A method and system for intelligent diagnosis and calibration of arc extinguishing characteristics of circuit breakers

Abstract

This invention relates to the field of circuit breaker testing technology, and discloses an intelligent diagnostic and calibration method and system for the arc-extinguishing characteristics of circuit breakers. The intelligent diagnostic and calibration method for the arc-extinguishing characteristics of circuit breakers includes the following steps: Step S1, generating a synchronization signal; Step S2, extracting feature parameters; Step S3, calculating the baseline clearance; Step S4, calculating the time window weight and clearance deficit; Step S5, calculating the basic compensation time; Step S6, calculating the residual deficit; Step S7, outputting the final compensation time, clearance deficit, and status label. This invention completes the diagnosis of the arc-extinguishing state and the calibration of the opening sequence of the circuit breaker by constructing a state variable group that matches the physical process of arc extinguishing. It also matches the operating conditions of offshore wind power parallel reactor disconnection, restoring the true insulation capacity of the break point, making the diagnostic results consistent with the actual state of the arc-extinguishing process. The calibration quantity can be adapted to the insulation gap of the break point, providing support for the control of the arc-extinguishing performance of the circuit breaker.

Description

Technical Field

[0001] This invention relates to the field of circuit breaker testing technology, and more specifically, to a method and system for intelligent diagnosis and calibration of the arc extinguishing characteristics of circuit breakers. Background Technology

[0002] The offshore wind power industry continues to expand, and offshore wind farm collection systems widely adopt 40.5kV vacuum circuit breakers to complete the switching operations of parallel reactors. Offshore wind farm collection systems contain a large number of long-distance submarine cables, whose inherent capacitance effect is significant. This results in a higher transient recovery voltage steepness across the circuit breaker terminals when the parallel reactors are disconnected, placing higher demands on the arc-extinguishing performance of the circuit breakers. The arc-extinguishing performance of the vacuum circuit breaker directly determines the safety of the switching operation and directly affects the stable operation of the entire collection system.

[0003] Existing diagnostic methods for the arc-extinguishing performance of vacuum circuit breakers mostly rely on the mechanical opening distance as the core criterion for judging the insulation capacity of the break, separating the diagnostic process from electrical and mechanical characteristics. These methods fail to consider the encroachment of residual conductive media on the insulation clearance before and after the current crosses zero, nor do they account for the differences in physical characteristics among different arc-extinguishing anomalies. Existing calibration methods often employ fixed opening sequence adjustment strategies, which cannot adaptively adjust based on the actual arc-extinguishing state of the circuit breaker.

[0004] Current methods directly equate the mechanical opening distance to the effective insulation clearance of the break, neglecting the constraint effect of the contact and shielding structures on the diffusion of residual particles, and also ignoring the continuous influence of arc energy on the post-arc dielectric recovery process. This approach results in a significant deviation between the arc-extinguishing capability obtained from the diagnosis and the actual state of the circuit breaker. The calculated calibration value cannot match the actual insulation gap at the break, ultimately leading to circuit breaker reignition and overvoltage faults, affecting the safe and stable operation of the equipment and the current collection system. Summary of the Invention

[0005] This invention provides a method and system for intelligent diagnosis and calibration of the arc extinguishing characteristics of circuit breakers, solving the technical problems mentioned in the background.

[0006] This invention provides an intelligent diagnostic and calibration method for the arc-extinguishing characteristics of circuit breakers, comprising the following steps: Step S1: Acquire the switching signal, extract the zero-point time, and shift the switching signal to align with the zero-point time to generate a synchronization signal; Step S2: Extract the opening gap increment, opening gap speed increase, arc energy, transferred charge, and dynamic resistance based on the synchronization signal; Step S3: Introduce the reference coefficient, calculate the reference displacement by combining the reference coefficient, arc energy and transferred charge, and subtract the reference displacement from the opening gap increment to obtain the reference clearance. Step S4: Calculate the status label by integrating the baseline clearance, dynamic resistance and synchronization signal; match the target coefficient according to the status label; calculate the target displacement by combining the target coefficient, arc energy and migrated charge; subtract the target displacement from the opening gap increment to obtain the target clearance; and calculate the time window weight and clearance deficit by integrating the target displacement and target clearance. Step S5: Calculate the critical speed by combining the time window weight and the opening gap growth rate, and divide the net air gap by the critical speed to obtain the basic compensation time. Step S6: Integrate the basic compensation time, the opening distance growth rate and the opening distance increment to synthesize the calibration opening distance, combine the target coefficient, the arc energy and the transferred charge to calculate the calibration displacement, and subtract the calibration displacement from the calibration opening distance to obtain the calibration net amount, subtract the target net amount from the calibration net amount to obtain the net amount increment, and subtract the net amount increment from the net amount deficit to obtain the residual deficit. Step S7: Divide the residual deficit by the critical speed and add the basic compensation time to generate the final compensation time. Adjust the circuit breaker tripping command issuance time according to the final compensation time, and output the final compensation time, net clearance deficit and status label.

[0007] This invention provides an intelligent diagnostic and calibration system for the arc-extinguishing characteristics of circuit breakers, comprising: The synchronization signal generation module acquires the on / off signal, extracts the zero-point time, and shifts the on / off signal to align with the zero-point time to generate the synchronization signal. The feature parameter extraction module extracts the opening gap increment, opening gap speed increase, arc energy, migrated charge, and dynamic resistance based on the synchronization signal. The reference clearance calculation module introduces a reference coefficient, combines the reference coefficient, arc energy and migrated charge to calculate the reference displacement, and subtracts the reference displacement from the opening gap increment to obtain the reference clearance. The standard clearance calculation module integrates the baseline clearance, dynamic resistance and synchronization signal to calculate the status label, matches the target coefficient according to the status label, calculates the target displacement by combining the target coefficient, arc energy and migrated charge, and obtains the target clearance by subtracting the target displacement from the opening gap increment. It also integrates the target displacement and target clearance to calculate the time window weight and clearance deficit. The basic compensation calculation module combines the time window weight and the opening distance growth rate to calculate the critical speed, and divides the net air volume deficit by the critical speed to obtain the basic compensation time. The residual deficit calculation module integrates the basic compensation time, the opening gap growth rate and the opening gap increment to synthesize the calibration opening gap, combines the target coefficient, arc energy and migrated charge to calculate the calibration displacement, and subtracts the calibration displacement from the calibration opening gap to obtain the calibration net volume, subtracts the target net volume from the calibration net volume to obtain the net volume increment, and subtracts the net volume deficit from the net volume increment to obtain the residual deficit. The final compensation time generation module divides the residual deficit by the critical speed and adds it to the basic compensation time to generate the final compensation time. It also adjusts the time of the circuit breaker tripping command based on the final compensation time and outputs the final compensation time, the net clearance deficit, and the status label.

[0008] The beneficial effects of this invention are as follows: This invention constructs a state quantity set matching the physical process of arc extinguishing by uniformly aligning multi-source signals, forming a feature calculation process adaptable to different anomaly types, and completing the diagnosis of the arc extinguishing state of the circuit breaker and the calibration of the opening timing. This invention can match the operating conditions of the disconnection of parallel reactors in offshore wind power, restore the true insulation capacity of the break, and make the diagnostic results closely match the actual state of the arc extinguishing process. The calibration quantity can be adapted to the insulation gap of the break. This invention can be implemented on the basis of existing circuit breaker hardware without additional equipment modification, adapting to the application environment of offshore wind power sites, and providing support for the control of the arc extinguishing performance of circuit breakers. Attached Figure Description

[0009] Figure 1 This is a flowchart of an intelligent diagnosis and calibration method for the arc extinguishing characteristics of a circuit breaker according to the present invention; Figure 2 This is a state diagnosis distribution diagram of the present invention; Figure 3 This is a schematic diagram of the confusion matrix of the present invention; Figure 4 This is a diagram showing the net clearance deficit and compensation before and after calibration according to the present invention. Detailed Implementation

