Early warning methods, systems, equipment and media for potential safety hazards of equipment electricity use
The use of a carbon nanotube array current detector and chaos theory algorithms, combined with deep learning, enhances the ability to detect subtle current and voltage anomalies, providing timely and accurate early warnings for equipment safety hazards, thereby preventing accidents and ensuring stable operation.
Patent Information
- Application Number
- CN202411356176.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-27
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-09-27
AI Technical Summary
Existing equipment power safety monitoring systems lack the ability to accurately detect subtle current and voltage anomalies, especially in complex environments, and fail to adapt to dynamic operating conditions, leading to inadequate early warning of potential safety hazards.
Implementing ultra-high frequency dynamic sampling using a carbon nanotube array current detector, combined with chaos theory algorithms to analyze current data and deep learning-based fusion assessment of current and voltage features, to identify nonlinear characteristics and potential safety hazards.
Enables timely and accurate detection of equipment power safety hazards by capturing subtle current changes and analyzing voltage status, providing early warning signals to prevent safety accidents and ensuring the stable operation of equipment, thereby reducing maintenance and operational costs and increasing the safety of the equipment. The system is designed to ensure the safety of the equipment.
Smart Images

Figure CN118884095B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and in particular to an early warning method, system, equipment and medium for hidden dangers of electrical safety in equipment. Background Art
[0002] In the field of traditional equipment power safety monitoring, simple current and voltage monitoring instruments are usually used, which can only display basic electrical parameter values and lack the ability to conduct in-depth analysis and early warning of equipment power safety hazards. For example, the sampling frequency of traditional current monitoring equipment is low, making it difficult to capture the rapidly changing current anomalies during equipment operation, and it is impossible to detect subtle current changes caused by aging of internal equipment components, short circuit precursors, etc.
[0003] In terms of voltage monitoring, traditional methods can only provide real-time voltage values, but cannot deeply explore the potential characteristic information in the voltage data. In complex electromagnetic environments and variable equipment working conditions, traditional monitoring methods are difficult to accurately determine whether there are safety hazards in the equipment.
[0004] At the same time, traditional monitoring systems have limitations in data processing and analysis capabilities. They can usually only perform simple threshold judgments on single current or voltage data, and only sound an alarm when the current or voltage exceeds a preset fixed threshold. However, this method cannot adapt to the complex operating conditions of the equipment and the dynamic characteristics of safety hazards.
[0005] In addition, as the intelligence and automation of equipment continue to improve, the accuracy and timeliness of early warning of equipment power safety hazards are also increasingly required. Existing technical means are difficult to meet the needs of detecting and early warning of potential safety hazards in complex equipment systems and large-scale equipment environments. Summary of the invention
[0006] The main purpose of the present invention is to provide an early warning method, system, device and medium for equipment power safety hazards, aiming to overcome the current defect that it is impossible to accurately judge whether the equipment has safety hazards.
[0007] To achieve the above object, the present invention provides an early warning method for hidden dangers of electrical safety of equipment, comprising the following steps:
[0008] The real-time current of the device is sampled dynamically at ultra-high frequency by a current detector based on a carbon nanotube array to obtain multiple current data sample points;
[0009] Calculating chaos characteristic indexes for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic indexes to form a risk current characteristic set;
[0010] The voltage monitoring module is used to synchronously monitor the real-time voltage of the equipment, and the monitored voltage data is processed by data regularization and feature extraction to obtain the key characteristic parameters of the voltage;
[0011] The risk current feature set and voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain equipment power safety hazard assessment results.
[0012] Furthermore, the current detector and the voltage monitoring module are both IoT devices. The data platform provides computing power to calculate the chaotic characteristic indicators of the current data sample points. The hidden danger investigation system performs fusion analysis to obtain the equipment power safety hidden danger assessment results, and sends them to the alarm system for abnormal alarm processing.
[0013] Furthermore, the chaos characteristic indicators include correlation dimension, Lyapunov index, and approximate entropy value.
[0014] Furthermore, the calculation of the chaos characteristic index for the current data sample points is performed, and based on the chaos characteristic index, nonlinear characteristic points in the current data sample points are obtained to form a risk current characteristic set, including:
[0015] The Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent;
[0016] Analyze the current data sample points based on the correlation integral method to obtain the correlation dimension;
[0017] The current data sample points are calculated using an approximate entropy algorithm based on information theory to obtain an approximate entropy value;
[0018] The correlation dimension, Lyapunov index, and approximate entropy value are calculated based on a weighted comprehensive determination function to obtain a determination value;
[0019] The current data sample points whose determination values are greater than the determination threshold are taken as nonlinear feature points and constitute a risk current feature set.
[0020] Furthermore, the Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent, including:
[0021] Performing data preprocessing based on empirical mode decomposition on the current data sample points to decompose them into a plurality of IMF components to obtain a plurality of IMF subsequences;
[0022] Perform phase space reconstruction on each IMF subsequence, determine the optimal embedding dimension and time delay parameter based on the autocorrelation function method and the false neighbor point method, and construct the phase space matrix;
[0023] Using the Wolf algorithm to perform a preliminary calculation of the maximum Lyapunov exponent of the phase space matrix to obtain a plurality of preliminary Lyapunov exponents;
[0024] The preliminary Lyapunov exponents were screened and optimized based on statistics, and the Lyapunov exponents were obtained after removing outliers.
[0025] Furthermore, the monitored voltage data is subjected to data regularization and feature extraction processing to obtain key voltage characteristic parameters, including:
[0026] De-noising the monitored voltage data to obtain de-noised voltage data;
[0027] Perform synchronous calibration processing on the denoised voltage data based on the timestamp, perform fast Fourier transform on the calibrated voltage data to the frequency domain, extract frequency domain features, and extract time domain features of the calibrated voltage data based on a sliding window;
[0028] The frequency domain features and time domain features are fused to obtain fused feature data; the key characteristic parameters of voltage are screened out from the fused feature data through a machine learning classifier.
[0029] Furthermore, the risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain equipment power safety hazard assessment results, including:
[0030] The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on chaotic neural network dynamics, and the chaotic excitation process is performed on them using the nonlinear dynamic characteristics of chaotic neurons to obtain a chaotic fusion feature state;
[0031] The feature extraction method based on strange attractor analysis is used to reduce the dimension of the chaotic fusion feature state and capture the key features, and the key fusion feature vector is extracted.
[0032] Through the evaluation algorithm based on the combination of fuzzy logic reasoning and probability statistics, the key fusion feature vectors are comprehensively evaluated and processed to determine the probability that the equipment is in different safety hazard states, and the equipment power safety hazard assessment results are obtained.
