A primary and secondary fused pole-mounted circuit breaker state collaborative analysis method

Abstract

The application provides a primary and secondary fusion column circuit breaker state collaborative analysis method, comprising: identifying the arc extinguishing current waveform symmetry type by analyzing the opening arc duration and pulse direction information data in the initial collected data set, the arc extinguishing current waveform symmetry type including an asymmetric type; if the arc extinguishing current waveform symmetry type is in the asymmetric type, evaluating the residual charge polarity neutralization state by analyzing the positive and negative transient electromagnetic pulse alternation law through the pulse direction information data; fusing the residual charge polarity neutralization state and the contact cumulative wear depth, using the random forest algorithm to predict whether the integral baseline zero position is abnormally drifted, and obtaining the integral baseline zero position abnormal drift prediction result; if the adjusted discrimination threshold exceeds the preset allowed range, identifying the actual contact wear pattern by jointly analyzing the cumulative wear depth and the opening arc duration.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for collaborative analysis of the status of pole-mounted circuit breakers that integrates primary and secondary circuit breaker functions. Background Technology

[0002] In the field of power system operation and maintenance, the condition monitoring and assessment of circuit breakers are of paramount importance to ensuring the safety and stability of the power grid. As a core component of power equipment, the performance of circuit breakers directly affects the reliability and security of power transmission. Especially in the context of integrated primary and secondary circuit breaker technology, how to achieve accurate perception and collaborative analysis of equipment status has become a key issue of concern in the industry. However, current methods for monitoring the condition of circuit breakers struggle to fully capture the interactive effects of various factors during equipment operation, particularly lacking in-depth comprehensive consideration of the complex relationship between contact wear and current measurement deviation.

[0003] In the prior art, an intelligent primary and secondary integrated pole-mounted circuit breaker, publication number CN212008850U, discloses a technology that integrates a sensor group, a sensor aggregation unit, and a comprehensive evaluation model to achieve real-time monitoring and self-sensing and self-diagnosis of the pole-mounted circuit breaker's operating status. However, this technology fails to fully consider the irregular interaction between contact wear accumulation and changes in arc characteristics, leading to current measurement baseline drift. Particularly when alternating positive and negative pulses cause charge cancellation, it is difficult to dynamically adjust the judgment criteria, resulting in inaccurate status assessment results. Contact wear affects the circuit breaker's breaking capacity, and the accuracy of current measurement is closely related to equipment status judgment. The dynamic changes between these two are often overlooked, leading to inaccurate status assessment results. Furthermore, accumulated contact wear alters the arc characteristics during circuit breaker breaking, thus affecting the current signal measurement baseline. However, this effect is not a simple linear accumulation but is disturbed by the alternating positive and negative pulses during the arc process.

[0004] When positive and negative charges cancel each other out in the measurement circuit, the baseline drift amplitude may exhibit an irregular trend, varying with the degree of contact wear. This irregular trend makes traditional condition assessment methods difficult to accurately reflect the true condition of the equipment. Specifically, in actual business scenarios, when a circuit breaker is in tripping operation, contact wear may lead to a prolonged arc duration, thus altering the zero-crossing point judgment time of the current signal. If the judgment criteria are not dynamically adjusted according to the symmetry of the arc, the degree of contact wear may be misjudged. For example, under certain operating conditions, the asymmetry of the positive and negative arc pulses can cause abnormal fluctuations in the measurement baseline, thereby affecting the assessment of the contact condition; this situation is particularly common during high-load operation.

[0005] Therefore, how to accurately identify the actual state of the contact and achieve coordinated state assessment amidst the complex interaction between contact wear and current measurement baseline drift has become a critical issue that urgently needs to be addressed. Summary of the Invention

[0006] This invention provides a method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers, the method comprising: The initial dataset is obtained by collecting data on the cumulative wear depth of the contacts, the duration of the opening arc, and the direction of the arc-extinguishing transient pulse through the vacuum interrupter wear monitoring unit and the Rogowski coil integral baseline acquisition unit. By analyzing the arc duration and pulse direction information in the initial data acquisition dataset, the symmetry type of the arc extinction current waveform is identified, which includes asymmetry type. If the symmetric type of the arc-extinguishing current waveform is in an asymmetric type, the alternation pattern of positive and negative transient electromagnetic pulses is analyzed by pulse direction information data to assess the polarity neutralization state of the residual charge. By fusing the residual charge polarity neutralization state with the cumulative wear depth of the contact, the random forest algorithm is used to predict whether there is abnormal drift of the integral baseline zero position, and the prediction result of abnormal drift of the integral baseline zero position is obtained. Based on the prediction results of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc extinction current waveform, the threshold adjustment direction and the adaptation section are determined. Based on the threshold adjustment direction and the adaptation section, the current zero-crossing point discrimination threshold is dynamically adjusted, and the adjusted discrimination threshold is determined. If the adjusted threshold exceeds the preset allowable range, the actual wear pattern of the contact will be identified by combining the cumulative wear depth and the duration of the opening arc. By fusing the actual wear pattern of the contacts with the prediction results of the abnormal drift of the integral baseline zero position, a neural network algorithm is used to generate a primary and secondary state collaborative assessment, and the state collaborative analysis results of the pole-mounted circuit breaker are obtained.

[0007] Furthermore, the initial dataset is obtained by acquiring data on the cumulative wear depth of the contacts, the duration of the opening arc, and the direction of the arc-extinguishing transient pulse through the vacuum interrupter wear monitoring unit and the Rogowski coil integral baseline acquisition unit, including: The cumulative wear depth of the contact is obtained by comparing the difference in stroke before and after the contact movement with the initial thickness of the contact, which is collected by a displacement sensor. The duration of the opening arc is determined by capturing the time span of the start and end times of the opening arc discharge signal using a current transformer. Record the order of appearance and amplitude direction of each pulse in the positive and negative polarity sequence at the moment of arc extinction, obtain the direction information data of the transient pulse at the moment of arc extinction, and aggregate and integrate them to obtain the initial acquisition dataset.

[0008] Furthermore, by analyzing the arc duration and pulse direction information data in the initial acquired dataset, the symmetry type of the arc extinction current waveform is identified. The symmetry type of the arc extinction current waveform includes asymmetric types, including: Extract the duration of the shunt arc and the pulse direction information from the initial acquisition dataset. Based on the amplitude distribution of the positive and negative pulses, calibrate the peak position and zero-crossing time of each pulse to obtain the positive half-wave period and negative half-wave period of the arc extinguishing current waveform. The waveform deviation index is obtained based on the ratio of the difference between the positive and negative half-wave durations to the duration of the opening arc. If the waveform offset index exceeds the preset symmetry judgment threshold, the arc extinction current waveform is determined to be of an asymmetrical type.

[0009] Furthermore, if the arc-extinguishing current waveform is asymmetrical, then the alternation pattern of positive and negative transient electromagnetic pulses is analyzed through pulse direction information data to assess the neutralization state of the residual charge polarity, including: Extract the alternating occurrence nodes of positive and negative transient electromagnetic pulses from the pulse direction information data, mark the timing positions of adjacent positive and negative polarity pulses, and obtain the polarity peak value of each pulse and the interval between adjacent pulses. The amplitude difference between adjacent positive and negative pulse peaks is calculated based on the polarity peak value and the interval duration, and the ratio is calculated with the interval duration to obtain the reverse cancellation strength. If the reverse cancellation strength exceeds the preset neutralization threshold, the residual charge polarity neutralization state is determined to be fully neutralized.

