Arc crosstalk identification method and device, computer readable storage medium and system

Abstract

This application discloses an arc crosstalk identification method, apparatus, computer-readable storage medium, and system. The method includes acquiring current sampling data of the target observation period of the circuit under test, and extracting features from the current sampling data to obtain a feature data set; constructing a stochastic process model from the current sampling data, fitting the current sampling data based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient; and determining whether the target observation period of the circuit under test is affected by crosstalk signals based on the feature data set and the correlation characterization parameters, using a preset discrimination strategy. By fusing multi-dimensional signal features with relevant parameters of the Fokker-Planck equation, rapid and accurate mechanistic identification of arc crosstalk signals is achieved, fundamentally improving the anti-interference capability and operational reliability of fault arc detection in complex electromagnetic environments.

Description

Technical Field

[0001] This application relates to the field of electrical safety technology, and in particular to an arc crosstalk identification method and apparatus, a computer-readable storage medium and system. Background Technology

[0002] In real-world multi-circuit, high-density electrical installations, a fault arc signal generated in one circuit can be transmitted to adjacent non-faulty circuits via electromagnetic induction, common impedance coupling, or other means. This is known as "arc crosstalk signal." This crosstalk signal is not a genuine fault on the circuit, but it can be detected by monitoring devices. Its characteristics are similar to a real arc, easily leading to misjudgments and malfunctions by protection devices. For example, crosstalk signals can cause protection devices on non-faulty circuits to misjudge a local fault, resulting in unnecessary false tripping (maloperation). Such maloperations not only interrupt normal power supply and affect system reliability but may also mask the true fault location, delaying the handling of real fire risks and severely reducing the practical value and user trust of the fault arc protection system.

[0003] Current methods for detecting fault arcs have the following problems when dealing with crosstalk arcs: they have a single feature dimension and weak crosstalk identification ability; they have not constructed a physical mechanism model for signal propagation, so they cannot determine from the root cause whether the abnormal signal is generated locally or coupled from external sources; high-precision calculation models are computationally complex, while lightweight and fast algorithms are difficult to guarantee reliability, thus there is a problem of not being able to balance real-time performance and reliability.

[0004] Therefore, effectively distinguishing between real fault arcs and crosstalk signals has become a key technical bottleneck that must be addressed to improve the application efficiency of existing fault arc detection technologies. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the purpose of this application is to provide an arc crosstalk identification method and device, a computer-readable storage medium and system, to solve the problems of weak crosstalk identification capability of current arc detection methods, inability to fundamentally identify crosstalk signals and inability to balance real-time performance and reliability.

[0006] The first aspect of this application provides a method for identifying arc crosstalk, including:

[0007] Acquire current sampling data of the target circuit under test during the observation period, and perform feature extraction on the current sampling data to obtain a feature data set; The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient. Based on the feature data set and the associated characterization parameters, a preset discrimination strategy is used to determine whether the observation period of the target circuit to be detected is affected by crosstalk signals.

[0008] In some embodiments of this application, the current sampling data is constructed as a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters between the drift coefficient and the diffusion coefficient, including: Obtain the probability density function of the signal amplitude in the current sampling data, and fit the probability density function based on the Fokker-Planck equation to obtain the drift coefficient function and the diffusion coefficient function; Relative strength and stability analysis are performed on the drift coefficient function and the diffusion coefficient function to obtain the correlation characterization parameters that characterize the drift coefficient and the diffusion coefficient.

[0009] In some embodiments of this application, the probability density function for obtaining the signal amplitude in the current sampling data includes: The probability density function of the signal amplitude in the current sampling data is estimated by using a sliding time window.

[0010] In some embodiments of this application, feature extraction of the current sampling data to obtain feature data includes: Time-domain features are extracted from the current sampling data to obtain a time-domain statistical feature dataset; Frequency domain features are extracted from the current sampling data to obtain a frequency domain statistical feature dataset; The time-domain statistical feature dataset and the frequency-domain statistical feature dataset are merged into a feature data set.

[0011] In some embodiments of this application, the time-domain statistical feature dataset includes at least one of peak factor, zero-crossing rate, skewness, and kurtosis; the frequency-domain statistical feature dataset includes at least one of dominant frequency, frequency center, bandwidth, harmonic ratio, and spectral entropy.

[0012] In some embodiments of this application, determining whether the observation period of the target circuit to be detected is affected by crosstalk signals based on the feature data set and the associated characterization parameters using a preset discrimination strategy includes: The feature data and the associated representation parameters in the feature data are weighted and fused to obtain a comprehensive score for crosstalk suspicion. Determine whether the crosstalk suspicion comprehensive score is greater than the preset crosstalk threshold. If it is, it means that the target observation period of the circuit under test is affected by the crosstalk signal; otherwise, it means that the target observation period of the circuit under test is not affected by the crosstalk signal.

[0013] In some embodiments of this application, the arc crosstalk identification method further includes: When it is determined that the target observation period of the circuit under test is affected by crosstalk signal, it is determined whether there is a strong local arc feature in the feature data set. If so, it indicates that there is a real fault arc in the target observation period of the circuit under test and it is accompanied by crosstalk signal interference. A trip command is sent to the protection actuator so that the protection actuator executes the trip command and controls the circuit under test to be de-energized. Otherwise, it indicates that the target observation period of the circuit under test is only affected by crosstalk signal. The strong local arc characteristics include at least one of a peak factor greater than a preset factor threshold and a preset arc harmonic.