[0033] Furthermore, after obtaining the equipment power safety hazard assessment result, the method further includes:
[0034] Acquire a first detection identifier of the current detector, and acquire a second detection identifier of the voltage monitoring module;
[0035] Based on the first detection identifier and the second detection identifier, generating an identifier sequence;
[0036] Acquire an operation cycle of the device, and change the identification sequence based on the operation cycle to obtain a changed sequence;
[0037] A key is generated based on the change sequence, and is used to encrypt and store the current data sample points, voltage data, and equipment power safety hazard assessment results.
[0038] The present invention also provides an early warning system for hidden dangers of electrical safety of equipment, comprising:
[0039] A collection unit, used to perform ultra-high frequency dynamic sampling of the real-time current of the device through a current detector based on a carbon nanotube array to obtain multiple current data sample points;
[0040] A calculation unit, used for calculating a chaos characteristic index for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current characteristic set;
[0041] The extraction unit is used to synchronously monitor the real-time voltage of the equipment through the voltage monitoring module, and perform data regularization and feature extraction on the monitored voltage data to obtain key characteristic parameters of the voltage;
[0042] The early warning unit is used to input the risk current feature set and the voltage key feature parameters into a fusion evaluation model based on deep learning for fusion analysis to obtain an equipment power safety hazard evaluation result.
[0043] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.
[0044] 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 steps of any of the above-mentioned methods are implemented.
[0045] The present invention provides an early warning method, system, device and medium for hidden dangers in the use of electricity by equipment, including: performing ultra-high frequency dynamic sampling of the real-time current of the equipment through a current detector based on a carbon nanotube array to obtain multiple current data sample points; calculating the chaotic characteristic index of the current data sample points, and obtaining the nonlinear characteristic points in the current data sample points based on the chaotic characteristic index to form a risk current feature set; synchronously monitoring and processing the real-time voltage of the equipment through a voltage monitoring module, and performing data regularization and feature extraction processing on the monitored voltage data to obtain key voltage characteristic parameters; inputting the risk current feature set and the key voltage characteristic parameters into a fusion evaluation model based on deep learning for fusion analysis to obtain the evaluation result of the hidden dangers in the use of electricity by equipment. In the present invention, the risk current feature set is obtained through current data, the key voltage characteristic parameters are obtained through voltage data, and fusion analysis is performed to obtain the evaluation result of the hidden dangers in the use of electricity by equipment, thereby realizing the detection and early warning of potential safety hazards, and overcoming the defect that the current equipment cannot be accurately judged whether there are safety hazards. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 It is a schematic diagram of the steps of a method for early warning of hidden dangers of electrical safety of equipment in one embodiment of the present invention;
[0047] Figure 2 It is a structural block diagram of an early warning system for hidden dangers of electrical safety of equipment in one embodiment of the present invention;
[0048] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.
[0049] The implementation, functional features and advantages of the present invention will be further described in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0050] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with 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 intended to limit the present invention.
[0051] Reference Figure 1 In one embodiment of the present invention, a method for early warning of hidden dangers of electrical safety in equipment is provided, comprising the following steps:
[0052] Step S1, performing ultra-high frequency dynamic sampling of the real-time current of the device through a current detector based on a carbon nanotube array to obtain a plurality of current data sample points;
[0053] Step S2, calculating the chaos characteristic index of the current data sample points, and obtaining the nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current characteristic set;
[0054] Step S3, synchronously monitoring the real-time voltage of the device through the voltage monitoring module, and performing data regularization and feature extraction on the monitored voltage data to obtain key voltage characteristic parameters;
[0055] Step S4: input the risk current feature set and voltage key feature parameters into a fusion assessment model based on deep learning for fusion analysis to obtain equipment power safety hazard assessment results.
[0056] In this embodiment, a warning method for hidden dangers of power safety in equipment is proposed, which combines data processing algorithms and intelligent fusion evaluation models to achieve accurate monitoring of the power safety status of equipment and early warning of hidden dangers. Through multi-dimensional data collection, complex feature extraction and deep fusion analysis, potential safety risks in equipment operation can be discovered in a timely manner, providing strong technical support for ensuring the stable operation of equipment and avoiding safety accidents.
[0057] As described in step S1 above, the use of a current detector based on a carbon nanotube array has significant advantages. A carbon nanotube array current detector is a device that uses a carbon nanotube array to detect current. Due to its unique nanostructure and excellent electrical properties, the carbon nanotube array can achieve ultra-high frequency dynamic sampling. This makes it possible to capture subtle changes in the device current in a very short time, providing a rich data basis for subsequent detailed analysis. For example, at the moment of device startup or when encountering a momentary high load, traditional current detectors will miss some key current spikes or fluctuation details, while carbon nanotube array current detectors can accurately record these changes, which is of great significance for the early detection of potential faults that may be caused by the sudden large current shock to the internal components of the device.
[0058] As described in step S2 above, by using a professional chaos theory algorithm to conduct an in-depth analysis of the collected current data sample points, the hidden nonlinear features can be excavated. The above nonlinear feature points are often closely related to the abnormal operating state of the equipment. For example, when certain electronic components of the equipment begin to age or show a tendency of local short circuit, the current signal will show changes in chaotic characteristics. By accurately identifying these nonlinear feature points and forming a risk current feature set, we can effectively focus on the key current data parts that may have safety hazards, providing an important basis for subsequent comprehensive evaluation.
[0059] As described in step S3 above, the voltage monitoring module can realize synchronous monitoring of the real-time voltage of the equipment. In the process of data regularization and feature extraction, data processing technology is used to clean, standardize and extract features of the voltage data to obtain key characteristic parameters that can accurately reflect the voltage status of the equipment. For example, by analyzing the time domain characteristics (such as mean, variance, rate of change, etc.) and frequency domain characteristics (such as spectrum distribution, main frequency components, etc.) of the voltage, the stability, periodicity and potential abnormal fluctuations of the equipment voltage can be fully understood.
[0060] As described in step S4 above, the risk current feature set and the voltage key feature parameters are input into the fusion assessment model based on deep learning, realizing the deep fusion analysis of multi-source data. The deep learning model has powerful self-learning and pattern recognition capabilities, and can automatically learn the complex correlation between current and voltage data, as well as the mapping rules between the equipment's power safety hazards. Through model optimization with a large amount of training data, the model can accurately assess the degree of power safety hazards of the equipment and generate detailed assessment results. According to the assessment results, early warning signals of different levels can be issued in a timely manner so that equipment managers can take corresponding measures, such as equipment maintenance, load adjustment or emergency shutdown, etc., thereby effectively preventing the occurrence of safety accidents.
[0061] In this embodiment, the chaos theory is used to deeply mine the current data, which improves the ability to identify early safety hazards of the equipment. The deep learning model is used to deeply integrate and analyze the current and voltage data, which enables a comprehensive and accurate assessment of the power safety status of the equipment. Based on accurate assessment results, early warning signals can be issued in a timely manner, providing a strong guarantee for the safe operation of the equipment.