[0010] Furthermore, the step of calculating the reverse cancellation strength by ratio with the interval duration includes: The alternating occurrence nodes of positive transient electromagnetic pulses and negative transient electromagnetic pulses are retrieved from the pulse direction information data. Adjacent positive pulses and negative pulses are paired according to the arrangement order of the alternating occurrence nodes in the time series, and the polarity peak value of each paired pulse group and the interval duration between adjacent pulses are obtained. The amplitude difference between the peak value of the positive pulse and the peak value of the negative pulse in each pair of pulse groups is calculated based on the polarity peak value of each pair of pulse groups, and the ratio is calculated with the interval duration to obtain the reverse cancellation strength of each pair of pulse groups. The polarity offset integrity index is obtained by statistically analyzing the ratio of the number of paired pulse groups whose reverse offset strength exceeds the preset offset threshold to the total number of all paired pulse groups. If the polarity offset integrity index exceeds the preset neutralization judgment threshold, the residual charge polarity neutralization state is determined to be a fully neutralized state; if it is below the preset neutralization judgment threshold, it is determined to be an insufficiently neutralized state.

[0011] Furthermore, by fusing the residual charge polarity neutralization state with the cumulative wear depth of the contact, and using the random forest algorithm to predict whether there is abnormal drift in the integral baseline zero position, the prediction result of abnormal drift in the integral baseline zero position is obtained, including: The residual charge polarity neutralization state is converted into a numerical feature, with sufficient neutralization assigned a high neutralization numerical label and insufficient neutralization assigned a low neutralization numerical label, which is combined with the contact cumulative wear depth to form a fused feature set; A drift prediction classifier is constructed using the random forest algorithm. The drift prediction classifier consists of several independent decision trees that form a decision path based on the wear depth distribution range and the neutral state category. The independent decision categories are statistically analyzed to obtain the integral baseline zero-position abnormal drift prediction result.

[0012] Furthermore, the step of determining the threshold adjustment direction and adaptation segment based on the prediction results of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc-extinguishing current waveform, dynamically adjusting the current zero-crossing discrimination threshold based on the threshold adjustment direction and adaptation segment, and determining the adjusted discrimination threshold includes: If the prediction result of the zero-position abnormal drift of the integral baseline is that there is abnormal drift and the symmetry type of the arc extinguishing current waveform is asymmetric, then the threshold adjustment direction is determined to be offset compensation of the corresponding polarity based on the polarity of the longer duration of the positive half-wave period and the negative half-wave period. If the drift prediction result is that there is no abnormal drift, then the threshold adjustment direction is determined to be to keep the original value. The appropriate segment for current zero-crossing point discrimination is determined based on the difference between the threshold adjustment direction and the duration of the positive and negative half-waves. An offset compensation amount is superimposed on a preset current zero-crossing point discrimination benchmark threshold. The offset compensation amount is determined by the product of the duration difference and the preset compensation mapping coefficient. The polarity of the offset compensation amount is consistent with the threshold adjustment direction, thus obtaining the adjusted discrimination threshold.

[0013] Furthermore, if the adjusted threshold exceeds a preset allowable range, the actual wear pattern of the contacts is identified through joint analysis of the cumulative wear depth and the duration of the opening arc, including: The wear arc correlation feature is obtained by comparing the cumulative wear depth of the contact with the duration of the opening arc. The wear arc correlation feature includes a combination category of wear depth level and arc time level. If the wear depth level and the arc time level are within the preset matching range, it is determined to be a uniform wear pattern; otherwise, it is determined to be a non-uniform wear pattern. Non-uniform wear morphology identifies the wear distribution area of ​​local ablation and the corresponding spalling depth based on whether the wear depth level is higher or lower than the arc time level.

[0014] Furthermore, by fusing the actual wear morphology of the contacts with the predicted abnormal drift of the integral baseline zero point, a neural network algorithm is used to generate a primary and secondary state collaborative assessment, resulting in the state collaborative analysis results of the pole-mounted circuit breaker, including: The wear distribution area and spalling depth in the actual wear morphology of the contact are converted into wear morphology feature vectors, and the prediction result of the zero-position abnormal drift of the integral baseline is converted into a drift state feature vector. The first and second fusion feature vectors are spliced ​​and fused together. A state collaborative assessment network is constructed using a neural network algorithm, which outputs three classification nodes: the remaining life status of the corresponding contacts, the risk level of zero drift of the integral baseline, and maintenance and disposal suggestions. These are then integrated to form the state collaborative analysis results of the pole-mounted circuit breaker.

[0015] Furthermore, the integrated results of the pole-mounted circuit breaker status collaborative analysis include: Extract the wear distribution area and spalling depth from the actual wear morphology of the contact, and the drift tendency level from the zero-position abnormal drift prediction result of the integral baseline, and analyze the remaining life range of the contact's opening and closing operations. If the drift state is abnormal and the lifetime range is short, the interference level is determined to be high interference; if the drift state is not abnormal and the lifetime range is long, the interference level is determined to be low interference; the other combinations are determined to be medium interference. Long cycles correspond to good conditions, medium cycles to average conditions, and short cycles to poor conditions. High interference corresponds to high risk, medium interference corresponds to medium risk, and low interference corresponds to low risk. Based on the combination of remaining life status and drift risk level, maintenance and disposal recommendations are matched and integrated to obtain the results of collaborative analysis of pole-mounted circuit breaker status.

[0016] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a collaborative analysis method for the status of pole-mounted circuit breakers, integrating primary and secondary data. Addressing the risks to protection operation caused by contact wear, zero-point drift of the integral baseline, and asymmetry in the arc-extinguishing current waveform during operation, this method collects information on cumulative contact wear depth, arc duration, and pulse direction. It analyzes the symmetry type of the arc-extinguishing current waveform and the neutralization state of residual charge polarity. A random forest algorithm is integrated to predict abnormal zero-point drift, and the current zero-crossing threshold is dynamically adjusted. Simultaneously, a neural network algorithm is used to generate collaborative analysis results including remaining contact life, drift risk level, and maintenance recommendations. This invention, through deep fusion of primary and secondary status data and the application of intelligent algorithms, accurately assesses the operating status of circuit breakers, effectively reducing the risk of protection maloperation or failure to operate, and improving the reliability of power grid operation. Attached Figure Description

[0017] Figure 1This is a flowchart of a method for coordinated analysis of the status of a primary and secondary integrated pole-mounted circuit breaker according to the present invention.

[0018] Figure 2 This is a schematic diagram of a primary and secondary integrated pole-mounted circuit breaker state collaborative analysis method according to the present invention.

[0019] Figure 3 This is another schematic diagram of the primary and secondary fusion pole-mounted circuit breaker state collaborative analysis method of the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0021] like Figures 1 to 3 This embodiment of the method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers may specifically include: S101. The vacuum interrupter wear monitoring unit and the Rogowski coil integral baseline acquisition unit collect data on the cumulative wear depth of the contacts, the duration of the opening arc, and the direction of the arc-extinguishing transient pulse to obtain the initial acquisition dataset.

[0022] The displacement sensor in the vacuum interrupter wear monitoring unit collects the stroke difference before and after contact movement. Based on the comparison between the stroke difference and the initial contact thickness, the cumulative wear depth of the contact is obtained. The current transformer in the Rogowski coil integral baseline acquisition unit captures the start and end times of the arc discharge signal during the opening operation. The duration of the opening arc is determined based on the time span of these start and end times. The Rogowski coil integral baseline acquisition unit records the positive and negative polarity sequence of transient electromagnetic pulses at the moment of arc extinguishing. Based on the order of appearance and amplitude direction of each pulse in the positive and negative polarity sequence, the direction information data of the arc-extinguishing transient pulses is obtained. The cumulative wear depth of the contact, the duration of the opening arc, and the direction information data of the arc-extinguishing transient pulses are aggregated and integrated to obtain the initial acquisition dataset.

[0023] In one embodiment, the vacuum interrupter wear monitoring unit has a built-in displacement sensor installed between the fixed end and the movable end of the contact actuation mechanism. When the circuit breaker performs an opening or closing action, the displacement sensor records the travel value of the contact from the initial position to the final position in real time. After each action, the travel value is compared with the initial travel value, and the difference between the two is calculated as the wear amount generated by this action. The cumulative wear depth of the contact is obtained by accumulating the wear amounts of each action. The specific calculation process is as follows: the initial travel value S0 is recorded when the circuit breaker closes for the first time, and the current travel value S is recorded after each action. The wear amount d is calculated as d = S0 - S, where S0 is the initial travel value, S is the current travel value, and d is the wear amount of this action. For example, if S0 = 10.0 mm and S = 9.8 mm, then d = 0.2 mm. Then, all d values ​​are accumulated to output the cumulative wear depth.

