Electrical fire monitoring method based on automatic compensation of inherent residual current

Abstract

The present invention discloses an electrical fire monitoring method based on automatic compensation of inherent residual current, which relates to the field of electrical safety monitoring technology and includes the following steps: S100, collecting residual current signals, performing multi-component time-frequency deconstruction processing, extracting energy distribution characteristics through wavelet packet transform, and extracting mutation indicators in combination with short-time spectral entropy analysis, constructing a preliminary distribution map of high-order harmonic interference, and determining the boundary characteristics of the interference signal in the time domain and frequency domain. The present invention accurately locates high-frequency interference signals through time-frequency deconstruction, wavelet packet analysis and spectral entropy indicators, effectively eliminates non-fault harmonic components by combining amplitude-frequency coupling identification and recursive elimination strategies, and introduces adaptive compensation and stability backtracking mechanisms to construct closed-loop control, thereby achieving accurate identification and dynamic correction of real leakage risks, significantly improving the recognition accuracy and safety reliability of the monitoring system in complex industrial environments, and having good engineering adaptability.

Description

Technical Field

[0001] This invention relates to the field of electrical safety monitoring technology, specifically to an electrical fire monitoring method based on automatic compensation of inherent residual current. Background Art

[0002] "Electrical fire monitoring based on automatic compensation of inherent residual current" refers to the ability to identify and model inherent residual current (i.e., non-faulty residual current generated under normal operating conditions due to factors such as cable structure, equipment leakage inductance, and capacitive coupling) in electrical systems during operation. This allows for the dynamic differentiation between normal residual current and abnormal leakage behavior, and the automatic elimination of inherent components through a compensation mechanism, thereby achieving accurate identification and early warning of actual leakage risks. This method effectively avoids the false alarms or missed alarms caused by inherent residual current in traditional monitoring systems, improving the sensitivity, accuracy, and reliability of fire hazard identification. It is particularly suitable for intelligent monitoring and early warning response to early signs of electrical fires (such as insulation damage and grounding faults) in complex power distribution scenarios.

[0003] The existing technology has the following shortcomings:

[0004] In industrial load environments, nonlinear loads such as frequency converters, welding equipment, and power rectifiers are widely used, easily introducing high-order harmonic signals with multiple frequency components into the power distribution system. These harmonic signals are superimposed on the normal operating current through channels such as conductors and ground coupling capacitors, and are mixed into the monitoring signal path during residual current acquisition, resulting in non-power frequency components in the residual current waveform acquired by the system. Because the spectral characteristics of high-order harmonics overlap to some extent with the waveform of the actual leakage current, current automatic compensation algorithms based on waveform amplitude and frequency threshold judgment are difficult to effectively distinguish between the two sources. This may lead to misidentification of high-order harmonic interference as inherent residual current components for compensation processing, or misidentification as abnormal leakage current signals triggering alarms, ultimately causing misjudgments and missed judgments, seriously affecting the accuracy and reliability of electrical fire monitoring systems.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure. Summary of the Invention

[0006] The purpose of this invention is to provide an electrical fire monitoring method based on automatic compensation of inherent residual current, so as to solve the problems in the background art mentioned above.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an electrical fire monitoring method based on automatic compensation of inherent residual current, comprising the following steps:

[0008] S100 acquires the residual current signal, performs multi-component time-frequency deconstruction processing, extracts energy distribution characteristics through wavelet packet transform, and extracts abrupt change indicators by combining short-time spectral entropy analysis, constructs a preliminary distribution map of high-order harmonic interference, and determines the boundary characteristics of the interference signal in the time and frequency domains.

[0009] S200, based on the preliminary distribution map of high-order harmonic interference, performs amplitude-frequency coupling relationship identification, analyzes the phase shift and response delay between the power frequency and the interference frequency band, identifies the dynamic correlation between atypical leakage and high-order harmonic interference in the residual current, and determines the location and change path of the interference frequency band.

[0010] S300 performs recursive harmonic interference removal based on the location and change path of the interference frequency band, uses differential filtering to remove interference components, and retains low-frequency non-steady-state components through local waveform reconstruction to obtain the residual current signal after structural purification.

[0011] S400, based on the residual current signal after structural purification, performs adaptive compensation control, combines historical operating conditions and the current rate of change to generate a compensation reference curve for the inherent residual current, and dynamically adjusts the compensation amplitude.

[0012] S500, based on the adaptive compensation control process, constructs a stability index backtracking mechanism to track the signal stability within multiple sampling periods, analyzes response lag, amplitude offset and error accumulation, and corrects the adaptive compensation control parameters.

[0013] The S600, based on the modified adaptive compensation control parameter system, performs joint closed-loop control, integrating time-frequency deconstruction, coupling identification, interference elimination, compensation control and backtracking mechanism to achieve accurate identification and alarm of real leakage risk under high-order harmonic interference conditions.

[0014] Preferably, step S100 includes:

[0015] The zero-sequence current signal in the three-phase power distribution line is collected, and after high-frequency interference is filtered out by an anti-aliasing low-pass filter, it is digitized by an analog-to-digital converter chip at a sampling rate of more than 10,000 times per second.

[0016] The digitized residual current signal is subjected to four-level wavelet packet decomposition. The db6-order function in the Daubechies wavelet function is used to decompose the residual current signal into 16 sub-bands, and the frequency domain energy in each sub-band is calculated.

[0017] Based on the frequency band energy distribution results, short-time spectral entropy analysis is introduced. The signal is windowed according to the power frequency cycle and the sliding window is used to extract the energy concentration and structural change characteristics of each frame of the signal.

[0018] A comprehensive two-dimensional time-frequency disturbance map is constructed to identify the frequency boundaries, duration, and disturbance intensity of high-frequency interference signals, thereby completing the construction of a preliminary distribution map of high-order harmonic interference.

[0019] Preferably, step S200 includes:

[0020] Extract frequency sub-bands with energy density higher than a preset threshold and spectral entropy lower than a preset threshold, and establish a pairing relationship between the power frequency band and the interference frequency band;

[0021] The instantaneous phase shift between corresponding frequency bands is calculated based on the Hilbert transform, and its phase coupling characteristics are determined.

[0022] Using the time location of the residual current mutation point as the anchor point, calculate the response delay of each frequency band signal and determine its time-domain response dependency.

[0023] Based on the above analysis results, a frequency band behavior identification result set is formed, and the coupling type and change path of the interference frequency band are marked.

[0024] Preferably, step S300 includes:

[0025] Construct a set of frequency templates for the frequency bands to be removed from interference, and extract the corresponding frequency band node coefficients to reconstruct the interference component signals;

[0026] After dynamic time warping, differential frequency filtering is performed to initially eliminate interference frequency band signals.