[0062] In one embodiment, the current detector and the voltage monitoring module are both IoT devices, and the data platform provides computing power to calculate the chaotic characteristic indicators of the current data sample points; the hidden danger investigation system performs fusion analysis to obtain the equipment power safety hidden danger assessment results, and sends them to the alarm system for abnormal alarm processing.
[0063] In this embodiment, the solution process in the above embodiment is completed based on the above-mentioned Internet of Things devices, data platform, hidden danger detection system and alarm system. The above-mentioned Internet of Things devices are connected to the Internet so that the data they collect can be transmitted through the network, which is convenient for rapid transmission and analysis. The above-mentioned data platform provides an efficient computing power foundation, and can calculate the chaos characteristic index for the current data sample points. Then, based on the chaos characteristic index, the nonlinear characteristic points in the current data sample points are obtained to form a risk current feature set. The above-mentioned hidden danger detection system is mainly used to identify hidden dangers for inspection. It is used to perform a fusion analysis of the risk current feature set and the key characteristic parameters of the voltage to obtain the equipment power safety hidden danger assessment results. Furthermore, the above-mentioned hidden danger detection system sends the equipment power safety hidden danger assessment results to the above-mentioned alarm system for abnormal alarm processing.
[0064] In one embodiment, the chaos characteristic indicators include correlation dimension, Lyapunov index, and approximate entropy value.
[0065] In one embodiment, the calculation of the chaos characteristic index for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current characteristic set include:
[0066] The Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent;
[0067] Analyze the current data sample points based on the correlation integral method to obtain the correlation dimension;
[0068] The current data sample points are calculated using an approximate entropy algorithm based on information theory to obtain an approximate entropy value;
[0069] The correlation dimension, Lyapunov index, and approximate entropy value are calculated based on a weighted comprehensive determination function to obtain a determination value;
[0070] The current data sample points whose determination values are greater than the determination threshold are taken as nonlinear feature points and constitute a risk current feature set.
[0071] In this embodiment, the Wolf algorithm is a method for calculating important characteristic indicators (maximum Lyapunov index, also known as Lyapunov index in this article) in a chaotic system. In the analysis of the current data sample points, it tracks the evolution trajectory of adjacent data points in the phase space and calculates the separation speed between them over time. Specifically, the current data sample points are first reconstructed in phase space, and the appropriate embedding dimension and time delay parameter are determined to convert the one-dimensional current data into points in a high-dimensional phase space. Then, an initial point and its neighboring points are selected in the phase space, and the distance change between them is observed as time goes by. According to the exponential growth or contraction rate of the distance, the maximum Lyapunov index is calculated step by step by iteration according to the calculation formula of the Wolf algorithm. This index reflects the degree of chaos of the current data in the phase space, that is, the sensitivity to the initial conditions. A larger positive Lyapunov exponent indicates chaotic characteristics, and the current data has potential nonlinear dynamic changes.
[0072] The correlation integral method is an important means to reveal the inherent structural characteristics of chaotic systems. For current data sample points, this method can be used to analyze the degree of correlation and distribution law between data points. First, the current data is preprocessed, such as normalization, so that its numerical range is unified within an interval that is easy to calculate. Then, a distance metric is defined in the phase space, such as Euclidean distance. For each data point, the distance between it and other data points is calculated. According to the set distance threshold, the number of point pairs with a distance less than the threshold is counted. By changing the distance threshold and the embedding dimension, the calculation is repeated to obtain a series of correlation integral values. Finally, the correlation dimension is determined by using the curve fitting method under double logarithmic coordinates. The correlation dimension reflects the geometric filling characteristics and complexity of the current data in the phase space. Different correlation dimension values may correspond to different current dynamic behavior patterns, which helps to discover potential nonlinear structures.
[0073] The approximate entropy algorithm based on information theory is a method to measure the complexity and randomness of time series. In the analysis of current data sample points, it can evaluate the irregularity and prediction difficulty of current data. For example, a suitable positive integer is selected as the embedding dimension, and a number of continuous data points are selected from the current data to form a vector. Then, the similarity measure between these vectors is calculated, and the distance function is usually used to determine whether two vectors are similar. The approximate entropy value is calculated based on the ratio of similar vectors and a certain calculation formula. The larger the approximate entropy value, the higher the complexity of the current data, and there may be more uncertainties and potential nonlinear characteristics.
[0074] In order to comprehensively utilize the three key indicators of Lyapunov index, correlation dimension and approximate entropy value to accurately determine the nonlinear characteristic points, a weighted comprehensive determination function was constructed. First, the weight coefficient of each indicator in the determination function was determined based on experience and experimental data. These weights reflect the relative importance of each indicator in evaluating the nonlinear characteristics of current data. Then, the Lyapunov index, correlation dimension and approximate entropy value corresponding to each current data sample point were substituted into the determination function for calculation to obtain a determination value. The determination value combines the information of the three indicators and can more comprehensively reflect the chaotic characteristics and nonlinearity of the current data sample point. By comparing the calculated determination value with a pre-set determination threshold, it is determined whether the current data sample point is a nonlinear characteristic point. If the determination value is greater than the determination threshold, it means that the data point has a high chaotic characteristic and nonlinearity, and it is related to the potential safety hazards of the equipment. These current data sample points that meet the conditions are screened out to form a risk current feature set. This set concentrates the current data with potential risks, providing a key information source for the subsequent evaluation and early warning of equipment power safety hazards.
[0075] In this embodiment, three different chaotic characteristic indices are used to describe the chaotic characteristics and nonlinear characteristics of the current data from different angles. Compared with the single indicator analysis method, the multi-indicator comprehensive analysis can reveal the complex information in the current data more comprehensively and accurately, and improve the ability to identify nonlinear characteristic points. The Lyapunov index reflects the sensitivity of the system, the correlation dimension describes the geometric structure of the data, and the approximate entropy value measures the complexity of the data. The combination of the three can provide a deeper understanding of the intrinsic characteristics of the current data and reduce the possibility of misjudgment and missed judgment.
[0076] Constructing a weighted comprehensive judgment function is an innovation of this solution. By setting the weight coefficient reasonably, the role of each indicator in the judgment can be flexibly adjusted according to the characteristics of different devices and actual application scenarios. This method can adapt to the current data characteristics of different types of equipment, improving the versatility and adaptability of the solution. At the same time, the use of judgment thresholds for screening can clearly divide linear and nonlinear feature points, providing a clear boundary for subsequent risk assessment.