[0014] A second aspect of this application provides an arc crosstalk identification device, including a feature extraction module, an association characterization parameter acquisition module, and a crosstalk identification module; The feature extraction module is used to acquire current sampling data of the target observation period of the circuit to be detected, and to perform feature extraction on the current sampling data to obtain a feature data set. The correlation characterization parameter acquisition module is used to construct a stochastic process model from the current sampling data and fit the current sampling data based on the Fokker-Planck equation to obtain the correlation characterization parameters between the drift coefficient and the diffusion coefficient. The crosstalk identification module is used to determine, based on the feature data set and the associated characterization parameters, whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy.

[0015] A third aspect of this application provides an arc crosstalk identification system, including a cloud service platform, an arc crosstalk identification device, and a protection actuator that are interconnected. The arc crosstalk identification device is the arc crosstalk identification device described above. The arc crosstalk identification device is also used to upload current sampling data to the cloud service platform; The protection actuator is used to control the circuit under test to disconnect power when a trip command is received, and to feed back the trip status data to the cloud service platform; The cloud service platform is used to store the current sampling data and the tripping status data.

[0016] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0017] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects: The arc crosstalk identification method applied in this application describes the coupling and propagation process of current signals by introducing the Fokker-Planck equation. By fusing multi-dimensional signal features with relevant parameters of the Fokker-Planck equation, rapid and accurate mechanistic identification of arc crosstalk signals is achieved, fundamentally improving the anti-interference capability and operational reliability of fault arc detection in complex electromagnetic environments. A dual identification mechanism of multi-dimensional data fusion and strong local arc feature recognition is employed to achieve dual detection of fault arcs and crosstalk signal interference, improving the accuracy of fault arc detection. This enables precise differentiation between fault and interference signals, fundamentally improving the reliability and intelligence level of the smart grid protection system.

[0018] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims, and drawings. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 The diagram shown is a flowchart illustrating the arc crosstalk identification method described in an embodiment of this application.

[0020] Figure 2 The diagram shows a process schematic of the arc crosstalk identification method described in the embodiments of this application.

[0021] Figure 3 This diagram illustrates the process of obtaining associated characterization parameters in the arc crosstalk identification method described in this application embodiment.

[0022] Figure 4 The diagram shows a comparison and difference of drift coefficients in the process of identifying fault arcs and crosstalk signals using the arc crosstalk identification method described in this application embodiment.

[0023] Figure 5 The diagram shows a comparison and difference of the diffusion coefficients during the identification of fault arcs and crosstalk signals using the arc crosstalk identification method described in this application embodiment.

[0024] Figure 6 The diagram shown is a structural schematic of the arc crosstalk identification device described in an embodiment of this application.

[0025] Figure 7 The diagram shown is a structural schematic of the arc crosstalk identification system described in an embodiment of this application.

[0026] Specific element symbol explanations: 10-Arc crosstalk identification device, 20-Cloud service platform, 30-Protection actuator, 1-Arc crosstalk identification system, 11-Feature extraction module, 12-Association characterization parameter acquisition module, 13-Crosstalk identification module, 14-Fault arc identification module. Detailed Implementation

[0027] To make the technical problems, technical solutions, and beneficial effects to be solved by this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the scope of this application.

[0028] It should be noted that when a component is referred to as being "set on" another component, it can be directly on or indirectly on that other component. When a component is referred to as being "connected to" another component, it can be directly connected to or indirectly connected to that other component.

[0029] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0030] Short circuits, poor connections, and aging in electrical circuits can all lead to arc faults. Arc faults, characterized by high temperatures and energy density, are a significant cause of electrical fires. With the widespread deployment of AFCI (Arc Fault Circuit Interrupter) and AFDD (Arc Fault Detection Devices) in smart grids, new energy power plants, energy storage systems, and daily life, the accuracy and reliability of detection are becoming increasingly critical. Existing arc fault detection technologies primarily rely on threshold judgments based on time-domain (e.g., periodic arc extinction) or frequency-domain (e.g., high-frequency noise) characteristics of the current waveform. However, crosstalk signals, as coupling attenuation products of real arcs, significantly overlap with real arc characteristics in both the time and frequency domains, rendering existing single-domain analysis-based techniques essentially ineffective in distinguishing them. While some methods employing complex artificial intelligence models have shown some effectiveness, they suffer from inherent drawbacks such as high computational burden, "black box" decision-making, and difficulty in real-time and reliable deployment on embedded devices.