[0027] A sliding time window is set to perform local waveform reconstruction, preserving low-frequency unsteady components and performing amplitude correction;

[0028] Based on the high-frequency energy ratio and reconstruction error, a threshold for elimination iteration is set, and multiple rounds of interference elimination are recursively performed to output the residual current signal after structural purification.

[0029] Preferably, step S400 includes:

[0030] Construct a working condition feature library containing residual current feature vectors and compensation reference waveforms from historical operating scenarios;

[0031] Extract the residual current characteristics in the current cycle and calculate the Euclidean distance between them and the historical template, determine the optimal matching template and generate the dynamic compensation target curve;

[0032] Compare the current signal with the target compensation curve point by point and perform amplitude compensation and delay adjustment operations within the set threshold range;

[0033] Statistical error indicators are analyzed over multiple periods, and the compensation baseline curve is reconstructed based on the error feedback results to achieve closed-loop optimization.

[0034] Preferably, step S500 includes:

[0035] Ten consecutive sampling periods are set as the analysis window. The maximum amplitude, minimum amplitude, root mean square value, average rise slope and energy center frequency of each period are extracted to construct a periodic feature vector set.

[0036] The range, standard deviation and mean square deviation of each parameter are calculated based on the periodic feature vector set, and the stability index of the current window is constructed by combining them according to the set weights.

[0037] When the stability index is lower than the set threshold, the compensation output is compared with the compensation reference curve on a cycle-by-cycle basis to extract the response delay point, amplitude offset and error accumulation value to identify the cause of instability.

[0038] Adjust the compensation control parameters based on the identification results and enable the correction value in the next cycle, while recording the correction information for subsequent tracking and analysis.

[0039] Preferably, step S600 includes:

[0040] A unified sampling time reference and data transmission link are set, and the frequency domain energy matrix and perturbation spectrum of the residual current signal are extracted by wavelet packet decomposition as input;

[0041] Based on the frequency domain spectrum, amplitude-frequency coupling identification processing is performed to extract the phase shift, response delay, and energy index of the interference frequency band, drive three rounds of recursive interference removal processing, and output a purified signal.

[0042] Based on the purification signal, the current compensation parameter system is loaded, and the target compensation curve is generated by combining the rate of change and the dynamic adjustment coefficient. The signal is adjusted point by point and amplitude control is executed.

[0043] At the end of the cycle, the compensation effect and stability index are analyzed. If the compensation accuracy decreases, the dynamic parameters and template matching are corrected, the parameter system is updated and loaded into the next cycle to complete the closed-loop control.

[0044] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0045] This invention achieves accurate localization of high-frequency interference signals by introducing multi-component time-frequency deconstruction, wavelet packet energy extraction, and spectral entropy abrupt change analysis. It effectively distinguishes and eliminates non-faulty harmonic components by establishing an amplitude-frequency coupling behavior recognition mechanism and a recursive interference elimination strategy. Furthermore, by combining historical operating conditions and dynamic trends, it implements adaptive compensation and stability backtracking adjustment, constructing a closed-loop feedback control logic that enables the system to continuously optimize judgment criteria and correct compensation strategies online. Overall, this solution solves the problem of misjudgment and missed reporting caused by the inability to distinguish between high-order harmonics and real leakage signals in traditional systems. It achieves precise modeling and intelligent response to abnormal current behavior in complex power distribution scenarios, exhibiting high engineering adaptability and safety protection. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0047] Figure 1 This is a flowchart of the electrical fire monitoring method based on automatic compensation of inherent residual current according to the present invention. Detailed Implementation

[0048] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0049] This invention provides, for example Figure 1 The electrical fire monitoring method based on automatic compensation of inherent residual current, as shown, includes the following steps:

[0050] S100 acquires the residual current signal and performs multi-component time-frequency deconstruction processing on the residual current signal. The processing includes extracting multi-band energy distribution characteristics through wavelet packet transform and extracting signal abrupt change index by combining short-time spectral entropy analysis, so as to construct a preliminary distribution map of high-order harmonic interference and determine the boundary characteristics of the interference signal in the time domain and frequency domain.

[0051] The process of acquiring the residual current signal and performing multi-component time-frequency decomposition processing on the signal includes the following steps:

[0052] High-precision acquisition of residual current in the electrical system is achieved. Specifically, a zero-sequence current sensor is installed in the three-phase power distribution line. This sensor monitors the unbalanced current differences between the three phases in real time and transmits the residual current signal to the signal acquisition circuit via analog signal transmission. The acquisition circuit is equipped with an anti-aliasing low-pass filter. The cutoff frequency of this filter is set below half of the system's maximum sampling frequency to suppress high-frequency interference noise outside the 50Hz frequency band. The filtered analog signal is then fed into a 16-bit high-precision analog-to-digital converter chip and digitized at a sampling rate of over 10,000 times per second, ensuring at least 200 sampling points are obtained within each 50Hz power frequency cycle, thus fully recording the detailed changes in the signal. During signal acquisition, shielded transmission cables with insulation are used to prevent external electromagnetic interference from distorting the signal.

[0053] The digitized residual current signal undergoes wavelet packet decomposition to analyze the energy characteristics of its different frequency bands from a frequency structure perspective. In this process, the Daubechies wavelet function, with its good symmetry and tight support characteristics, is selected as the base function. Through four layers of wavelet packet decomposition, the original time series signal is divided into 16 sub-bands. These sub-bands cover different frequency ranges from 0Hz to the sampling bandwidth. For example, the first sub-band might cover 0 to 312.5Hz, the second sub-band 312.5 to 625Hz, and so on, until the last sub-band covers the frequency range closest to the sampling bandwidth. Within each sub-band, the energy of the corresponding signal, i.e., the sum of the squared amplitude values ​​of the sample points, is calculated to quantify the activity level of the signal in that frequency range. The resulting frequency band energy spectrum clearly reflects the intensity distribution of the residual current in each frequency range. Particularly in the 250Hz to 3000Hz band, abnormal energy concentration indicates that this band may be subject to interference from higher harmonic signals.

[0054] Daubechies wavelet functions are a class of wavelet basis functions with good time-frequency localization capabilities and orthogonality, often used for multi-scale decomposition of non-stationary signals. In this step, they are used as basis functions for wavelet packet decomposition to decompose the acquired residual current signal into multiple frequency bands, thereby extracting the local energy features of different frequency components. Daubechies wavelets can effectively preserve transient changes and high-frequency interference features in signals, and are particularly suitable for identifying high-order harmonics introduced by equipment such as frequency converters and welding machines in industrial electrical systems. Specific selectable functions include db4, db6, and db8, among which db6 achieves a good balance between signal fidelity and computational complexity, making it the preferred function for constructing the multi-component time-frequency decomposition model in this step.