[0077] Accurately identifying nonlinear feature points in current data and forming a risk current feature set provides a key basis for early warning of hidden dangers in equipment power use. By focusing on and analyzing these current data with potential risks, possible equipment failures or abnormal operating conditions can be discovered in advance. For example, in the early stages of equipment failure, current data may show slight nonlinear changes. This technical solution can capture these changes in a timely manner and provide early warning signals to equipment managers so that they can take appropriate maintenance measures to avoid further deterioration of equipment failures, thereby reducing equipment maintenance costs and production losses and ensuring safe and stable operation of equipment.
[0078] In one embodiment, the Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent, including:
[0079] Performing data preprocessing based on empirical mode decomposition on the current data sample points to decompose them into a plurality of IMF components to obtain a plurality of IMF subsequences;
[0080] Perform phase space reconstruction on each IMF subsequence, determine the optimal embedding dimension and time delay parameter based on the autocorrelation function method and the false neighbor point method, and construct the phase space matrix;
[0081] Using the Wolf algorithm to perform a preliminary calculation of the maximum Lyapunov exponent of the phase space matrix to obtain a plurality of preliminary Lyapunov exponents;
[0082] The preliminary Lyapunov exponents were screened and optimized based on statistics, and the Lyapunov exponents were obtained after removing outliers.
[0083] In this embodiment, Empirical Mode Decomposition (EMD) is an adaptive data processing method, which is particularly suitable for processing nonlinear and non-stationary signals, and the current data of the device usually has such characteristics. For the current data sample point, the EMD method decomposes it into several intrinsic mode function (IMF) components. Each IMF component represents the local characteristics of the original signal at different time scales. Through this decomposition, the complex current signal can be decomposed into a series of relatively simple sub-signals, namely IMF subsequences. For example, during the operation of the equipment, it may be affected by various factors, such as load changes, electromagnetic interference, etc., causing the current signal to contain fluctuation components of different frequencies and amplitudes. EMD can separate these different components and provide a clearer data structure for subsequent analysis.
[0084] Furthermore, the phase space reconstruction of each IMF subsequence is to transform the one-dimensional time series data into points in a high-dimensional phase space in order to reveal its inherent dynamic characteristics and chaotic features. First, the optimal embedding dimension and time delay parameter are determined according to the autocorrelation function method and the false neighbor point method. The autocorrelation function method is used to analyze the autocorrelation of the time series and find a suitable time delay so that the points in the reconstructed phase space can retain the information of the original signal to the greatest extent. The false neighbor point method is used to determine the embedding dimension. By detecting the false neighbor ratio of adjacent points in the phase space, a suitable embedding dimension is found so that the trajectory in the phase space can accurately describe the dynamic behavior of the system. Once the embedding dimension and time delay parameter are determined, the IMF subsequence can be constructed into a phase space matrix. For example, assuming that the embedding dimension is m and the time delay is τ, then data points at every τ time interval can be selected from the IMF subsequence to form an m-dimensional vector, and multiple such vectors constitute the phase space matrix.
[0085] The Wolf algorithm is a classic method for calculating the maximum Lyapunov exponent. Based on the reconstructed phase space matrix, the Wolf algorithm calculates the separation speed between adjacent points by tracking the evolution trajectory of adjacent points in the phase space. Specifically, an initial point and its neighboring points are selected in the phase space, and the change in the distance between them is calculated over time. According to the exponential growth or contraction rate of the distance, multiple preliminary Lyapunov exponents are calculated step by step according to the preset calculation formula of the Wolf algorithm. These preliminary exponents reflect the degree of chaos of each IMF subsequence in the phase space and its sensitivity to initial conditions.
[0086] Since there may be noise, measurement errors and other factors in the actual data, the Lyapunov exponents obtained by preliminary calculation may contain some outliers. Therefore, statistical screening and optimization are required. First, the mean and standard deviation of the preliminary Lyapunov exponents are calculated. Then, according to certain rules, such as the 3σ principle (that is, data points are considered outliers when they deviate from the mean by more than 3 times the standard deviation), outliers are identified and removed. After screening and optimization, the final Lyapunov exponent is obtained, which can more accurately reflect the chaotic characteristics and potential nonlinear dynamic changes of current data.
[0087] In this embodiment, EMD is used for data preprocessing, which can adaptively decompose current data without presetting basis functions, and is suitable for various complex equipment current signals. Compared with traditional methods such as Fourier transform, EMD can better handle nonlinear and non-stationary signals, retain the local characteristics of the original signal, and provide a more accurate data basis for subsequent analysis. Combining the autocorrelation function method and the false neighbor point method to determine the optimal parameters, the phase space matrix can be constructed more accurately to capture the inherent dynamic structure of the current data. Compared with the empirical selection of parameters, this method has stronger scientificity and adaptability, and can improve the accuracy and reliability of subsequent Lyapunov exponent calculations. By preprocessing the current data and optimizing the phase space reconstruction, it can be better applied to the analysis of equipment current data. At the same time, the stability and accuracy of the calculation results are further improved through statistical screening and optimization processing. The accurately calculated Lyapunov exponent can be used as a key characteristic indicator to evaluate the degree of chaos and potential safety hazards of equipment current data. For example, a large positive Lyapunov exponent may mean that the equipment is in an unstable or potentially faulty state, providing important early warning information to equipment managers, helping them to take timely measures for maintenance and overhaul to avoid equipment failures.
[0088] In one embodiment, the monitored voltage data is subjected to data regularization and feature extraction processing to obtain key voltage feature parameters, including:
[0089] De-noising the monitored voltage data to obtain de-noised voltage data;
[0090] Perform synchronous calibration processing on the denoised voltage data based on the timestamp, perform fast Fourier transform on the calibrated voltage data to the frequency domain, extract frequency domain features, and extract time domain features of the calibrated voltage data based on a sliding window;
[0091] The frequency domain features and time domain features are fused to obtain fused feature data; the key characteristic parameters of voltage are screened out from the fused feature data through a machine learning classifier.
[0092] In this embodiment, since the voltage data monitored by the equipment may be interfered by various noises during the acquisition process, such as electromagnetic interference, environmental noise, etc., these noises will affect the accuracy of subsequent feature extraction and analysis. The original voltage data is processed using a data denoising algorithm. Common denoising methods include wavelet transform denoising, median filtering, etc. Taking wavelet transform denoising as an example, it decomposes the voltage signal into different frequency scales, removes the high-frequency components corresponding to the noise by thresholding the coefficients of the high-frequency part, and retains the low-frequency and partial high-frequency information of the useful signal, and finally reconstructs the denoised voltage data. This can improve the quality of the data and provide a more accurate basis for subsequent processing.
[0093] Furthermore, synchronous calibration of the denoised voltage data based on timestamp is to solve the time asynchronization problem that may exist in multi-source data acquisition or distributed monitoring systems. By referring to a unified time standard, the voltage data of each monitoring point is aligned and adjusted on the time axis to ensure the consistency of the data in the time series. Fast Fourier transform of the calibrated voltage data to the frequency domain is to analyze the frequency components of the voltage signal. In the frequency domain, the amplitude, frequency, and phase characteristics of the main frequency components can be extracted, which reflect the energy distribution and periodic changes of the voltage signal at different frequencies. For example, the failure of some equipment will show abnormal voltage fluctuations at specific frequencies.