[0024] Specifically, the Rogowski coil integral baseline acquisition unit uses a Rogowski coil wrapped around the main circuit conductor of the circuit breaker as a current sensing element. When the tripping action triggers arc discharge, the Rogowski coil senses the change in the magnetic field generated by the arc current and outputs a corresponding induced voltage signal. The induced voltage signal is converted into an output signal proportional to the arc current through an integral circuit. The complete waveform of the arc discharge signal from the start time to the extinguishing time is recorded, and the duration of the tripping arc is determined based on the time span between the start and end times.

[0025] In one possible implementation, the Rogowski coil integrating baseline acquisition unit performs high-speed sampling of the transient electromagnetic pulse at the moment of arc extinction, identifies the amplitude direction of each pulse as positive or negative, and arranges them in chronological order to form a positive and negative polarity sequence. Based on the order of occurrence of positive and negative pulses in the positive and negative polarity sequence and the directional characteristics of the amplitude of each pulse, the direction information data of the transient pulse during arc extinction is extracted.

[0026] For example, the cumulative wear depth of the contact, the duration of the opening arc, and the direction information of the arc extinguishing transient pulse are aligned and matched according to a unified timestamp identifier, and then aggregated to form an initial collection dataset containing multi-dimensional state information.

[0027] S102. By analyzing the arc duration and pulse direction information data in the initial acquisition dataset, the symmetry type of the arc extinction current waveform is identified. The symmetry type of the arc extinction current waveform includes the asymmetry type.

[0028] The duration of the arc tripping arc and the direction of the transient pulses during arc extinction are extracted from the initial dataset. Based on the amplitude distribution of the positive and negative pulses in the transient pulse direction information, the peak position and zero-crossing time of each pulse are calibrated, and the positive and negative half-wave periods of the arc extinction current waveform are obtained. The duration difference between the positive and negative half-wave periods is calculated based on their time spans. Combined with the ratio of the arc tripping arc duration to the duration difference, a waveform offset index is obtained. If the waveform offset index is lower than a preset symmetry threshold, the arc extinction current waveform is determined to be symmetrical; if the waveform offset index exceeds the preset symmetry threshold, the arc extinction current waveform is determined to be asymmetrical.

[0029] In one embodiment, the duration of the opening arc and the direction information of the arc-extinguishing transient pulse are read from the initial acquisition dataset. The duration of the opening arc reflects the complete time span from the generation to the extinguishing of the arc after the circuit breaker contacts separate. The direction information of the arc-extinguishing transient pulse includes the polarity direction and timing arrangement of each pulse in the current waveform at the moment of arc extinguishing.

[0030] Specifically, based on the amplitude distribution characteristics of positive and negative pulses in the transient pulse direction information data of arc extinction, the peak position of the arc extinction current waveform is calibrated and the zero-crossing moment is identified. The zero-crossing moment refers to the time point when the current value crosses zero when the current waveform transitions from the positive half-cycle to the negative half-cycle or from the negative half-cycle to the positive half-cycle. By detecting the instant when the current value drops or rises to zero between the peak values ​​of adjacent positive and negative pulses, the time coordinates of each zero-crossing moment are obtained, and the positive and negative half-wave intervals of the arc extinction current waveform are divided accordingly. The time spans of the positive and negative half-wave periods are measured respectively.

[0031] In one possible implementation, the time span of the positive half-wave period is subtracted from the time span of the negative half-wave period to obtain the duration difference between the positive and negative half-waves. This duration difference reflects the degree of uniformity in the time distribution of the arc-extinguishing current waveform in the positive and negative directions. When the positive and negative half-wave periods are similar, the duration difference approaches zero, indicating that the waveform has high time symmetry. Further, the duration difference is compared with the duration of the opening arc to obtain a waveform offset index. The waveform offset index characterizes the proportion of the duration difference relative to the overall arc duration. This index eliminates the influence of differences in the absolute value of the arc duration under different opening conditions, making the waveform symmetry comparable under different operating conditions.

[0032] For example, the waveform offset index is compared with a preset symmetry judgment threshold. If the waveform offset index is lower than the symmetry judgment threshold, the arc extinguishing current waveform is determined to be symmetrical, indicating that the positive and negative half-waves are basically balanced in time distribution. If the waveform offset index exceeds the symmetry judgment threshold, the arc extinguishing current waveform is determined to be asymmetrical, indicating that there is a significant deviation in the duration of the positive and negative pulses during the arc extinguishing process.

[0033] S103. If the arc-extinguishing current waveform is in an asymmetrical type instead of a symmetrical type, the alternation pattern of positive and negative transient electromagnetic pulses is analyzed by pulse direction information data to assess the polarity neutralization state of the residual charge.

[0034] If the symmetrical type of the arc-extinguishing current waveform is asymmetrical, then the alternating occurrence nodes of positive and negative transient electromagnetic pulses are extracted from the arc-extinguishing transient pulse direction information data. The timing positions of adjacent positive and negative polarity pulses are calibrated based on these alternating occurrence nodes, and the polarity peak value of each pulse and the interval between adjacent pulses are obtained. Based on the polarity peak value and the interval, the amplitude difference between the peak values ​​of adjacent positive and negative pulses is calculated, and the ratio of the amplitude difference to the interval is calculated to obtain the reverse cancellation strength between the positive and negative pulses. The reverse cancellation strength is compared with a preset neutralization threshold. If the reverse cancellation strength exceeds the preset neutralization threshold, the residual charge polarity neutralization state is determined to be sufficiently neutralized; if the reverse cancellation strength is lower than the preset neutralization threshold, the residual charge polarity neutralization state is determined to be insufficiently neutralized.

[0035] In one embodiment, when the symmetry type of the arc-extinguishing current waveform is determined to be asymmetric, in-depth processing of the alternation pattern of positive and negative transient electromagnetic pulses is triggered. The alternation points of positive and negative transient electromagnetic pulses on the time axis are identified from the arc-extinguishing transient pulse direction information data. These alternation points refer to the time inflection points between the end of an adjacent positive pulse and the beginning of a negative pulse, or between the end of a negative pulse and the beginning of a positive pulse.

[0036] Specifically, based on the distribution of the alternating occurrence nodes in the time series, the temporal position relationship of each pair of adjacent positive and negative polarity pulses is determined. For each positive pulse and its adjacent negative pulse, the amplitude value at the peak value of each pulse is extracted as the polarity peak value, and the time span between the peak values ​​of two adjacent pulses is measured as the interval duration.

[0037] In one possible implementation, the calculation process for the reverse cancellation strength is as follows: The absolute values ​​of adjacent positive and negative pulse peaks are taken, and the difference is calculated to obtain an amplitude difference. This amplitude difference reflects the degree of imbalance in energy release between the positive and negative pulses. The amplitude difference is then divided by the interval duration to obtain the reverse cancellation strength. When the peak amplitudes of the positive and negative pulses are close and the interval duration is short, the reverse cancellation strength is large, indicating that the positive and negative charges have completed relatively sufficient mutual cancellation in a short time. When the peak amplitude difference is significant or the interval duration is long, the reverse cancellation strength is small, indicating that the charge cancellation process is delayed or incomplete.

[0038] It should be noted that the residual charge polarity neutralization state reflects the degree to which positive and negative charges cancel each other out in the measurement circuit at the moment of arc extinction. The reverse cancellation strength is compared with a preset neutralization judgment threshold. If the reverse cancellation strength exceeds the neutralization judgment threshold, it indicates that the energy of the positive and negative pulses has achieved relatively balanced mutual cancellation at the moment of arc extinction, and the residual charge polarity neutralization state is determined to be fully neutralized.

