Arcing monitoring system and method for high-voltage circuit breaker

Abstract

The invention discloses a high-voltage circuit breaker arcing monitoring system and method, and relates to the technical field of power equipment operation state detection, and the method comprises the following steps: S1, synchronously collecting arcing spectral signals and electromagnetic interference vector data under a unified time reference, and carrying out primary differential processing to generate stable base plane signals; s2, inputting the stable base plane signal into the interference identification model, generating a multi-dimensional spectrogram through time-frequency mapping, extracting an interference time sequence boundary, and outputting an interference weight matrix; the arcing monitoring method based on spectrum and electromagnetic signal synchronous acquisition is constructed, interference identification, signal purification, energy analysis and control instruction generation are fused, parameter self-optimization is realized through closed-loop feedback, the arcing identification accuracy and the arc quenching response efficiency are effectively improved, and the arcing monitoring method is suitable for large-scale popularization and application. And the operation safety and the intelligent level of the circuit breaker in a strong interference environment are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment operation state detection, and particularly relates to a high-voltage circuit breaker arc monitoring system and method. BACKGROUND

[0002] High-voltage circuit breaker arc monitoring refers to a process of capturing, identifying and analyzing high-temperature plasma arc generated between circuit breaker contacts in real time through optical, acoustic, electromagnetic, thermal imaging, gas composition analysis and other sensing means during high-voltage circuit breaker opening and closing operations or abnormal operation states. Arcing is a transient discharge phenomenon accompanied by strong light radiation, shock sound waves, severe temperature rise and electromagnetic pulses. If it cannot be detected and evaluated in time, the high temperature and strong current may damage the contacts, insulating medium and mechanical components, and even cause equipment failure or power grid accidents. The arc monitoring technology can determine the time, location, duration and energy intensity of arc occurrence by collecting arc signal characteristic parameters such as spectral distribution, sound pressure waveform, current distortion rate, infrared temperature rise curve and partial discharge intensity, combining data processing and pattern recognition algorithms, realizing health assessment of high-voltage circuit breaker working state and early warning of potential faults, and thus improving the safety and reliability of the power system.

[0003] The prior art has the following disadvantages: In the process of high-voltage circuit breaker arc monitoring, when lightning strike, electromagnetic pulse (EMP) radiation impact or large current short circuit accident occurs in the external environment, a transient strong interference signal with extremely high amplitude will be generated in a very short time. This kind of transient signal enters the arc monitoring link through electromagnetic coupling, conductor induction or ground reverse conduction, and superimposes on the real signal generated by the arc in the time domain and frequency domain, thereby causing distortion of the waveform characteristics, amplitude distribution and energy density of the monitoring signal. Since the instantaneous amplitude of the interference signal is usually much higher than that of the arc signal, and the duration is short but enough to cover the early feature interval of the arc, the monitoring system may not be able to correctly identify that the arc has occurred in the feature extraction and event discrimination stage, resulting in delayed monitoring response or even complete loss of arc identification ability. In this case, the arc will continue to exist and release high-temperature plasma and strong electromagnetic radiation, causing the metal on the contact surface to melt and weld together, the insulation and metal structure inside the arc chamber to be continuously ablated, and the cavity pressure to rise sharply, which may eventually induce shell rupture or explosive failure, resulting in not only the destruction of the circuit breaker, but also serious power grid accidents such as bus voltage loss and system cascading trip.

[0004] The above information disclosed in the background section is only used to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The application aims to provide a high-voltage circuit breaker arc monitoring system and method to solve the problems in the background art.

[0006] To achieve the above-mentioned purpose, the application provides the following technical solution: a high-voltage circuit breaker arc monitoring method, comprising the following steps: S1, synchronously collecting arc spectrum signals and electromagnetic interference vector data under a unified time reference, and performing primary difference processing to generate stable base surface signals; S2, inputting the stable base surface signals into an interference recognition model, generating a multi-dimensional spectrum graph through time-frequency mapping, extracting interference time sequence boundaries and outputting an interference weight matrix; S3, adaptively adjusting filter kernel coefficients according to the interference weight matrix, and performing enhancement and suppression processing on the signals to obtain a dynamic purified arc characteristic signal sequence; S4, performing continuous energy analysis on the arc characteristic signal sequence, and completing early trigger estimation of protection action based on energy growth trend and a preset threshold; S5, generating a control instruction according to the trigger estimation result, adjusting the opening pressure and arc extinguishing gas flow to realize rapid arc extinguishing; S6, returning the whole-process operation data to the recognition model and the controller, updating the recognition parameters and the control strategy, and constructing a closed-loop optimized arc monitoring mechanism.

[0007] Preferably, step S1 comprises: A miniature fiber optic spectrometer is installed at the arc-extinguishing chamber observation window of the high-voltage circuit breaker to collect spectrum signals during the arc process, the spectrum response range is set to 350 nanometers to 1050 nanometers, and the sampling frequency is set to 100 kilohertz; A three-axis electric field induction antenna is installed outside the circuit breaker shell, and a current transformer is installed at the incoming and outgoing line circuits to collect electromagnetic interference vector data, and the three-axis electric field antenna measurement frequency band covers 10 hertz to 10 megahertz; A GPS time chip is used to mark the collected data with a unified time, and the dual-channel synchronous collection and alignment of the spectrum signals and the electromagnetic vector data are realized based on the unified time reference; Primary difference processing is performed, non-coordinated electromagnetic signals are smoothed and corrected or removed using linear interpolation within a local time window, time reconstruction and amplitude normalization are completed, and stable base surface signals are generated.

[0008] Preferably, step S2 comprises: The stable base surface signals are time-sequentially divided using a sliding time window of 500 microseconds per segment, and a frequency distribution graph in the range of 1 kilohertz to 5 megahertz is obtained through fast Fourier transform; The amplitude relationship between the energy of each frequency point and the reference energy is calculated based on the sliding window spectrum diagram to identify the starting time and ending time of the high-amplitude transient interference signal. The timing and frequency range of the interference event are mapped to the time axis and frequency band axis to construct an interference weight matrix, and the interference level is set according to the interference amplitude and duration and a weight coefficient is assigned, which is used to guide subsequent signal processing.

[0009] Preferably, step S3 comprises: Each time point of the interference weight matrix is matched with the sampling point of the stable base surface signal, and the filter window length and the weighting distribution corresponding to different interference levels are assigned; The window length and center weight are adjusted in the sliding weighted average filtering according to the interference level to realize the smooth suppression of the high interference signal segment and the detail enhancement of the low interference signal segment; The amplitude weighted amplification is performed on the 250 kHz to 2 MHz frequency band signal, the attenuation processing is performed on the non-target frequency band signal, and the amplitude normalization and dynamic overlap reconstruction are performed after filtering in each frequency band; The processed signal is kept with the original timestamp for sorting and reconstruction to form a continuous dynamic purified arc feature signal sequence.