[0055] Building upon the frequency band energy distribution analysis, short-time spectral entropy analysis is introduced to further reveal the variation characteristics of the residual current signal along the time axis. Specifically, the complete signal sequence is windowed according to the length of one power frequency cycle, with each analysis window sliding forward by a fixed sample interval to obtain multiple overlapping signal frames. The frequency band energy distribution contained in each frame is standardized, and the changes in energy concentration between frequency bands are then analyzed. The trend of spectral entropy values ​​reveals the stability and structural complexity of the signal. If a frequency band consistently exhibits high energy concentration and strong structural regularity across multiple sliding windows, it indicates that the signal components in that band may originate from periodic harmonic interference generated by electrical equipment. Conversely, if the energy distribution of a frequency band is unstable, changes drastically, and exhibits strong transient abrupt changes, it is often closely related to abnormal leakage current processes. By integrating the spectral variation results from different time windows, a two-dimensional time-frequency disturbance distribution map is constructed to describe the persistence of the interference signal within the time range and its distribution pattern in the frequency space.

[0056] Finally, by combining the frequency energy distribution characteristics obtained from wavelet packet decomposition with the time-varying perturbation characteristics obtained from short-time spectral entropy analysis, a preliminary distribution map of higher harmonic interference is constructed. This map is presented as a two-dimensional matrix, where each row corresponds to an analysis period, and each column corresponds to a frequency sub-interval. The values ​​in the matrix represent the perturbation intensity of that frequency band during that period. Based on this map, frequency bands with high perturbation persistence and concentrated energy can be identified, and further parameters such as the frequency boundaries, duration, energy peak frequency, and coupling degree between the higher harmonic signals and adjacent frequency bands can be extracted.

[0057] The main purpose of this step is to provide high-resolution, structured time-frequency information support for subsequent identification and compensation of higher harmonic interference. Specifically, it involves multi-dimensional detailed analysis of the acquired residual current signal to accurately reveal the hidden non-power frequency interference characteristics. Through wavelet packet transform, the original current signal can be decomposed into multiple sub-signals with different frequency ranges, thereby extracting the local energy distribution of each frequency band and identifying whether there is abnormal energy accumulation concentrated in the high-frequency range. This lays the foundation for frequency domain localization of harmonic interference. Simultaneously, combined with short-time spectral entropy analysis, the trend of energy change over time in each frequency band can be captured, determining whether the signal exhibits abrupt changes or regularity, thus distinguishing between periodic harmonic disturbances and non-periodic leakage behavior. The preliminary distribution map of higher harmonic interference formed by combining these two methods not only clarifies the distribution boundaries of the interference signal on the time and frequency axes but also extracts its key behavioral characteristics, such as the frequency band center, disturbance duration, and frequency jump regions. The establishment of this map provides accurate, detailed, and traceable basic data support for subsequent amplitude-frequency coupling analysis, interference behavior identification, misjudgment avoidance, and dynamic compensation control. It is a key preliminary step in the entire process of intelligent identification of real leakage risks in this invention.

[0058] S200, based on the preliminary distribution map of high-order harmonic interference, performs amplitude-frequency coupling relationship identification processing. By analyzing the phase shift trend and response delay characteristics between the power frequency band and the interference frequency band, it identifies the dynamic behavior correlation between atypical leakage components and high-order harmonic interference components in the residual current signal, and determines the frequency band location and change path of harmonic interference.

[0059] The process of identifying amplitude-frequency coupling relationships based on the preliminary distribution map of higher harmonic interference can specifically include the following steps:

[0060] Based on the preliminary distribution map of high-order harmonic interference constructed in the previous steps, frequency component identification and structured partitioning are performed. Specifically, frequency sub-bands in the map that consistently exhibit energy density higher than a set threshold (e.g., 80th percentile) and spectral entropy lower than a set stability threshold (e.g., 0.3) across multiple consecutive sampling periods are defined as interference frequency bands to be judged. This region is typically concentrated in the sub-band between 150Hz and 4500Hz, including common industrial harmonic source frequency ranges such as the 3rd, 5th, 7th, and 11th harmonics. Simultaneously, the wavelet packet sub-band corresponding to the 50Hz power frequency signal is used as a reference frequency band, and its position and variation characteristics in the map are extracted to establish an analytical pairing relationship between the power frequency band and each interference frequency band, clarifying the frequency domain position and temporal distribution basis of the analyzed objects.

[0061] For the paired frequency sub-bands, the instantaneous phase changes between the power frequency band and each interference frequency band are calculated. Specifically, Hilbert transform is used to extract the analytical signal representation of each frequency band, and then the corresponding instantaneous phase trajectory is calculated. For each pair of power frequency and interference frequency bands, the phase difference change trend is statistically analyzed within four adjacent complete sampling periods, using a sliding window step of 10 milliseconds. If an interference frequency band maintains a stable phase shift of less than 30 degrees with the power frequency band within most sampling windows and exhibits a consistent synchronous change over time, it is determined to be an interference component with coupling behavior. Conversely, if the phase difference fluctuates randomly or shows a significant asynchronous trend in different periods, the interference frequency band is determined to have no strong correlation with the power frequency component in the phase dimension and is suspected to be an independent harmonic source signal.

[0062] Further time delay response analysis was performed on frequency bands exhibiting phase coupling characteristics. Abrupt change points were identified in the residual current signal using the energy surge threshold method; a surge point was defined as one where the signal energy increased by more than 30% above the average level in any sampling period. Using the time of the abrupt change point as an anchor point, the response start times of the power frequency signal and each interfering frequency signal before and after that location were detected, and the time difference was calculated as the interference response delay value. If the delay time was between 5 and 20 milliseconds, and the delay value remained consistent across multiple abrupt change point scenarios, it indicated a time-domain response dependency between the interfering frequency band and the power frequency band, classifying it as a subordinate coupled frequency band. If the delay time differed by more than 10 milliseconds between one abrupt change point, or if the response start time fluctuated drastically, it indicated that the frequency band was not driven by the main frequency band and was considered a harmonic disturbance with independent behavior.