[0010] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0011] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0012] like Figures 1-4 As shown, a method for intelligent diagnosis and calibration of the arc-extinguishing characteristics of a circuit breaker includes the following steps: Step S1: Acquire the switching signal, extract the zero-point time, and shift the switching signal to align with the zero-point time to generate a synchronization signal; Step S2: Extract the opening gap increment, opening gap speed increase, arc energy, transferred charge, and dynamic resistance based on the synchronization signal; Step S3: Introduce the reference coefficient, calculate the reference displacement by combining the reference coefficient, arc energy and transferred charge, and subtract the reference displacement from the opening gap increment to obtain the reference clearance. Step S4: Calculate the status label by integrating the baseline clearance, dynamic resistance and synchronization signal; match the target coefficient according to the status label; calculate the target displacement by combining the target coefficient, arc energy and migrated charge; subtract the target displacement from the opening gap increment to obtain the target clearance; and calculate the time window weight and clearance deficit by integrating the target displacement and target clearance. Step S5: Calculate the critical speed by combining the time window weight and the opening gap growth rate, and divide the net air gap by the critical speed to obtain the basic compensation time. Step S6: Integrate the basic compensation time, the opening distance growth rate and the opening distance increment to synthesize the calibration opening distance, combine the target coefficient, the arc energy and the transferred charge to calculate the calibration displacement, and subtract the calibration displacement from the calibration opening distance to obtain the calibration net amount, subtract the target net amount from the calibration net amount to obtain the net amount increment, and subtract the net amount increment from the net amount deficit to obtain the residual deficit. Step S7: Divide the residual deficit by the critical speed and add the basic compensation time to generate the final compensation time. Adjust the circuit breaker tripping command issuance time according to the final compensation time, and output the final compensation time, net clearance deficit and status label.

[0013] In one embodiment of the present invention, acquiring an on / off signal, extracting the zero-point time, and shifting the on / off signal to align with the zero-point time to generate a synchronization signal includes: The formula for calculating the breaking signal is as follows, based on the arc voltage, breaking current, transient recovery voltage, post-arc current, trip coil current, and contact displacement: ; in, For on / off signal, Arc voltage To interrupt the current, For transient recovery voltage, For the back-arc current, For the trip coil current, For contact displacement, Time is the independent variable; The formula for calculating the zero-point time, based on the interruption current, the time of the tripping command, and the end time of the observation window, is as follows: ; in, It is the time of midnight. To find the sign of the independent variable corresponding to the minimum value, The time when the tripping command is issued. The end time of the observation window. For interrupting current; Based on the switching signal and the zero-point time, the formula for calculating the synchronization signal is as follows: ; in, For synchronization signal, This is the translation time variable.

[0014] It should be noted that the interruption signal is a time-series set of multi-dimensional electrical and mechanical parameters synchronously acquired during the circuit breaker's opening process, used to comprehensively characterize the state changes throughout the entire process of circuit breaker opening and arc extinguishing. Arc voltage is the voltage drop across the arc between the contacts during circuit breaker opening, a key electrical parameter characterizing the arc's development and extinction state. Breaking current is the interrupted current flowing through the main circuit during circuit breaker opening, a key parameter for determining the current zero point and opening condition. Transient recovery voltage is the recovery voltage re-established across the break after the main current crosses zero and extinguishes, a key parameter characterizing the pressure the medium at the break point withstands. Post-arc current is the weak conductive current formed by the movement of residual charged particles between the breaks after the main current crosses zero, a key parameter characterizing the post-arc medium recovery process. The opening coil current is the current flowing through the coil of the opening operating mechanism during circuit breaker opening, a key parameter characterizing the timing and health status of the operating mechanism. Contact displacement is the change in position of the moving contact relative to the stationary contact during the circuit breaker opening process, and it is a key parameter characterizing the change in mechanical opening distance.

[0015] It should be noted that the time independent variable is an independent variable characterizing the timing changes during the tripping process, used to uniformly describe the time axis of all timing signals. Zero-point time is the moment when the absolute value of the main circuit interrupting current is at its minimum and a positive / negative crossover occurs during the tripping process; it serves as the unified benchmark for aligning the timing of all signals. The symbol corresponding to the minimum value of the independent variable is a mathematical operation symbol used to locate the independent variable corresponding to the minimum value of the function, used to accurately locate the zero-point time from the interrupting current timing sequence. The tripping command issuance moment is the moment when the circuit breaker control system issues the tripping action command to the tripping operating mechanism; it is the starting time benchmark for the tripping process. The observation window end moment is the end moment of the observation time window used to cover the entire tripping and arc extinguishing process, used to limit the time range for calculating all signals. The synchronization signal is the tripping signal after time shifting and alignment based on the zero-point time, used to eliminate timing misalignments between electrical and mechanical processes. The shifted time variable is a relative time variable with the zero-point time as its zero point, used to uniformly describe the relative timing of all signals after alignment.

[0016] It should be noted that the selection criterion for the end time of the observation window is that it must completely cover the entire process of the circuit breaker's tripping action, the entire arcing process, and the entire dielectric recovery process after the current crosses zero. Specifically, the preferred time range is 200 milliseconds after the tripping command is issued. For example, the total tripping time of a 40.5kV vacuum circuit breaker is typically no more than 60 milliseconds, the arcing time is typically no more than 5 milliseconds, and the dielectric recovery process after the current crosses zero is typically no more than 1 millisecond. A 200-millisecond observation window can completely cover all critical processes while reserving sufficient redundancy time to avoid signal truncation. For special long-arcting conditions, the end time of the observation window can be extended to 300 milliseconds after the tripping command is issued to ensure complete coverage of the tripping and arc extinguishing process.

[0017] It should be noted that the synchronous acquisition of all parameters of the switching signal must be completed using the same high-speed synchronous data acquisition device for all channels. The sampling frequency should preferably be set to 10 MHz, and the synchronous acquisition accuracy of all channels should not be less than 100 nanoseconds. For example, signals such as the post-arc current and transient recovery voltage during the arc extinguishing process of a vacuum circuit breaker exhibit abrupt changes at the microsecond level. A sampling frequency of 10 MHz can achieve a time resolution of 0.1 microseconds, which can completely capture the details of the abrupt changes in the signal. A synchronization accuracy within 100 nanoseconds can ensure that the signal timing deviation of all channels does not exceed 0.1 microseconds, fully meeting the accuracy requirements for current zero-point alignment. For special requirements of field operating conditions, the minimum sampling frequency should not be lower than 1 MHz, and the minimum synchronization accuracy should not be lower than 1 microsecond; otherwise, the integrity of signal acquisition and the accuracy of timing alignment cannot be guaranteed.

[0018] Specifically, traditional circuit breaker diagnostic methods typically separate electrical arc-extinguishing diagnosis from mechanical characteristic diagnosis. Electrical diagnosis only collects voltage and current signals, while mechanical diagnosis only collects travel and coil current signals, failing to fully characterize the electromechanical coupling characteristics of the arc-extinguishing process. This invention employs a multi-dimensional parameter combination of the breaking signal to simultaneously cover key information of the arc-extinguishing process. Arc voltage and breaking current characterize the arc development process; transient recovery voltage and post-arc current characterize the dielectric recovery process after the current crosses zero; trip coil current characterizes the timing characteristics of the operating mechanism; and contact displacement characterizes the change in mechanical opening distance. This fully reconstructs the coupling influence relationship between electrical and mechanical parameters during the arc-extinguishing process, providing complete foundational data for subsequent effective clearance calculations, which will not be elaborated upon here.

[0019] In one embodiment of the present invention, the extraction of the opening gap increment, opening gap speed increase, arc energy, migrated charge, and dynamic resistance based on the synchronization signal includes: Based on the contact displacement and zero-point time in the synchronization signal, the formula for calculating the opening gap increment is as follows: ; in, For the opening distance increment, For contact displacement, The contact displacement is at zero time. The formula for calculating the growth rate of the opening gap is as follows, based on the opening gap increment: ; in, The opening distance is the growth rate; The formula for calculating arc energy is as follows, based on arc voltage, breaking current, arc initiation time, and zero-point time: ; in, For electric arc energy, For the moment of arc establishment, It is the time of midnight. Arc voltage For interrupting current; The formula for calculating the transferred charge is as follows, based on the back-arc current and the zero-point time: ; in, For the transfer of charge, For the back-arc current, For integration time;

[0020] Based on the arc voltage, breaking current, and minimum positive value, the formula for calculating dynamic resistance is as follows: ; in, For dynamic resistance, It is a very small positive number.