[0031] Currently, the mainstream technical approaches to solving the problems of arc fault detection and interference fall into three categories. The first category comprises detection methods based on time-domain, frequency-domain, and time-frequency-domain signal analysis. These methods have clear physical meaning and can effectively capture key phenomena such as high-frequency oscillations and non-stationarity caused by arcs; however, they lack the ability to specifically identify crosstalk signals, and the accuracy of feature extraction is greatly affected by the load type (such as motors and switching power supplies), easily producing false alarms under complex loads. The second category consists of intelligent detection models based on artificial intelligence and machine learning. These models possess powerful automatic feature extraction and complex pattern classification capabilities, are highly adaptable to diverse loads, and have high detection accuracy; however, their computational complexity is usually high, and because they are essentially "black box" models, their decision-making process is difficult to interpret, and they cannot provide clear criteria for crosstalk identification from a physical mechanism perspective. One type is the suppression algorithm for specific noise and interference. This type of algorithm is highly targeted and can effectively filter out known types of interference in specific scenarios. However, the solution is local and isolated. It is mainly used to deal with "noise" related to the operation of the line itself, rather than "crosstalk signals" "coupled" from faults in adjacent lines. Therefore, it is difficult to effectively deal with the impact of crosstalk signals.

[0032] Based on this, this application improves upon the arc crosstalk identification methods and devices, computer-readable storage media, and systems in related technologies. These improvements are applicable to the fields of electrical engineering and electrical safety technology, and are particularly suitable for applications with stringent requirements for power supply reliability, system integration, and electrical safety, such as data centers, industrial automation control systems, communication infrastructure, and high-end commercial / civilian intelligent power distribution systems.

[0033] refer to Figure 1 As shown, the arc crosstalk identification method of this application embodiment includes the following steps.

[0034] Step S101: Obtain current sampling data of the target observation period of the circuit to be tested, and perform feature extraction on the current sampling data to obtain a feature data set.

[0035] The current data of the circuit under test is sampled in real time using a current sampling device to obtain current sampling data for each observation cycle. This embodiment uses crosstalk identification for a single observation cycle as an example. Repeating the arc crosstalk identification method in this embodiment can achieve crosstalk identification for each observation cycle. The single observation cycle targeted in this embodiment is set as the target observation cycle, and the target observation cycle can be any observation cycle of the circuit under test. The duration of the observation cycle can be set based on the actual situation, and this application does not impose too many limitations on it.

[0036] After acquiring the current sampling data of the target observation period of the circuit to be tested, time-domain feature extraction is performed on the current sampling data to obtain a time-domain statistical feature dataset, and frequency-domain feature extraction is performed on the current sampling data to obtain a frequency-domain statistical feature dataset. The time-domain statistical feature dataset and the frequency-domain statistical feature dataset are then merged into a feature data set.

[0037] Step S102: The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient.

[0038] The current sampling data is considered as a macroscopic manifestation of the motion of a large number of charged particles (including electrons and ions), whose state changes follow a stochastic process. The probability density function of the current sampling data is estimated, and then fitted to the probability density function based on the Fokker-Planck equation to obtain the drift coefficient function and the diffusion coefficient function. The relative strength and stability of the two are analyzed to determine the nature of the signal source from a physical mechanism perspective, and thus obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient.

[0039] Step S103: Based on the feature data set and associated characterization parameters, determine whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy.

[0040] The feature data and associated characterization parameters in the feature dataset are weighted and fused to obtain a comprehensive score for crosstalk suspicion. The magnitude of the comprehensive score for crosstalk suspicion is used to determine whether the observation period of the target circuit under test is affected by crosstalk signals.

[0041] This embodiment introduces Fokker-Planck analysis and organically combines it with time-frequency domain feature data to fundamentally distinguish between electric arcs and crosstalk from a mechanistic perspective. This significantly improves the accuracy and reliability of identification, and fundamentally enhances the anti-interference capability and operational reliability of fault arc detection in complex electromagnetic environments. It can be applied to any scenario that requires the deployment of fault arc protection but suffers from "false operation" problems due to complex electromagnetic environments and the coexistence of multiple circuits. It upgrades traditional "arc detection" to "intelligent arc diagnosis" that can identify the effects of crosstalk, fundamentally improving the reliability and intelligence level of smart grid protection systems.

[0042] The steps S101 to S103 of the arc crosstalk identification method in this embodiment will be described in detail below.

[0043] Step S101: Obtain current sampling data of the target observation period of the circuit to be tested, and perform feature extraction on the current sampling data to obtain a feature data set.

[0044] A current sampling device is used to continuously monitor the current of the circuit under test with high fidelity, thereby obtaining raw current data. For example, the current sampling device can sample the current at a rate of more than 10,000 times per second to ensure that the monitored data includes even the slightest fluctuations. After obtaining the raw current data, it needs to be preprocessed to obtain current sampling data. Preprocessing specifically includes filtering and amplitude normalization of the raw current data; filtering removes background "noise" such as stable power frequency components from the raw circuit data; amplitude normalization standardizes the signal amplitude.

[0045] The current sampling data of the circuit under test for each observation cycle can be obtained by means of the above method. The arc crosstalk identification method in this embodiment is illustrated by taking the arc crosstalk identification of a single observation cycle of the circuit under test as an example. The arc crosstalk identification of other observation cycles can be repeated by repeating the arc crosstalk identification method in this embodiment.