[0063] By combining phase shift analysis and time delay analysis results, a frequency band behavior identification result set is constructed. This result set includes dimensions such as interfering frequency band number, frequency range, phase coupling score (e.g., quantization value between 0 and 1), time delay consistency index (e.g., standard deviation), and comprehensive coupling credibility label (e.g., strong coupling, medium coupling, no coupling). Based on the identification results, all interfering frequency bands in the spectrum are classified and labeled, and an interference behavior spectrum with behavior labels is drawn. This spectrum can be used to clearly mark the coupling type, response path, and behavior change trajectory over time for each interfering frequency band, thus providing an accurate input basis for subsequent recursive interference removal and identification of real leakage behavior.

[0064] The purpose of this step is to conduct an in-depth analysis of the behavioral correlations between different frequency bands in the residual current signal, based on the preliminary distribution map of high-order harmonic interference constructed in the previous stage. This allows for the accurate differentiation between high-order harmonic components that are behaviorally coupled with the power frequency signal and uncoupled, independently distributed atypical leakage components. In complex industrial power environments, high-order harmonics introduced by nonlinear load devices (such as frequency converters, welding machines, and rectifiers) are often embedded in the normal current with a certain phase delay or characteristic synchronization. Although their frequency components are higher than the power frequency range, they often exhibit a response mode that is synchronous or lagging with changes in the power frequency in the time domain. This step, by quantitatively analyzing the phase shift trend between the power frequency band and each interference frequency band, and combining it with the delay calculation of the signal abrupt change response time, can effectively identify whether the interference frequency band is excited by the main frequency signal, thereby determining its source attribute. If a certain frequency band is highly synchronized with the power frequency band in the phase dimension, and its response delay is stable and predictable across multiple sampling periods, then the interfering frequency band can be determined to be an electromagnetic harmonic highly coupled with power frequency behavior. Conversely, if it exhibits strong behavioral independence, further attention should be paid to whether it may originate from abnormal leakage paths or non-periodic discharge signals caused by insulation deterioration. Through this step, the system can not only more accurately locate the specific position of the interfering frequency band in the spectrum, but also track its change path based on the time evolution law, providing necessary behavioral boundary division and frequency domain dynamic mapping support for subsequent high-precision harmonic elimination and extraction of real leakage signals. This identification mechanism breaks through the limitations of traditional static frequency threshold and amplitude judgment methods, enabling residual current monitoring to have stronger intelligent judgment capabilities and adaptability in complex electromagnetic interference backgrounds.

[0065] S300 performs recursive harmonic interference removal processing based on the location and change path of harmonic interference frequency bands. The removal processing includes using a different frequency differential filtering method to remove the identified interference frequency band signals and using a local waveform reconstruction method to retain low-frequency non-steady-state components, thereby obtaining the residual current signal after structural purification.

[0066] Based on the frequency band location and variation path of higher harmonic interference, the recursive harmonic interference removal process includes the following steps:

[0067] A frequency template set for the interference bands to be eliminated is constructed. Based on the frequency band behavior identification results extracted during the amplitude-frequency coupling relationship identification process in the previous stage, frequency sub-bands with strong coupling behavior in multiple sampling periods are selected. Taking a sampling frequency of 10kHz and a wavelet packet decomposition level of 4 layers as an example, 16 equal-bandwidth sub-bands can be obtained. By statistically analyzing the phase offset, response delay, and energy density ratio between the power frequency reference band (0–312.5Hz) and high-frequency bands (such as 625–937.5Hz, 937.5–1250Hz, and 1250–1562.5Hz), the above three frequency ranges are selected as significant high-order harmonic interference bands. These frequency bands, along with their respective start and end frequencies, center frequencies, average energy values, and trend markers, form an interference frequency template set, which serves as the target frequency set for subsequent differential filtering elimination processing.

[0068] Differential filtering is performed on the interference components corresponding to the aforementioned frequency templates. The specific steps are as follows: Within each sampling period, the original residual current signal is decomposed using four layers of wavelet packets to extract the node coefficients corresponding to the aforementioned interference frequency bands (e.g., nodes 5, 6, and 7); these node coefficients are then reconstructed separately into an interference component signal sequence; subsequently, a point-to-point subtraction method is used to subtract the reconstructed interference signal from the original residual current signal sample by sample to obtain the initial filtering result. To prevent phase alignment errors from causing overlapping frequency attenuation errors, a dynamic time warping method is used to locally align the starting and peak points of the original and interference signals, ensuring maximum synchronization of the differential operation at key waveform feature points, thereby improving the accuracy of interference removal.

[0069] Based on the initially filtered signal, local waveform reconstruction is performed to protect the low-frequency non-steady-state waveform structure that may contain real leakage characteristics. This step uses a sliding energy detection method to identify the location of abrupt changes in the signal by setting a local time window (e.g., 500 sampling points). The detection logic is as follows: when the signal energy surge within the local window exceeds 1.5 times the average value of the previous period's window and lasts for more than 10 milliseconds, this segment is considered a non-steady-state segment that may contain abnormal leakage behavior. Within the non-steady-state segment, a forward predictive interpolation algorithm is used to compensate for any energy peaks that may have been mistakenly deleted from the local waveform, and a backward amplitude recalibration mechanism is used to perform amplitude correction by referring to the historical peak range of the previous 10 power frequency cycles. This method ensures that even when the boundaries of the interference frequency band and the abnormal signal frequency band overlap, the sudden behavior characteristics of the real leakage signal can still be preserved to the maximum extent.

[0070] A recursive elimination process is constructed to perform multiple rounds of iterative purification on the remaining interference. The high-frequency energy retention rate and waveform reconstruction error after two consecutive rounds of elimination are used as the judgment criteria. Specifically, if the energy proportion above 625Hz in the signal after elimination is still greater than 5%, or the local window reconstruction error (i.e., the mean square error of the difference between the reconstructed waveform and the original signal waveform) exceeds 3% of the average amplitude of the original signal, then a second round of elimination is executed, repeating the above differential and reconstruction steps. After each iteration, the energy statistics parameters of the interference template are updated, and the frequency band coverage is dynamically adjusted to adapt to the actual interference change path. The iteration process is set to a maximum of three rounds. When the purification criteria are met—high-frequency energy proportion less than 5%, reconstruction error less than 2%, and abrupt change detection accuracy higher than 90%—purification is considered complete, and the final purified residual current signal is output.