[0021] It should be noted that the opening gap increment is the change in contact displacement at any given moment relative to the contact displacement at zero point, used to characterize the increase in the mechanical opening gap of the contacts after the current crosses zero. The contact displacement at zero point is the position of the moving contact at zero point, used as the reference value for calculating the opening gap increment. The opening gap growth rate is the first derivative of the opening gap increment with respect to time, used to characterize the speed at which the contacts separate. Arc energy is the integral of arc power with respect to time from the moment the arc is established to the zero point, used to characterize the total energy injected into the break gap before the current crosses zero. The moment the arc is established is the moment when the contacts have just separated and the arc voltage has steadily increased during the opening process, serving as the starting time reference for calculating arc energy. The migrated charge is the integral of the absolute value of the subsequent arc current with respect to time from the zero point to the current moment, used to characterize the total amount of residual charged particles migrating between the breaks after the current crosses zero. The dynamic resistance is the ratio of the arc voltage to the breaking current at any given moment, used to characterize the changes in the arc development process and the contact state of the contacts. Minimal positive numbers are fixed constants used to prevent the denominator from being zero in division operations, thus avoiding numerical overflow during the calculation process. They are preferably set to 10 to the power of -6.

[0022] It should be noted that the arc establishment moment is determined using a dual-threshold joint judgment rule. The moment when both conditions are simultaneously met is considered the arc establishment moment. The first condition is that the contact displacement signal reaches the just-separated position, i.e., the position where the moving contact and stationary contact have just separated. This position can be calibrated through the stroke curve of the circuit breaker's factory test. The second condition is that the arc voltage signal exhibits a step increase, and the rate of change of the arc voltage exceeds a preset threshold, which is preferably set to 1 kV per microsecond. The earliest moment when both conditions are simultaneously met is the arc establishment moment. In other words, the contact just separating is a prerequisite for arc generation, and the step increase in arc voltage is a key indicator of stable arc establishment. The dual-threshold joint judgment avoids misjudgments caused by single-threshold judgment, ensuring the accuracy of arc establishment moment positioning.

[0023] In one embodiment of the present invention, a reference coefficient is introduced, and the reference displacement is calculated by combining the reference coefficient, arc energy, and transferred charge. The reference clearance is obtained by subtracting the reference displacement from the opening gap increment, including: Based on the baseline geometric constraint gain coefficient, the equivalent evacuation channel perimeter length, and the equivalent evacuation channel area, the formula for calculating the baseline perimeter constraint amplification factor is as follows: ; in, As the reference perimeter constraint magnification factor, The reference geometric constraint gain coefficient, This is the equivalent evacuation channel perimeter length. Pi This is the equivalent evacuation channel area; Based on the reference perimeter constraint magnification factor, reference charge mapping coefficient, migrated charge, reference energy mapping coefficient, arc energy, opening distance increment, and attenuation characteristic length, the formula for calculating the reference displacement is as follows: ; in, As the reference displacement, The reference charge mapping coefficient, For the transfer of charge, As the baseline energy mapping coefficient, For electric arc energy, For the opening distance increment, For the attenuation characteristic length, It is a natural exponential function; The formula for calculating the reference clearance amount is as follows, based on the opening distance increment and the reference displacement: ; in, This is the baseline net air volume.

[0024] It should be noted that the reference perimeter constraint amplification factor is used to characterize the degree to which the perimeter structure of the contact and shield amplifies the headroom hysteresis effect. The reference geometric constraint gain coefficient is a fixed coefficient used in the pre-diagnosis process, characterizing the degree of influence of the perimeter structure on the headroom hysteresis effect. The equivalent evacuation channel perimeter length is a geometric parameter characterizing the perimeter size of the residual particle diffusion channel between the contact and shield. The equivalent evacuation channel area is a geometric parameter characterizing the cross-sectional area of ​​the residual particle diffusion channel between the contact and shield. The reference displacement is the perimeter constraint back-arc headroom hysteresis displacement calculated based on the reference coefficient set, used for the reference feature construction in the pre-diagnosis process. The reference charge mapping coefficient is a fixed coefficient used in the pre-diagnosis process to convert the back-arc migrated charge into equivalent headroom loss. The reference energy mapping coefficient is a fixed coefficient used in the pre-diagnosis process to convert the arc energy into equivalent headroom loss. The attenuation characteristic length is a characteristic parameter characterizing the rate attenuation of the arc energy memory effect with the increase of the mechanical opening distance. The reference headroom amount is obtained by subtracting the reference displacement from the mechanical opening distance increment, used as the reference effective insulation headroom in the pre-diagnosis process.

[0025] It should be noted that the offline identification of the reference geometric constraint gain coefficient, reference charge mapping coefficient, and reference energy mapping coefficient was completed using a type test platform for the same type of circuit breaker. The specific test design and identification process consisted of four steps. The first step was to build the test platform. The test circuit adopted a 40.5kV parallel reactor breaking circuit consistent with the offshore wind power site, which can simulate breaking conditions with different arcing energies and different reignition times. At the same time, a signal acquisition device consistent with the field was installed to synchronously acquire all electrical and mechanical signals. The second step was to conduct multiple sets of breaking tests, covering breaking scenarios of different severity, such as normal operation, slight reignition, and multiple reignition. Each set of tests was repeated 10 times, and the complete signal data and arc extinguishing results of all tests were recorded. The third step was data preprocessing. The signals of each set of tests were zero-point aligned, and the corresponding intermediate state quantities such as opening distance increment, arc energy, and migrated charge were calculated. At the same time, the actual net clearance hysteresis displacement corresponding to each set of tests was calibrated by using the arc extinguishing results and dielectric recovery simulation data. The fourth step is parameter identification. The least squares method is used, with the calibrated real net clearance hysteresis displacement as the target value, to fit and identify the reference geometric constraint gain coefficient, reference charge mapping coefficient, and reference energy mapping coefficient to obtain the optimal coefficient values.

[0026] Specifically, the reference coefficients consist of the reference geometric constraint gain coefficient, the reference charge mapping coefficient, and the reference energy mapping coefficient, forming a general fixed coefficient set used in the pre-diagnosis stage. The reference geometric constraint gain coefficient is dimensionless, with an optimal value range of 0.02 to 0.20. It adapts to the geometric characteristics of the 40.5kV vacuum interrupter contacts and shield, controlling the perimeter constraint amplification factor within a reasonable range and preventing displacement calculation results from exceeding the circuit breaker's rated opening distance. The reference charge mapping coefficient has dimensions of mm / mC, with an optimal value range of 0.05 to 0.50 mm / mC. It accurately maps the post-arc migration charge to the equivalent insulation headroom loss, and the calculation results match the measured post-arc charge range under the condition of disconnection of offshore wind power parallel reactors. The reference energy mapping coefficient has dimensions in mm / J, with an optimal range of 0.001 to 0.010 mm / J. This quantifies the continuous impact of arc energy on the recovery process of the post-arc medium. The calculated energy-related displacement loss is perfectly matched to the measured arc energy range of a 40.5kV vacuum circuit breaker. The attenuation characteristic length has dimensions in mm, maintaining the same dimension as the opening distance increment. The optimal range is 5 to 15 mm, with 8 mm being the preferred value. This aligns with the physical laws of metal vapor diffusion in a 40.5kV vacuum circuit breaker and perfectly matches the industry-measured safe travel of 6 mm.

[0027] It should be noted that the equivalent evacuation channel perimeter length and equivalent evacuation channel area are directly extracted and calculated from the two-dimensional engineering drawings or three-dimensional geometric model of the vacuum interrupter. The equivalent evacuation channel is the annular diffusion channel between the contact end face and the main shield. The perimeter length is the average of the circumference of the contact's outer diameter and the circumference of the main shield's inner diameter, calculated as π multiplied by the sum of the contact's outer diameter and the main shield's inner diameter. The equivalent evacuation channel area is the cross-sectional area of ​​the annular channel between the contact end face and the main shield, calculated as π multiplied by the square of half the main shield's inner diameter, minus π multiplied by the square of half the contact's outer diameter.