[0010] Preferably, the specific steps of performing amplitude weighted amplification on the 250 kHz to 2 MHz frequency band signal, performing attenuation processing on the non-target frequency band signal, and performing amplitude normalization and dynamic overlap reconstruction after filtering in each frequency band are as follows: The 250 kHz to 2 MHz frequency band signal in the stable base surface signal is extracted, and the amplification coefficient between 1.25 and 1.5 is set according to the frequency interval to perform amplitude weighted processing; The 1 kHz to 250 kHz frequency band and 2 MHz to 5 MHz frequency band signals are extracted, and the attenuation coefficients of 0.5, 0.7 and 0.9 are set according to the interference level of the corresponding time point in the interference weight matrix to perform amplitude reduction; The processed signals of different frequency bands are normalized and superimposed and reconstructed respectively to form a frequency selective enhanced signal for subsequent arc feature extraction and analysis.

[0011] Preferably, step S4 comprises: The arc feature signal sequence is divided into a plurality of overlapping sliding windows according to a 1 millisecond time length, the absolute value integral energy of the signal amplitude in each window is calculated to obtain the energy value of each segment of the electric signal; An energy time sequence is constructed, the energy change trend of the continuous window is extracted, and the amplitude ratio before and after signal enhancement is corrected to obtain an accurate energy growth curve; When the current energy value exceeds the energy threshold value of 2.5 millijoules and the growth ratio in the three consecutive windows all exceeds 1.5 times, an early trigger estimation instruction is output.

[0012] Preferably, the step S5 comprises: The arcing energy value, the growth rate and the risk level information in the early trigger estimation result are received and transmitted to the control response structure in a period not exceeding 1 millisecond; According to the risk level, the corresponding opening driving pressure and arc extinguishing gas flow control parameters are searched from the preset response strategy table; The corresponding control instructions are output to the opening driving execution unit and the gas flow control unit respectively, so that the contacts are separated within 10 milliseconds, and the high-pressure gas flow is controlled to form a high-speed gas flow along the arcing direction; The opening action time, the gas flow start time, the control instruction issuing time and the feedback state are recorded and fed back, which are used for subsequent correction and optimization of the control strategy.

[0013] Preferably, the step S6 comprises: The operation data including the opening time, the execution pressure, the gas flow rate, the arcing energy and the waveform are collected, and are packaged into a feedback data frame after being uniformly attached with a time stamp; The relevant data in the feedback data frame are input to the identification structure, and if the identification deviation exceeds the set threshold value, the spectrum window width, the interference energy threshold value and the slope factor are adjusted and updated as the default enabled parameters; The control related data in the feedback data frame are input to the response control structure, and if the response does not reach the predetermined effect, the driving pressure is increased, the gas flow is increased, and the action time delay is shortened, and the modified version is set as the default control strategy next time.

[0014] A high-voltage circuit breaker arcing monitoring system comprises a synchronous sensing and base surface construction module, an interference identification and timing extraction module, a signal purification and feature extraction module, an energy analysis and early warning judgment module, a response control execution module, and a data feedback and strategy optimization module: The synchronous sensing and base surface construction module synchronously collects arcing spectrum signals and electromagnetic interference vector data under a unified time reference, and performs primary differential processing to generate stable base surface signals; The interference identification and timing extraction module inputs the stable base surface signals into an interference identification model, generates a multi-dimensional frequency spectrum graph through time-frequency mapping, extracts interference timing boundaries and outputs an interference weight matrix; The signal purification and feature extraction module adaptively adjusts the filter kernel coefficient according to the interference weight matrix, performs enhancement and suppression processing on the signals, and obtains a dynamic purified arcing feature signal sequence; The energy analysis and early warning judgment module performs continuous energy analysis on the arc burning characteristic signal sequence, and triggers the protection action in advance based on the energy growth trend and a preset threshold value; The response control execution module generates a control instruction according to the trigger estimation result, adjusts the opening pressure and the arc extinguishing gas flow, and realizes rapid extinguishing of the arc burning; The data feedback and strategy optimization module feeds back the whole-process operation data to the identification model and the controller, updates the identification parameters and the control strategy, and constructs a closed-loop optimized arc burning monitoring mechanism.

[0015] In the above technical solution, the present application provides technical effects and advantages: The present application is based on the synchronous acquisition of spectral signals and electromagnetic vector signals, extracts stable base signals through primary difference, effectively removes noise bottom fluctuations; combined with an interference identification model and a multi-dimensional spectrum diagram, accurately extracts the timing boundary of transient interference and constructs an interference weight matrix, realizing fine separation of arc burning and interference in the signal processing stage; further based on the dynamic purification of the arc burning characteristic sequence, energy progressive analysis is carried out in each period, and the prediction result is converted into a driving control instruction in advance, so as to ensure that the arc extinguishing action is completed before the critical arc energy; finally, through feedback correction of the whole-process operation data, continuous iterative optimization of the identification parameters and the control strategy is realized. This closed-loop mechanism breaks through the technical link of "perception-identification-control-optimization", and under complex working conditions such as multi-source strong interference and high load of power grid, it can still realize early perception, accurate identification, active response and dynamic self-learning adjustment of arc burning risk, effectively avoid the damage of circuit breaker and systemic power accident caused by identification failure or response delay, and significantly enhance the operation safety and intelligent level of power equipment. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0017] Figure 1 The method flowchart of the arc burning monitoring method of the high-voltage circuit breaker of the present application.

[0018] Figure 2 The module schematic diagram of the arc burning monitoring system of the high-voltage circuit breaker of the present application. DETAILED DESCRIPTION

[0019] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example implementations to those skilled in the art.

[0020] The application provides a high-voltage circuit breaker arc monitoring method as shown in the following formula: Figure 1 The application provides a high-voltage circuit breaker arc monitoring method as shown in the following formula: S1, a synchronous dual-domain perception link is constructed, the spectrum signal of the internal arc of the high-voltage circuit breaker and the vector data of the external electromagnetic interference are synchronously collected under the same time reference, and the collected spectrum signal and vector data are subjected to primary differential separation processing to generate stable base surface signals for subsequent analysis; To realize high-precision and high-stability identification and monitoring of the arc behavior of the high-voltage circuit breaker in the running process, a dual-channel synchronous perception mode is adopted to collect the spectrum signal and the vector signal of the external electromagnetic interference when the arc is generated, and the signal differential and separation processing is completed under the unified time reference to construct the stable base surface signals required for subsequent data analysis, and the specific process includes the following steps: A miniature fiber-optic spectrometer is installed at the observation window of the arc-extinguishing chamber of the high-voltage circuit breaker to collect the strong light radiation signal generated in the arc discharge process. The spectral response range of the spectrometer is set to 350 nanometers to 1050 nanometers to cover the common light-emitting waveband of the arc, and a quartz lens with an incident light aperture of 1.0 millimeter is configured to enhance the incident light intensity. The sampling frequency of the spectrum data is set to 100 kilohertz to ensure that the mutation waveform characteristics in the initial stage of the arc are captured. At the same time, a three-axis electric field induction antenna is installed outside the circuit breaker shell to collect the electric field changes caused by external events such as lightning strikes, electromagnetic pulses, and cable transient short circuits; and a current transformer is arranged at the inlet and outlet line circuit of the circuit breaker to detect the conductor coupling interference signal generated by the large current mutation in the line. The measurement frequency band of the three-axis electric field antenna covers 10 hertz to 10 megahertz, and the current transformer adopts a high-frequency type ring-shaped ferrite structure with excellent high-frequency response characteristics.