[0071] The purpose of this step is to implement targeted signal purification processing based on the identification of the location of harmonic interference frequency bands and their time-varying paths. This effectively removes non-leakage interference components introduced by high-order harmonic coupling from the residual current signal, while preserving the non-steady-state and non-periodic characteristics of potential abnormal leakage behavior to the greatest extent possible. In industrial power distribution systems, high-order harmonics introduced by nonlinear loads such as frequency converters, welding equipment, and power rectifiers often superimpose on the normal power frequency current to form periodic disturbances. These disturbance signals have specific frequency band concentrations and time persistence. If they are not identified and eliminated, they are easily misjudged as leakage phenomena, triggering false alarms. This step first uses the interfering frequency bands identified in the previous stage to accurately separate these frequency bands from the original signal using a differential frequency filtering method. Unlike traditional fixed-bandwidth filters, this method can dynamically construct frequency templates based on actual interference behavior, ensuring the accuracy and range control of the removal. Subsequently, to avoid the loss of genuine abnormal current signals due to false filtering, a local waveform reconstruction method was introduced to reconstruct and compensate the amplitude of abrupt waveforms in the low-frequency band in the time domain, ensuring that the genuine leakage signal maintains structural integrity during interference removal. Through this recursive processing mechanism of "removal + reconstruction," the final output is a structurally purified residual current signal. This signal not only significantly reduces noise interference but also possesses highly faithful original abnormal characteristics, providing a precise and reliable signal foundation for subsequent compensation control and risk assessment. This processing achieves fine coordination in the frequency domain, time domain, and behavioral characteristics, significantly improving the electrical fire monitoring system's response capability and false alarm suppression effect against genuine leakage risks under complex interference backgrounds.

[0072] S400 performs adaptive compensation control processing based on the residual current signal after structural purification. The processing includes combining historical operating condition characteristics with the current signal change rate to dynamically generate a compensation reference curve for the inherent residual current, and adjusting the residual current compensation amplitude in real time according to the compensation reference curve.

[0073] The process of performing adaptive compensation control based on the residual current signal after structural purification includes the following steps:

[0074] A historical operating condition feature library was constructed for generating the compensation benchmark curve, and typical inherent residual current behavior templates related to the power frequency operating state were extracted. During the initial stage of system operation or periodic inspections, multiple time periods under normal operating conditions without abnormal leakage events were selected. The residual current signal after structural purification was sampled point-by-point within each sampling period, and five parameters were extracted, including the peak amplitude, periodic mean, slope of change, frequency domain energy concentration rate, and zero-crossing point position within the period, forming a feature vector. These feature vectors were categorized and organized according to time period, load type, and current level. For example, the high-load operation phase during the day, the low-load standby phase at noon, and the maintenance operation phase at night were categorized separately, and the corresponding extracted average residual current waveforms were saved as benchmark templates. Each template is stored in the feature database in vector form as a reference for subsequent comparisons with the current operating conditions.

[0075] The system acquires the residual current signal after structural purification, collected in real time. Within a complete power frequency cycle, it extracts the latest waveform features at 200 sampling points and compares them item by item with historical operating condition templates in the feature library. The Euclidean distance algorithm is used to calculate the distance between the current cycle feature vector and all templates in the database, and the template with the smallest distance is selected as the current operating condition matching reference. For example, if the peak value of the current residual current waveform is 38mA, the average period is 12mA, and the average rate of change is 5mA / ms, then it has the highest matching degree with the template corresponding to the "medium load stable operation scenario" in the database. The corresponding compensation reference waveform in this template is used as the initial compensation reference curve. Furthermore, based on the changing trend of the current waveform in 10 consecutive sampling points, the instantaneous adjustment coefficient is calculated to correct the response characteristics of the compensation curve at local abrupt changes or gradual transitions, forming the final usable dynamic compensation target curve.

[0076] Based on the generated compensation target curve, the current residual current signal is adjusted point-by-point. At each sampling point, the current signal amplitude is compared in real time with the reference value at the same time position on the compensation target curve. The amplitude difference is calculated, and it is determined whether it exceeds the dynamic adjustment threshold. For example, if the amplitude at the current point is 42mA and the reference value is 35mA, and the difference exceeds the set error upper limit of 5mA, an amplitude reduction operation is performed; if the difference is within ±3mA and the waveform change is stable, standard proportional compensation is executed. During this adjustment process, the upper and lower limits of amplitude adjustment are set at ±15% to avoid overcompensation caused by misjudgment due to short-term fluctuations. At the same time, a compensation buffer is set at local waveform abrupt change points (such as slope change rate greater than 10mA / ms), delaying the adjustment by two sampling points to prevent response imbalance caused by instantaneous and drastic changes in the compensation curve. The adjustment process adopts real-time sliding window logic to ensure that the compensation output in each cycle remains continuous, smooth, and closely follows the dynamic characteristics of the original waveform.

[0077] Short-term statistical analysis and self-correction are performed on the compensation execution results over multiple consecutive power frequency cycles. A feedback mechanism based on error accumulation and response deviation is constructed to correct the construction parameters of the compensation reference curve for the next cycle. Within a time window of 10 cycles, the average deviation, maximum error, mean square error, and slope deviation of the current compensation residual current signal are statistically analyzed and compared with the data from the previous window. For example, if the mean square error exceeds twice the value of the previous window, or the proportion of high-frequency energy in the residual signal increases by more than 10%, it is determined that the current compensation curve has a structural mismatch problem and the compensation reference needs to be reconstructed. At this time, the feature matching process is re-executed, the template curve with the closest second matching degree is used, and the dynamic adjustment coefficient is recalculated. Simultaneously, the error feedback value is written into the historical record and used in the weight allocation of the next parameter tuning strategy, thus forming a closed-loop adjustment mechanism with continuous learning and adaptive capabilities.

[0078] The purpose of this step is to further identify and compensate for the inherent non-faulty residual current components that still exist in the residual current signal obtained after structural purification. This ensures that the electrical fire monitoring system can more accurately identify abnormal current fluctuations caused by actual leakage behavior, thereby effectively reducing the risk of false alarms and missed alarms. In industrial power distribution systems, inherent residual current is widely present in various load conditions under normal operating conditions. It is usually caused by factors such as cable-to-ground capacitance, leakage inductance of electrical equipment, and zero deviation of current transformers. Its manifestation is stable but overlaps with early abnormal leakage signals in amplitude and waveform. This step constructs a historical operating condition feature database, statistically models the residual current under typical load operating scenarios, extracts representative non-faulty current behavior curves as compensation benchmarks, and combines the change rate and waveform characteristics of the current signal after structural purification acquired in real time to dynamically select the reference model that best matches the current operating condition, generating a target compensation curve that changes over time. Subsequently, the system adjusts the amplitude of the current residual current signal in real time according to this curve, realizing the dynamic elimination of inherent background current components, thus making the remaining current change part more directional and valuable for anomaly identification. This step achieves personalized, adaptive, and dynamic optimization of residual current compensation processing at both the time and behavioral levels, effectively avoiding the response lag or misjudgment problems caused by traditional fixed threshold or static template methods. It provides a purer and more reliable basic signal for subsequent leakage risk warning algorithms and is a key link in realizing intelligent monitoring and accurate alarm.