[0039] For example, if the reverse cancellation strength is lower than the preset neutralization determination threshold, it indicates that there is a single polarity charge accumulation residue during the arc extinguishing process, and the positive and negative pulses fail to effectively cancel each other out. The residual charge polarity neutralization state is determined to be insufficient neutralization. In this state, the integral baseline of the measurement circuit is at risk of drifting due to charge accumulation.

[0040] By retrieving the alternating occurrence nodes, polarity peaks, and interval durations of positive and negative transient electromagnetic pulses from the pulse direction information data, analyzing the corresponding connection and reverse cancellation strength between adjacent positive and negative polarity pulses, identifying the completeness of polarity offsetting of residual charge at the moment of arc extinguishing, and obtaining the polarity neutralization state of residual charge.

[0041] The alternating occurrence nodes of positive and negative transient electromagnetic pulses are retrieved from the arc-extinguishing transient pulse direction information data. Based on the time sequence of these alternating occurrence nodes, adjacent positive and negative pulses are paired, and the interval between the polarity peak value of each paired pulse group and the adjacent pulse is obtained. Based on the polarity peak value of each paired pulse group, the amplitude difference between the positive and negative pulse peak values ​​in each paired pulse group is calculated. This amplitude difference is then compared to the interval duration to obtain the reverse cancellation strength of each paired pulse group. Based on the reverse cancellation strength of each paired pulse group, the ratio of the number of paired pulse groups with reverse cancellation strength exceeding a preset cancellation threshold to the total number of all paired pulse groups is calculated to obtain the polarity offset integrity index. If the polarity offset integrity index exceeds a preset neutralization judgment threshold, the residual charge polarity neutralization state is determined to be a fully neutralized state; if the polarity offset integrity index is below the preset neutralization judgment threshold, the residual charge polarity neutralization state is determined to be an insufficiently neutralized state.

[0042] In one embodiment, alternating occurrence nodes of positive and negative transient electromagnetic pulses on the time axis are identified from the arc-extinguishing transient pulse direction information data. Alternating occurrence nodes refer to the time inflection points during the arc-extinguishing process where the current signal transitions from a positive pulse to a negative pulse or vice versa; these nodes mark the boundary positions between positive and negative polarity pulses.

[0043] Specifically, based on the sequential arrangement of the alternating occurrence nodes in the time series, adjacent positive and negative pulses are paired. The pairing rule follows the principle of temporal proximity, that is, a positive pulse and a negative pulse that are temporally adjacent are grouped into a pairing pulse group. During the arc extinguishing process of the pole-mounted circuit breaker, the arc current usually exhibits alternating positive and negative oscillation characteristics, thus forming multiple pairing pulse groups. Each pairing pulse group represents one alternating release process of positive and negative polarity charges.

[0044] In one possible implementation, for each paired pulse group, the amplitude values ​​at the peak values ​​of the positive and negative pulses are extracted and used as the polarity peak values ​​of that paired pulse group. Simultaneously, the time interval between the occurrence times of the positive and negative pulse peaks is measured and used as the interval duration of that paired pulse group. The polarity peak values ​​reflect the energy intensity carried by the positive and negative pulses respectively, and the interval duration reflects the temporal tightness between the positive and negative pulses.

[0045] It should be noted that the calculation process of the reverse cancellation strength reflects the assessment of the ability of positive and negative pulses to cancel each other out. For each paired pulse group, the difference between the absolute value of the positive pulse peak and the absolute value of the negative pulse peak is calculated to obtain the amplitude difference. The smaller the amplitude difference, the more balanced the energy release of the positive and negative pulses in the paired pulse group. The reverse cancellation strength is obtained by dividing the amplitude difference by the interval duration. When the peak amplitudes of the positive and negative pulses are close and the interval duration is short, the reverse cancellation strength value is large, indicating that the positive and negative charges have completed relatively sufficient mutual cancellation in a very short time; when the peak amplitude difference is significant or the interval duration is long, the reverse cancellation strength value is small, indicating that there is energy imbalance or timing delay in the charge cancellation process. Furthermore, the statistical method for the polarity offset integrity index is as follows: Traverse all paired pulse groups, compare the reverse cancellation strength of each paired pulse group with a preset cancellation threshold, and count the number of paired pulse groups whose reverse cancellation strength exceeds the preset cancellation threshold. Divide this number by the total number of all paired pulse groups to obtain the polarity offset integrity index. The value of this indicator ranges from zero to one. The closer the value is to one, the more sufficient the positive and negative polarities of most paired pulse groups are canceled out during the arc extinguishing process. The closer the value is to zero, the less sufficient the positive and negative polarities of most paired pulse groups are canceled out.

[0046] For example, under high-load tripping conditions of a pole-mounted circuit breaker, due to the large arc energy, dozens of paired pulse groups may be generated at the moment of arc extinction. If more than a preset proportion of these paired pulse groups achieve sufficient reverse cancellation, the polarity offset integrity index will show a high value, reflecting that the offsetting of positive and negative charges during the arc extinction process is relatively sufficient.

[0047] Understandably, the determination of the residual charge polarity neutralization state is directly related to the stability of the Rogowski coil integrator circuit baseline. The polarity offset integrity index is compared with a preset neutralization threshold. If the polarity offset integrity index exceeds the neutralization threshold, it indicates that the positive and negative transient electromagnetic pulses achieved relatively balanced mutual cancellation during the arc extinguishing process, and the residual charge has essentially completed polarity neutralization in the measurement circuit. Therefore, the residual charge polarity neutralization state is determined to be a fully neutralized state.

[0048] Preferably, if the polarity offset integrity index is lower than a preset neutralization threshold, it indicates that there is a residual accumulation of a single polarity charge during the arc extinguishing process, and the positive and negative pulses have failed to effectively cancel each other out. The residual charge polarity neutralization state is then determined to be insufficient neutralization. In this insufficient neutralization state, the accumulated residual charge will have a continuous impact on the baseline zero point of the Rogowski coil integrator circuit, causing a shift in the baseline of subsequent current measurements, thereby affecting the accuracy of the circuit breaker condition assessment.

[0049] S104. By fusing the residual charge polarity neutralization state with the cumulative wear depth of the contact, the random forest algorithm is used to predict whether there is abnormal drift of the integral baseline zero position, and the prediction result of abnormal drift of the integral baseline zero position is obtained.

[0050] The residual charge polarity neutralization state and the cumulative wear depth of the contact are obtained. The residual charge polarity neutralization state is converted into a numerical feature, where a fully neutralized state is assigned a high neutralization numerical label, and an insufficiently neutralized state is assigned a low neutralization numerical label. The numerical features are combined with the cumulative wear depth of the contact to form a fusion feature set. Based on the fusion feature set, a drift prediction classifier is constructed using a random forest algorithm. The random forest algorithm constructs several independent decision trees. Each decision tree forms a decision path from the root node to the leaf node based on the distribution range of the cumulative wear depth of the contact and the category of the residual charge polarity neutralization state. The fusion feature set is input into the drift prediction classifier. Each decision tree traverses the decision path according to the wear depth value and neutralization state category in the fusion feature set, and outputs an independent judgment category for whether there is abnormal drift at the zero point of the integral baseline. The independent judgment category includes whether there is abnormal drift or not. Based on the independent decision categories output by each decision tree, the number of decision trees that are determined to have abnormal drift and the number of decision trees that are determined not to have abnormal drift are counted to obtain the prediction result of the zero position of the integral baseline abnormal drift.

[0051] In one embodiment, the residual charge polarity neutralization state and the cumulative wear depth of the contact are obtained, and the residual charge polarity neutralization state is numerically converted. A fully neutralized state indicates that the positive and negative charges are evenly canceled during the arc extinguishing process, with minimal impact on the integrator circuit baseline, and is assigned a high neutralization value label; an insufficiently neutralized state indicates the presence of residual charge of a single polarity, which has the potential to interfere with the integrator circuit baseline, and is assigned a low neutralization value label.