[0021] To realize synchronous collection of the two signals, GPS time chip is configured in the spectrometer and the electromagnetic induction equipment respectively, and the unified nanosecond-level time stamp is written into each frame of signal data through a special time reference control unit. The data collected by each signal channel is sent to a high-throughput edge computing terminal in real time, and is stored in the form of "spectrum signal-time stamp" and "electromagnetic vector-time stamp" pairs. The time reference adopts UTC satellite time to realize microsecond-level synchronization accuracy and ensure complete alignment of the subsequent signals in the time dimension. Under this structure, even if the starting points of the spectrum signal and the electromagnetic interference signal are slightly different, complete synchronization processing can be realized through the unified time reference.

[0022] The differentially processed data is separated. The arc spectrum signal appears a rapid rise in intensity, a main peak shift and a multi-order spectrum line expansion at the moment of discharge occurrence, while the electromagnetic interference signal shows a pulse wave packet, a short period high amplitude jump or a wideband harmonic enhancement. By comparing the time period of the continuous rising interval in the spectrum signal with the electromagnetic signal in the corresponding time period, the electromagnetic signal change part with non-corresponding slope change, amplitude maximum but duration extremely short is identified. These changes usually represent external interference pulses rather than real arc behavior. In the specific processing, the electromagnetic vector signal is projected on different axes, and the correlation between the energy density change rate in each direction and the spectrum change trend is calculated to determine whether it constitutes an actual arc coordination feature. The non-coordinated signal part identified is smoothed or removed by linear interpolation based on a local time window, so as to retain the original observation data strongly related to the arc behavior.

[0023] The spectrum signal and the electromagnetic signal after differential processing are time reconstructed and amplitude normalized to form a stable base signal. The stable signal has the following three key features: first, it is completely aligned with the original signal in the time dimension, avoiding data jitter or out of step; second, it eliminates high-frequency sudden interference components in amplitude, showing a smoother feature transition; third, it retains the real change trend of the initial arc, including the initial growth of the spectrum intensity, the peak value shift process and the electromagnetic disturbance duration characteristics. The stable base signal will be used as the basis for subsequent interference identification, feature extraction, energy calculation and control response, providing high-quality, low-interference and continuous data for arc monitoring.

[0024] The core role of this step is to provide high-quality, low-interference and time-consistent signal input basis for high-voltage circuit breaker arc monitoring, thereby significantly improving the accuracy and reliability of subsequent identification, analysis and response links. High-voltage circuit breakers may produce high-temperature plasma arc during opening and closing operations or abnormal operation, which is accompanied by strong light radiation, transient electromagnetic disturbance and complex nonlinear signal characteristics. However, in the actual operating environment, external disturbances such as lightning, electromagnetic pulses, power system short circuits will produce transient strong interference signals with large amplitude and wide spectrum in the same period. Once these interference signals are superimposed on the arc sensing link, they can easily mask the key features of the arc real signal, making it difficult for the monitoring system to accurately identify whether the arc occurs, resulting in response delay or even misjudgment.

[0025] The step is to construct a synchronous dual-domain perception link, respectively collect original signals from the spectral domain (representing the arc) and the electromagnetic vector domain (representing the interference), and realize microsecond-level synchronous sampling depending on a unified time reference, so as to effectively retain the timing characteristics of the respective signals. On this basis, through preliminary differential separation processing, based on the correlation analysis between the spectral intensity change trend and the electromagnetic signal amplitude jump characteristic, the external interference components that do not belong to the arc process can be effectively eliminated, so as to extract the real and pure arc characteristic signal. The finally generated stable base signal has good continuity, anti-interference ability and alignment capability, can be used as the core data source for subsequent energy analysis, protection trigger estimation and response control decision, and is the key starting point and technical basis of the entire arc monitoring process.

[0026] S2, input the stable base signal into the transient interference recognition model, generate the corresponding multi-dimensional spectrum graph through time-frequency mapping, determine the timing boundary of the transient interference signal according to the change characteristics in the spectrum graph, and output the interference weight matrix consistent with the time axis; In order to realize accurate recognition and quantitative analysis of the transient interference signal generated in the operation process of the high-voltage circuit breaker, improve the effectiveness and anti-interference ability of the arc monitoring signal, the stable base signal after preliminary differential processing is analyzed in frequency and time dimensions to extract the distribution range of the interference signal on the time axis, and the interference weight information corresponding to the signal data is constructed for dynamic enhancement and screening of the subsequent characteristic signal. The processing process includes the following steps: The stable base signal is segmented and input according to continuous timing, and frequency expansion processing is performed. In specific implementation, the stable base signal is divided into multiple non-overlapping time periods by using a sliding time window of 500 microseconds, and each signal is converted into a frequency domain form by fast Fourier transform to obtain a frequency distribution graph in the range of 1 kilohertz to 5 megahertz. When converting, a Hamming window function is used for window function smoothing to suppress the frequency spectrum leakage phenomenon and maintain good frequency resolution. In order to improve the signal continuity, a 250-microsecond overlap area is set between adjacent processing windows to realize sliding analysis in the time domain. The finally output spectrum result is arranged in time sequence to form a two-dimensional spectrum graph, in which the horizontal axis represents time elapse and the vertical axis represents frequency distribution, and the gray or color intensity in the spectrum graph represents the signal energy value of each frequency point, in units of decibel milliwatts (dBm).

[0027] The time series analysis of the energy distribution change in the two-dimensional spectrum diagram is performed to identify the occurrence period and impact range of the transient interference. In actual processing, a reference energy benchmark value is first determined, which can be taken from the average spectral intensity of the signal 5 milliseconds before the disturbance occurs, and then the energy amplification multiple is calculated based on the benchmark value. When the energy value of a certain frequency at a certain time point exceeds twice the benchmark value and lasts more than 300 microseconds, the time period is determined as a suspected interference period. If the spectral coverage width of the mutation region exceeds 20 kilohertz, and similar amplification occurs in more than three frequency bands at the same time, it is further determined that the event is a high-amplitude transient interference. By marking the start time, end time and frequency coverage range of each interference event, all suspected interference events are recorded and numbered. To verify the actual impact range of the interference, the energy change trend of the arc spectrum signal can be combined for superposition analysis to confirm whether there is a corresponding mutation. The above method can effectively distinguish the spectral disturbance caused by typical interference behaviors such as lightning strike, current mutation, electromagnetic radiation impact, etc.

[0028] Based on the identified interference period and frequency range, an interference weight matrix is constructed to guide the dynamic processing of subsequent signals. The horizontal axis of the matrix is consistent with the time sequence of the stable base signal sampling points, and the vertical axis is the frequency band division interval, for example, 1 kilohertz to 5 megahertz is divided into 50 equal width frequency bands. Each cell in the matrix represents the interference intensity level of a certain frequency band in a certain time segment. According to the actual measured energy mutation amplitude and duration, the interference intensity is divided into three levels: low interference (amplitude increase less than 1.5 times the benchmark value), medium interference (amplitude increase between 1.5 times and 3 times), and high interference (amplitude increase more than 3 times and duration more than 1 millisecond). Each level is assigned a different weight coefficient, for example, 0.2, 0.5, 0.8, to adjust the processing degree of subsequent signal filtering and enhancement. The final interference weight matrix will change dynamically with time, covering the energy evolution process of each transient interference event, providing spectral distribution characteristics and time correspondence, providing accurate basic data support for subsequent arc feature signal extraction, energy evaluation and protection trigger decision.