[0046] Set any observation period of the circuit under test as the target observation period. Acquire current sampling data during the target observation period and perform multi-dimensional feature extraction on the current sampling data. (Reference) Figure 2 As shown, multi-dimensional feature extraction specifically includes extracting time-domain statistical features from current sampling data. Further, based on the shape of the current waveform changing over time, at least one of the following parameters is calculated: peak factor, zero-crossing rate, skewness, and kurtosis. These parameters are then aggregated into a time-domain statistical feature dataset. The peak factor can be used to determine the impact intensity of the current sampling data, the zero-crossing rate can be used to determine the rhythm of current changes, and the skewness and kurtosis are used to determine the symmetry and spike-like nature of the current waveform. These features can outline the behavior of the current signal on the time axis, thereby revealing a significantly different and abnormally intense fluctuation pattern compared to a stable load.

[0047] Multi-dimensional feature extraction of current sampling data also includes frequency domain statistical feature extraction. This involves converting the current sampling data into frequency domain data using a Fast Fourier Transform (FFT), and then obtaining at least one of the following: dominant frequency, frequency center, bandwidth, harmonic ratio, and spectral entropy. These are then compiled into a frequency domain statistical feature dataset. The dominant frequency and frequency center are primarily used to determine the energy concentration location of the current frequency domain data; bandwidth is used to determine the distribution of the current frequency domain data; the harmonic ratio is used to determine whether there are regular harmonic components in the current frequency domain data; and the spectral entropy is used to determine whether the overall spectrum of the current frequency domain data is orderly or chaotic. In the process of arc crosstalk identification, real arcs produce specific high-frequency oscillation characteristics, while the spectrum of crosstalk signals often becomes diffuse and chaotic during coupling. Therefore, the frequency domain statistical feature dataset is mainly used to identify the energy characteristics of related crosstalk in the current sampling data.

[0048] After obtaining the time-domain statistical feature dataset and the frequency-domain statistical feature dataset, the two can be merged into a feature data set containing all feature data. Through the extraction of the above multi-dimensional basic features, a comprehensive primary descriptive data of the electric arc phenomenon can be constructed.

[0049] Step S102: The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient.

[0050] refer to Figure 2 and Figure 3 As shown, the current sampling data is specifically treated as a macroscopic manifestation of the motion of a large number of charged particles (including electrons and ions), whose state changes follow a stochastic process. The probability density function P(x,t) of the current sampling data amplitude is estimated using a sliding time window. The window duration, sliding step size, and number of samples within the window are all set based on practical considerations and will not be elaborated upon here. The probability density function for obtaining the current sampling data amplitude using a sliding time window adopts a conventional approach in this field and will not be elaborated upon here.

[0051] Then, the probability density function P(x,t) obtained by fitting the Fokker-Planck equation is used to obtain the drift coefficient function μ(x,t) and the diffusion coefficient function D(x,t). The drift coefficient function μ(x,t) represents the deterministic trend of the current sampling data in the state space, such as directional motion influenced by the system voltage; while the diffusion coefficient function D(x,t) represents the intensity of random fluctuations in the current sampling data, such as disordered motion caused by collisions or ionization.

[0052] The Fokker-Planck equations are as follows:

[0053] Where P(x,t) represents the probability density function, μ(x,t) represents the drift coefficient function, D(x,t) represents the diffusion coefficient function, x represents the instantaneous amplitude of the current sampling data, and t represents time.

[0054] Experiments revealed that real fault arcs originate locally, with a definite driving force (system voltage) for particle motion. Therefore, their process exhibits a relatively strong drift term and a weak diffusion term. In contrast, crosstalk signals, as externally coupled signals, are more influenced by the random parameters of the coupling path, resulting in a significantly enhanced diffusion term and a weakened and unstable drift term. Therefore, after obtaining the drift coefficient function μ(x,t) and the diffusion coefficient function D(x,t), the correlation parameters characterizing the drift and diffusion coefficients can be estimated using numerical substitution. Based on these correlation parameters, local arcs and external crosstalk signals can be effectively distinguished.

[0055] For example, a series of drift coefficients and a series of diffusion coefficients can be estimated based on numerical substitution. Then, it is determined whether all drift coefficients are less than a preset drift threshold and whether the standard deviation of all drift coefficients is greater than a preset stability threshold. If so, it indicates that the drift term of the current sampling data of the circuit under test in the current observation period has weakened and unstable characteristics. At the same time, it is determined whether all diffusion coefficients are greater than a preset threshold. If so, it indicates that the diffusion term of the current sampling data of the circuit under test in the current observation period is enhanced. At this time, the correlation characterization parameter representing the drift coefficient and the diffusion coefficient is set to A. If a series of drift coefficients and a series of diffusion coefficients do not meet the above requirements, the correlation characterization parameter representing the drift coefficient and the diffusion coefficient is set to B, where A is greater than B. The above numerical substitution method can also be replaced by the moment estimation method, the maximum likelihood estimation method, etc., and this embodiment does not impose fixed restrictions on this. At the same time, the corresponding values ​​of the correlation characterization parameters representing the drift coefficient and the diffusion coefficient can also be obtained based on other determination methods, and these are not fixedly restricted here either.

[0056] Step S103: Based on the feature data set and associated characterization parameters, determine whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy.

[0057] refer to Figure 2 As shown, the time-domain statistical feature data and frequency-domain statistical feature data in the feature data set are used as evidence to represent the phenomenon, and the correlation feature parameters representing the drift coefficient and diffusion coefficient are used as evidence to represent the physical mechanism. The above three types of data are fused through a preset discrimination strategy to obtain a comprehensive score for crosstalk suspicion that represents whether there is a crosstalk signal.