[0028] It should be noted that the determination of the attenuation characteristic length adopts a combination of simulation of metal vapor diffusion in a vacuum interrupter and offline test calibration, with a preferred value of 8 mm. The specific calibration process consists of three steps. The first step is to conduct metal vapor diffusion simulation. Based on the actual geometry of the same type of interrupter, a fluid simulation model is established to simulate the diffusion and attenuation process of metal vapor at different opening distances after the current crosses zero, obtaining the attenuation curve of metal vapor concentration as the opening distance increases, and initially determining the range of the attenuation characteristic length. The second step is to conduct offline breaking tests. Breaking tests of parallel reactors at different opening distances are carried out on a type test platform, recording the reignition probability and dielectric recovery characteristics at different opening distances, obtaining the influence curve of the opening distance on the arc extinguishing success rate. The third step is parameter fitting calibration. Based on the attenuation curve obtained from the simulation and using the arc extinguishing success rate curve obtained from the test as the calibration target, the attenuation characteristic length is fitted and calibrated to obtain the optimal value. For circuit breakers of different voltage levels, the range of the attenuation characteristic length is 5 mm to 15 mm, with a larger value for higher voltage levels.

[0029] Specifically, this invention abstracts the space between the contact and the shield as an equivalent evacuation channel for residual particles. The perimeter constraint amplification factor is calculated by the perimeter length and area of ​​the channel. The larger the ratio of the square of the perimeter length to 4 times pi multiplied by the channel area, the more complex the perimeter of the channel is, the higher the degree of narrowing, the stronger the constraint on particle diffusion, and the larger the corresponding perimeter constraint amplification factor. This can accurately quantify the degree of amplification of the net clearance hysteresis effect by the geometric structure, which will not be elaborated here.

[0030] Specifically, this invention decomposes the reference displacement into the superposition of two factors with clear physical meanings. The post-arc charge mapping term represents the insulation clearance occupied by residual charged particles at the current moment, and the arc energy memory decay term represents the continuous influence of the metal vapor generated by the arc energy before zero crossing on the insulation clearance. Moreover, this influence decays exponentially with the increase of the mechanical opening distance. The superposition of the two factors can completely quantify the clearance lag effect at different times, which will not be elaborated here.

[0031] Specifically, existing technologies generally equate mechanical clearance directly with effective insulation clearance, ignoring the encroachment of residual conductive media on the insulation clearance, leading to a systematic overestimation of arc-extinguishing capability. This invention, however, clarifies that effective clearance is the net insulation gap that can truly withstand the recovery voltage after deducting the encroachment from residual conductive media and structural edge effects within the mechanical clearance. This accurately characterizes the true insulation withstand capability of the fracture surface, thus resolving the issue of the inequivalence between mechanical clearance and effective insulation clearance, which will not be elaborated upon here.

[0032] In one embodiment of the present invention, a state label is calculated by fusing the baseline clearance, dynamic resistance, and synchronization signal; a target coefficient is matched based on the state label; the target displacement is calculated by combining the target coefficient, arc energy, and migrated charge; and the target clearance is obtained by subtracting the target displacement from the opening gap increment. The calculation of the time window weight and clearance deficit is then performed by fusing the target displacement and target clearance.

[0033] The formula for calculating electrical concealment characteristics based on arc voltage, breaking current, transient recovery voltage, post-arc current, and dynamic resistance is as follows: ; in, For electrical concealment features, For electrical branch network operation functions; Based on the trip coil current, opening gap increment, opening gap growth rate, and reference clearance, the calculation formula for the reference mechanical hidden characteristics is as follows: ; in, As the baseline mechanical hiding feature, For mechanical branch network operation functions; The formula for calculating the pre-diagnosis time window weight is as follows, based on the reference displacement, reference clearance, first weighting coefficient, second weighting coefficient, and offset constant: ; in, Weighting of the pre-diagnosis time window, For activation function, As the first weighting coefficient, This is the second weighting coefficient. It is a bias constant; Based on the pre-diagnosis time window weight, electrical hidden features, and baseline mechanical hidden features, the calculation formula for the pre-diagnosis fusion features is as follows: ; in, For the preliminary diagnosis of fusion features, It is the time of midnight. The end time of the observation window. For feature concatenation operations; Based on the pre-diagnostic fusion features, the classification label matrix, and the classification bias constant, the formula for calculating the state label is as follows: ; in, For status labels, To find the maximum value function, For class probability transformation function, For the category label matrix, For classification bias constants; Based on the status label and the preset mapping relationship, the formula for calculating the coefficient number corresponding to the target coefficient is as follows: ; in, Number the coefficients. This is a preset mapping relationship;

[0034] Based on the target geometric constraint gain coefficient, the equivalent evacuation channel perimeter length, and the equivalent evacuation channel area, the formula for calculating the target perimeter constraint amplification factor is as follows: ; in, The target perimeter constraint magnification factor. The gain coefficient is the target geometric constraint. The formula for calculating the target displacement is as follows, based on the target perimeter constraint magnification factor, target charge mapping coefficient, migrated charge, target energy mapping coefficient, arc energy, opening distance increment, and attenuation characteristic length: ; in, For the target displacement, For target charge mapping coefficients, The target energy mapping coefficient; The formula for calculating the target clearance amount is as follows, based on the opening distance increment and the target displacement: ; in, For target net air volume; The calculation formula for the mechanical concealment characteristics of the target is as follows, based on the trip coil current, opening gap increment, opening gap speed increase, and target clearance: ; in, To hide the mechanical features of the target; The formula for calculating the time window weight is as follows, based on the target displacement, target clearance, first weighting coefficient, second weighting coefficient, and offset constant: ; in, Assign time window weights; Based on the time window weight, electrical concealment features, and target mechanical concealment features, the formula for calculating the formal fusion features is as follows: ; in, Formal fusion features; Based on the formal fusion characteristics, the deficit weight matrix, and the deficit bias vector, the formula for calculating the net air deficit is as follows: ; in, This is due to a net air gap. This is the deficit weight matrix. This is the shortfall bias vector.

[0035] It should be noted that the electrical hidden features are deep features extracted by the electrical branch network from the input multi-dimensional electrical timing signals, used to characterize the abnormal electrical state during the arc extinguishing process. The electrical branch network operation function is a temporal convolutional network operation rule used to extract deep features from the electrical timing signals. The baseline mechanical hidden features are deep features extracted by the mechanical branch network from the input multi-dimensional mechanical timing signals and the baseline clearance, used to characterize the state of the operating mechanism during the pre-diagnosis process. The mechanical branch network operation function is a gated recurrent unit network operation rule used to extract deep features from the mechanical timing signals. The pre-diagnosis time window weights are temporal weight coefficients used to characterize the arc extinguishing risk level at different times during the pre-diagnosis process, used to focus on the feature information of high-risk periods. The activation function is a function used to perform nonlinear transformation on the intermediate calculation results of the network, used to enhance the nonlinear fitting ability of the network; it is preferentially set to the Sigmoid activation function, with the output value fixed between 0 and 1. The first weight coefficient is used to calculate the time window weights, corresponding to the network weight parameters of the baseline displacement term. The second weight coefficient is used to calculate the time window weights, corresponding to the network weight parameters of the baseline clearance term. The bias constant is a network bias parameter used to calculate the time window weights, serving as a baseline value for adjusting the time window weights.

[0036] It should be noted that the pre-diagnosis fusion feature is the fusion result of the electrical hidden features (weighted by time window) and the baseline mechanical hidden features during the pre-diagnosis process, used to generate the pre-diagnosis state label. Feature concatenation is an operation that merges two feature vectors of the same time length along the feature dimension, used to fuse the feature information of the electrical and mechanical branches. The state label is the classification result used to characterize the type and severity of the circuit breaker's arc-extinguishing anomaly, and is a key output of the pre-diagnosis process. The category probability transformation function is a function that converts multiple values ​​output by the network into probability values ​​between 0 and 1, used to generate the corresponding probabilities for different state labels. The classification label matrix is ​​the network weight matrix used to map the pre-diagnosis fusion features to the corresponding values ​​of different state labels. The classification bias constant is the network bias parameter used for calculating the classification labels, used to adjust the baseline probabilities of different state labels. The coefficient number is a one-to-one correspondence with the state label, used to index the target coefficient group. The preset mapping relationship is a one-to-one correspondence rule between the state label and the coefficient group number, used to match the corresponding target coefficient group based on the state label.