[0079] S500, based on the execution process of adaptive compensation control processing, constructs a residual current stability index backtracking mechanism to track the stability of the residual current signal after structural purification in multiple sampling periods, analyzes the response lag, amplitude offset and error accumulation in the compensation results, and corrects the adaptive compensation control processing parameter system based on the analysis results;

[0080] Based on the execution process of adaptive compensation control, a residual current stability index backtracking mechanism is constructed, which includes the following steps:

[0081] A fixed number of sampling periods were set as the stability evaluation window, and the residual current signal after structural purification was collected in segments. Using a 50Hz power frequency cycle as the basic unit, 200 sampling points were collected per cycle, and 10 consecutive cycles were selected as the analysis window, processing a total of 2000 data points. Within each cycle, five characteristic parameters of the purified residual current signal were extracted: maximum amplitude (mA), minimum amplitude (mA), root mean square value (mA), average rise slope within the cycle (mA / ms), and frequency energy center distribution value (Hz). Taking the example data, period 1 has a maximum amplitude of 42.3 mA, a minimum amplitude of 18.5 mA, a root mean square (RMS) of 30.1 mA, an average rise slope of 6.2 mA / ms within the period, and a frequency energy center distribution value of 140 Hz; period 2 has a maximum amplitude of 41.7 mA, a minimum amplitude of 17.9 mA, an RMS of 30.5 mA, an average rise slope of 5.9 mA / ms within the period, and a frequency energy center distribution value of 135 Hz; and so on up to period 10. These parameters are organized into 10 sets of periodic feature vectors to form a behavioral benchmark dataset within the current time window, providing a quantitative basis for subsequent stability analysis.

[0082] Based on the aforementioned characteristic parameters, the fluctuation amplitude and behavioral consistency of the residual current signal within the current window are calculated to construct a comprehensive stability index. Specifically, for each parameter, the range (maximum minus minimum), standard deviation (reflecting periodic fluctuations), and mean square deviation (reflecting the average deviation) are calculated. For example, for the root mean square value sequence {30.1, 30.5, 30.3, ...}, if its standard deviation is 0.7 mA and its range is 1.6 mA, it belongs to the moderate fluctuation level. The standardized results of the five parameters are then combined according to set weights to calculate the comprehensive index, for example, amplitude fluctuation 0.35, slope fluctuation 0.25, energy distribution stability 0.25, root mean square stability 0.1, and minimum value stability 0.05. A comprehensive weighted calculation yields the stability score for the current window. The score range is 0 to 1; a score ≥ 0.8 is considered highly stable, between 0.6 and 0.8 is considered moderately stable, and < 0.6 is considered unstable, triggering a parameter backtracking process.

[0083] After determining that the state is unstable, a cycle-by-cycle response difference analysis is performed on the current compensation execution process to identify the specific causes of instability. Specifically, the actual compensation output signal of each cycle is aligned with the adaptive compensation reference curve of the previous cycle, and the response delay point (time difference of the first rising edge, in the number of sampling points), amplitude offset (difference between the actual peak value and the target peak value in each cycle, in mA), and cumulative error (sum of the absolute values ​​of errors over the past three cycles, in mA·cycle) are compared. For example, if the first rising edge of cycle 8 lags by 4 sampling points, the amplitude deviation reaches 7.5 mA, and the cumulative error over three consecutive cycles exceeds 20 mA·cycle, then the current compensation strategy is considered to have insufficient configuration in lag control and amplitude prediction, and the compensation control parameter system needs to be corrected immediately.

[0084] Based on the identified specific deviation types and numerical feedback, the compensation control strategy parameters are corrected, and the correction results are applied to the next sampling period. If the main problem is identified as response lag, the compensation adjustment slope is increased from the current setting of 6mA / ms to 8mA / ms to accelerate the response speed. If amplitude deviation is identified, the weight of the dynamic adjustment coefficient is adjusted, changing the weight of the waveform change rate in the current period from 0.3 to 0.45 to enhance the ability to follow local abrupt changes. If the deviation is due to error accumulation, a suboptimal operating condition template is used in the template matching stage, and the matching threshold is relaxed from 0.9 to 0.85 to improve adaptation flexibility. The corrected parameters are immediately applied in the next period, and the lag value, deviation value, and correction amount during this correction process are recorded in the stability tracking log for subsequent backtracking analysis and long-term model optimization. The entire backtracking correction logic runs continuously in a sliding window manner during the compensation control process, ensuring that compensation accuracy and output stability are dynamically maintained even under complex operating conditions such as frequent load switching and increased external disturbances.

[0085] The purpose of this step is to establish a dynamic monitoring and feedback optimization mechanism for the residual current signal after structural purification. This mechanism continuously evaluates and adjusts the execution effect of adaptive compensation control, ensuring that the electrical fire monitoring system maintains high accuracy and stability even under long-term operation and complex interference environments. Due to frequent load state switching and dynamic changes in the power grid harmonic environment in industrial power applications, although the adaptive compensation algorithm has strong real-time response capabilities, its parameter settings may deviate from the current optimal state during actual operation, potentially leading to problems such as response lag, insufficient amplitude compensation, or overcompensation, thus affecting the identification of abnormal leakage current behavior. This step, by setting multiple continuous sampling period analysis windows, tracks the stability changes of the purified signal in real time across multiple indicators (such as root mean square value, maximum amplitude, slope of change, and energy distribution trend), and further calculates the fluctuation amplitude, standard deviation, response time offset, and error accumulation between periods, forming a quantifiable stability index. When the index falls below a set stability threshold (e.g., a score below 0.6), a backtracking analysis process is triggered to pinpoint key parameter points in the current compensation strategy that cause a decrease in stability. These include insufficient response rate, accumulated errors in the dynamic adjustment factor, and incorrect selection of operating conditions. Based on the backtracking results, the compensation parameter system is automatically corrected, such as adjusting the dynamic compensation slope, reselecting the compensation curve template, and optimizing the sampling comparison strategy within the cycle. Ultimately, this mechanism enables online self-optimization of the compensation strategy without human intervention. Through continuous learning and feedback adjustment, it maintains high adaptability and reliability of compensation control. It is a crucial supporting element for achieving intelligent, closed-loop, and accurate residual current monitoring, possessing significant technical necessity and practical value.

