding fault diagnosis result. For example, it can judge whether there are different types of faults such as short circuit faults, open circuit faults, abnormal power fluctuations, etc., and determine information such as the severity of the fault and the possible location.

[0091] In this embodiment, by extracting amplitude, frequency, phase, and time-frequency domain features of wavelet transform, multiple important dimensions of the electrical signal are covered, enabling the comprehensive characterization of the characteristics of the decomposed signal. This multi-dimensional feature representation method can more accurately capture various subtle changes related to faults in complex electrical signals in the new energy access scenario, reducing misdiagnosis cases caused by insufficient feature information. The fast learning and efficient classification capabilities of the extreme learning machine give it significant advantages in processing complex feature vector sets. Compared with traditional diagnostic methods, it can process a large amount of feature data faster and has a high diagnostic accuracy. At the same time, the extreme learning machine has strong adaptability to data and can well handle the diversity and complexity of electrical signal features in the new energy access scenario, providing reliable results for fault diagnosis.

[0092] In one embodiment, the parameters of the extreme learning machine are optimized by the sparrow search algorithm, specifically as follows:

[0093] Individuals of the sparrow population are randomly generated in the sparrow search algorithm; each individual of the sparrow population represents a set of parameters of the extreme learning machine;

[0094] The diagnostic accuracy of the extreme learning machine corresponding to each individual of the sparrow population on the training set is calculated as the fitness value;

[0095] According to the discoverer-joiner-scout rule of the sparrow search algorithm, the individuals of the sparrow population are updated and iterated to the maximum number of iterations to obtain the optimal parameters of the extreme learning machine.

[0096] In this embodiment, by simulating the behavior of the sparrow group, the optimal parameter combination of the extreme learning machine is searched in the parameter space, thereby improving the performance of the extreme learning machine in fault diagnosis.

[0097] First, individuals of the sparrow population are randomly generated in the sparrow search algorithm; each individual of the sparrow population represents a set of parameters of the extreme learning machine. Among them, the sparrow search algorithm is a swarm intelligence-based optimization algorithm, and a certain scale of sparrow population needs to be generated at the beginning stage. Each sparrow individual in here corresponds to a set of parameters of the extreme learning machine in the algorithm. For example, the parameters of the extreme learning machine include the number of hidden layer neurons, input weights, biases, etc. By randomly generating these parameter combinations (i.e., individuals of the sparrow population), points can be widely scattered in the parameter space, covering different possible solution regions, laying a foundation for finding the optimal parameters subsequently. This random generation method ensures the initial diversity of the search and avoids falling into local optimal solutions.

[0098] Then, calculate the diagnostic accuracy rate of the extreme learning machine corresponding to each individual in the sparrow population on the training set as the fitness value. For each generated sparrow population individual (i.e., a set of extreme learning machine parameters), apply it to the extreme learning machine and perform fault diagnosis operations on the training set. The training set is a large amount of sample data containing known fault types and normal states. By calculating the diagnostic accuracy rate of the extreme learning machine on this training set, the goodness of this set of parameters for the fault diagnosis task can be measured. This diagnostic accuracy rate is the fitness value of the sparrow individual in the sparrow search algorithm. The higher the fitness value, the stronger the diagnostic ability of this set of extreme learning machine parameters under the current training set, which also means that this set of parameters is more likely to be the optimal parameter combination we are looking for.

[0099] Furthermore, update the sparrow population individuals according to the discoverer-joiner-scout rule of the sparrow search algorithm and iterate to the maximum number of iterations to obtain the optimal extreme learning machine parameters. In the sparrow group, the discoverer usually has better search ability and resource acquisition ability. In the algorithm, the discoverer is responsible for searching in a wider area and guiding the whole group towards a more promising area. According to the discoverer rule, some sparrow individuals (discoverers) will update their positions (i.e., update the represented extreme learning machine parameters) according to the current search situation, such as exploring new parameter combinations towards areas with higher fitness values. The joiners are the individuals who follow the discoverers to search. They will adjust their positions according to the positions of the discoverers and their own fitness situations. This simulates the information sharing and cooperation behaviors among individuals in the sparrow group, enabling the entire population to more effectively explore the parameter space. The scouts are responsible for monitoring the dangers in the surrounding environment (understood as preventing getting stuck in local optima in the algorithm). When potential dangers are detected (such as the population being too concentrated in a certain local area), the scouts will prompt some sparrow individuals to change their positions, jump out of the local optimal area, and search in a wider space again.