[0058] This embodiment can construct an interpretable, low-computational-complexity weighted decision model, specifically by weighting and summing the time-domain statistical feature data, frequency-domain statistical feature data, and the associated characterization parameters representing the drift coefficient and diffusion coefficient in the feature data set to obtain a comprehensive score for crosstalk suspicion.

[0059] Furthermore, since the associated characterization parameters representing the drift coefficient and the diffusion coefficient reveal the essential physical difference between local fault arcs and crosstalk signal interference, the associated characterization parameters are given a greater weight than the other two types of data in the above data fusion process. The specific weight values ​​can be determined based on the actual situation, and no fixed restrictions are imposed on them here.

[0060] After obtaining the crosstalk suspicion comprehensive score, it is determined whether the crosstalk suspicion comprehensive score is greater than the preset crosstalk threshold. If it is, it indicates that the target observation period of the circuit under test is affected by the crosstalk signal; otherwise, it indicates that the target observation period of the circuit under test is not affected by the crosstalk signal, meaning that the circuit under test is normal within the target observation period. Based on this, it can be determined whether the target observation period of the circuit under test is affected by crosstalk signal interference.

[0061] The aforementioned method weighted fusion of time-domain features, frequency-domain features, and the correlation parameters between drift and diffusion coefficients overcomes the limitation of single-dimensional analysis failing in complex scenarios. Furthermore, this method does not rely on the absolute stability of any single feature; even if a certain type of feature is disturbed under specific extreme conditions, other features can still provide effective information, ensuring the robustness of the overall system decision-making and enabling it to adapt to a wider range of industrial and civilian scenarios; thus enhancing the robustness and adaptability of the recognition framework.

[0062] After determining whether the observation period of the target circuit under test is affected by crosstalk signal interference, it is possible to further determine whether there is a fault arc in the observation period of the target circuit under test based on the feature data in the current feature data set. Therefore, the following embodiment also includes step S104.

[0063] In one embodiment of this application, the arc crosstalk identification method further includes step S104.

[0064] Step S104: When it is determined that the target observation period of the circuit under test is affected by crosstalk signal, it is determined whether there is a strong local arc feature in the feature data set. If so, it means that there is a real fault arc in the target observation period of the circuit under test and it is accompanied by crosstalk signal interference. Then, a trip command is sent to the protection actuator so that the protection actuator executes the trip command and controls the circuit under test to be de-energized. Otherwise, it means that the target observation period of the circuit under test is only affected by crosstalk signal.

[0065] refer to Figure 3As shown in this embodiment, after determining that the target observation period of the circuit under test is affected by crosstalk signals, it further determines whether there are strong local features in the acquired corresponding feature data set. If so, it indicates that there is a real fault arc in the target observation period of the circuit under test. At this time, a trip command needs to be sent to the protection actuator, which then executes the trip command, thereby controlling the circuit under test to disconnect and avoid safety accidents such as fires caused by fault arcs. When there are no strong local features in the corresponding feature data set, it indicates that no fault arc occurred in the target observation period of the circuit under test, and it was only affected by crosstalk signals.

[0066] Furthermore, the strong local arc characteristic includes at least one of a peak factor greater than a preset threshold factor and a preset arc harmonic. The strong local arc characteristic may also include other characteristic data that meet certain requirements, which are not subject to fixed restrictions here.

[0067] The thresholds and logic in the identification method of this embodiment are based on physical mechanisms, making the judgment results more accurate. This method can clearly distinguish between three states: pure crosstalk, crosstalk accompanied by a real arc, and a pure real arc. For the pure crosstalk state, it adopts the principle of recording only and not malfunctioning, which greatly improves the continuity of power supply while ensuring safety. At the same time, the above decisions can be traced back to three specific characteristic values ​​and physical parameters, making "why it is judged as crosstalk" or "why it is judged as a real fault" transparent and explainable. This greatly facilitates product debugging, on-site fault diagnosis, and responsibility determination, and is conducive to passing strict safety certification.

[0068] To verify the effectiveness of the arc crosstalk identification method of this application, the following simulation is performed using a circuit to be verified as an example, and the process of identifying fault arcs and crosstalk signals in the circuit is simulated using the arc crosstalk identification method of this application. Figure 4 This is a schematic diagram showing the comparison and differences of drift coefficients in the process of identifying fault arcs and crosstalk signals using the arc crosstalk identification method described in this application embodiment. Figure 5 This diagram illustrates the comparison and differences in diffusion coefficients during the identification of fault arcs and crosstalk signals using the arc crosstalk identification method described in this application. (Reference) Figure 4 and Figure 5 As shown, the drift coefficient and diffusion coefficient of fault arc and crosstalk signal are quite different. Therefore, this embodiment uses the drift coefficient and diffusion coefficient to distinguish fault arc and crosstalk signal interference, which has strong accuracy and reliability.