[0086] The S600, based on the modified adaptive compensation control processing parameter system, performs joint closed-loop control processing, which combines multi-component time-frequency deconstruction processing, amplitude-frequency coupling relationship identification processing, recursive harmonic interference elimination processing, adaptive compensation control processing, and stability backtracking processing to form a control closed loop, thereby achieving accurate identification and alarm of real leakage risks under high-order harmonic interference conditions.

[0087] Based on the modified adaptive compensation control processing parameter system, the process of executing joint closed-loop control processing includes the following steps, which aim to integrate the previous processing links into a complete control process with real-time response and self-adjustment capabilities, so as to realize the identification and alarm of real leakage signals under high-order harmonic interference conditions.

[0088] A unified sampling time reference and data transmission link are constructed at the structural level to ensure that multi-component time-frequency decomposition processing, amplitude-frequency coupling relationship identification processing, recursive harmonic interference elimination processing, adaptive compensation control processing, and stability backtracking processing can be executed sequentially and efficiently within the same cycle. Specifically, the sampling frequency is set to 10,000 points per second, corresponding to 200 sampling points collected per 50Hz power frequency cycle. After signal acquisition, the db6-order wavelet function in the Daubechies wavelet function is immediately invoked to perform 4-level wavelet packet decomposition, splitting the residual current signal into 16 fixed-width sub-bands. The energy density and short-time spectral entropy value of each band are calculated, outputting a frequency domain energy distribution matrix and a disturbance identification map. This map is expressed in two-dimensional matrix form, with each row corresponding to a frequency band and each column corresponding to a periodic window. The values ​​in the matrix are normalized energy and disturbance intensity indices, providing accurate and continuous input data for subsequent interference identification.

[0089] The process involves identifying amplitude-frequency coupling relationships and using the output to drive recursive harmonic interference removal. This process extracts the phase shift, response delay (expressed as the number of sampling points), burst duration, and center frequency identifier for each frequency band as a parameter set. A list of frequency bands to be removed is constructed using the identifier "Frequency band number = 8, center frequency = 1250Hz, phase shift = 27°, response delay = 4 points". Upon entering the recursive interference removal process, the frequency band undergoes three rounds of differential filtering by matching wavelet packet coefficient numbers. The first round subtracts the energy of the frequency band to reconstruct the signal; if the energy of this frequency band still accounts for more than 5% of the total energy of the purified signal in the remaining signal, the second round of processing begins; if interference energy remains after two rounds, the coefficients of adjacent frequency band signals are weighted and removed in the third round. All parameters retain 6 bits of valid numerical precision during the filtering process to ensure that removal does not result in false deletions. The final purified signal output will significantly eliminate periodic interference in the frequency domain while fully preserving low-frequency non-steady-state abnormal signals.

[0090] The system loads the adaptive compensation control parameter system corresponding to the current cycle and performs precise compensation based on the purified signal. This parameter system includes five key control factors: the optimal matching template number (e.g., the load stability state in template T3), the dynamic adjustment coefficient (range set to 0.8–1.2), the upper and lower limits of the compensation amplitude (±10mA respectively), the response time lag correction value (fixed to the difference between the sampling points of the previous cycle), and the width of the local abrupt change tracking window (set to 20 sampling points). By analyzing the rate of change, peak amplitude position, and slope of the current cycle signal compared to the reference value of the previous cycle through a sliding window, the compensation target curve for the current cycle is dynamically generated. Then, compensation adjustment is performed for each sampling point: if the current amplitude is more than 5mA higher than the target value, the compensation is reduced to the target value; if it is less than 3mA lower than the target value, the compensation is increased, with the compensation amplitude not exceeding the ±10mA limit. Linear interpolation is used during the adjustment process to ensure consistent response between abrupt and gradual change segments.

[0091] At the end of each cycle, stability backtracking is performed to analyze the consistency of the compensation effect with historical cycles, and parameter correction is triggered based on the stability score. From the residual signals of 10 consecutive cycles, indicators such as the root mean square value sequence, slope trend, spectral centroid offset rate, and burst point volatility are extracted. If the calculated comprehensive stability index is below 0.65, the compensation accuracy is considered to have decreased. If the lag response continuously exceeds 4 sampling points and the cumulative error exceeds 30% of the average total error of the previous 3 cycles, the dynamic adjustment coefficient for the current cycle is adjusted from 1.0 to 1.1, the abrupt change tracking window is shortened to 16 sampling points, and the compensation template is replaced with the second-best model closest to the current cycle (e.g., switching from T3 to T2). After correction, the parameters are immediately loaded into the next cycle, forming a complete closed loop of parameter-compensation-feedback-correction. By executing this sequence in each cycle, the system can dynamically adapt to grid noise, load changes, and environmental disturbances during continuous operation, achieving self-learning and self-optimization.

[0092] The purpose of this step is to organically integrate multiple key processing flows involved in the entire electrical fire monitoring process—including multi-component time-frequency deconstruction processing, amplitude-frequency coupling relationship identification processing, recursive harmonic interference elimination processing, adaptive compensation control processing, and stability backtracking processing—into a closed-loop control system according to a unified data structure, time window, and execution sequence. This enables the system to continuously, dynamically, and accurately identify real leakage signals and issue precise alarms in complex power distribution environments where high-order harmonic interference is significant. Traditional electrical fire monitoring often relies on single criteria, such as amplitude thresholds or frequency domain templates, which are difficult to maintain a high recognition rate in situations with overlapping spectra or high noise backgrounds. Through the closed-loop mechanism constructed in this step, the time-frequency intrinsic features of high-frequency disturbances are first extracted from the original residual current signal, and then their behavioral boundaries and change paths are dynamically identified. Subsequently, non-leakage interference components are accurately eliminated through recursive processing. Based on the structurally purified signal, an adaptive compensation algorithm is used to dynamically correct the inherent background current. Finally, the stability backtracking mechanism continuously evaluates the compensation accuracy and corrects the compensation strategy parameters in a closed loop to avoid misjudgment and response lag. More importantly, this closed-loop structure is not a static series connection, but rather a linkage between parameter feedback and time-domain response, ensuring that each processing stage is not only effective for the current signal, but also adaptable to non-stationary conditions such as load changes and interference frequency shifts in the next cycle. This control closed-loop mechanism endows the system with self-learning, self-adaptation, and self-correction capabilities, enabling it to maintain high sensitivity and accuracy in identifying real electrical fire hazards even under long-term operation, complex conditions, and multi-source interference. It is a core supporting element for ensuring the practicality, reliability, and deployability of the monitoring system.