[0037] It should be noted that the target perimeter constraint amplification factor is calculated based on the target coefficient set matched by the state labels, and is used as a perimeter constraint amplification parameter in the formal diagnostic process. The target geometric constraint gain coefficient corresponds to the state labels and characterizes the degree of influence of the perimeter structure on the clearance hysteresis effect. The target displacement is the perimeter constraint back-arc clearance hysteresis displacement calculated based on the target coefficient set matched by the state labels, and is a key feature in the formal diagnostic process. The target charge mapping coefficient corresponds to the state labels and converts the amount of back-arc migrated charge into equivalent clearance loss. The target energy mapping coefficient corresponds to the state labels and converts the arc energy into equivalent clearance loss. The target clearance amount is obtained by subtracting the target displacement from the mechanical gap increment, and represents the effective insulation clearance that can truly withstand the recovery voltage during the formal diagnostic process. The target mechanical hidden feature is a deep feature extracted by the mechanical branch network from the input multi-dimensional mechanical timing signal and the target clearance amount, and is used for the state representation of the operating mechanism during the formal diagnostic process.

[0038] Specifically, the target coefficients consist of target geometric constraint gain coefficients, target charge mapping coefficients, and target energy mapping coefficients, forming multiple sets of dedicated coefficients that correspond one-to-one with the status labels. Under normal operating conditions, the target coefficients are preferably consistent with the reference coefficients. Under anomalies dominated by insufficient clearance, the target geometric constraint gain coefficient is preferably 1.2 to 1.8, the target charge mapping coefficient is preferably 0.15 to 0.30 mm / mC, and the target energy mapping coefficient is preferably 0.008 to 0.015 mm / J. Under anomalies dominated by mechanism delay, the target geometric constraint gain coefficient is preferably 0.8 to 1.1, the target charge mapping coefficient is preferably 0.04 to 0.18 mm / mC, and the target energy mapping coefficient is preferably 0.002 to 0.010 mm / J. Under anomalies dominated by localized ablation, the target geometric constraint gain coefficient is preferably 1.3 to 2.0, the target charge mapping coefficient is preferably 0.20 to 0.35 mm / mC, and the target energy mapping coefficient is preferably 0.010 to 0.018 mm / J. The above range reflects the differences in the amplification of the headroom lag effect by different anomaly types. Among them, the mechanism delay-dominated anomaly is mainly manifested as mechanical timing deviation, so the target coefficient is close to the benchmark coefficient. The local ablation-dominated anomaly will enhance the local electric field distortion and medium recovery degradation, so the target coefficient is higher than the benchmark coefficient.

[0039] It should be noted that the optimal range of all coefficients was calibrated by fitting multi-condition data from the 40.5kV vacuum circuit breaker type test. This ensures that the calculated results of the reference displacement, target displacement, and calibration displacement are all in length dimension, which is completely consistent with the dimensions of the opening distance increment, clearance, and clearance deficit. It also conforms to the physical laws of the medium recovery process in the vacuum interrupter, which will not be elaborated here.

[0040] It should be noted that the time window weight is a temporal weight coefficient used in the formal diagnostic process to characterize the degree of arc extinguishing risk at different times, focusing on the characteristic information of high-risk periods. The formal fusion feature is the fusion result of the electrical hidden features weighted by the time window and the target mechanical hidden features, used to calculate the final clearance deficit. The clearance deficit is the output of the formal diagnostic process, representing the amount of effective insulation clearance that needs to be supplemented to achieve stable arc extinguishing, and is one of the key diagnostic results. The deficit weight matrix is ​​the network weight matrix used to map the formal fusion feature to the clearance deficit. The deficit bias vector is the network bias parameter used to calculate the clearance deficit, adjusting the baseline value of the clearance deficit.

[0041] It should be noted that the electrical branch network operation function adopts a network structure of 3 layers of causal temporal convolutional layers combined with 1 layer of global average pooling layer. The specific hyperparameters are: kernel size of 3 for each convolutional layer, dilation rates of 1, 2, and 4 respectively, 64 hidden channels, ReLU activation function, and causal padding to ensure no leakage of temporal information. The mechanical branch network operation function adopts a network structure of 2 layers of gated recurrent units combined with 1 fully connected layer. The specific hyperparameters are: hidden layer dimension of 32 for each gated recurrent unit layer, dropout ratio of 0.2 to prevent overfitting, output dimension of 32 for the fully connected layer, and ReLU activation function. For small sample training scenarios, the number of convolutional layer channels and the hidden layer dimension of the gated recurrent unit can be halved to improve the model's generalization ability; this will not be elaborated upon here.

[0042] It should be noted that the feature concatenation operation adopts a direct vector dimension concatenation method. Specifically, for electrical and mechanical hidden feature vectors at the same time, the first and last parts are concatenated along the feature dimension to form a new fused feature vector. For example, if the dimension of the electrical hidden feature vector is 64 and the dimension of the mechanical hidden feature vector is 32, the dimension of the concatenated fused feature vector will be 96.

[0043] It should be noted that the specific implementation logic of the category probability transformation function is to perform exponential operations on the original values ​​of multiple dimensions of the network output, and then divide them by the sum of the exponents of all dimension values, so that the output value of each dimension is between 0 and 1, and the sum of the output values ​​of all dimensions is 1. Each dimension's output value corresponds to a probability of a classification label. The specific categories of classification labels are set to four: normal operating condition, insufficient clearance-dominated anomaly, mechanism delay-dominated anomaly, and local ablation-dominated anomaly. These four labels cover all common arc-extinguishing anomaly types under the operation of parallel reactor interruption. In addition, the sample training method adopts a supervised training approach. The training dataset consists of more than 1000 sets of interruption data collected from offline tests of the same type of circuit breaker. Each set of data is labeled with a corresponding classification label. The cross-entropy loss function is used as the training loss function, and the Adam optimizer is used for model optimization. The training batch size is set to 32, and the training epochs are set to 100. Training is stopped early when the validation set loss no longer decreases for 10 consecutive epochs, which will not be elaborated here.

[0044] It should be noted that the preset mapping relationship is constructed using a fixed mapping table with a one-to-one correspondence between status labels and coefficient group numbers. The specific construction and calibration process consists of three steps. The first step is to determine the number of coefficient groups corresponding to each category label. There are four categories of labels, each with four independent coefficient groups. Each coefficient group contains the corresponding geometric constraint gain coefficient, charge mapping coefficient, and energy mapping coefficient. The second step is to conduct specific offline calibration tests for each type of anomaly. For example, for anomalies dominated by insufficient clearance, breaking tests are conducted at different clearance levels to calibrate the optimal coefficient group for this anomaly type. For anomalies dominated by localized ablation, breaking tests are conducted at different ablation degrees to calibrate the optimal coefficient group for this anomaly type. The third step is to establish a one-to-one mapping table, binding each status label to the corresponding calibrated coefficient group number to form a fixed preset mapping relationship. During the diagnostic process, the corresponding coefficient group number can be directly matched based on the status label.

[0045] It should be noted that the first weight coefficient, second weight coefficient, bias constant, classification label matrix, classification bias constant, deficit weight matrix, and deficit bias vector were all obtained through supervised end-to-end training. The training dataset consists of over 2000 sets of interruption data collected from offline tests of the same type of circuit breaker. Each set of data contains complete synchronous acquisition signals and is labeled with the corresponding actual net clearance deficit and classification label. The dataset is divided into training, validation, and test sets in a 7:2:1 ratio. The training process consists of three steps: the first step is data preprocessing, which involves zero-point alignment of all training data, calculation of corresponding intermediate state quantities and baseline features, and formation of standardized training samples. The second step is multi-task joint training, employing a multi-task loss function. The classification task uses the cross-entropy loss function, and the regression task uses the mean squared error loss function. The two loss functions are superimposed with a 1:1 weight. End-to-end training is performed using the Adam optimizer, with a training batch size of 32 and an initial learning rate of 0.001. Every 10 training epochs, the learning rate decays to 0.5 times its original value. The third step is model validation and optimization. After each training round, the model accuracy is verified on the validation set, and the parameters of the model with the highest accuracy on the validation set are saved. The convergence condition is that the classification accuracy on the validation set no longer improves for 10 consecutive rounds, and the mean squared error of the regression task no longer decreases for 10 consecutive rounds. At this point, the model is considered to have converged, and training is stopped. After training, the model's generalization ability is verified on the test set to ensure that the classification accuracy is not lower than 95% and the mean relative error of the regression task does not exceed 5%. This will not be elaborated further here.

[0046] In one embodiment of the present invention, when calculating the critical speed by combining the time window weight and the opening gap growth rate, and dividing the net air gap by the critical speed to obtain the basic compensation, the method includes: Based on the time window weight and the opening distance growth rate, the formula for calculating the critical velocity is as follows: ; in, The critical velocity, It is a very small positive number; Based on the net air gap and critical velocity, the calculation formula for basic compensation is as follows: ; in, When providing basic compensation.