[0094] At the same time, the time domain features of the calibrated voltage data are extracted based on the sliding window. In the time domain, statistical features such as mean, variance, maximum, minimum, and rate of change can be calculated. These features describe the basic characteristics and change trends of the voltage signal in the time dimension. For example, sudden changes in voltage and continuous stable states can be reflected through time domain features.
[0095] Then, the extracted frequency domain features and time domain features are fused in order to comprehensively utilize the information in different domains and more comprehensively describe the characteristics of the voltage data. Methods such as feature splicing and weighted fusion can be used. For example, the frequency domain and time domain features are spliced into a high-dimensional feature vector in a certain order, or different weights are assigned to the features according to their importance and then combined. The fused feature data obtained in this way contains richer voltage information and can more accurately reflect the essential characteristics of the voltage data.
[0096] Finally, the key characteristic parameters of voltage are screened out from the fused feature data using machine learning classifiers. Common machine learning classifiers include support vector machines and random forests. First, a large amount of labeled historical voltage data is used to train the classifier, allowing the classifier to learn the association patterns between different voltage features and the operating status of the equipment or potential safety hazards. Then, the fused feature data is input into the trained classifier, which evaluates and screens the features based on the knowledge it has learned, and selects the key characteristic parameters of voltage that are important indicators of the equipment's power safety status. These key characteristic parameters will serve as an important basis for the subsequent evaluation of equipment power safety hazards.
[0097] In this embodiment, voltage data can be processed and analyzed comprehensively. Each step is carefully designed according to the characteristics of voltage data and actual application requirements, and cooperates with each other to form an efficient data processing system. Compared with the traditional simple voltage data analysis method, this scheme can more deeply explore the potential information in voltage data and improve the accuracy of the assessment of the safety status of equipment power consumption. Fast Fourier transform and sliding window are used to extract frequency domain and time domain features respectively, and the characteristics of voltage signals in different domains are fully considered. At the same time, through innovative feature fusion methods, frequency domain and time domain features are organically combined to make up for the limitations of single domain feature analysis. This multi-domain feature fusion method can more comprehensively and accurately describe the characteristics of voltage data and provide richer information for the assessment of equipment power safety hazards. The introduction of machine learning classifiers for the screening of key feature parameters makes the entire technical solution intelligent and adaptive. The machine learning classifier can automatically learn and identify complex patterns and associations. After training based on a large amount of historical data, it can accurately select the most representative and important voltage key feature parameters from the fused feature data. This method not only improves the accuracy and efficiency of screening, but also can adapt to changes in different equipment and different operating environments, and has strong versatility and scalability. By accurately extracting and screening key characteristic parameters of voltage, this technical solution can provide an accurate basis for the assessment of hidden safety hazards of equipment power use. These key characteristic parameters can reflect abnormal changes in equipment voltage and potential safety risks, helping equipment managers to promptly identify problems and take appropriate measures to prevent and repair them. For example, when certain key characteristic parameters exceed the normal range, early warning signals can be issued in time to remind managers to inspect and maintain the equipment, thereby effectively avoiding equipment failures and safety accidents and ensuring the safe and stable operation of the equipment.
[0098] In one embodiment, the risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain an equipment power safety hazard assessment result, including:
[0099] The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on chaotic neural network dynamics, and the chaotic excitation process is performed on them using the nonlinear dynamic characteristics of chaotic neurons to obtain a chaotic fusion feature state;
[0100] The feature extraction method based on strange attractor analysis is used to reduce the dimension of the chaotic fusion feature state and capture the key features, and the key fusion feature vector is extracted.
[0101] Through the evaluation algorithm based on the combination of fuzzy logic reasoning and probability statistics, the key fusion feature vectors are comprehensively evaluated and processed to determine the probability that the equipment is in different safety hazard states, and the equipment power safety hazard assessment results are obtained.
[0102] In this embodiment, the fusion assessment model based on chaotic neural network dynamics takes advantage of the advantages of chaos theory and neural networks. Chaotic neurons have unique nonlinear dynamic characteristics and can perform complex interactions and processing on the input risk current feature set and voltage key feature parameters. When these feature data are input into the model, the chaotic neurons start to work and perform chaotic excitation processing on the data through their internal dynamic mechanisms. This excitation process is similar to introducing a small disturbance in a dynamic system and then observing the response and evolution of the system. In this process, the potential associations and complex patterns between different features begin to emerge. For example, risk current features and voltage key features appear independent under normal circumstances, but under the action of chaotic excitation, the hidden relationships and mutual influences between them are exposed. This processing method can provide richer information and deeper insights for subsequent analysis.
[0103] The feature extraction method based on strange attractor analysis is suitable for processing complex data states after chaotic excitation, namely chaotic fusion characteristic states. Strange attractor is an important concept in chaotic systems, which represents a special dynamic structure of the system in phase space. By performing strange attractor analysis on chaotic fusion characteristic states, key structures and patterns in the data can be found. In actual operation, mathematical methods are first used to reconstruct the phase space of chaotic fusion characteristic states and map high-dimensional data into a suitable phase space. Then, the main characteristic directions and key feature points of the data are determined by calculating the relevant parameters of the attractor, such as fractal dimension, Lyapunov index, etc. Next, feature extraction algorithms, such as variants of principal component analysis (PCA) or feature extraction algorithms specifically for chaotic data, are used to reduce the dimension of the data and capture key features. In this way, the complex chaotic fusion characteristic state can be transformed into a more concise and representative key fusion feature vector, while retaining the most relevant information about the hidden dangers of equipment power use.
[0104] Finally, the evaluation algorithm based on the combination of fuzzy logic reasoning and probability statistics is an innovative evaluation method that can make full use of the flexibility of fuzzy logic and the accuracy of probability statistics. For the extracted key fusion feature vectors, first, the fuzzy logic reasoning part fuzzifies the feature vectors according to the pre-defined fuzzy rules and fuzzy sets. These fuzzy rules are established based on expert knowledge and historical data to describe the fuzzy relationship between different feature combinations and the safety hazard status of the equipment. Then, the probability statistics part performs probability calculation on the fuzzified results. Through the learning and statistical analysis of a large amount of historical data, the probability distribution of different feature combinations corresponding to different safety hazard states is determined. Finally, the results of fuzzy logic reasoning and probability statistics are combined to obtain the probability of the equipment being in different safety hazard states. For example, the probability that the equipment is in a low-risk state may be 30%, the probability that it is in a medium-risk state may be 50%, and the probability that it is in a high-risk state may be 20%. These probability values are the evaluation results of the equipment's electrical safety hazards, providing intuitive risk assessment information for equipment managers.