[0069] This application's arc crosstalk identification method, by analyzing the correlation between the drift coefficient μ(t) and the diffusion coefficient D(t), can directly distinguish at the physical level whether a signal originates from a deterministic process driven by a local strong voltage field (a real arc) or from a random disturbance (crosstalk) originating from an external coupling path. This judgment is unaffected by surface interference from changes in load type, significantly improving the crosstalk identification accuracy (reaching over 95%) in complex electrical environments (such as those containing numerous frequency converters and switching power supplies), while maintaining an extremely high detection rate for real arcs (>98%), fundamentally reducing the risk of false tripping and failure to trip of protection devices.

[0070] like Figure 6 As shown in the figure, this application embodiment also provides an arc crosstalk identification device 10, including a feature extraction module 11, an association characterization parameter acquisition module 12, and a crosstalk identification module 13.

[0071] The feature extraction module 11 is used to acquire current sampling data of the target observation period of the circuit to be detected, and to extract features from the current sampling data to obtain a feature data set.

[0072] The correlation characterization parameter acquisition module 12 is used to construct a stochastic process model from the current sampling data and fit the current sampling data based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient.

[0073] The crosstalk identification module 13 is used to determine whether the observation period of the target circuit under test is affected by crosstalk signals based on the feature data set and associated characterization parameters, and through a preset discrimination strategy.

[0074] Furthermore, the correlation characterization parameter acquisition module 12 is specifically used for: The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient, including: Obtain the probability density function of the signal amplitude in the current sampling data, and fit the probability density function based on the Fokker-Planck equation to obtain the drift coefficient function and the diffusion coefficient function; Relative strength and stability analysis are performed on the drift coefficient function and the diffusion coefficient function to obtain the correlation characterization parameters that characterize the drift coefficient and the diffusion coefficient.

[0075] Furthermore, the correlation characterization parameter acquisition module 12 is specifically used for: The probability density function of the signal amplitude in the current sampling data is estimated by using a sliding time window.

[0076] Furthermore, the feature extraction module 11 is specifically used for: Feature extraction of the current sampling data to obtain a feature data set includes: Time-domain features are extracted from the current sampling data to obtain a time-domain statistical feature dataset; Frequency domain features are extracted from the current sampling data to obtain a frequency domain statistical feature dataset; The time-domain statistical feature dataset and the frequency-domain statistical feature dataset are merged into a feature data set.

[0077] Furthermore, the feature extraction module 11 is specifically used for: The time-domain statistical feature dataset includes at least one of peak factor, zero-crossing rate, skewness, and kurtosis; the frequency-domain statistical feature dataset includes at least one of dominant frequency, frequency center, bandwidth, harmonic ratio, and spectral entropy.

[0078] Furthermore, the crosstalk identification module 13 is specifically used for: Based on the feature data set and the associated characterization parameters, determining whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy includes: The feature data in the feature data set and the associated characterization parameters are weighted and fused to obtain a comprehensive score for crosstalk suspicion. If the crosstalk suspicion comprehensive score is greater than the preset crosstalk threshold, it is determined that the observation period of the target circuit under test is affected by the crosstalk signal; otherwise, it is determined that the observation period of the target circuit under test is not affected by the crosstalk signal.

[0079] Furthermore, the arc crosstalk identification method also includes a fault arc identification module 14; The fault arc identification module 14 is used to determine whether there is a strong local arc feature in the feature data set when the target observation period of the circuit under test is affected by crosstalk signal. If so, it indicates that there is a real fault arc and crosstalk signal interference in the target observation period of the circuit under test, and sends a trip command to the protection actuator so that the protection actuator executes the trip command and controls the circuit under test to be de-energized. Otherwise, it indicates that the target observation period of the circuit under test is only affected by crosstalk signal. The strong local arc characteristics include at least one of a peak factor greater than a preset factor threshold and the presence of a preset arc harmonic.

[0080] This application's arc crosstalk identification device fundamentally distinguishes between local arcs and crosstalk from a mechanistic perspective, significantly improving identification accuracy and reliability, achieving interpretable intelligent decision-making, and constituting an evolvable networked security system. These effects collectively solve the long-standing problem of malfunction in arc crosstalk identification devices, demonstrating significant technological advancement and commercial application value.

[0081] like Figure 7 As shown in the figure, this application embodiment also provides an arc crosstalk identification system 1, which includes a cloud service platform 20, an arc crosstalk identification device 10 and a protection actuator 30 that are interconnected, wherein the arc crosstalk identification device is the arc crosstalk identification device 10 in the above embodiment.

[0082] The arc crosstalk identification device 10 is mainly used to perform the steps in the above-described arc crosstalk identification method to identify crosstalk in the circuit under test. Simultaneously, it packages multiple types of data, including at least current sampling data, into a defined data packet. The arc crosstalk identification device 10 can periodically or event-triggeredly send the aforementioned data packet to the cloud service platform 20 via a communication interface.

[0083] The protection actuator 30 can be a circuit breaker or relay controller containing a tripping mechanism, installed in the circuit under test. The protection actuator 30 is primarily used to receive tripping commands sent by the arc crosstalk identification device 10 when a fault arc is detected. Upon receiving the tripping command, its tripping steps can be: immediately or after a very short configurable delay to initiate the tripping process; or illuminating a warning indicator or recording the event, but without direct tripping, continuously monitoring subsequent commands. The protection actuator 30 can also receive parameter configuration information from the cloud service platform 20 and update parameters based on this information; or send its own important event information to the cloud service platform 20 for information reporting. Once the arbitration logic of the protection actuator 30 determines to initiate the tripping process, it must drive a high-current switch (such as a MOSFET, relay, or mechanical tripping mechanism) to forcibly disconnect the circuit under test.