[0047] It should be noted that the critical speed is the average opening distance growth rate after being weighted by the time window, used to characterize the effective separation speed of the contacts during high-risk periods. Basic compensation is obtained by converting the clearance deficit from the length domain to the time domain, and the initial time amount required for the tripping command is calculated in advance.

[0048] It should be noted that the integration interval for critical velocity calculation is fixed from the zero point to the end of the observation window. A minimum threshold needs to be set for the time window weight to filter the effective integration interval; the minimum threshold is preferably set to 0.1. Specifically, only moments with a time window weight greater than or equal to 0.1 are included in the integration calculation; moments with a time window weight less than 0.1 are excluded. When the time window weight is less than 0.1, it represents a low-risk period with minimal impact on arc extinguishing success or failure. Including it in the calculation would lower the true value of the critical velocity, leading to an overestimation of the calibration value. Setting a minimum threshold of 0.1 effectively filters irrelevant data from low-risk periods, ensuring that the critical velocity only reflects the effective opening distance growth rate during high-risk periods. For conditions with extremely high reignition risk, the minimum threshold can be increased to 0.3 to further focus on the highest-risk period.

[0049] Specifically, this invention uses time window weights to perform a weighted average of the opening gap growth rate, focusing only on the opening gap growth rate during high-risk periods. The resulting critical speed is not a common mechanical indicator, but an effective opening gap speed directly related to the arc extinguishing risk, which can characterize the contact's ability to provide effective clearance during high-risk periods.

[0050] In one embodiment of the present invention, the calibration opening distance is synthesized by integrating the basic compensation, the opening distance increase rate, and the opening distance increment; the calibration displacement is calculated by combining the target coefficient, arc energy, and migrated charge; the calibration clearance is obtained by subtracting the calibration displacement from the calibration opening distance; the clearance increment is obtained by subtracting the target clearance from the calibration clearance; and the residual clearance is obtained by subtracting the clearance increment from the clearance deficit. Based on the basic compensation time, the opening distance growth rate, and the opening distance increment, the calculation formula for the synthetic calibration opening distance is as follows: ; in, To calibrate the opening distance; The formula for calculating the calibration displacement is as follows, based on the target perimeter constraint magnification factor, target charge mapping coefficient, migrated charge, target energy mapping coefficient, arc energy, calibration opening distance, and attenuation characteristic length: ; in, For calibrating displacement; The formula for calculating the calibration clearance amount is as follows, based on the calibration opening distance and calibration displacement: ; in, To calibrate the headroom; The formula for calculating the increase in net air volume, based on the calibrated net air volume and the target net air volume, is as follows: ; in, For the increase in net air volume, This is a function operation to extract the maximum value of an interval; The formula for calculating the residual deficit, based on the net air gap and the net air gap increment, is as follows: ; in, This is a residual shortfall. This is a function operation to extract the maximum value.

[0051] It should be noted that the calibration clearance is obtained after time-series translation during basic compensation, and is the trajectory of the calibrated mechanical clearance increment, used to simulate the change process of the mechanical clearance after calibration. The calibration displacement is recalculated based on the calibration clearance, and is the hysteresis displacement of the perimeter constraint back arc after calibration, used to simulate the change of the clearance hysteresis effect after calibration. The calibration clearance amount is obtained by subtracting the calibration displacement from the calibration clearance, and is the trajectory of the calibrated effective insulation clearance, used to simulate the change process of the effective clearance after calibration. The clearance amount increment is the maximum value of the difference between the effective clearance amount before and after calibration, used to characterize the effective clearance improvement brought about by the initial calibration. The residual deficit is the effective clearance amount that has not been replenished after the initial calibration, used for secondary correction of the initial calibration amount.

[0052] It should be noted that the translational reconstruction method for calibrating the opening distance is based on the first-order linearization assumption of a small calibration amount, and its applicable boundaries are divided into two parts. The first part is the value range during basic compensation. The absolute value of the basic compensation should not exceed 5 milliseconds and should not exceed 10% of the total circuit breaker opening time. That is, when the basic compensation time does not exceed 5 milliseconds, the opening distance growth rate at the same moment during the opening process can be approximated as a constant value, and the error introduced by the linearization assumption does not exceed 5%, which will not affect the accuracy of the calibration effect. The second part is the applicable conditions of the linearization assumption. This method is only applicable to the opening distance growth process after the circuit breaker's opening action enters the uniform speed stage, that is, after the acceleration stage following the contact separation, the opening distance growth rate enters the stable uniform speed stage. In other words, the opening distance growth rate in the uniform speed stage remains basically constant, and the linearization assumption of translational reconstruction is fully valid. However, the opening distance growth rate in the acceleration stage changes significantly, and the linearization assumption will introduce a large error. For scenarios with large calibration amounts exceeding 5 milliseconds during basic compensation, a piecewise linearization approach is required for reconstruction. The reconstruction is divided into multiple time periods based on the change in the opening distance growth rate. Within each time period, the corresponding average growth rate is used for translational reconstruction to ensure the accuracy of the reconstruction. This will not be elaborated upon here.

[0053] In one embodiment of the present invention, the residual deficit is divided by the critical speed and the basic compensation is added to generate the final compensation time. The timing of the circuit breaker tripping command is adjusted according to the final compensation time, and the final compensation time, the clearance deficit, and the status label are output, including: Based on the residual deficit, critical velocity, and basic compensation, the calculation formula for the final compensation is as follows: ; in, When it is the final compensation; The final merged output includes the final compensation time, net air gap, and status label.

[0054] It should be noted that the final compensation time is the advance time of the tripping command after residual deficit correction, and it is a key calibration result. The final compensation time output directly affects the tripping phase control system of the circuit breaker, and the specific implementation process consists of four steps: The first step is operating condition identification. The circuit breaker control system collects the main circuit voltage and current signals in real time, identifies the operating condition of the parallel reactor being disconnected, and determines the target time for the tripping operation; the second step is diagnostic calibration. The tripping data of the previous tripping operation under the same operating condition is input into the diagnostic and calibration process of this method to calculate the corresponding final compensation time; the third step is tripping timing adjustment. The control system advances the originally planned tripping command issuance time by the corresponding time length of the final compensation time to generate a new tripping command issuance time; the fourth step is tripping execution. At the new tripping command issuance time, the control system issues a tripping command to the circuit breaker operating mechanism to complete the tripping operation.

[0055] In one embodiment of the present invention, a circuit breaker arc extinguishing characteristic intelligent diagnosis and calibration system includes: a synchronization signal generation module, which acquires the interruption signal, extracts the zero-point time, and shifts the interruption signal to align with the zero-point time to generate a synchronization signal; The feature parameter extraction module extracts the opening gap increment, opening gap speed increase, arc energy, migrated charge, and dynamic resistance based on the synchronization signal. The reference clearance calculation module introduces a reference coefficient, combines the reference coefficient, arc energy and migrated charge to calculate the reference displacement, and subtracts the reference displacement from the opening gap increment to obtain the reference clearance. The standard clearance calculation module integrates the baseline clearance, dynamic resistance and synchronization signal to calculate the status label, matches the target coefficient according to the status label, calculates the target displacement by combining the target coefficient, arc energy and migrated charge, and obtains the target clearance by subtracting the target displacement from the opening gap increment. It also integrates the target displacement and target clearance to calculate the time window weight and clearance deficit. The basic compensation calculation module combines the time window weight and the opening distance growth rate to calculate the critical speed, and divides the net air volume deficit by the critical speed to obtain the basic compensation time. The residual deficit calculation module integrates the basic compensation time, the opening gap growth rate and the opening gap increment to synthesize the calibration opening gap, combines the target coefficient, arc energy and migrated charge to calculate the calibration displacement, and subtracts the calibration displacement from the calibration opening gap to obtain the calibration net volume, subtracts the target net volume from the calibration net volume to obtain the net volume increment, and subtracts the net volume deficit from the net volume increment to obtain the residual deficit. The final compensation time generation module divides the residual deficit by the critical speed and adds it to the basic compensation time to generate the final compensation time. It also adjusts the time of the circuit breaker tripping command based on the final compensation time and outputs the final compensation time, the net clearance deficit, and the status label.