[0029] The transient interference identification model refers to an interference judgment structure built based on time series spectrum analysis. Its input is the frequency distribution data of the stable base signal in a unit time window, and its output is the interference intensity weight value corresponding to the time axis. The identification process relies on the following physical logic: the signal is segmented according to a fixed time window, and the frequency energy spectrum of each signal segment is obtained through fast Fourier transform; then, the energy change rate of adjacent time windows in the spectrum within a specific frequency band is compared, and it is determined whether there is a sudden increase in amplitude. If the energy jumps significantly in a short period of time and occurs simultaneously in multiple frequency bands, it is determined to be transient interference behavior. This identification structure does not need to rely on deep learning or machine learning model training process. It is built entirely based on frequency analysis and time continuity laws, and has good repeatability and engineering feasibility.

[0030] The main purpose of this step is to identify and quantify the impact of transient electromagnetic interference during the arc monitoring process of high-voltage circuit breakers, and to map this impact in a time-consistent manner into weight parameters that can be used for subsequent signal processing, thereby providing a clear boundary basis and control basis for enhancing the arc characteristic signal and suppressing the interference component. In the actual operating environment of high-voltage circuit breakers, disturbance events such as lightning strikes, electromagnetic pulses, and high-current short circuits in power systems can introduce high-amplitude, high-frequency transient interference signals in a very short period of time. These interference signals are often superimposed on the original arcing signal in the form of pulse waves, harmonic clusters, or broadband mutations, causing spectral energy distortion and time characteristic drift, which seriously interferes with the accuracy of arc monitoring judgment.

[0031] By inputting the stable base plane signal after differential processing into the transient interference identification model and performing spectrum analysis under a unified time reference, the performance characteristics of the above-mentioned interference in the time domain and frequency domain can be effectively captured. Specifically, indicators such as the energy mutation area, frequency band hopping trend, and peak drift rate of the spectrum graph can accurately reflect the starting time, duration, and frequency coverage of the interference signal. By further locating and marking these mutation features, the boundary position of the interference signal on the time axis can be extracted. This boundary information is then matched with the specific frequency band to construct an "interference weight matrix" synchronized with the original signal. This matrix not only clearly marks when and at what frequency the interference occurred, but also quantifies the interference intensity level, forming a "risk guidance map" for subsequent signal processing.

[0032] Therefore, this step is not limited to identifying interference itself. More importantly, it expresses interference events in a structured, quantified, and aligned manner, enabling precise support for subsequent dynamic filtering, adaptive enhancement, and protection action triggering of arcing signals. This proactive interference identification mechanism significantly enhances the overall arcing monitoring system's immunity and practicality in complex electromagnetic environments, and is a key step in the transition from "perception" to "trusted judgment."

[0033] S3, according to the interference weight matrix, the filter kernel coefficient in the signal processing process is adaptively and dynamically adjusted, the enhancement processing of the arc signal and the suppression processing of the interference signal are performed under the time resolution of microseconds, and a continuous and dynamic purified arc characteristic signal sequence is obtained; In order to realize the accurate extraction of the arc signal of the high-voltage circuit breaker under the complex interference background, the filter kernel parameters used in the signal processing process are dynamically adjusted according to the interference weight matrix constructed in the previous step and the time sequence and frequency domain characteristics of the signal, the cooperative processing of the arc signal enhancement and the interference signal suppression is realized, and finally a continuous, clear and low-noise arc characteristic signal sequence is obtained. The technical process includes the following steps: The two-dimensional structure of the interference weight matrix is strictly matched with the stable base surface signal. In the specific implementation, the time sampling frequency of the stable base surface signal is set to 1 megahertz, forming 1 million sampling points per second; the interference weight matrix is generated with the same time step, so that each sampling point corresponds to an interference intensity coefficient. The frequency axis is divided into 100 continuous bandwidth intervals from 1 kilohertz to 5 megahertz, each interval is 49 kilohertz, and the interference weight values are divided into three levels of 0.1, 0.5 and 0.9, representing low, medium and high interference states. These weight values will be used as a reference for processing and input for the kernel function parameter adjustment in the subsequent filtering process.

[0034] A plurality of fixed structure sliding weighted average filters are configured in the signal processing channel for real-time processing of the arc signal. The basic structure of the filter is a symmetric window structure, which is initially set to a 11-point length smooth weighted structure. According to the interference weight value of the corresponding time point, the kernel length and the weight distribution in each time period are dynamically adjusted: when the interference level of a certain frequency band corresponding to the time interval is 0.9, the window length is expanded to 21 points, and the center weight value is relatively reduced to enhance the filtering smoothness; when the interference level is 0.1, the window length remains at 11 points, and the center weight value is increased to improve the detail retention capability. Before each filtering process, the kernel parameters are updated once according to the interference level corresponding to the current time point, so as to ensure that the signal processing has high real-time and adaptability.

[0035] The arc signal is enhanced and processed at the same time, and the frequency band selective suppression operation is performed. According to the field experiment, the arc signal energy is mainly concentrated between 250 kHz and 2 MHz, so the signal in this frequency band is processed by amplitude weighting and amplification based on filtering, and the amplification coefficient is set to 1.25-1.5; and for the two frequency bands of 1 kHz-250 kHz and 2 MHz-5 MHz, the attenuation processing is performed according to the interference weight value classification, and the attenuation coefficients are set to 0.5, 0.7 and 0.9 respectively. In specific implementation, all frequency bands are extracted by a band-pass filter, and then amplitude scaling and normalization are performed respectively before being superimposed again to ensure smooth transition of the enhanced and suppressed segments, and no cross-frequency segment mutation distortion is introduced.

[0036] The processed signal data is reconstructed into an arc characteristic signal sequence in time series form, and an amplitude normalization processing is performed before output to standardize the signal between 0 and 1, which is convenient for subsequent quantitative analysis and feature extraction. In order to ensure that the signal is consistent with the original sampling in time sequence, all data are kept with the original time stamp information during filtering processing, and the signal is reconstructed according to time axis after processing to ensure that there is no phase drift or sampling misplacement. The final output signal has the following technical characteristics: 1) the mutation edge and peak position of the arc signal are retained; 2) high-amplitude interference in non-target frequency band is removed; 3) the dynamic response capability is high, and it is suitable for real-time identification of microsecond-level events. The arc characteristic signal sequence will be used as the direct input data for energy progressive analysis and protection action estimation.

[0037] The purpose of this step is to use the interference weight matrix extracted in the previous step to dynamically adjust the filter kernel function parameters of the signal processing link on the microsecond time scale, so as to enhance the real arc signal and selectively suppress the superimposed interference signal, and finally obtain a continuous, stable and low-noise arc characteristic signal sequence. In the operation process of the high-voltage circuit breaker, the arc is a typical transient discharge phenomenon, and its signal itself has the characteristics of fast rising edge, high-frequency component mutation and spectrum concentration. However, due to external interference sources such as lightning, electromagnetic pulse, cable induction interference, etc., a large number of high-amplitude and high-energy noise signals will be generated in the same time scale, which are easy to overlap with the real arc signal in time and frequency domain, thus causing great interference to the arc identification.