[0100] By continuously updating the positions (parameter combinations) of the sparrow population individuals according to the discoverer-joiner-scout rule, the entire population continuously searches and evolves in the parameter space. Keep iterating until the preset maximum number of iterations is reached. After the iteration ends, the extreme learning machine parameters represented by the sparrow individual with the highest fitness value in the population are the optimized optimal parameters. The above optimal parameters can enable the extreme learning machine to show better performance in fault diagnosis, improving the diagnostic accuracy rate and reliability.

[0101] In this embodiment, the sparrow search algorithm has strong global search ability by simulating the complex behaviors of a sparrow population. Compared with traditional gradient-based optimization algorithms, it is not easily trapped in local optimal solutions. In the complex high-dimensional space search problem of extreme learning machine parameter optimization, this global search ability can explore the parameter space more comprehensively, find a better parameter combination, and thus improve the performance of the extreme learning machine within the entire fault diagnosis range. Using the sparrow search algorithm to optimize the extreme learning machine parameters avoids the cumbersome process and uncertainty of manual parameter adjustment. Through automatic search and iterative update, it can quickly find the parameters that maximize the diagnosis accuracy of the extreme learning machine under a given training set, improving the efficiency and accuracy of parameter optimization. The parameters of the extreme learning machine optimized by the sparrow search algorithm can significantly improve the diagnosis accuracy of the extreme learning machine in fault diagnosis.

[0102] In one embodiment, the new energy access scenario is an application scenario where a new energy power system is connected to a power grid.

[0103] In this embodiment, the new energy access scenario refers to the application environment where a new energy power system is connected to a traditional power grid to achieve power transmission and distribution. The new energy power system here mainly includes systems that generate electricity using renewable energy such as solar photovoltaic power generation systems, wind power generation systems, biomass power generation systems, and tidal power generation systems. The power grid covers the entire power transmission and distribution system from high-voltage transmission grids to medium- and low-voltage distribution grids.

[0104] In one embodiment, after obtaining the fault diagnosis result and the oscillogram data with time stamps, it includes:

[0105] Perform discrete processing on the oscillogram data to obtain discrete data points; perform curve fitting based on the discrete data points to generate a signal curve;

[0106] Perform hash calculation on the fault diagnosis result and the oscillogram data with time stamps to obtain corresponding hash values;

[0107] Generate a communication key based on the hash value and the signal curve;

[0108] Encrypt the fault diagnosis result and the oscillogram data based on the communication key and transmit them to the management terminal.

[0109] In this embodiment, first, the recorded wave data is discretized to obtain discrete data points; curve fitting is performed according to the discrete data points to generate a signal curve. The recorded wave data is a record of the electrical signal waveforms before and after the fault occurrence time, with a large amount of continuous data. Discretizing it means selecting data points from the continuous waveform data according to certain rules and intervals to obtain discrete data points. This can simplify the data representation without losing key information, facilitating subsequent calculations and processing. For example, data points can be selected at fixed time intervals, or the selection method of discrete points can be adaptively determined according to the change characteristics of the signal. After obtaining the discrete data points, a signal curve is generated through curve fitting technology. Curve fitting is to use mathematical methods to find a curve that best represents the distribution law of these discrete data points. Commonly used curve fitting methods include polynomial fitting, spline fitting, etc. Through curve fitting, the change trend of the electrical signal can be presented more intuitively, and in the subsequent key generation process, the characteristic information of the curve can be utilized. This conversion from discrete data to a continuous curve not only retains the core information of the original recorded wave data but also provides a suitable data form for subsequent security processing.

[0110] Furthermore, hash calculation is performed on the fault diagnosis result with time stamps and the recorded wave data to obtain the corresponding hash value. Hash calculation is a one-way function that maps data of any length to a hash value of a fixed length. Performing hash calculation on the fault diagnosis result with time stamps and the recorded wave data can obtain a hash value that uniquely identifies these data. The hash value has the following characteristics: one is that different data inputs will not produce the same hash value (collision resistance); the other is that the hash value of a given data is uniquely determined and the original data cannot be deduced from the hash value in reverse (one-wayness). By calculating the hash value, it can provide basic data for subsequent generation of communication keys and also ensure the integrity of the data, because any tampering with the data will cause the hash value to change.

[0111] Furthermore, a communication key is generated based on the hash value and the signal curve. Generating a communication key using the hash value and the signal curve together is one of the cores of this solution. The hash value represents the characteristic identifier of the fault diagnosis result and the recorded wave data, while the signal curve contains the waveform information of the recorded wave data. Combining the two can generate a communication key with high security and uniqueness. For example, the hash value and certain characteristic parameters of the signal curve (such as the peak value, slope, etc. of the curve) can be combined and transformed according to a preset algorithm to generate a complex communication key. The key generated in this way is closely related to the original data and at the same time has sufficient randomness and complexity, making it difficult to be cracked.