[0052] Specifically, the numerical features are combined with the cumulative wear depth of the contacts to form a fused feature set. The cumulative wear depth of the contacts reflects the degree of physical wear of the circuit breaker contacts, while the residual charge polarity neutralization state reflects the charge balance characteristics of the arc extinction process. Together, they constitute a multi-dimensional feature vector describing the operating state of the circuit breaker in the fused feature set. The deeper the contact wear, the more unstable the arc energy release during tripping, and there is a correlation between this and the residual charge neutralization state.

[0053] In one possible implementation, a drift prediction classifier is constructed based on the fused feature set using the random forest algorithm. The random forest algorithm is an ensemble learning method that forms a forest structure by constructing several independent decision trees, each trained based on different sampling combinations of the fused feature set. When constructing a single decision tree, a portion of samples is randomly selected from the historical samples of the fused feature set, and the splitting criteria are chosen from two feature dimensions: the cumulative wear depth of the contact and the neutralization state of the residual charge polarity. A branching structure extending downwards from the root node is constructed layer by layer.

[0054] It should be noted that the decision path formation process for each decision tree follows feature splitting rules. The root node first divides the data based on the numerical range of the cumulative wear depth of the contacts. This division uses a preset wear splitting threshold, which is statistically calculated based on historical sample data. The input is the contact wear depth dataset D, which comes from the actual measurement records of 1000 high-voltage circuit breaker contacts over the past year. The average value μ is calculated using the following formula: , where d i This represents the wear depth value for each sample, and n represents the total number of samples.

[0055] For example, if the sample data are 0.2mm, 0.3mm, 0.4mm, 0.6mm, and 0.7mm, then n=5. =2.2mm, μ=0.44mm, and the threshold is set to the calculated μ, with a value range of 0.1-1.0mm. Samples with wear depth below the threshold are assigned to the left child node, and samples with wear depth exceeding the threshold are assigned to the right child node. At the child node level, further division is performed based on the category of residual charge polarity neutralization state. Samples in the fully neutralized state and samples in the insufficiently neutralized state are guided to different branch paths, eventually reaching the leaf node. Each leaf node corresponds to a judgment category of zero-point anomalous drift of the integral baseline, including two categories: presence of anomalous drift and absence of anomalous drift.

[0056] For example, when the decision path of a decision tree is "the cumulative wear depth of the contact exceeds the wear splitting threshold and the residual charge polarity neutralization state is insufficiently neutralized", the leaf node corresponding to this path is classified as having abnormal drift; when the decision path is "the cumulative wear depth of the contact is below the wear splitting threshold and the residual charge polarity neutralization state is sufficiently neutralized", the leaf node corresponding to this path is classified as not having abnormal drift. Further, the fused feature set is input into each decision tree in the drift prediction classifier for judgment. Each decision tree, based on the specific value of the cumulative wear depth of the contact and the specific category of the residual charge polarity neutralization state in the fused feature set, traverses downwards from the root node along the matched branch path until reaching the corresponding leaf node, outputting the independent judgment category of the decision tree regarding whether there is abnormal drift at the integral baseline zero. Since each decision tree is constructed based on different sample subsets and feature partitioning methods, inputting the same fused feature set into different decision trees may yield different independent judgment categories.

[0057] Understandably, the majority voting mechanism is the core component of the random forest algorithm's final prediction. The algorithm counts the number of decision trees identified as exhibiting abnormal drift and the number identified as not exhibiting abnormal drift, then compares the number of votes for each category. If the number of decision trees identified as exhibiting abnormal drift is greater, the final prediction result for the zero-position abnormal drift of the integral baseline is that abnormal drift exists; conversely, if the number of decision trees identified as not exhibiting abnormal drift is greater, the final prediction result for the zero-position abnormal drift of the integral baseline is that abnormal drift does not exist.

[0058] Preferably, the prediction result of the abnormal zero-point drift of the integral baseline reflects the baseline stability state of the Rogowski coil integrator circuit. When the prediction result indicates the presence of abnormal drift, it suggests that under the combined conditions of current contact wear and residual charge neutralization, the zero-point reference line of the integrator circuit is at risk of shifting, affecting the accuracy of subsequent current measurement data. This prediction result provides data support for the comprehensive evaluation of the circuit breaker's condition.

[0059] S105. Based on the prediction results of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc extinction current waveform, determine the threshold adjustment direction and the adaptation section. Based on the threshold adjustment direction and the adaptation section, dynamically adjust the current zero-crossing point discrimination threshold and determine the adjusted discrimination threshold.

[0060] Based on the predicted result of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc-extinguishing current waveform, the threshold adjustment direction is determined. If the drift prediction result indicates abnormal drift and the waveform symmetry type is asymmetrical, the threshold adjustment direction is determined to be offset compensation of the corresponding polarity based on the polarity of the longer duration of the positive half-wave period and the negative half-wave period. If the drift prediction result indicates no abnormal drift, the threshold adjustment direction is determined to maintain the original value. Based on the threshold adjustment direction and the difference in duration between the positive and negative half-wave periods in the symmetry type of the arc-extinguishing current waveform, an adaptation segment for current zero-crossing point discrimination is defined. The starting boundary of the adaptation segment is the moment when the current waveform drops to the preset zero-crossing detection starting point, and the ending boundary of the adaptation segment is the moment when the current waveform rises to the preset zero-crossing detection ending point. Based on the threshold adjustment direction and the adaptation segment, an offset compensation amount is superimposed on the preset current zero-crossing point discrimination benchmark threshold. The offset compensation amount is determined by the product of the duration difference and the preset compensation mapping coefficient. The polarity of the offset compensation amount is consistent with the threshold adjustment direction, resulting in the adjusted discrimination threshold.

[0061] In one embodiment, the threshold for judging the zero-crossing point of the current is dynamically adjusted based on the prediction result of the abnormal drift of the integral baseline zero point and the symmetry type of the arc-extinguishing current waveform. When the drift prediction result indicates that there is abnormal drift and the waveform symmetry type is asymmetrical, it indicates that the baseline of the integral circuit has a tendency to shift and the positive and negative half-wave distributions of the arc-extinguishing current waveform are unbalanced. In this case, the threshold adjustment direction is determined to be offset compensation of the corresponding polarity based on the polarity of the longer duration of the positive half-wave period and the negative half-wave period.

[0062] Specifically, if the duration of the positive half-wave period is longer than that of the negative half-wave period, the threshold adjustment direction is positive offset compensation; if the duration of the negative half-wave period is longer than that of the positive half-wave period, the threshold adjustment direction is negative offset compensation. When the drift prediction result indicates that there is no abnormal drift, it means that the baseline of the integrator circuit remains stable, and the threshold adjustment direction is to maintain the original value, without the need for compensation adjustment.

[0063] In one possible implementation, an adaptation segment for current zero-crossing detection is defined based on the difference in duration between the threshold adjustment direction and the positive half-wave period and the negative half-wave period. The starting boundary of the adaptation segment is set at the moment when the current waveform amplitude drops to a preset zero-crossing detection start point, which corresponds to a preset percentage position of the current peak value; the ending boundary of the adaptation segment is set at the moment when the current waveform amplitude rises to a preset zero-crossing detection end point, which also corresponds to a preset percentage position of the current peak value. The adaptation segment defines the effective time window range for zero-crossing detection.

[0064] It should be noted that the offset compensation amount is determined based on the product of the duration difference and a preset compensation mapping coefficient. The compensation mapping coefficient is a pre-calibrated scaling factor that reflects the proportional relationship between the duration difference and the degree of integral baseline offset. The larger the duration difference, the more severe the waveform asymmetry, and the larger the corresponding offset compensation amount; the smaller the duration difference, the smaller the offset compensation amount.

[0065] For example, the offset compensation amount is superimposed on a preset current zero-crossing point discrimination reference threshold, and the polarity of the offset compensation amount is consistent with the threshold adjustment direction. If the threshold adjustment direction is positive offset compensation, the positive offset compensation amount is superimposed on the reference threshold; if the threshold adjustment direction is negative offset compensation, the negative offset compensation amount is superimposed on the reference threshold to obtain the adjusted discrimination threshold.