[0105] In this embodiment, a fusion evaluation model based on chaotic neural network dynamics is adopted to break through the limitations of traditional data analysis methods. The introduction of chaos theory enables the model to better handle the nonlinear and complex dynamic characteristics in the equipment power consumption data, and improves the ability to identify potential safety hazards. Compared with traditional neural network models, chaotic neural networks can more accurately capture potential patterns and associations in data, and provide more valuable intermediate results for subsequent analysis. Compared with traditional feature extraction techniques, the feature extraction method based on strange attractor analysis has stronger adaptability and accuracy. It can find key features in chaotic data, avoid information loss and redundancy, and improve the quality and representativeness of features. At the same time, through feature dimensionality reduction, the computational complexity is reduced and the efficiency of subsequent evaluation algorithms is improved.
[0106] In one embodiment, the core component of the fusion evaluation model based on chaotic neural network dynamics is a neuron with chaotic characteristics. Compared with neurons in traditional neural networks, chaotic neurons have more complex dynamic behaviors. They can show chaotic characteristics such as sensitivity to initial conditions, non-periodicity and long-term unpredictability. These chaotic neurons form a network structure through complex connection methods. The adjustment of connection weights is based on a learning algorithm, such as a variant of the back propagation algorithm or a training algorithm specifically for chaotic neural networks. During the training process, the model continuously adjusts the connection weights according to the input data and the expected output to optimize the performance of the model.
[0107] When input data (such as risk current feature sets and voltage key feature parameters) enter the model, chaotic neurons begin to be activated. The chaotic excitation mechanism causes the state of neurons to change rapidly in a short period of time, and this change is highly nonlinear. This chaotic excitation process helps the model discover hidden patterns and complex associations in the input data. Through the action of chaotic dynamics, the potential relationships between different features can be propagated and evolved in the network, thus providing a rich information basis for fusion analysis. The fusion evaluation model based on chaotic neural network dynamics has strong dynamic learning and adaptation capabilities. Due to the characteristics of chaotic systems, the model can self-adjust and optimize under changing environmental and data conditions. For example, when the operating state of the equipment changes or new safety hazard patterns emerge, the model can quickly adapt to these changes through the dynamic adjustment of chaotic neurons, and relearn and identify new features and patterns. This ability enables the model to be effectively applied to the assessment of equipment power safety hazards in the long term without losing accuracy due to data changes. At the input layer, the risk current feature set and voltage key feature parameters are received simultaneously and mapped to the chaotic neural network through connection weights. These feature data are fused at multiple levels in the network. First, at a low level, chaotic neurons perform a preliminary fusion of adjacent feature data through local connections and interactions. This local fusion helps to extract local features and associations. Then, at a high level, through global connections and feedback mechanisms, information from different local areas is integrated to achieve global fusion analysis. This multi-level fusion method can make full use of the complementarity and correlation between different features to improve the accuracy and reliability of the evaluation. Chaotic dynamics plays a key role in the fusion analysis process. The nonlinear dynamic behavior of chaotic neurons enables different feature data to interact and evolve in a complex manner in the network. For example, chaotic excitation can lead to the propagation and diffusion of feature data in the network, thereby promoting information exchange and fusion between different features. At the same time, the sensitivity of the chaotic system enables the model to respond to small feature changes, improving the sensitivity to safety hazards. In addition, the dynamic characteristics of chaotic neural networks can also help the model adapt to feature data of different types and sizes, and achieve flexible fusion analysis.
[0108] In this embodiment, chaotic neural network dynamics can effectively process the nonlinear and complex dynamic characteristics in the equipment power consumption data, which is difficult to achieve with traditional data analysis methods. Through chaotic excitation and complex network structure, the model can discover subtle patterns and associations hidden in the data and improve the ability to identify safety hazards. The model has strong adaptive capabilities and can self-adjust and optimize under different equipment operating conditions and environments. At the same time, the robustness of the chaotic system makes the model resistant to noise and interference, and can maintain stable performance in practical applications.
[0109] In one embodiment, after obtaining the equipment power safety hazard assessment result, the method further includes:
[0110] Acquire a first detection identifier of the current detector, and acquire a second detection identifier of the voltage monitoring module;
[0111] Based on the first detection identifier and the second detection identifier, generating an identifier sequence;
[0112] Acquire an operation cycle of the device, and change the identification sequence based on the operation cycle to obtain a changed sequence;
[0113] A key is generated based on the change sequence, and is used to encrypt and store the current data sample points, voltage data, and equipment power safety hazard assessment results.
[0114] In this embodiment, each current detector and voltage monitoring module has its own unique identification, which can be the serial number, MAC address or other unique identifier of the device. The first detection identification of the current detector and the second detection identification of the voltage monitoring module are obtained in order to establish an association with these data acquisition devices. The first detection identification and the second detection identification are combined in a preset order to generate an identification sequence. This sequence not only contains the information of the data acquisition device, but also provides basic identity recognition information for subsequent data encryption and management. For example, the first detection identification can be placed in front and the second detection identification can be placed in the back, separated by a specific separator in the middle to form a continuous identification sequence.
[0115] The operation cycle of a device refers to the time interval from the start to the stop or completion of a complete work task of the device. Understanding the operation cycle of the device is very important for data management and encryption. Change the identification sequence according to the operation cycle of the device. A common method is to insert information related to the operation cycle into the identification sequence, such as the start time, end time or number of the operation cycle. In this way, the changed sequence not only contains the device detection identification, but also is associated with the operation status and time information of the device. For example, if the operation cycle of the device is in hours, the number of hours of the operation cycle can be converted into a specific coding format and inserted into a specific position in the identification sequence. In this way, each time the device runs a new cycle, the identification sequence will be updated and changed accordingly according to the changes in the operation cycle.
[0116] A key is generated based on the changed sequence, which will be used to encrypt the current data sample points, voltage data, and equipment power safety hazard assessment results. The key can be generated using a variety of encryption algorithms and technologies, such as symmetric encryption algorithms (such as AES) or asymmetric encryption algorithms (such as RSA).
[0117] When generating a key, the change sequence can be used as an input parameter and processed through an encryption algorithm to generate a key with sufficient complexity and security. This key should have sufficient length and randomness to ensure the security of the data. Then, the generated key is used to encrypt the current data sample points, voltage data, and equipment power safety hazard assessment results. The encryption process can be performed during data transmission or before the data is stored in a local or cloud server. The encrypted data will be stored in ciphertext, and only authorized users with the correct key can decrypt and access the data.