[0038] This step uses the time-frequency dual-dimensional interference intensity information provided by the interference weight matrix to dynamically adjust the parameter settings of the filter corresponding to each time point, such as adjusting the length of the sliding window, the distribution of weighted coefficients, the frequency band selection range, etc. For time periods with strong interference, the system will increase the smoothness of the filter to suppress sudden non-target fluctuations; while for time periods with weak interference, the window size will be kept smaller to preserve signal details. This adaptive filtering adjustment mechanism, different from the traditional static filtering method, can not only identify and respond to sudden and nonlinear changes in interference behavior, but also maximize the retention of arcing intrinsic characteristics at key signal structures such as rising edges, main peak areas, energy-increasing segments, etc., to avoid excessive suppression of target signals.

[0039] The resulting continuous, dynamically purified arcing characteristic signal sequence accurately reflects the arcing event's onset, duration, and energy intensity changes. It boasts the combined advantages of high temporal accuracy, high spectral fidelity, and low interference error, providing a highly reliable data foundation for subsequent arcing energy analysis, fault level assessment, and active protection response. Therefore, this step is the core link in transforming raw, noisy sensor data into clear, interpretable arcing information, and serves as a key technical bridge for transitioning from perception to identification in high-voltage arc monitoring.

[0040] S4, performing continuous energy analysis on the arcing characteristic signal sequence obtained through continuous dynamic purification, and calculating the arcing energy growth trend in real time based on the signal enhancement results. Based on the relationship between this trend and the preset energy threshold, the early triggering prediction of the protection action is completed; To achieve dynamic energy assessment and risk trend identification of arcing processes inside high-voltage circuit breakers, after obtaining a continuously dynamically purified arcing characteristic signal sequence, continuous energy analysis is performed on it time period by time, and judgment is made based on the energy growth trend and the set energy threshold to achieve early triggering estimation of protection actions. This technical process includes the following steps: The arcing characteristic signal sequence is segmented into 1-millisecond time windows, each containing 1000 sampling points. The windows are updated using a sliding method with 50% overlap, meaning an evaluation is performed every 500 microseconds. Within each window, the absolute value of the signal amplitude is summed and multiplied by the time interval between each point (i.e., 1 microsecond). The integrated energy of the electrical signal within that time period is calculated. The resulting energy is expressed in volt-seconds (V·s), which can be converted to a relative energy value. The energy value for each segment serves as the arcing energy reference point within that time period, enabling continuous tracking of energy changes.

[0041] Build the energy time series, record the past 10 consecutive sliding window energy values, form an actual measurement-based energy change trend chart. When each new window enters, update the latest energy data, and remove the earliest data point to maintain a constant length of sliding window. By calculating the energy difference between two consecutive time points, the energy growth per unit time is extracted. At the same time, record the change direction and rate of the growth. If the energy growth in the last three windows is positive and gradually increases, it indicates that the arc burning state is in a rapid growth stage.

[0042] Compare the energy value of each window with the average energy value of the previous 10 windows to calculate the relative growth ratio. If the ratio exceeds 1.5 times and the number of signal peaks in the window is more than three, it indicates that the signal has appeared typical behavior of arc burning aggravation. On this basis, in order to avoid numerical errors caused by signal enhancement processing, the energy value of the signal enhancement section is corrected. The correction method is: the ratio of the average amplitude of the original signal before enhancement to the average amplitude after enhancement is used as a correction factor, which is multiplied by the energy value obtained after enhancement. The corrected energy reference value of this section ensures the accuracy of the actual evaluation.

[0043] Early trigger judgment of protective action. Set the critical criterion for energy change as: 1) The energy growth ratio in the last three sliding windows exceeds 1.5 times; 2) The total energy value of the current window exceeds the set absolute energy threshold value of 2.5 millijoules. The threshold value is determined based on the melting energy threshold of the circuit breaker contact metal material, ensuring that the value has engineering basis. When the above two conditions are met at the same time, it is determined that the current arc burning has the risk of continuous aggravation and may cause structural damage, and an early trigger estimation instruction is issued to provide a basis for subsequent control structure to issue opening or enhanced arc extinguishing action.

[0044] The current judgment result is output in the form of a signal, and the original signal of the current time window, the calculated energy value, the growth ratio, the growth rate, and the correction factor are recorded and stored in the operation record table. This data table serves as a basis for subsequent operation and maintenance systems, and is also used for future optimization and adjustment of energy threshold parameters and judgment criteria. Through continuous updating and verification, this method can build a dynamic energy recognition system with self-adjusting and engineering adaptability, further improving the intelligent level and reliability of the entire arc burning monitoring and protection mechanism.

[0045] The main role of this step is to realize the real-time dynamic monitoring and trend judgment of the internal arc state of the high-voltage circuit breaker, and to make an estimated judgment of the protection action in advance before the arc energy reaches the disaster level, to provide decision basis for the response measures such as fast contact opening, arc extinguishing gas flow enhancement. Since arc is essentially a transient high-temperature and high-energy discharge process, its energy release behavior has obvious stages and progression, and the arc intensity is often within the controllable range in the initial stage, but if it is not timely identified and processed, its continuous growth may lead to contact melting, insulation breakdown, and even cavity explosion and other serious faults. Therefore, it is not enough to rely on the judgment of whether the arc has occurred, and the key is whether the trend of rapid growth of arc energy can be captured.

[0046] This step divides the arc characteristic signal sequence after continuous dynamic purification into time windows, performs energy integration calculation at the millisecond level, and thus constructs an arc energy change curve evolving with time. At the same time, by calculating the energy growth rate between adjacent time periods and combining the amplitude amplification ratio in the signal enhancement process, the signal is energy corrected to ensure that the final judgment result has high physical accuracy and engineering reference value. Then, the growth trend is compared with the preset energy threshold (such as the energy corresponding to the melting point of the material, the critical bearing value of the equipment), to determine whether the current arc state has a trend of rapid evolution in a dangerous direction.

[0047] If the judgment result shows that the arc energy continues to rise in a short time and exceeds the set growth threshold and absolute energy threshold, the system will give an estimated signal to trigger the protection action in advance. This signal is not a simple passive response, but is generated based on the "trend" and "prediction" logic, has a feedforward control capability, and helps the power system to perform fast current breaking, arc extinguishing and other operations in advance before the arc causes substantial damage. In summary, this step not only enhances the ability of arc monitoring from identification to prediction, but also builds a bridge from data to behavior control, which is a key link to realize intelligent and highly reliable power protection.

[0048] S5, the estimated result is transmitted in advance to the hierarchical response controller, which generates corresponding drive control instructions according to the real-time arc energy threshold, and controls the contact opening to execute pressure and arc extinguishing gas flow output flow, to realize the rapid extinction of arc; To realize the fast response and active control of the arc state, the estimated result is transmitted in real time to the hierarchical response control structure, and specific control instructions are dynamically generated according to the level evaluation result of the arc energy, so as to adjust the drive pressure of the contact opening and the output flow of the arc extinguishing gas, to quickly extinguish the arc and prevent further energy accumulation. The whole control response process includes the following steps: The pre-ignition estimation result output by the arc energy analysis module is transmitted to the control response structure, which includes the current arc duration (in milliseconds), the current cumulative energy value (in millijoules), the energy growth rate (in millijoules / millisecond), and the risk level determination result (marked as level I, level II, or level III). Information transmission is completed through an industrial Ethernet communication interface, with a transmission period of no more than 1 millisecond, ensuring real-time data. The risk level determination criteria are as follows: level I is when the energy does not exceed 60% of the total threshold, level II is when it reaches between 60% and 90%, and level III is when it exceeds 90%. This level division is used to determine the intensity and priority of the response strategy.