[0112] Finally, encrypt the fault diagnosis result and the waveform recording data based on the communication key and transmit them to the management terminal. After generating the communication key, use this key to encrypt the fault diagnosis result and the waveform recording data. The encryption process combines the original data with the communication key through a specific encryption algorithm (such as a symmetric encryption algorithm or an asymmetric encryption algorithm) to convert the original data into ciphertext form. In this way, even if the data is intercepted during the data transmission process, the attacker cannot obtain the content therein. The encrypted ciphertext data is then transmitted to the management terminal, and the management terminal can restore the original fault diagnosis result and waveform recording data only by using the corresponding decryption key (the same as the communication key in symmetric encryption and paired with the communication key in asymmetric encryption), thus ensuring the security of data transmission.

[0113] In this embodiment, through operations such as hash calculation, curve fitting, and key generation, the generated communication key has high security. Encrypting the data based on this key can effectively prevent the data from being stolen, tampered with, or forged during the transmission process, ensuring the confidentiality and integrity of the fault diagnosis result and the waveform recording data, which is crucial for the security management of the power system. The generation method of the communication key is closely related to the fault diagnosis result and the waveform recording data itself, and this correlation makes the key unique and unpredictable. Compared with traditional fixed keys or simple generation methods, it can better resist attacks against the key and enhance the reliability of the entire data transmission security mechanism.

[0114] In one embodiment, generating the communication key based on the hash value and the signal curve includes:

[0115] Extract numerical characters from the hash value and combine them into multiple discrete numbers according to a preset combination rule; perform curve simulation on the multiple discrete numbers to obtain a digital curve;

[0116] Add the digital curve and the signal curve to the same coordinate system;

[0117] Translate the signal curve according to a preset rule so that the signal curve has the most intersections with the digital curve, and use the intersections at this time as the target intersections;

[0118] Obtain the ordinate values of each target intersection on the digital curve and combine them in sequence to obtain the communication key.

[0119] In this embodiment, by extracting numerical characters from the hash value and performing curve simulation, interacting with the signal curve in the same coordinate system, finally determining the target intersections and obtaining the combination of ordinate values to form the communication key, it provides a highly secure encryption guarantee for data transmission.

[0120] First, extract the numeric characters from the hash value and combine them into multiple discrete numbers according to a preset combination rule; perform curve simulation on the multiple discrete numbers to obtain a digital curve. A hash value is usually a string composed of numbers and letters. First, extract the numeric characters from the hash value, and these numeric characters represent a part of the characteristic information of the hash value. Then, according to the preset combination rule, combine these numeric characters into multiple discrete numbers. This combination rule can be designed according to specific application requirements and security requirements. For example, it can be combined according to a specific order, interval, or algorithm. The obtained set of discrete numbers provides the basic data for subsequent curve simulation.

[0121] Perform curve simulation on the multiple combined discrete numbers, that is, regard these numbers as the ordinate values of points in a two-dimensional coordinate system, and the abscissa can be assigned values in a certain order (such as the natural number sequence). By connecting these points, a digital curve can be obtained. This digital curve represents a visual representation form of the hash value and also provides a reference curve for the interactive operation with the signal curve.

[0122] Furthermore, add the digital curve and the signal curve to the same coordinate system so that the two can be compared and operated in the same space. The purpose of doing this is to generate a communication key using the relative position relationship between the two curves. In the same coordinate system, the shapes, trends, and intersection points of the two curves can be clearly observed. The above coordinate system can be set according to specific requirements. For example, appropriate axis ranges, scales, etc. can be selected to better display the characteristics of the two curves.

[0123] Furthermore, translate the signal curve according to a preset rule so that the signal curve and the digital curve generate the most intersection points, and use the intersection points at this time as the target intersection points. The above preset rule is to find the position where the signal curve and the digital curve generate the most intersection points. By translating the signal curve in a certain direction and step size, the relative position relationship between the two curves can be changed. During the translation process, continuously calculate the number of intersection points of the two curves. This translation operation can be realized through a mathematical algorithm. For example, it can be translated according to a fixed pixel unit or a certain ratio.