[0118] In this embodiment, by associating the identifiers of the current detector and voltage monitoring module with the data, data traceability is achieved. In the subsequent data analysis and troubleshooting process, the source of the data and the acquisition device can be quickly determined based on the detection identifier, which is helpful for accurately analyzing the problem and locating the fault. At the same time, the identification sequence is changed in combination with the operation cycle of the equipment, which further enhances the association between the data and the operation status of the equipment, so that when analyzing the data, the background and context information of the data generation can be better understood, thereby improving the accuracy and effectiveness of the data analysis. Compared with the traditional fixed key encryption method, the key generated by this scheme is dynamic and personalized, and different keys are generated according to different devices and operation cycles, which increases the difficulty of cracking. Even if the data key of a device is cracked, it will not affect the data security of other devices or different operation cycles.
[0119] In one embodiment, the step of changing the identification sequence based on the operation cycle to obtain a changed sequence includes:
[0120] Obtaining the target character type corresponding to the operation cycle in the database; searching for target characters of all target character types in the identification sequence;
[0121] Add the target characters to a matrix of two rows and two columns according to a preset rule to generate a first matrix; add the digital characters in the operation cycle to the matrix of two rows and two columns according to a preset rule to generate a second matrix; calculate the product of the first matrix and the second matrix to obtain a third matrix;
[0122] Obtain the total number m of target characters, combine the characters in the third matrix in sequence, and separate them into m component character combinations;
[0123] Each target character is sequentially combined with each sub-character to form a one-to-one mapping, and in the identification sequence, each target character is replaced with the mapped sub-character combination to obtain the changed sequence.
[0124] In this embodiment, various information related to the operation cycle is stored in the database, including target character types corresponding to different operation cycles. These target character types are predefined. After the target character type corresponding to the current operation cycle is obtained from the database, a search and search are performed in the identification sequence to find all target characters belonging to the target character type. For example, if the target character type is a specific letter or symbol, all characters that meet this type are found in the identification sequence.
[0125] The found target characters are added to a matrix of two rows and two columns according to a preset rule to generate a first matrix. The preset rule may be to determine the position of each character in the matrix according to the order of appearance of the characters or other logic related to the device.
[0126] For the numeric characters in the operation cycle, they are also added to another two-row and two-column matrix to generate a second matrix according to the preset rules. For example, the key digital information such as the start time and end time of the operation cycle can be mapped to each position of the second matrix in a certain order and manner. The product of the first matrix and the second matrix is calculated to obtain the third matrix. The matrix multiplication operation follows certain mathematical rules, and the product result here contains the information combination between the target character and the operation cycle numeric character after a specific operation.
[0127] Get the total number of target characters m, and then combine the characters in the third matrix in order. Then, according to the number of target characters m, separate the combined characters into m sub-character combinations. This process is actually a further processing and division of the matrix operation results so as to perform subsequent mapping operations with the target characters. Then, each target character is sequentially combined with each sub-character to form a one-to-one mapping relationship. This means that each target character has a corresponding processed sub-character combination. Finally, in the identification sequence, the original target characters are replaced with the mapped sub-character combination to obtain a changed identification sequence, that is, a changed sequence. In this way, the dynamic change of the identification sequence based on the operation cycle is realized, so that it contains richer information related to the operation cycle.
[0128] In this embodiment, it is an innovative method to process the identification sequence and the operation cycle information by using matrix operation. By mapping the character and digital information into the matrix for multiplication operation, a complex data fusion and conversion process can be realized. This method breaks through the traditional method of simple character replacement or direct addition of numbers, and provides a more logical and operable means for data change. After this complex change process, the identification sequence contains more detailed information related to the operation cycle, which not only improves the security of the data, but also makes the data more traceable. In the subsequent data management and analysis, the operation cycle information of the device and the association with the target character can be quickly understood by parsing the sub-character combination in the change sequence, which helps to locate the problem more accurately and conduct security assessment. The above-mentioned change method based on matrix operation can provide a richer foundation for subsequent data encryption. By closely combining the operation cycle information with the identification sequence, a more complex and difficult-to-crack encryption key or input parameter of the encryption algorithm can be generated. At the same time, for a large amount of device data management, this method can achieve efficient data classification and organization, and improve the efficiency and accuracy of data management.
[0129] Reference Figure 2 In one embodiment of the present invention, there is also provided an early warning system for potential safety hazards of electrical equipment, including:
[0130] A collection unit, used to perform ultra-high frequency dynamic sampling of the real-time current of the device through a current detector based on a carbon nanotube array to obtain multiple current data sample points;
[0131] A calculation unit, used for calculating a chaos characteristic index for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current characteristic set;
[0132] The extraction unit is used to synchronously monitor the real-time voltage of the equipment through the voltage monitoring module, and perform data regularization and feature extraction on the monitored voltage data to obtain key characteristic parameters of the voltage;
[0133] The early warning unit is used to input the risk current feature set and the voltage key feature parameters into a fusion evaluation model based on deep learning for fusion analysis to obtain an equipment power safety hazard evaluation result.
[0134] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.
[0135] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3As shown. 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 designed by the computer 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. When the computer program is executed by the processor, the above method is implemented.
[0136] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure 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.
[0137] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and 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.
[0138] In summary, the early warning method, system, device and medium for equipment power safety hazards provided in the embodiments of the present invention include: performing ultra-high frequency dynamic sampling of the real-time current of the equipment through a current detector based on a carbon nanotube array to obtain multiple current data sample points; calculating the chaos characteristic index of the current data sample points, and obtaining the nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current feature set; synchronously monitoring and processing the real-time voltage of the equipment through a voltage monitoring module, and performing data regularization and feature extraction processing on the monitored voltage data to obtain voltage key characteristic parameters; inputting the risk current feature set and the voltage key characteristic parameters into a fusion evaluation model based on deep learning for fusion analysis to obtain the equipment power safety hazard evaluation result. In the present invention, the risk current feature set is obtained through current data, the voltage key characteristic parameters are obtained through voltage data, and fusion analysis is performed to obtain the equipment power safety hazard evaluation result, which realizes the detection and early warning of potential safety hazards and overcomes the current defect that it is impossible to accurately judge whether the equipment has safety hazards.
[0139] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and 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-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.
[0140] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.