[0056] It should be noted that after obtaining the final compensation time, the circuit breaker control system advances the original scheduled tripping command issuance time by the corresponding time length of the final compensation time, generating a calibrated tripping command issuance time, and then issues the tripping command to the circuit breaker operating mechanism at the calibrated tripping command issuance time. After the tripping is executed, the calibrated interruption signal is collected and used as the data input for the next diagnostic and calibration under the same operating conditions, used to update the status label, clearance deficit, and final compensation time.

[0057] Specifically, the deployment and implementation of this invention is divided into three stages: on-site data acquisition, offline model calibration, and online diagnosis and calibration. In the on-site data acquisition stage, a single high-speed synchronous data acquisition device is used to synchronously acquire all signals. The acquired signals include arc voltage, breaking current, transient recovery voltage, post-arc current, trip coil current, and contact displacement. During acquisition, the sampling frequency is set to 10 MHz, and the synchronous acquisition accuracy of all channels is no less than 100 nanoseconds. To meet the specific requirements of on-site operating conditions, the minimum sampling frequency is no less than 1 MHz, and the minimum synchronization accuracy is no less than 1 microsecond. The acquisition device is installed in the local control cabinet of the offshore wind farm circuit breaker, and the acquired signals are uploaded to the diagnosis and calibration system of the wind farm monitoring system via the industrial ring network of the wind farm.

[0058] The offline model calibration phase was completed using a type test platform with the same type of circuit breaker. The test circuit adopted a 40.5kV parallel reactor breaking circuit consistent with that used in offshore wind power sites, simulating breaking conditions with different arcing energies and different reignition times. Multiple sets of breaking tests were conducted, covering different scenarios such as normal operation, slight reignition, and multiple reignition. Each set of tests was repeated 10 times, and complete signal data and arc extinguishing results of all tests were recorded. Based on the test data, the model parameters were calibrated and the network model was trained. The calibrated model and parameters were then deployed to the diagnostic and calibration system of the site monitoring system.

[0059] During the online diagnostic and calibration phase, the site monitoring system identifies the operating condition of the circuit breaker and parallel reactor being disconnected in real time, retrieves the collected data from the previous trip under the same operating condition, inputs it into the diagnostic and calibration system, completes the entire process calculation, and outputs the corresponding results.

[0060] Taking the disconnection operation of a 40.5kV vacuum circuit breaker parallel reactor in an offshore wind farm as an example, the output of this invention includes three items. The first item is the clearance deficit, which is calculated to be 2 mm in this operation. The second item is the status label, which classifies this operation as a clearance deficiency-dominated anomaly. The third item is the final compensation time, which is calculated to be 1.5 milliseconds in this operation.

[0061] Specifically, the clearance deficit represents the difference between the effective insulation clearance of the circuit breaker's break point that can withstand the recovery voltage during the critical dielectric recovery period after the current crosses zero under the current parallel reactor disconnection condition, and the minimum insulation clearance required to achieve stable arc extinguishing. The unit is millimeters. A larger value indicates a more significant insulation capacity gap in the break point. Final compensation indicates the time interval (milliseconds) required for the circuit breaker's tripping command to be advanced relative to the original timing sequence in order to compensate for the clearance deficit. This value can be directly input into the circuit breaker's tripping phase control system to adjust the timing of the tripping command.

[0062] Specifically, normal operating condition means that during the circuit breaker's opening process, the dielectric recovery process at the break point and the mechanical action sequence both meet design requirements, with no risk of reignition and no need for opening sequence calibration. Insufficient clearance-dominated anomaly indicates that while the circuit breaker's mechanical action sequence is normal, residual conductive dielectric in the break point occupies the insulation gap after the current crosses zero, resulting in insufficient effective insulation clearance to meet arc extinguishing requirements. This is the core cause of the reignition risk. Mechanism delay-dominated anomaly indicates that the circuit breaker's opening mechanism experiences an action delay, causing a lag in the increase of the contact mechanical opening distance during the critical period of current zero crossing. This prevents the effective insulation clearance from being established in time, thus triggering the arc extinguishing risk. Localized ablation-dominated anomaly indicates that the circuit breaker contact surface has experienced localized ablation due to repeated reignitions, leading to distortion of the electric field distribution at the break point, a decrease in dielectric recovery speed, and a significant reduction in effective insulation clearance under the same mechanical opening distance.

[0063] It should be noted that, as Figure 2 As shown, the condition diagnosis distribution diagram illustrates the correspondence between condition diagnosis results and key electrical quantities. The horizontal axis represents arc energy in J, indicating the total energy injected into the break gap from arc establishment to current zero. The vertical axis represents migrated charge in mC, representing the cumulative migration of residual charged particles after current zero. The scatter plot categories represent condition labels, distinguishing between normal operating conditions, insufficient clearance, mechanism delay, and localized ablation. The bubble size represents the clearance deficit in mm; a larger value indicates a greater amount of effective insulation clearance still needed to achieve stable arc extinguishing. This diagram essentially reflects the combined output of condition labels and clearance deficit after the fusion of electrical and mechanical branches in this invention.

[0064] It should be noted that, as Figure 3The diagram shows a confusion matrix, used to illustrate the consistency and misclassification distribution of the status label recognition results. The horizontal axis represents the predicted label, and the vertical axis represents the true label. The label categories are normal operating condition, insufficient clearance-dominated anomaly, mechanism delay-dominated anomaly, and local ablation-dominated anomaly. The first row of each cell represents the number of samples, and the second row represents the proportion of the true label in that row. The larger the value on the main diagonal, the more accurate the corresponding category recognition. The off-diagonal values ​​represent the number of misclassifications. Based on 100 sample data points, the overall accuracy rate is 98.0%. Among them, two local ablation-dominated anomalies were identified as insufficient clearance-dominated anomalies, indicating that the two have similar responses in some feature intervals.

[0065] It should be noted that, as Figure 4 The diagram shown illustrates the changes in clearance deficit before and after calibration, as well as the output results during final compensation. The horizontal axis represents samples numbered 1 to 15, indicating different operating conditions or different interruption cycles. The left vertical axis represents the deficit in mm, where the deficit before calibration represents the clearance deficit output in the formal diagnostics, and the residual deficit represents the clearance that was not fully compensated after basic compensation. The right vertical axis represents the final compensation time in milliseconds, indicating the amount of time the control system should advance the tripping command. If the residual deficit is significantly smaller than the deficit before calibration, it indicates that the calibrated opening distance and calibrated clearance have effectively improved the insulation capacity of the break. This diagram corresponds to the critical speed calculation, basic compensation determination, calibration opening distance reconstruction, residual deficit correction, and final compensation output processes in this invention.

[0066] It should be noted that the interval and threshold sizes are set for ease of comparison. The threshold size depends on the amount of sample data and the base number set by those skilled in the art for each group of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, derived from software simulations using a large amount of collected data to obtain the most recent realistic results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. The embodiments of this example have been described above, but this example is not limited to the specific implementation described above. The specific implementation described above is merely illustrative and not restrictive. Those skilled in the art can make many variations based on the guidance of this example, all of which fall within the protection scope of this example.

Claims

1. A circuit breaker arc quenching characteristic intelligent diagnosis and calibration method, characterized in that, Includes the following steps: Step S1: Acquire the switching signal, extract the zero-point time, and shift the switching signal to align with the zero-point time to generate a synchronization signal; Step S2: Extract the opening gap increment, opening gap speed increase, arc energy, transferred charge, and dynamic resistance based on the synchronization signal; Step S3: Introduce the reference coefficient, calculate the reference displacement by combining the reference coefficient, arc energy and transferred charge, and subtract the reference displacement from the opening gap increment to obtain the reference clearance. Step S4: Calculate the status label by integrating the baseline clearance, dynamic resistance and synchronization signal; match the target coefficient according to the status label; calculate the target displacement by combining the target coefficient, arc energy and migrated charge; subtract the target displacement from the opening gap increment to obtain the target clearance; and calculate the time window weight and clearance deficit by integrating the target displacement and target clearance. Step S5: Calculate the critical speed by combining the time window weight and the opening gap growth rate, and divide the net air gap by the critical speed to obtain the basic compensation time. Step S6: Integrate the basic compensation time, the opening distance growth rate and the opening distance increment to synthesize the calibration opening distance, combine the target coefficient, the arc energy and the transferred charge to calculate the calibration displacement, and subtract the calibration displacement from the calibration opening distance to obtain the calibration net amount, subtract the target net amount from the calibration net amount to obtain the net amount increment, and subtract the net amount increment from the net amount deficit to obtain the residual deficit. Step S7: Divide the residual deficit by the critical speed and add the basic compensation time to generate the final compensation time. Adjust the circuit breaker tripping command issuance time according to the final compensation time, and output the final compensation time, net clearance deficit and status label.

2. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, The arc voltage, breaking current, transient recovery voltage, post-arc current, trip coil current, and contact displacement are collected and combined to generate a breaking signal. Within the time interval from the issuance of the tripping command to the end of the observation window, locate the moment corresponding to the minimum absolute value of the interrupting current and determine it as the zero point time. Using the zero point time as a reference, the switching signal is time-shifted and aligned to generate a synchronization signal.

3. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, Extract the contact displacement from the synchronization signal, subtract the contact displacement corresponding to the zero point time from the contact displacement to generate the opening gap increment, and use the derivative of the opening gap increment with respect to time to generate the opening gap speed increase. The arc voltage and breaking current are extracted from the synchronization signal. The absolute value of the product of the arc voltage and breaking current is integrated over time over the interval from the moment the arc is established to the zero point to generate arc energy. Extract the back-arc current from the synchronization signal, and integrate the absolute value of the back-arc current over the interval from zero time to the current time to generate migration charge; The dynamic resistance is generated by dividing the arc voltage by the sum of the breaking current and a very small positive number.

4. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, Determine the reference coefficients, which include the reference geometric constraint gain coefficient, the reference charge mapping coefficient, and the reference energy mapping coefficient; First, calculate the square of the equivalent evacuation channel perimeter length, divide it by the product of the equivalent evacuation channel area and π (pi), subtract one from the result, multiply it by the reference geometric constraint gain coefficient, and finally add one to obtain the reference perimeter constraint amplification factor. Divide the opening gap increment by the attenuation characteristic length and take the opposite number. Substitute the result into the natural exponential function. Multiply the result by the reference energy mapping coefficient and the arc energy in turn. Then multiply the reference charge mapping coefficient and the migration charge. Add the two products together and multiply the sum by the reference perimeter constraint amplification factor to generate the reference displacement. Subtract the reference displacement from the opening gap increment to generate the reference clearance.

5. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, The arc voltage, breaking current, transient recovery voltage, post-arc current and dynamic resistance are combined and input into the electrical branch network operation function to extract electrical hidden features. The trip coil current, opening gap increment, opening gap growth rate and reference clearance are combined and input into the mechanical branch network operation function to extract the reference mechanical hidden features; Multiply the baseline displacement by the first weighting coefficient, add the baseline net clearance by the second weighting coefficient, add the bias constant, and substitute into the activation function to generate the pre-diagnosis time window weights. The electrical hidden features and the baseline mechanical hidden features are concatenated and the result is multiplied by the pre-diagnosis time window weight. The result is accumulated within the interval from the zero point time to the end of the observation window to generate the pre-diagnosis fusion feature. Multiply the pre-diagnostic fusion features by the classification label matrix, add the classification bias constant, substitute into the category probability transformation function, take the category corresponding to the highest probability, and generate the state label. Input the status label into the preset mapping relationship to obtain the coefficient number.

6. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 5, characterized in that, First, calculate the square of the equivalent evacuation channel perimeter length, divide it by the product of the equivalent evacuation channel area and π four times, subtract one from the result, multiply it by the target geometric constraint gain coefficient corresponding to the coefficient number, and finally add one to obtain the target perimeter constraint amplification factor. Divide the opening distance increment by the attenuation characteristic length and take the opposite number. Substitute the result into the natural exponential function. Multiply the result by the target energy mapping coefficient corresponding to the coefficient number and the arc energy. Then multiply the target charge mapping coefficient corresponding to the coefficient number and the migration charge. Add the two products together and multiply the sum by the target perimeter constraint amplification factor to generate the target displacement. Subtract the target displacement from the opening gap increment to generate the target clearance. The trip coil current, opening gap increment, opening gap speed increase and target clearance are combined and input into the mechanical branch network operation function to extract the target mechanical hidden features; Multiply the target displacement by the first weight coefficient, add the target net volume by the second weight coefficient, add the bias constant, and substitute into the activation function to generate the time window weight. The electrical concealment features and the target mechanical concealment features are concatenated, and the result is multiplied by the time window weight. The result is accumulated within the interval from the zero point to the end of the observation window to generate the formal fused features. The net air gap is generated by multiplying the formal fusion features by the gap weight matrix and adding the gap bias vector.

7. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, Multiply the time window weight by the opening distance growth rate, and integrate over time over the interval from the zero point to the end of the observation window to obtain the velocity numerator. The weight of the time window is integrated over time from the zero point to the end of the observation window, and then a very small positive number is added to obtain the denominator of the weight. Divide the numerator of the velocity by the denominator of the weight to generate the critical velocity. Add a very small positive number to the critical speed to get the division denominator, then divide the net air gap by the division denominator to generate the basic compensation.

8. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 5, characterized in that, When multiplying the opening gap growth rate by the basic compensation, add the opening gap increment to synthesize the calibrated opening gap; Divide the calibration opening distance by the attenuation characteristic length and take the opposite number. Substitute the result into the natural exponential function. Multiply the result by the target energy mapping coefficient corresponding to the coefficient number and the arc energy in turn. Then multiply the target charge mapping coefficient corresponding to the coefficient number and the migration charge. Add the two products together and multiply the sum by the target perimeter constraint amplification factor to generate the calibration displacement. Subtract the calibration displacement from the calibration clearance to generate the calibration net clearance. Subtract the target net air volume from the calibrated net air volume, extract the maximum difference within the interval from the zero point time to the end of the observation window, and generate the net air volume increment. Subtract the net air volume increment from the net air volume deficit to obtain the intermediate calculation result. Extract the zero value and the maximum value in the intermediate calculation result to generate the residual deficit.

9. The intelligent diagnosis and calibration method for the arc-extinguishing characteristics of a circuit breaker according to claim 1, characterized in that, Add a very small positive number to the critical speed to obtain the correction divisor term, divide the residual deficit by the correction divisor term, and then add the result to the basic compensation to generate the final compensation term. When merging the final compensation output, the net air volume deficit and status label are included.

10. A system for intelligent diagnosis and calibration of the arc-extinguishing characteristics of a circuit breaker, characterized in that, The method for intelligent diagnosis and calibration of the arc-extinguishing characteristics of a circuit breaker as described in any one of claims 1 to 9 includes: The synchronization signal generation module acquires the on / off signal, extracts the zero-point time, and shifts the on / off signal to align with the zero-point time to generate the synchronization signal. The feature parameter extraction module extracts the opening gap increment, opening gap speed increase, arc energy, migrated charge, and dynamic resistance based on the synchronization signal. The reference clearance calculation module introduces a reference coefficient, combines the reference coefficient, arc energy and migrated charge to calculate the reference displacement, and subtracts the reference displacement from the opening gap increment to obtain the reference clearance. The standard clearance calculation module integrates the baseline clearance, dynamic resistance and synchronization signal to calculate the status label, matches the target coefficient according to the status label, calculates the target displacement by combining the target coefficient, arc energy and migrated charge, and obtains the target clearance by subtracting the target displacement from the opening gap increment. It also integrates the target displacement and target clearance to calculate the time window weight and clearance deficit. The basic compensation calculation module combines the time window weight and the opening distance growth rate to calculate the critical speed, and divides the net air volume deficit by the critical speed to obtain the basic compensation time. The residual deficit calculation module integrates the basic compensation time, the opening gap growth rate and the opening gap increment to synthesize the calibration opening gap, combines the target coefficient, arc energy and migrated charge to calculate the calibration displacement, and subtracts the calibration displacement from the calibration opening gap to obtain the calibration net volume, subtracts the target net volume from the calibration net volume to obtain the net volume increment, and subtracts the net volume deficit from the net volume increment to obtain the residual deficit. The final compensation time generation module divides the residual deficit by the critical speed and adds it to the basic compensation time to generate the final compensation time. It also adjusts the time of the circuit breaker tripping command based on the final compensation time and outputs the final compensation time, the net clearance deficit, and the status label.

Images