[0084] In one embodiment of this application, the protection actuator 30 may also feed back its own tripping status data (such as "tripped", "not tripped", "coil fault" etc.) to the cloud service platform 20 after the tripping action, so that the cloud service platform 20 can make subsequent reminders or operations.

[0085] The cloud service platform 20 is mainly used to receive current sampling data sent by the arc crosstalk identification device and tripping status data sent by the protection actuator. After acquiring the data, it parses all incoming data, structures it, and stores it in a database or time-series database to form a long-term "electrical fingerprint" archive.

[0086] Meanwhile, the cloud service platform 20 can also perform the following operations based on the input data. For example, it can run advanced analysis algorithms to identify the distribution patterns of crosstalk and arcing events in specific regions or load types; verify and fine-tune the weight and threshold parameters in Fokker-Planck analysis based on massive amounts of real data to form a more universal or more suitable "optimized algorithm package" for specific scenarios; draw regional electrical safety heat maps to predict potential high-risk nodes; and generate analysis reports, optimized algorithm parameter packages, and regional risk warning information.

[0087] The cloud service platform 20 can also analyze the optimization parameter packages generated by the analysis module; batch management commands initiated by maintenance personnel; process and formulate secure distribution strategies (such as batch distribution or distribution during off-peak electricity periods), push commands or update packages to target device groups; and output configuration update commands or new firmware packages for arc crosstalk identification devices. Simultaneously, it pushes regional risk reports to the mobile devices of relevant responsible persons.

[0088] To illustrate how the arc crosstalk identification system in this embodiment works, three practical scenarios are used to explain it below.

[0089] Scenario 1: Crosstalk from an electric arc is sensed in adjacent lighting circuits.

[0090] The lighting circuit itself is intact and fault-free. The abnormal fluctuations in its current are entirely due to electromagnetic induction coupling. This induced voltage is random and unstable, and is not a direct product of the lighting circuit's own voltage field. Therefore, the directional drift force on the current "particles" is very weak (μ(t)≈0), and their movement mainly manifests as violent and disordered oscillations (strong diffusion) caused by the random parameter changes of the coupling path. The diffusion coefficient D(t) calculated by the algorithm will be abnormally prominent, and both its value and variation lack regularity.

[0091] Scenario 2: A high-power inverter air conditioner starts up in a smart building data center.

[0092] The air conditioner compressor generates a huge surge current and a lot of electrical noise when it starts up. This noise is coupled to the power supply lines of the adjacent server rack through the power distribution bus.

[0093] Based on the arc crosstalk identification device of this application, the algorithm initially determined that the device was "suspected of being affected by crosstalk." Further analysis was conducted to determine if there was a flat-shoulder current (a characteristic of a series arc) lasting more than half a cycle, if there was a periodic arc-extinguishing point strictly synchronized with the power supply frequency, and if the harmonic components exhibited a specific proportion of the arc. Since the air conditioner's startup noise is broadband, it does not possess the aforementioned deterministic arc characteristics. Therefore, the algorithm ultimately concluded that the output was pure crosstalk signal interference. Simultaneously, it did not trigger a trip, ensuring uninterrupted power supply to the server, and recorded the detailed feature vector and decision process of this event in the local log. After receiving the event report, the cloud service platform marked the "coupling relationship" between the air conditioner and the server line in its big data. If similar events occur frequently in the future, the cloud service platform can issue a warning of "line layout optimization suggestions" to the maintenance personnel, rather than a fire alarm.

[0094] Scenario 3: A photovoltaic power station maintenance worker handles a frequent alarm.

[0095] Continuous operation of the device described in this application revealed that the Fokker-Planck characteristics of each alarm exhibited a "strong diffusion" pattern, and secondary checks failed to uncover evidence of strong local arcing. Therefore, it was classified as "persistent suspected crosstalk," and the reporting priority of this event was increased by packaging and uploading the characteristic data of 10 consecutive events.

[0096] The cloud service platform invoked its big data analytics module to horizontally compare with other strings in the same power station, finding no similar phenomenon and ruling out common-mode interference. Vertical analysis of the string's historical data revealed that "strong diffusion" events consistently occurred during periods of rapid change in sunlight intensity (such as when clouds passed). Combining this with the power station topology map, the platform discovered that the alarm string was laid parallel to a DC bus trunk cable for a considerable distance. It was inferred that other branches on the trunk cable experienced significant power fluctuations due to changes in sunlight, generating numerous switching harmonics that coupled to this string via the long parallel cabling. Therefore, a diagnostic report was output, concluding that "there is a high probability of electromagnetic crosstalk caused by cabling coupling; it is recommended to check the spacing and shielding between line A and trunk cable B." Simultaneously, the platform issued a set of finely adjusted parameters to the device: slightly increasing the crosstalk threshold for this specific location to enhance immunity and reduce invalid alarms. An alarm notification and a detailed diagnostic report were also sent to the mobile device.