[0141] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A method for early warning of hidden dangers of electrical safety in equipment, characterized in that: The following steps are involved: The real-time current of the device is sampled dynamically at ultra-high frequency by a current detector based on a carbon nanotube array to obtain multiple current data sample points; Calculating chaos characteristic indexes for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic indexes to form a risk current characteristic set; The voltage monitoring module is used to synchronously monitor the real-time voltage of the equipment, and the monitored voltage data is processed by data regularization and feature extraction to obtain the key characteristic parameters of the voltage; The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain an equipment power safety hazard assessment result; The chaos characteristic indicators include correlation dimension, Lyapunov index, and approximate entropy value; The calculation of the chaos characteristic index of the current data sample points is performed, and based on the chaos characteristic index, nonlinear characteristic points in the current data sample points are obtained to form a risk current characteristic set, including: The Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent; Analyze the current data sample points based on the correlation integral method to obtain the correlation dimension; The current data sample points are calculated using an approximate entropy algorithm based on information theory to obtain an approximate entropy value; The correlation dimension, Lyapunov index, and approximate entropy value are calculated based on a weighted comprehensive determination function to obtain a determination value; The current data sample points whose determination values are greater than the determination threshold are used as nonlinear feature points to form a risk current feature set; The monitored voltage data is subjected to data regularization and feature extraction processing to obtain key voltage feature parameters, including: De-noising the monitored voltage data to obtain de-noised voltage data; Perform synchronous calibration processing on the denoised voltage data based on the timestamp, perform fast Fourier transform on the calibrated voltage data to the frequency domain, extract frequency domain features, and extract time domain features of the calibrated voltage data based on a sliding window; The frequency domain features and time domain features are integrated to obtain integrated feature data; the key characteristic parameters of voltage are screened out from the integrated feature data through a machine learning classifier; The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain equipment power safety hazard assessment results, including: The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on chaotic neural network dynamics, and the chaotic excitation process is performed on them using the nonlinear dynamic characteristics of chaotic neurons to obtain a chaotic fusion feature state; The feature extraction method based on strange attractor analysis is used to reduce the dimension of the chaotic fusion feature state and capture the key features, and the key fusion feature vector is extracted. Through the evaluation algorithm based on the combination of fuzzy logic reasoning and probability statistics, the key fusion feature vectors are comprehensively evaluated and processed to determine the probability that the equipment is in different safety hazard states, and the equipment power safety hazard assessment results are obtained.
2. The early warning method for hidden dangers of electrical safety in equipment according to claim 1 is characterized in that: The current detector and voltage monitoring module are both IoT devices. The data platform provides computing power to calculate the chaotic characteristic indicators of the current data sample points. The hidden danger investigation system performs fusion analysis to obtain the equipment power safety hidden danger assessment results, and sends them to the alarm system for abnormal alarm processing.
3. The early warning method for hidden dangers of electrical safety in equipment according to claim 1 is characterized in that: The Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent, including: Performing data preprocessing based on empirical mode decomposition on the current data sample points to decompose them into a plurality of IMF components to obtain a plurality of IMF subsequences; Perform phase space reconstruction on each IMF subsequence, determine the optimal embedding dimension and time delay parameter based on the autocorrelation function method and the false neighbor point method, and construct the phase space matrix; Using the Wolf algorithm to perform a preliminary calculation of the maximum Lyapunov exponent of the phase space matrix to obtain a plurality of preliminary Lyapunov exponents; The preliminary Lyapunov exponents were screened and optimized based on statistics, and the Lyapunov exponents were obtained after removing outliers.
4. The early warning method for hidden dangers of electrical safety in equipment according to claim 1 is characterized in that: After obtaining the equipment power safety hazard assessment result, the method further includes: Acquire a first detection identifier of the current detector, and acquire a second detection identifier of the voltage monitoring module; Based on the first detection identifier and the second detection identifier, generating an identifier sequence; Acquire an operation cycle of the device, and change the identification sequence based on the operation cycle to obtain a changed sequence; A key is generated based on the change sequence, and is used to encrypt and store the current data sample points, voltage data, and equipment power safety hazard assessment results.
5. An early warning system for hidden dangers of electrical safety in equipment, characterized in that: include: A collection unit, used to perform ultra-high frequency dynamic sampling of the real-time current of the device through a current detector based on a carbon nanotube array to obtain multiple current data sample points; A calculation unit, used for calculating a chaos characteristic index for the current data sample points, and obtaining nonlinear characteristic points in the current data sample points based on the chaos characteristic index to form a risk current characteristic set; The extraction unit is used to synchronously monitor the real-time voltage of the equipment through the voltage monitoring module, and perform data regularization and feature extraction on the monitored voltage data to obtain key characteristic parameters of the voltage; An early warning unit, used for inputting the risk current feature set and the voltage key feature parameters into a fusion assessment model based on deep learning for fusion analysis, and obtaining an assessment result of equipment power safety hazards; The chaos characteristic indicators include correlation dimension, Lyapunov index, and approximate entropy value; The calculation of the chaos characteristic index of the current data sample points is performed, and based on the chaos characteristic index, nonlinear characteristic points in the current data sample points are obtained to form a risk current characteristic set, including: The Wolf algorithm is used to analyze the current data sample points to obtain the Lyapunov exponent; Analyze the current data sample points based on the correlation integral method to obtain the correlation dimension; The current data sample points are calculated using an approximate entropy algorithm based on information theory to obtain an approximate entropy value; The correlation dimension, Lyapunov index, and approximate entropy value are calculated based on a weighted comprehensive determination function to obtain a determination value; The current data sample points whose determination values are greater than the determination threshold are used as nonlinear feature points to form a risk current feature set; The monitored voltage data is subjected to data regularization and feature extraction processing to obtain key voltage feature parameters, including: De-noising the monitored voltage data to obtain de-noised voltage data; Perform synchronous calibration processing on the denoised voltage data based on the timestamp, perform fast Fourier transform on the calibrated voltage data to the frequency domain, extract frequency domain features, and extract time domain features of the calibrated voltage data based on a sliding window; The frequency domain features and time domain features are integrated to obtain integrated feature data; the key characteristic parameters of voltage are screened out from the integrated feature data through a machine learning classifier; The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on deep learning for fusion analysis to obtain equipment power safety hazard assessment results, including: The risk current feature set and the voltage key feature parameters are input into a fusion assessment model based on chaotic neural network dynamics, and the chaotic excitation process is performed on them using the nonlinear dynamic characteristics of chaotic neurons to obtain a chaotic fusion feature state; The feature extraction method based on strange attractor analysis is used to reduce the dimension of the chaotic fusion feature state and capture the key features, and the key fusion feature vector is extracted. Through the evaluation algorithm based on the combination of fuzzy logic reasoning and probability statistics, the key fusion feature vectors are comprehensively evaluated and processed to determine the probability that the equipment is in different safety hazard states, and the equipment power safety hazard assessment results are obtained.
6. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.
7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.
Citation Information
Patent Citations
GIS fault detection system and method based on multi-source information fusion and deep learning network
CN106597231A
Original position measurement device of electrical characteristics of carbon nanotube arrays
CN107402333A
Real-time monitoring method and system for potential safety hazard of electricity utilization of electrical equipment
CN118015798A
AC fault arc detection method, system, device and medium
CN118625073A
Project disaster warning method and system based on collaborative fusion of multi-physics monitoring data
US20230410012A1