[0049] According to the received risk level and current energy value, the control structure selects the corresponding control scheme in the preset response strategy table. The control strategy table presets 3 fixed response levels: at level I risk level, the system maintains the conventional response parameters, the contact opening drive pressure is set to the standard rated value (such as 3.5 megapascals), and the arc extinguishing gas flow is maintained at a normal flow (such as 25 liters / second); at level II, the contact opening pressure is increased to 1.4 times the rated value (such as 4.9 megapascals), and the arc extinguishing gas flow is increased to 35 liters / second; at level III, the contact opening pressure is set to 2.0 times the rated value (such as 7.0 megapascals), and the gas flow is increased to 50 liters / second. All response strategies are triggered by table lookup, and the response parameters are immediately loaded to the control output, and the command signal is transmitted to the opening execution unit and the gas flow control unit through the digital output interface.

[0050] After receiving the control signal, the opening drive execution unit starts the high-voltage electromagnetic thruster or servo motor located in the contact axis direction. The actuator is connected to the transmission assembly, which starts to generate a separation force within 1 millisecond of receiving a high-level signal and completes the opening action of 8 millimeters of contact opening within 10 milliseconds. The opening speed during this process is not less than 1 meter / second, ensuring that the arc channel is quickly lengthened, the arc voltage is increased, and eventually leading to natural extinction of the arc. At the same time, after receiving the flow instruction, the arc extinguishing gas flow unit controls the rapid release of high-pressure gas from the gas tank to the nozzle through the electromagnetic valve. The nozzle diameter is set to 2.5 millimeters, and the guide channel length is 100 millimeters. The released high-pressure gas forms a high-speed gas flow in the arc extinguishing chamber along the arc direction, with a flow rate of more than 80 meters / second, which can immediately reduce the arc temperature and destroy the plasma maintenance environment after the contact arc.

[0051] During the whole control process, the control structure records the contact opening time, the arc extinguishing gas flow response start time, the command issuing time, the feedback confirmation signal and the response result. The above data is archived through a time sequence record file and uploaded to the central monitoring device every 10 milliseconds for long-term archiving. If the feedback signal shows that the contacts have not separated within the set time or the gas flow has not reached the set flow value, the response is marked as "failed" and a fault troubleshooting prompt process is started. All response data will be used to modify the parameter reference value of the control strategy table for subsequent optimization, improving the pertinence and success rate of the next round of arc burning event response.

[0052] The role of this step is to establish a real-time linkage control mechanism from the arc burning energy prediction result to the arc extinguishing execution action, to achieve the proactive arc extinguishing response of the high-voltage circuit breaker before the arc burning risk has evolved into a fault disaster, thereby avoiding equipment damage, system trip and even power grid cascading accidents. During the operation of the high-voltage circuit breaker, once the arc is formed, it will rapidly act on the contacts, insulating medium and metal components with extremely high temperature and energy density. Although the previous step has judged that the arc is in a state of continuous growth through signal recognition and energy evaluation, if this information cannot be timely converted into an explicit executable control action, effective protection cannot be achieved. The traditional overcurrent or overvoltage protection usually has a certain delay, and the action logic takes "fault has been formed" as the premise, while this step emphasizes triggering the execution response in advance based on "trend estimation", with stronger feedforward control capability.

[0053] This step inputs the "advance trigger prediction result" into the response control structure in the form of data communication, selects the corresponding driving control strategy in real time according to the determined arc energy level (such as level I, level II, level III), and outputs explicit control parameter instructions, including the required gas pressure value, electromagnetic drive strength and specific flow value of the arc extinguishing gas flow. For example, when a high-risk level III arc is determined, the control structure can quickly generate a driving signal that is boosted to twice the rated pressure, ensuring that the contacts are separated at high speed within a very short time, and simultaneously releasing a high-speed high-pressure gas flow to form an arc extinguishing gas field that destroys the plasma arc channel. This linkage not only ensures that the opening and arc extinguishing actions are triggered simultaneously, but also automatically adjusts the action strength according to the risk level, achieving "on-demand response".

[0054] In addition, this step also has the function of monitoring and feedback of the response process. The control structure can judge whether the response action is accurately completed according to the actual opening time, gas flow response time, execution feedback signal and other parameters, and mark and record the control strategy, providing a data basis for subsequent optimization. In summary, this step is the key link in the whole arc monitoring and control chain that efficiently converts "early warning information" into "execution behavior", realizes the closed loop between monitoring, judgment and response, and is one of the core technologies to ensure the operation safety of high-voltage equipment and the efficiency of protection action.

[0055] S6, synchronously return the operation data in the whole process of contact opening and arc extinguishing to the transient disturbance identification model and the hierarchical response controller, respectively update the disturbance identification parameters and the response control strategy, and form a continuous closed-loop optimization high-voltage circuit breaker arc monitoring mechanism; In order to improve the accuracy of the high-voltage circuit breaker arc monitoring process and the effectiveness of the execution response, a closed-loop optimization mechanism is designed for the operation data generated in the whole process of contact opening and arc extinguishing, which is used for identifying parameter updating and response control strategy correction. The mechanism continuously evolves the monitoring link through operation process feedback driving, so that the monitoring structure can adapt to the changes of the field working conditions, and improve the stability and adaptability of the system under multiple disturbance and multiple scene conditions. The optimization process includes the following steps: All key operation data from the start of the early trigger judgment to the end of the arc extinguishing process are collected, and the data are packaged as feedback data frames in a unified timestamp format. The specific collection contents include the following nine types of data: 1) Contact opening start time (unit: microsecond); 2) Contact opening completion time (unit: microsecond); 3) Actual output pressure of the driving actuator (unit: megapascal); 4) Arc extinguishing gas flow start release time; 5) Actual flow rate of the arc extinguishing gas (unit: liter / second); 6) Arc extinguishing judgment time; 7) Total arc extinguishing time (unit: millisecond); 8) Total value of the current arc signal energy (unit: millijoule); 9) Final arc signal waveform curve segment.

[0056] The above data are synchronously recorded by the local collection equipment, and uploaded to the data processing area of the main control structure within 30 milliseconds after the action is completed. In order to ensure the sequence consistency and matching of the data, each piece of data is attached with a 13-bit timestamp, which is identified in the year-month-day-hour-minute-second-millisecond format, to ensure the correct timing in subsequent processing.

[0057] The part of the feedback data frame related to arc recognition, mainly including signal waveform, signal strength, arc extinction time and trigger point time difference, is input into the arc recognition structure for comparison of the matching conditions of recognition accuracy and interference separation parameters. If the time difference between the actual arc extinction time and the recognition prediction result exceeds 3 milliseconds, or the recognition starting point deviates from the actual breaking point by more than 5 milliseconds, the system will adjust the spectral window width, interference energy threshold and time mutation slope factor according to the actual arc extinction time. The specific adjustment operations include: expanding the sliding window of spectral analysis from the original 1000 microseconds to 1200 microseconds; increasing the interference energy trigger threshold by 20% from the original average value; extending the transient mutation judgment section from two consecutive points to three points. The above adjustment items will be stored as "corrected recognition parameters" in the recognition parameter configuration table and updated to the default enabled parameters before the next arc event.