[0097] This embodiment upgrades arc fault protection from a single device function to a comprehensive safety system capable of global monitoring, intelligent analysis, and remote management. Maintenance personnel can monitor the entire network's security situation in real time, accurately locate high-risk lines or areas, and shift from passive response to proactive prevention, providing crucial technical support for the safe operation of the smart grid. Furthermore, it enables continuous self-evolution of protection performance: local data accumulated by individual devices during operation (especially feature vectors of false alarms and missed alarms) is uploaded to the cloud service platform and used for global model optimization. The optimized parameter model can then be redistributed to the device cluster, allowing the entire product network's identification capabilities to continuously improve and enhance with deployment scale and operating time—something that static, isolated traditional protection devices cannot achieve.

[0098] This application also provides a readable storage medium storing a computer program, which, when executed by a processor, implements the steps described in the above-described method embodiments.

[0099] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0100] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0101] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0102] In the embodiments provided in this application, it should be understood that the disclosed apparatus / device and method can be implemented in other ways. For example, the apparatus / device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0103] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0104] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for identifying arc crosstalk, characterized in that, include: Acquire current sampling data of the target circuit under test during the observation period, and perform feature extraction on the current sampling data to obtain a feature data set; The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient. Based on the feature data set and the associated characterization parameters, a preset discrimination strategy is used to determine whether the observation period of the target circuit to be detected is affected by crosstalk signals.

2. The arc crosstalk identification method according to claim 1, characterized in that, The current sampling data is constructed into a stochastic process model, and the current sampling data is fitted based on the Fokker-Planck equation to obtain the correlation characterization parameters of the drift coefficient and the diffusion coefficient, including: Obtain the probability density function of the signal amplitude in the current sampling data, and fit the probability density function based on the Fokker-Planck equation to obtain the drift coefficient function and the diffusion coefficient function; Relative strength and stability analysis are performed on the drift coefficient function and the diffusion coefficient function to obtain the correlation characterization parameters that characterize the drift coefficient and the diffusion coefficient.

3. The arc crosstalk identification method according to claim 2, characterized in that, The probability density function for obtaining the signal amplitude in the current sampling data includes: The probability density function of the signal amplitude in the current sampling data is estimated by using a sliding time window.

4. The arc crosstalk identification method according to claim 1, characterized in that, Feature extraction of the current sampling data to obtain a feature data set includes: Time-domain features are extracted from the current sampling data to obtain a time-domain statistical feature dataset; Frequency domain features are extracted from the current sampling data to obtain a frequency domain statistical feature dataset; The time-domain statistical feature dataset and the frequency-domain statistical feature dataset are merged into a feature data set.

5. The arc crosstalk identification method according to claim 4, characterized in that, The time-domain statistical feature dataset includes at least one of peak factor, zero-crossing rate, skewness, and kurtosis; the frequency-domain statistical feature dataset includes at least one of dominant frequency, frequency center, bandwidth, harmonic ratio, and spectral entropy.

6. The arc crosstalk identification method according to claim 1, characterized in that, Based on the feature data set and the associated characterization parameters, determining whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy includes: The feature data in the feature data set and the associated characterization parameters are weighted and fused to obtain a comprehensive score for crosstalk suspicion. Determine whether the crosstalk suspicion comprehensive score is greater than the preset crosstalk threshold. If it is, it means that the target observation period of the circuit under test is affected by the crosstalk signal; otherwise, it means that the target observation period of the circuit under test is not affected by the crosstalk signal.

7. The arc crosstalk identification method according to claim 6, characterized in that, Also includes: When it is determined that the target observation period of the circuit under test is affected by crosstalk signal, it is determined whether there is a strong local arc feature in the feature data set. If so, it indicates that there is a real fault arc in the target observation period of the circuit under test and it is accompanied by crosstalk signal interference. A trip command is sent to the protection actuator so that the protection actuator executes the trip command and controls the circuit under test to be de-energized. Otherwise, it indicates that the target observation period of the circuit under test is only affected by crosstalk signal. The strong local arc characteristics include at least one of a peak factor greater than a preset factor threshold and a preset arc harmonic.

8. An arc crosstalk identification device, characterized in that, It includes a feature extraction module, an association representation parameter acquisition module, and a crosstalk identification module; The feature extraction module is used to acquire current sampling data of the target observation period of the circuit to be detected, and to perform feature extraction on the current sampling data to obtain a feature data set. The correlation characterization parameter acquisition module is used to construct a stochastic process model from the current sampling data and fit the current sampling data based on the Fokker-Planck equation to obtain the correlation characterization parameters between the drift coefficient and the diffusion coefficient. The crosstalk identification module is used to determine, based on the feature data set and the associated characterization parameters, whether the observation period of the target circuit to be detected is affected by crosstalk signals through a preset discrimination strategy.

9. An arc crosstalk identification system, characterized in that, It includes a cloud service platform, an arc crosstalk identification device, and a protection actuator that are interconnected, wherein the arc crosstalk identification device is the arc crosstalk identification device as described in claim 8; The arc crosstalk identification device is also used to upload current sampling data to the cloud service platform; The protection actuator is used to control the circuit under test to disconnect power when a trip command is received, and to feed back the trip status data to the cloud service platform; The cloud service platform is used to store the current sampling data and the tripping status data.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images