[0066] S106. If the adjusted judgment threshold exceeds the preset allowable range, the actual wear pattern of the contacts is identified by joint analysis of the cumulative wear depth and the duration of the opening arc.

[0067] If the adjusted threshold exceeds a preset allowable range, the cumulative wear depth of the contact and the duration of the opening arc are obtained. The cumulative wear depth and the duration of the opening arc are then compared together. Based on their numerical ratio, a wear arc correlation feature is obtained, which includes a combination of wear depth level and arc time level. The uniformity of contact wear distribution is determined based on the matching degree between the wear depth level and the arc time level in the wear arc correlation feature. If the wear depth level and the arc time level are within a preset matching range, the actual wear pattern of the contact is determined to be uniform. If the wear depth level and the arc time level exceed the preset matching range, the actual wear pattern of the contact is determined to be non-uniform. If the actual wear pattern of the contact is non-uniform, the wear distribution area with localized ablation and the corresponding peeling depth on the contact surface are identified based on whether the wear depth level is higher or lower than the arc time level in the wear arc correlation feature, thus obtaining the actual wear pattern of the contact.

[0068] In one embodiment, when the adjusted discrimination threshold exceeds a preset allowable range, it indicates that the current operating state of the circuit breaker is abnormal, triggering a deeper identification of the actual wear pattern of the contacts. The cumulative wear depth of the contacts and the duration of the opening arc are obtained and compared together. Wear depth levels are classified according to the magnitude of the cumulative wear depth, and arc time levels are classified according to the length of the opening arc duration. The wear depth levels and arc time levels are combined to form a wear-arc correlation feature.

[0069] Specifically, the wear depth level is classified according to the ratio of the cumulative wear depth d to the initial contact thickness t, r=d / t: low level is r<0.3, medium level is 0.3≤r≤0.6, and high level is r>0.6. The arc time level is classified according to the ratio of the opening arc duration a to the standard opening cycle s, p=a / s: low level is p<0.5, medium level is 0.5≤p≤0.8, and high level is p>0.8. The wear-arc correlation characteristic reflects the correspondence between the physical wear degree of the contact and the arc discharge characteristics. Under normal wear conditions, the two should show a matching level relationship; for example, the same level is considered a match, a difference of one level is considered a slight mismatch, and a difference of two levels is considered a severe mismatch.

[0070] In one possible implementation, the determination is based on the degree of matching between the wear depth level and the arc time level in the wear arc correlation characteristics. If the wear depth level and the arc time level are within a preset matching range, it indicates that the contact wear distribution is relatively uniform, and the cumulative wear depth and arc duration show a normal correspondence, thus the actual wear pattern of the contact is determined to be a uniform wear pattern. If the wear depth level and the arc time level exceed the preset matching range, it indicates that the contact wear distribution is uneven, thus the actual wear pattern of the contact is determined to be a non-uniform wear pattern.

[0071] It should be noted that when the actual wear pattern of the contact is non-uniform, the localized ablation characteristics of the contact surface are further identified based on whether the wear depth level is higher or lower than the arc time level in the wear arc correlation characteristics. If the wear depth level is higher than the expected range corresponding to the arc time level, it indicates that there is a localized deep ablation area on the contact surface, and the spalling depth in this area exceeds the normal level of uniform wear. If the wear depth level is lower than the expected range corresponding to the arc time level, it indicates that there is an area on the contact surface with concentrated arc ablation but insufficient material spalling.

[0072] For example, based on the high or low state, the wear distribution area with local ablation on the contact surface and the corresponding peeling depth are determined to obtain the actual wear morphology of the contact, providing morphological feature information of the contact layer for comprehensive evaluation of the circuit breaker condition.

[0073] S107. By fusing the actual wear pattern of the contacts with the prediction results of the zero-point abnormal drift of the integral baseline, a neural network algorithm is used to generate a primary and secondary state collaborative assessment, resulting in a collaborative analysis result of the pole-mounted circuit breaker state, including the remaining life status of the contacts, the risk level of the zero-point drift of the integral baseline, and maintenance and disposal recommendations.

[0074] The actual wear morphology of the contacts and the predicted abnormal drift of the integral baseline zero point are obtained. The wear distribution area and spalling depth in the actual wear morphology of the contacts are converted into wear morphology feature vectors, and the predicted abnormal drift of the integral baseline zero point is converted into drift state feature vectors. The wear morphology feature vectors and the drift state feature vectors are concatenated and fused to obtain a primary and secondary fused feature vector. A state collaborative evaluation network is constructed using a neural network algorithm. The input of the neural network algorithm is the primary and secondary fused feature vector, and the output is three classification nodes, corresponding to the remaining life status of the contacts, the risk level of the zero point drift of the integral baseline, and maintenance and treatment recommendations, respectively. The primary and secondary fused feature vectors are input into the state collaborative evaluation network. After layer-by-layer transmission and nonlinear mapping within the network, each output node outputs the corresponding classification result, obtaining the category of the remaining life status of the contacts, the category of the risk level of the zero point drift of the integral baseline, and the category of maintenance and treatment recommendations. Based on the category of the remaining life status of the contacts, the category of the risk level of the zero point drift of the integral baseline, and the category of maintenance and treatment recommendations, the state collaborative analysis result of the pole-mounted circuit breaker is formed.

[0075] In one embodiment, the actual wear morphology of the contact and the predicted result of the zero-position anomaly drift of the integral baseline are obtained, and both are processed into feature vectors. The actual wear morphology of the contact includes information in two dimensions: wear distribution area and spalling depth. The wear distribution area is divided into four regions according to the position of the contact surface and encoded as follows: top region is encoded as 1, bottom region as 2, left region as 3, and right region as 4. The spalling depth is divided into three levels according to the depth value range: 0 to 0.5 mm is level 1, 0.5 to 1 mm is level 2, and above 1 mm is level 3. The wear image data is input, and the distribution area and depth value are extracted using a region segmentation algorithm such as K-means clustering. Then, they are combined to form a wear morphology feature vector. For example, the vector [1,4,2] indicates that the top region and the right region have a level 2 depth. The predicted result of the zero-position anomaly drift of the integral baseline includes two categories: presence of anomaly drift and absence of anomaly drift. It is converted into a drift state feature vector using a binary encoding method, i.e., presence is 1 and absence is 0.

[0076] Specifically, the wear morphology feature vector and the drift state feature vector are concatenated and fused to form a primary-secondary fused feature vector. The wear morphology feature vector reflects the physical state of the contacts of the primary equipment of the circuit breaker, while the drift state feature vector reflects the electrical state of the secondary measurement circuit. The fusion of the two achieves a coordinated representation of the primary equipment state and the secondary measurement state. The dimension of the primary-secondary fused feature vector is the sum of the dimensions of the wear morphology feature vector and the drift state feature vector.

[0077] In one possible implementation, a state collaborative evaluation network is constructed using a neural network algorithm. A neural network algorithm is a computational method that simulates the connection structure of biological neurons, mapping input to output through the connections and weight adjustments of multiple layers of neuron nodes. The number of neurons in the input layer of the state collaborative evaluation network matches the dimension of the primary and secondary fused feature vectors, with each input layer neuron receiving the value at the corresponding position in the fused feature vector. Hidden layers are set in the middle of the network, and the hidden layer neurons abstract and extract the input features through weighted summation and nonlinear transformation. The output layer has three neuron nodes, corresponding to the three evaluation dimensions: the remaining lifespan state of the contact, the risk level of zero-point drift of the integral baseline, and maintenance and disposal recommendations.

[0078] It should be noted that after the primary and secondary fusion feature vectors are input into the state collaborative evaluation network, the input layer neurons pass the values ​​in the fusion feature vectors to the hidden layer. Each neuron in the hidden layer weights and summarizes the received signals, and transforms them using a nonlinear mapping function to extract the correlation pattern between the contact wear state and the baseline drift state in the fusion features. After passing through the hidden layers, the signal reaches the three neuron nodes in the output layer.