[0058] The part of the running data frame related to control execution effect is fed back to the response control structure for verifying whether the actual action achieves the predetermined execution effect, and correcting the control parameter table accordingly. The specific judgment basis is: if the contact breaking time exceeds 15 milliseconds, or the actual flow rate of the arc extinguishing gas flow does not reach 90% of the set value or more, or the arc energy does not decrease by 50% or more within 10 milliseconds after the response action, the system will determine that the current response strategy is "low efficiency response", and the corresponding parameters will be upgraded. For example, the breaking drive pressure is increased from the original set value of 5.0 MPa to 5.5 MPa; the arc extinguishing gas flow is increased from 35 liters / second to 42 liters / second; the electromagnetic valve opening time delay is shortened from 2 milliseconds to 1 millisecond. The corrected response parameters are named with version number "R_time stamp", and the correction basis, original value and adjustment amplitude are recorded. This version will be automatically loaded as the default strategy for the next event response, and if the response effect is good for three times in a row, it will be solidified as the basic control strategy.

[0059] The purpose of this step is to realize the data-driven adaptive optimization mechanism of the high-voltage circuit breaker arc monitoring system. Through the structured collection and feedback of the running data generated during the breaking and arc extinguishing process, the recognition accuracy of the arc signal and the control strategy of the response action can be continuously updated according to the real running results, so as to build a monitoring and control system with closed-loop learning ability. In actual application, the signal pattern and energy release characteristics generated by the arc will be influenced by factors such as device structure, power grid disturbance, climate environment, etc. These factors have certain uncertainty and dynamic change characteristics. If the monitoring and recognition parameters and the control response strategy rely on static setting for a long time, they may not adapt to the changes on site, resulting in recognition deviation or response lag, thereby reducing the stability and protection efficiency of the overall system.

[0060] The step is achieved by collecting the contact opening starting time, the action time, the arc extinguishing gas flow response delay, the peak flow, the arc extinguishing completion time, the arc extinguishing curve slope, and the total energy dissipation during each circuit breaker action, and uploading these data to the transient disturbance identification structure and the response control structure, so as to realize the automatic correction of key identification parameters (such as the spectrum window width, the disturbance judgment threshold, and the spectrum mutation boundary) and the dynamic modification of the response strategy (such as the driving pressure, the gas flow intensity, and the action rhythm). The feedback mechanism ensures that the system can optimize the future monitoring identification and control execution logic based on the “actual performance” rather than the “theoretical preset”.

[0061] For example, if it is found that the arc duration during the arc extinguishing process is significantly longer than the estimation, the system will backtrack to identify whether the starting point is delayed, and correct the window starting position or extend the identification duration for spectrum identification. Similarly, if the arc extinguishing gas flow fails to reach the target flow within the limited time during the response, the control strategy will automatically increase the gas flow pressure setting value or shorten the electromagnetic valve response time. The “two-way update” mechanism based on the measured results not only improves the success rate of single event handling, but also forms a self-correcting monitoring control closed-loop system that continuously evolves and enhances robustness during operation, significantly enhancing the response speed and extinguishing efficiency of the high-voltage circuit breaker under complex working conditions, and improving the overall safety and reliability of the power grid operation.

[0062] Through the above-mentioned high-voltage circuit breaker arc monitoring method, high-precision identification and rapid response control of the arc signal under strong electromagnetic interference background can be achieved, and the stability, accuracy and real-time of arc monitoring can be significantly improved. Based on the synchronous collection of spectrum signals and electromagnetic vector signals, the stable base signal is extracted through primary difference, and the noise bottom fluctuation is effectively stripped. Combined with the interference identification model and the multi-dimensional spectrum diagram, the time sequence boundary of transient interference is accurately extracted and the interference weight matrix is constructed, realizing fine separation of arc and interference in the signal processing stage. Further, based on the dynamic purification of the arc feature sequence, energy progressive analysis is carried out for each period, and the prediction result is converted into a driving control instruction in advance, so as to ensure that the arc extinguishing action is completed before the critical arc energy. Finally, through the feedback correction of the whole process operation data, the identification parameters and control strategies are continuously iteratively optimized. The closed-loop mechanism breaks through the technical link of “perception-identification-control-optimization”, and under complex working conditions such as multi-source strong interference and high load of power grid, it can still realize early perception, accurate identification, active response and dynamic self-learning adjustment of arc risk, effectively avoid the damage of circuit breaker and systemic power accident caused by identification failure or response delay, and significantly enhance the operation safety and intelligent level of power equipment.

[0063] The present application provides a method for monitoring arc of high-voltage circuit breaker, comprising the following steps: Figure 2The high-voltage circuit breaker arc monitoring system shown comprises a synchronous sensing and base surface construction module, an interference identification and time sequence extraction module, a signal purification and feature extraction module, an energy analysis and early warning judgment module, a response control execution module, and a data back transmission and strategy optimization module. The synchronous sensing and base surface construction module synchronously collects arc spectrum signals and electromagnetic interference vector data under a unified time reference, and performs primary differential processing to generate stable base surface signals. The interference identification and time sequence extraction module inputs the stable base surface signals into an interference identification model, generates a multi-dimensional frequency spectrum graph through time-frequency mapping, extracts interference time sequence boundaries, and outputs an interference weight matrix. The signal purification and feature extraction module adaptively adjusts filter kernel coefficients according to the interference weight matrix, performs enhancement and suppression processing on the signals, and obtains a dynamic purified arc feature signal sequence. The energy analysis and early warning judgment module performs continuous energy analysis on the arc feature signal sequence, and completes early trigger estimation of protection action based on energy growth trend and a preset threshold. The response control execution module generates a control instruction according to the trigger estimation result, adjusts the opening pressure and arc extinguishing gas flow, and realizes rapid arc extinguishing. The data back transmission and strategy optimization module transmits the whole process operation data to the identification model and the controller, updates the identification parameters and the control strategy, and constructs a closed-loop optimized arc monitoring mechanism.

[0064] The high-voltage circuit breaker arc monitoring method provided by the embodiment of the present application is realized through the high-voltage circuit breaker arc monitoring system described above, and the specific method and process of the high-voltage circuit breaker arc monitoring system are described in the embodiment of the high-voltage circuit breaker arc monitoring method, which will not be repeated here.

[0065] The above only describes certain exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above drawings and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

Claims

1. A method for monitoring arcing of a high-voltage circuit breaker, characterized in that: The following steps are involved: S1, synchronously collects arc spectrum signals and electromagnetic interference vector data under a unified time base, and performs primary differential processing to generate a stable base plane signal; S2, inputs the stable base plane signal into the interference recognition model, generates a multi-dimensional spectrum through time-frequency mapping, extracts the interference time sequence boundary and outputs the interference weight matrix; S3, adaptively adjust the filter kernel coefficient according to the interference weight matrix, enhance and suppress the signal, and obtain a dynamically purified arcing characteristic signal sequence; S4, continuously analyzing the energy of the arcing characteristic signal sequence, and completing the early triggering prediction of the protection action based on the energy growth trend and the preset threshold; S5, generates control instructions based on the trigger prediction results, adjusts the trip pressure and arc extinguishing airflow, and realizes rapid arc extinguishing; S6, transmits the whole process operation data back to the identification model and controller, updates the identification parameters and control strategy, and builds a closed-loop optimized arc monitoring mechanism.