[0124] When the signal curve is translated to the position where it generates the most intersection points with the digital curve, determine these intersection points as the target intersection points. These target intersection points represent the coincidence points of the two curves at a specific position, and they contain the common information of the hash value and the signal curve. By selecting the position with the most intersection points, the characteristics of the two curves can be utilized to the greatest extent to generate a communication key, improving the complexity and security of the key.

[0125] Finally, obtain the ordinate values of each target intersection on the digital curve and combine them in sequence to obtain the communication key. For each target intersection, obtain its ordinate value on the digital curve. The above ordinate value is a specific value of the digital curve at the intersection, representing the characteristics of the hash value at that position. Since the digital curve is generated from the hash value, these ordinate values are closely related to the hash value. Combine the ordinate values of each target intersection in a certain order to obtain the final communication key. The above combination order can be determined according to a preset rule. For example, it can be combined in ascending order of the abscissa value of the intersection. The communication key generated in this way not only contains the information of the hash value but also is related to the positional relationship of the signal curve, with high complexity and uniqueness, making it difficult to be cracked.

[0126] In this embodiment, by deeply integrating the hash value and the signal curve, the generated communication key has extremely high security. It is difficult for an attacker to deduce the communication key by analyzing a single hash value or signal curve, greatly increasing the security of data transmission. Due to the uniqueness of the hash value, the complexity of the signal curve, and the random translation and intersection selection operations in the process of generating the communication key, the generated communication key has uniqueness and unpredictability. Even for the same fault diagnosis result and waveform recording data, the communication key generated each time may be different, further enhancing the security of the data.

[0127] Refer to Figure 2 , the present invention also provides a method for monitoring waveform recording and fault diagnosis applicable to the new energy access scenario, including the following steps: Collect the electrical signals in the new energy access scenario in real time to obtain the original electrical signals; Determine the decomposition layer number and wavelet basis through an adaptive threshold, and decompose the original electrical signals to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signal; Perform fault diagnosis on the decomposed signals based on the extreme learning machine to obtain the fault diagnosis result; wherein, optimize the parameters of the extreme learning machine through the sparrow search algorithm; Mark the time of fault occurrence in the fault diagnosis result, and record the waveform of the electrical signal in the preset time period before and after the time of fault occurrence to obtain the fault diagnosis result and waveform recording data with time marks.

[0128] In this embodiment, for the specific implementation of each step in the above method embodiment, please refer to that described in the above device embodiment, and details will not be repeated here.

[0129] Refer to Figure 3 , the embodiment of the present invention also provides a computer device, which can be a server, and its internal structure can be as Figure 3As shown in the figure. The computer device includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program, when executed by the processor, implements the above method.

[0130] Those skilled in the art can understand that Figure 3 the structure shown in the figure is only a block diagram of some structures related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0131] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0132] In summary, for the device for monitoring and recording waves and fault diagnosis applicable to the new energy access scenario provided in the embodiments of the present invention, it is used for: real-time collecting electrical signals in the new energy access scenario to obtain original electrical signals; determining the decomposition layer number and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the signals; performing fault diagnosis on the decomposed signals based on an extreme learning machine to obtain a fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized through a sparrow search algorithm; marking the time of fault occurrence in the fault diagnosis result, and recording the waveforms of electrical signals in a preset period before and after the time of fault occurrence to obtain a fault diagnosis result with a time mark and recording wave data. In the present invention, by determining the decomposition layer number and wavelet basis through an adaptive threshold and decomposing the original electrical signals, the key information of the electrical signals can be effectively extracted; by performing fault diagnosis on the decomposed signals through an extreme learning machine, the obtained fault diagnosis result can improve the accuracy rate.

[0133] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided by the present invention and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0134] It should be noted that in this article, the terms "including", "comprising", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that includes a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, apparatus, article, or method that includes the element.

[0135] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A device for monitoring and recording waves and fault diagnosis applicable to new energy access scenarios, characterized in that, The device is used for: Collecting electrical signals in the new energy access scenario in real time to obtain the original electrical signals; the original electrical signals include voltage and current; Determining the decomposition layer number and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals; wherein, the adaptive threshold is dynamically adjusted according to the energy distribution of the original electrical signals; Performing fault diagnosis on the decomposed signals based on the extreme learning machine to obtain a fault diagnosis result; wherein, the parameters of the extreme learning machine are optimized by the sparrow search algorithm; Marking the time of fault occurrence in the fault diagnosis result, and recording the electrical signal waveforms in a preset time period before and after the time of fault occurrence to obtain a fault diagnosis result with time marking and oscillogram data.