[0079] For example, the contact remaining life status output node outputs the contact remaining life status category based on the combined features of wear distribution area and spalling depth in the fusion features. The category includes three levels: sufficient life, moderate life, and insufficient life. The integral baseline zero-point drift risk level output node outputs the drift risk level category based on the interaction between drift status and wear morphology in the fusion features. The category includes three levels: low risk, medium risk, and high risk. The maintenance and handling recommendation output node outputs the maintenance and handling recommendation category by integrating the status features of the first two dimensions. The category includes four types: normal operation, enhanced monitoring, planned maintenance, and emergency handling. Furthermore, the classification results of each output node are obtained through training within the network. During the training phase, the state collaborative evaluation network first preprocesses the historical operating data, including normalizing the feature vectors of contact wear status and baseline drift status, scaling the values ​​to the range of 0 to 1. Then, labels are generated: based on actual maintenance records, the remaining contact life status is mapped as follows: sufficient life corresponds to an operating time greater than 1000 hours, medium life corresponds to 500 to 1000 hours, and insufficient life corresponds to less than 500 hours; drift risk level is mapped as follows: low risk corresponds to a drift value less than 0.1, medium risk corresponds to 0.1 to 0.5, and high risk corresponds to greater than 0.5; maintenance and disposal recommendations are based on a combination of the above two, such as sufficient life and low risk being mapped to normal operation. The cross-entropy loss function is used, where L = -sum(y i ×log(p i ), where L is the loss value, y i For real labels, p iTo predict probabilities, the Adam optimizer with a learning rate of 0.001 is used to adjust parameters so that the network can accurately output the corresponding state category and maintenance suggestions based on the input primary and secondary fusion feature vectors.

[0080] Understandably, the collaborative analysis results of the pole-mounted circuit breaker status are formed by integrating the remaining contact life status category, the integral baseline zero-point drift risk level category, and the maintenance and handling recommendation category. This collaborative analysis result correlates and evaluates the contact wear status of the primary circuit breaker with the baseline drift status of the secondary measurement circuit, outputting comprehensive assessment information covering the remaining equipment life, measurement risk level, and maintenance and handling recommendations.

[0081] Preferably, the results of the collaborative analysis of the status of the pole-mounted circuit breaker are output in the form of structured data, including the specific level of the remaining life status of the contacts, the specific level of the risk level of the zero drift of the integral baseline, and the specific type of maintenance and disposal recommendations, so as to provide power operation and maintenance personnel with a comprehensive basis for judging the status of the circuit breaker and a reference for maintenance decisions.

[0082] The wear distribution area and spalling depth in the actual wear morphology of the contacts are extracted, as well as the drift tendency level in the zero-point abnormal drift prediction results of the integral baseline. The remaining life range that the contacts can carry for opening and closing operations and the interference level of zero-point drift on protection actions are analyzed. The remaining life status and drift risk level are assessed, and maintenance and disposal suggestions are given to obtain the status collaborative analysis results of the pole-mounted circuit breaker.

[0083] Based on the stated lifespan range and the stated drift state category, the interference level of the integral baseline zero-point drift on the protection operation is assessed. If the drift state category indicates abnormal drift and the lifespan range is a short-cycle lifespan range, the interference level is determined to be high. If the drift state category indicates no abnormal drift and the lifespan range is a long-cycle lifespan range, the interference level is determined to be low. Other combinations are determined to be medium-level interference. The remaining lifespan state is determined based on the stated lifespan range: long-cycle lifespan range corresponds to a good state, medium-cycle lifespan range corresponds to a medium state, and short-cycle lifespan range corresponds to a poor state. The drift risk level is determined based on the stated interference level: high interference level corresponds to high risk, medium interference level corresponds to medium risk, and low interference level corresponds to low risk. Based on the combination of the remaining lifespan status and the drift risk level, corresponding maintenance and handling recommendations are matched. The specific matching logic is as follows: Good status + low risk corresponds to continued monitoring without intervention; Good status + medium risk corresponds to periodic inspection; Medium status + low risk corresponds to optimized operating parameters; Good status + high risk corresponds to immediate repair; Medium status + medium risk corresponds to planned maintenance; Medium status + high risk corresponds to emergency replacement; Poor status + low risk corresponds to enhanced monitoring; Poor status + medium risk corresponds to time-limited repair; Poor status + high risk corresponds to immediate decommissioning and replacement. Integrating the remaining lifespan status, the drift risk level, and the maintenance and handling recommendations yields the pole-mounted circuit breaker status collaborative analysis results.

[0084] In one embodiment, feature information in two dimensions—wear distribution area and spalling depth—is extracted from the actual wear morphology of the contact. The wear distribution area reflects the spatial distribution range of wear on the contact surface, and the spalling depth reflects the longitudinal degree of contact material loss. Simultaneously, a drift state category is extracted from the integral baseline zero-point abnormal drift prediction result, including two states: presence of abnormal drift and absence of abnormal drift.

[0085] Specifically, the remaining lifespan range that the contact can withstand for opening and closing operations is determined based on the ratio of the wear distribution area coverage to the spalling depth relative to the initial contact thickness. The calculation process is as follows: Input the spalling depth D (in millimeters), the initial contact thickness T (in millimeters), and the percentage of the wear distribution area coverage A, where the percentage R = D / T. According to the IEC 60947 standard and material fatigue test data, when R is less than 0.2 and A is less than 0.3, the lifespan range is determined to be a long-cycle lifespan range, indicating that the contact still has the capacity to withstand a relatively large number of opening and closing operations. According to accelerated life tests, when R is greater than 0.5 or A is greater than 0.7, the lifespan range is determined to be a short-cycle lifespan range, indicating that the remaining capacity of the contact to withstand opening and closing operations is limited. The situation between these two is determined to be a medium-cycle lifespan range.

[0086] In one possible implementation, the interference level of the integral baseline zero-point drift on protection operation is evaluated based on the combination of the lifespan range and the drift state category. The interference level reflects the degree to which the measurement circuit baseline offset affects the accuracy of the circuit breaker's protection function. If the drift state category indicates abnormal drift and the lifespan range is a short-cycle lifespan range, the combination of these two factors indicates that the circuit breaker faces the dual risks of severe contact wear and inaccurate measurement baseline, and the interference level is determined to be high. If the drift state category indicates no abnormal drift and the lifespan range is a long-cycle lifespan range, it indicates that the overall condition of the circuit breaker is good, and the interference level is determined to be low.

[0087] It should be noted that the lifespan range is defined as a percentage segment of the equipment's remaining lifespan relative to its design lifespan. For example, a remaining lifespan less than 20% of the design lifespan is considered a low range, 20% to 50% a medium range, and greater than 50% a high range. The determination process involves calculating R = (DU) / D, where R is the remaining lifespan ratio, D is the design lifespan in years, and U is the years already used. Then, the corresponding range is matched based on the R value to determine the remaining lifespan status. The interference level is defined as the intensity of external noise interference with the measured signal. The calculation process involves collecting signal samples and calculating I = (N / S) × 100, where I is the interference level percentage, N is the noise amplitude, and S is the signal amplitude. If I is less than 30, it is a low level; 30 to 60 is a medium level; and greater than 60 is a high level. This determines the drift risk level. The remaining lifespan status describes the circuit breaker's operating status at the primary equipment level, while the drift risk level describes the circuit breaker's operating status at the secondary measurement level.