[0093] The aforementioned electrical fire monitoring method based on automatic compensation of inherent residual current significantly improves the accuracy and reliability of the monitoring system in identifying real leakage risks under high-order harmonic interference environments. This method achieves accurate localization of high-frequency interference signals by introducing multi-component time-frequency deconstruction, wavelet packet energy extraction, and spectral entropy abrupt change analysis; it effectively distinguishes and eliminates non-faulty harmonic components by establishing an amplitude-frequency coupling behavior recognition mechanism and a recursive interference elimination strategy; further, by combining historical operating conditions and dynamic trends, it implements adaptive compensation and stability backtracking adjustment, constructing a closed-loop feedback control logic that enables the system to continuously optimize judgment criteria and correct compensation strategies online. Overall, this solution solves the problem of misjudgment and missed reporting caused by the inability to distinguish between high-order harmonics and real leakage signals in traditional systems, achieving refined modeling and intelligent response to abnormal current behavior in complex power distribution scenarios, and possessing high engineering adaptability and safety protection capabilities.

[0094] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An electrical fire monitoring method based on automatic compensation of inherent residual current, characterized in that, Includes the following steps: S100 acquires the residual current signal, performs multi-component time-frequency deconstruction processing, extracts energy distribution characteristics through wavelet packet transform, and extracts abrupt change indicators by combining short-time spectral entropy analysis, constructs a preliminary distribution map of high-order harmonic interference, and determines the boundary characteristics of the interference signal in the time and frequency domains. S200, based on the preliminary distribution map of high-order harmonic interference, performs amplitude-frequency coupling relationship identification, analyzes the phase shift and response delay between the power frequency and the interference frequency band, identifies the dynamic correlation between atypical leakage and high-order harmonic interference in the residual current, and determines the location and change path of the interference frequency band. S300 performs recursive harmonic interference removal based on the location and change path of the interference frequency band, uses differential filtering to remove interference components, and retains low-frequency non-steady-state components through local waveform reconstruction to obtain the residual current signal after structural purification. S400, based on the residual current signal after structural purification, performs adaptive compensation control, combines historical operating conditions and the current rate of change to generate a compensation reference curve for the inherent residual current, and dynamically adjusts the compensation amplitude. S500, based on the adaptive compensation control process, constructs a stability index backtracking mechanism to track the signal stability within multiple sampling periods, analyzes response lag, amplitude offset and error accumulation, and corrects the adaptive compensation control parameters. The S600, based on the modified adaptive compensation control parameter system, performs joint closed-loop control, integrating time-frequency deconstruction, coupling identification, interference elimination, compensation control and backtracking mechanism to achieve accurate identification and alarm of real leakage risk under high-order harmonic interference conditions.

2. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S100 includes: The zero-sequence current signal in the three-phase power distribution line is collected, and after high-frequency interference is filtered out by an anti-aliasing low-pass filter, it is digitized by an analog-to-digital converter chip at a sampling rate of more than 10,000 times per second. The digitized residual current signal is subjected to four-level wavelet packet decomposition. The db6-order function in the Daubechies wavelet function is used to decompose the residual current signal into 16 sub-bands, and the frequency domain energy in each sub-band is calculated. Based on the frequency band energy distribution results, short-time spectral entropy analysis is introduced. The signal is windowed according to the power frequency cycle and the sliding window is used to extract the energy concentration and structural change characteristics of each frame of the signal. A comprehensive time-frequency two-dimensional perturbation map is constructed to identify the frequency boundaries, duration, and perturbation intensity of high-frequency interference signals, thereby completing the construction of a preliminary distribution map of high-order harmonic interference.

3. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S200 includes: Extract frequency sub-bands with energy density higher than a preset threshold and spectral entropy lower than a preset threshold, and establish a pairing relationship between the power frequency band and the interference frequency band; The instantaneous phase shift between corresponding frequency bands is calculated based on the Hilbert transform, and its phase coupling characteristics are determined. Using the time location of the residual current mutation point as the anchor point, calculate the response delay of each frequency band signal and determine its time-domain response dependency. Based on the above analysis results, a frequency band behavior identification result set is formed, and the coupling type and change path of the interference frequency band are marked.

4. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S300 includes: Construct a set of frequency templates for the frequency bands to be removed from the interference, and extract the corresponding frequency band node coefficients to reconstruct the interference component signals; After dynamic time warping, differential frequency filtering is performed to initially eliminate interference frequency band signals. A sliding time window is set to perform local waveform reconstruction, preserving low-frequency unsteady components and performing amplitude correction; Based on the high-frequency energy ratio and reconstruction error, a threshold for elimination iteration is set, and multiple rounds of interference elimination are recursively performed to output the residual current signal after structural purification.

5. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S400 includes: Construct a working condition feature library containing residual current feature vectors and compensation reference waveforms from historical operating scenarios; Extract the residual current characteristics in the current cycle and calculate the Euclidean distance between them and the historical template, determine the optimal matching template and generate the dynamic compensation target curve; Compare the current signal with the target compensation curve point by point and perform amplitude compensation and delay adjustment operations within the set threshold range; Statistical error indicators are analyzed over multiple periods, and the compensation baseline curve is reconstructed based on the error feedback results to achieve closed-loop optimization.

6. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S500 includes: Ten consecutive sampling periods are set as the analysis window. The maximum amplitude, minimum amplitude, root mean square value, average rise slope and energy center frequency of each period are extracted to construct a periodic feature vector set. The range, standard deviation and mean square deviation of each parameter are calculated based on the periodic feature vector set, and the stability index of the current window is constructed by combining them according to the set weights. When the stability index is lower than the set threshold, the compensation output is compared with the compensation reference curve on a cycle-by-cycle basis to extract the response delay point, amplitude offset and error accumulation value to identify the cause of instability. Adjust the compensation control parameters based on the identification results and enable the correction value in the next cycle, while recording the correction information for subsequent tracking and analysis.

7. The electrical fire monitoring method based on automatic compensation of inherent residual current according to claim 1, characterized in that, Step S600 includes: A unified sampling time reference and data transmission link are set, and the frequency domain energy matrix and perturbation spectrum of the residual current signal are extracted by wavelet packet decomposition as input; Based on the frequency domain spectrum, amplitude-frequency coupling identification processing is performed to extract the phase shift, response delay, and energy index of the interference frequency band, drive three rounds of recursive interference removal processing, and output a purified signal. Based on the purification signal, the current compensation parameter system is loaded, and the target compensation curve is generated by combining the rate of change and the dynamic adjustment coefficient. The signal is adjusted point by point and amplitude control is executed. At the end of the cycle, the compensation effect and stability index are analyzed. If the compensation accuracy decreases, the dynamic parameters and template matching are corrected, the parameter system is updated and loaded into the next cycle to complete the closed-loop control.

Images