2. A high voltage circuit breaker arcing monitoring method according to claim 1, characterized in that: Step S1 includes: A micro-fiber spectrometer is installed at the observation window of the arc extinguishing chamber of the high-voltage circuit breaker to collect spectral signals during the arcing process. The spectral response range is set to 350 nm to 1050 nm, and the sampling frequency is set to 100 kHz. A three-axis electric field sensing antenna is installed on the outside of the circuit breaker housing, and a current transformer is installed at the incoming and outgoing line loops to collect electromagnetic interference vector data. The three-axis electric field antenna measurement frequency band covers 10 Hz to 10 MHz. The GPS timing chip is used to uniformly time-stamp the collected data, and a unified time base is used to achieve dual-channel synchronous collection and alignment of spectral signals and electromagnetic vector data; Perform primary differential processing, smooth or eliminate non-cooperative electromagnetic signals using linear interpolation within a local time window, complete time reconstruction and amplitude normalization, and generate a stable base surface signal.

3. A high voltage circuit breaker arcing monitoring method according to claim 1, characterized in that: Step S2 includes: The stable base plane signal is divided into time series using a sliding time window of 500 microseconds per segment, and the frequency distribution diagram in the range of 1 kHz to 5 MHz is obtained by fast Fourier transform. Based on the sliding window spectrum, the relationship between the energy of each frequency point and the reference energy is calculated to identify the start and end time of high-amplitude transient interference signals. The timing and frequency range of the interference events are mapped to the time axis and frequency band axis to construct an interference weight matrix. The interference level is set according to the interference amplitude and duration and a weight coefficient is assigned to guide subsequent signal processing.

4. A high voltage circuit breaker arcing monitoring method according to claim 1, characterized in that: Step S3 includes: Match each time point of the interference weight matrix with the sampling point of the stable base plane signal, and assign filter window lengths and weighted distributions corresponding to different interference levels; Adjust the window length and center weight in the sliding weighted average filter according to the interference level to achieve smooth suppression of high-interference signal segments and detail enhancement of low-interference signal segments; Perform amplitude-weighted amplification on signals in the 250 kHz to 2 MHz frequency band, perform attenuation on signals in non-target frequency bands, and perform amplitude normalization and dynamic overlap reconstruction after filtering in each frequency band; The processed signals are sequenced and reconstructed with their original timestamps to form a continuous and dynamically purified arcing characteristic signal sequence.

5. A high voltage circuit breaker arcing monitoring method according to claim 4, characterized in that: The specific steps for performing amplitude-weighted amplification on signals in the 250 kHz to 2 MHz frequency band, performing attenuation processing on signals in non-target frequency bands, and performing amplitude normalization and dynamic overlap reconstruction after filtering in each frequency band are as follows: Extract the 250 kHz to 2 MHz frequency band signal from the stable base plane signal, and set the amplification factor between 1.25 and 1.5 according to the frequency range for amplitude weighting processing; Extracting signals in the 1 kHz to 250 kHz frequency band and the 2 MHz to 5 MHz frequency band, and performing amplitude reduction by setting attenuation coefficients of 0.5, 0.7, and 0.9 based on the interference levels at the corresponding time points in the interference weight matrix; The signals processed in different frequency bands are normalized and superimposed to reconstruct the signals after frequency selective enhancement for subsequent arcing feature extraction and analysis.

6. A high voltage circuit breaker arcing monitoring method according to claim 1, characterized in that: Step S4 includes: The arcing characteristic signal sequence is divided into multiple overlapping sliding windows of 1 millisecond length, and the absolute value integral energy of the signal amplitude is calculated in each window to obtain the electrical signal energy value of each segment; Construct an energy time series, extract the energy change trend of continuous windows, and make corrections based on the amplitude ratio before and after signal enhancement to obtain an accurate energy growth curve; When it is determined whether the current energy value exceeds the energy threshold of 2.5 millijoules and the growth ratio in three consecutive windows exceeds 1.5 times, an early trigger estimation instruction is output.

7. A method for monitoring arcing of a high-voltage circuit breaker according to claim 1, characterized in that: Step S5 includes: Receive the arcing energy value, growth rate and risk level information in the early trigger prediction result, and transmit it to the control response structure in a period of no more than 1 millisecond; According to the risk level, the corresponding tripping drive pressure and arc extinguishing air flow control parameters are retrieved from the preset response strategy table; Output corresponding control instructions to the opening drive execution unit and the airflow control unit respectively, driving the contacts to separate within 10 milliseconds and controlling the high-pressure airflow to form a high-speed airflow along the arcing direction; Record and transmit back the tripping action time, airflow start time, control command issuance time and feedback status for subsequent correction and optimization of the control strategy.

8. A high voltage circuit breaker arcing monitoring method according to claim 1, characterized in that: Step S6 includes: Collect operating data including tripping time, actuation pressure, airflow velocity, arcing energy and waveform, and pack them into feedback data frames after adding timestamps; Input the identification-related data in the feedback data frame into the identification structure. If the identification deviation exceeds the set threshold, adjust the spectrum window width, interference energy threshold and slope factor and update them to the default enabled parameters; The control-related data in the feedback data frame is input into the response control structure. If the response does not achieve the expected effect, the driving pressure is increased, the airflow rate is increased, and the action delay is shortened, and the revised version is set as the default control strategy for the next time.

9. A high-voltage circuit breaker arcing monitoring system, used to implement a high-voltage circuit breaker arcing monitoring method according to any one of claims 1 to 8, characterized in that: It includes synchronous perception and base plane construction module, interference identification and timing extraction module, signal purification and feature extraction module, energy analysis and early warning judgment module, response control execution module and data feedback and strategy optimization module: The synchronous sensing and base plane construction module collects arc spectrum signals and electromagnetic interference vector data synchronously under a unified time base, and performs primary differential processing to generate a stable base plane signal; The interference identification and timing extraction module inputs the stable base plane signal into the interference identification model, generates a multi-dimensional spectrum through time-frequency mapping, extracts the interference timing boundary and outputs the interference weight matrix; The signal purification and feature extraction module adaptively adjusts the filter kernel coefficient according to the interference weight matrix, enhances and suppresses the signal, and obtains a dynamically purified arcing feature signal sequence; The energy analysis and early warning judgment module performs continuous energy analysis on the arcing characteristic signal sequence and completes the early triggering prediction of the protection action based on the energy growth trend and the preset threshold; The response control execution module generates control instructions based on the trigger prediction results, adjusts the trip pressure and arc extinguishing airflow, and realizes rapid arc extinguishing; The data feedback and strategy optimization module transmits the entire process operation data back to the identification model and controller, updates the identification parameters and control strategy, and builds a closed-loop optimized arc monitoring mechanism.

Images