2. The device for monitoring, recording and fault diagnosis applicable to new energy access scenarios according to claim 1, characterized in that The process of determining the decomposition layer number and wavelet basis through an adaptive threshold, and decomposing the original electrical signals to obtain decomposed signals includes: Calculating the energy distribution of the original electrical signals to obtain energy distribution data; Setting an initial threshold range according to the energy distribution data, and equally dividing the initial threshold range into multiple sub-threshold intervals; Performing wavelet packet simulation decomposition on each sub-threshold interval based on the original electrical signals to obtain a decomposition effect evaluation index corresponding to each sub-threshold interval; wherein, the decomposition effect evaluation index includes information entropy and root mean square error; Selecting the threshold in the sub-threshold interval corresponding to the optimal decomposition effect evaluation index as the adaptive threshold; Determining the decomposition layer number and wavelet basis according to the adaptive threshold, and performing wavelet packet decomposition on the original electrical signals to obtain decomposed signals.

3. The device for monitoring, recording and fault diagnosis applicable to the new energy access scenario according to claim 2, wherein The process of determining the decomposition layer number and wavelet basis according to the adaptive threshold, and performing wavelet packet decomposition on the original electrical signals to obtain decomposed signals includes: Performing discretization processing on the adaptive threshold to obtain multiple discrete threshold points; For each discrete threshold point, performing wavelet packet decomposition on the original electrical signals with different layer numbers, and successively increasing the layer number until the preset maximum layer number is reached; At each layer number, selecting different wavelet bases according to the discrete threshold points to perform wavelet packet decomposition to obtain multiple decomposition results; Calculating the energy ratio set under each decomposition result; Selecting the layer number and wavelet basis corresponding to the highest energy ratio set as the final decomposition layer number and wavelet basis; Using the final decomposition layer number and wavelet basis to perform wavelet packet decomposition on the original electrical signals to obtain the decomposed signals.

4. The device for monitoring, recording and fault diagnosis applicable to new energy access scenarios according to claim 1, wherein The process of performing fault diagnosis on the decomposed signals based on the extreme learning machine to obtain a fault diagnosis result includes: Extracting features from the decomposed signals based on the extreme learning machine to obtain a feature vector set; the feature vector set includes amplitude features, frequency features, phase features, and time-frequency domain features extracted by wavelet transform; Performing fault diagnosis on the feature vector set based on the extreme learning machine to obtain a fault diagnosis result.

5. The device for monitoring, recording waves and fault diagnosis applicable to new energy access scenarios according to claim 1, characterized in that, Optimizing the parameters of the extreme learning machine by the sparrow search algorithm, specifically: Randomly generating sparrow population individuals in the sparrow search algorithm; each sparrow population individual represents a set of extreme learning machine parameters; Calculating the diagnostic accuracy rate of the extreme learning machine corresponding to each sparrow population individual on the training set as the fitness value; Update the individuals of the sparrow population according to the discoverer-joiner-scout rule of the sparrow search algorithm, and iterate to the maximum number of iterations to obtain the optimal parameters of the extreme learning machine.

6. The device for monitoring, recording and fault diagnosis applicable to new energy access scenarios according to claim 1, wherein, The new energy access scenario is the application scenario where a new energy power system is connected to the power grid.

7. The device for monitoring, recording and fault diagnosis applicable to the new energy access scenario according to claim 1, wherein, After obtaining the fault diagnosis result with time stamp and the oscillogram data, it includes: Perform discrete processing on the oscillogram data to obtain discrete data points; perform curve fitting based on the discrete data points to generate a signal curve; Perform hash calculation on the fault diagnosis result with time stamp and the oscillogram data to obtain the corresponding hash value; Generate a communication key based on the hash value and the signal curve; Encrypt the fault diagnosis result and the oscillogram data based on the communication key and transmit them to the management end.

8. A method for monitoring, recording waveforms and fault diagnosis applicable to new energy access scenarios, characterized in that, It includes the following steps: Collect the electrical signals in the new energy access scenario in real time to obtain the original electrical signals; The original electrical signals include voltage and current; Determine the decomposition level and wavelet basis through an adaptive threshold, and decompose the original electrical signals to obtain decomposed signals; among them, the adaptive threshold is dynamically adjusted according to the energy distribution of the original electrical signals; Perform fault diagnosis on the decomposed signals based on the extreme learning machine to obtain a fault diagnosis result; among them, the parameters of the extreme learning machine are optimized through the sparrow search algorithm; Mark the fault occurrence time in the fault diagnosis result, and record the electrical signal waveforms in the preset time period before and after the fault occurrence time to obtain the fault diagnosis result with time stamp and the oscillogram data.

Images