[0088] For example, based on the combination of remaining life status and drift risk level, corresponding maintenance and handling recommendations are matched. The specific matching rules are as follows: remaining life status is divided into high (greater than 80%), medium (50% to 80%), and low (less than 50%); drift risk level is divided into low (less than 0.3), medium (0.3 to 0.7), and high (greater than 0.7). If the remaining life is high and the risk is low, normal operation is matched; if the remaining life is high and the risk is medium, enhanced monitoring is matched; if the remaining life is low-medium and the risk is high, planned maintenance is matched; if the remaining life is low and the risk is high, emergency handling is matched. After matching according to these rules, the remaining life status, drift risk level, and maintenance and handling recommendations are integrated to form a collaborative analysis result of the pole-mounted circuit breaker status. The specific integration method is as follows: input the remaining life value L (percentage) and the risk value R, output a structured report, including a status description such as L=70%, R=0.5 corresponding to enhanced monitoring, and recommendation text, forming a complete analysis document.

[0089] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers, characterized in that, The method includes: The initial dataset is obtained by collecting data on the cumulative wear depth of the contacts, the duration of the opening arc, and the direction of the arc-extinguishing transient pulse through the vacuum interrupter wear monitoring unit and the Rogowski coil integral baseline acquisition unit. By analyzing the arc duration and pulse direction information in the initial data acquisition dataset, the symmetry type of the arc extinction current waveform is identified, which includes asymmetry type. If the symmetric type of the arc-extinguishing current waveform is in an asymmetric type, the alternation pattern of positive and negative transient electromagnetic pulses is analyzed by pulse direction information data to assess the polarity neutralization state of the residual charge. By fusing the residual charge polarity neutralization state with the cumulative wear depth of the contact, the random forest algorithm is used to predict whether there is abnormal drift of the integral baseline zero position, and the prediction result of abnormal drift of the integral baseline zero position is obtained. Based on the prediction results of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc extinction current waveform, the threshold adjustment direction and the adaptation section are determined. Based on the threshold adjustment direction and the adaptation section, the current zero-crossing point discrimination threshold is dynamically adjusted, and the adjusted discrimination threshold is determined. If the adjusted threshold exceeds the preset allowable range, the actual wear pattern of the contact will be identified by combining the cumulative wear depth and the duration of the opening arc. By fusing the actual wear pattern of the contacts with the prediction results of the abnormal drift of the integral baseline zero position, a neural network algorithm is used to generate a primary and secondary state collaborative assessment, and the state collaborative analysis results of the pole-mounted circuit breaker are obtained.

2. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, The initial dataset is obtained by collecting data on the cumulative wear depth of the contacts, the duration of the opening arc, and the direction of the arc-extinguishing transient pulse through the vacuum interrupter wear monitoring unit and the Rogowski coil integral baseline acquisition unit. This dataset includes: The cumulative wear depth of the contact is obtained by comparing the difference in stroke before and after the contact movement with the initial thickness of the contact, which is collected by a displacement sensor. The duration of the opening arc is determined by capturing the time span of the start and end times of the opening arc discharge signal using a current transformer. Record the order of appearance and amplitude direction of each pulse in the positive and negative polarity sequence at the moment of arc extinction, obtain the direction information data of the transient pulse at the moment of arc extinction, and aggregate and integrate them to obtain the initial acquisition dataset.

3. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, The process involves analyzing the arc duration and pulse direction information data in the initial acquired dataset to identify the symmetry type of the arc extinction current waveform. Symmetry types of the arc extinction current waveform include asymmetric types, including: Extract the duration of the shunt arc and the pulse direction information from the initial acquisition dataset. Based on the amplitude distribution of the positive and negative pulses, calibrate the peak position and zero-crossing time of each pulse to obtain the positive half-wave period and negative half-wave period of the arc extinguishing current waveform. The waveform deviation index is obtained based on the ratio of the difference between the positive and negative half-wave durations to the duration of the opening arc. If the waveform offset index exceeds the preset symmetry judgment threshold, the arc extinction current waveform is determined to be of an asymmetrical type.

4. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, If the arc-extinguishing current waveform is asymmetrical, then the alternation pattern of positive and negative transient electromagnetic pulses is analyzed through pulse direction information data to assess the polarity neutralization state of the residual charge, including: Extract the alternating occurrence nodes of positive and negative transient electromagnetic pulses from the pulse direction information data, mark the timing positions of adjacent positive and negative polarity pulses, and obtain the polarity peak value of each pulse and the interval between adjacent pulses. The amplitude difference between adjacent positive and negative pulse peaks is calculated based on the polarity peak value and the interval duration, and the ratio is calculated with the interval duration to obtain the reverse cancellation strength. If the reverse cancellation strength exceeds the preset neutralization threshold, the residual charge polarity neutralization state is determined to be fully neutralized.

5. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 4, characterized in that, The process of calculating the reverse cancellation strength by ratio with the interval duration includes: The alternating occurrence nodes of positive transient electromagnetic pulses and negative transient electromagnetic pulses are retrieved from the pulse direction information data. Adjacent positive pulses and negative pulses are paired according to the arrangement order of the alternating occurrence nodes in the time series, and the polarity peak value of each paired pulse group and the interval duration between adjacent pulses are obtained. The amplitude difference between the peak value of the positive pulse and the peak value of the negative pulse in each pair of pulse groups is calculated based on the polarity peak value of each pair of pulse groups, and the ratio is calculated with the interval duration to obtain the reverse cancellation strength of each pair of pulse groups. The polarity offset integrity index is obtained by statistically analyzing the ratio of the number of paired pulse groups whose reverse offset strength exceeds the preset offset threshold to the total number of all paired pulse groups. If the polarity offset integrity index exceeds the preset neutralization judgment threshold, the residual charge polarity neutralization state is determined to be a fully neutralized state; if it is below the preset neutralization judgment threshold, it is determined to be an insufficiently neutralized state.

6. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, The method involves fusing the residual charge polarity neutralization state with the cumulative wear depth of the contact, and using a random forest algorithm to predict whether there is abnormal drift in the integral baseline zero point, thus obtaining the prediction result of abnormal drift in the integral baseline zero point, including: The residual charge polarity neutralization state is converted into a numerical feature, with sufficient neutralization assigned a high neutralization numerical label and insufficient neutralization assigned a low neutralization numerical label, which is combined with the contact cumulative wear depth to form a fused feature set; A drift prediction classifier is constructed using the random forest algorithm. The drift prediction classifier consists of several independent decision trees that form a decision path based on the wear depth distribution range and the neutral state category. The independent decision categories are statistically analyzed to obtain the integral baseline zero-position abnormal drift prediction result.

7. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, The process of determining the threshold adjustment direction and adaptation segment based on the prediction results of the zero-point abnormal drift of the integral baseline and the symmetry type of the arc-extinguishing current waveform, dynamically adjusting the current zero-crossing discrimination threshold based on the threshold adjustment direction and adaptation segment, and determining the adjusted discrimination threshold includes: If the prediction result of the zero-position abnormal drift of the integral baseline is that there is abnormal drift and the symmetry type of the arc extinguishing current waveform is asymmetric, then the threshold adjustment direction is determined to be offset compensation of the corresponding polarity based on the polarity of the longer duration of the positive half-wave period and the negative half-wave period. If the drift prediction result is that there is no abnormal drift, then the threshold adjustment direction is determined to be to keep the original value. The appropriate segment for current zero-crossing point discrimination is defined based on the difference between the threshold adjustment direction and the duration of the positive and negative half-waves. An offset compensation amount is superimposed on a preset current zero-crossing point discrimination benchmark threshold. The offset compensation amount is determined by the product of the duration difference and the preset compensation mapping coefficient. The polarity of the offset compensation amount is consistent with the threshold adjustment direction, thus obtaining the adjusted discrimination threshold.

8. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, If the adjusted threshold exceeds the preset allowable range, the actual wear pattern of the contacts is identified by jointly analyzing the cumulative wear depth and the duration of the opening arc, including: The wear arc correlation feature is obtained by comparing the cumulative wear depth of the contact with the duration of the opening arc. The wear arc correlation feature includes a combination category of wear depth level and arc time level. If the wear depth level and the arc time level are within the preset matching range, it is determined to be a uniform wear pattern; otherwise, it is determined to be a non-uniform wear pattern. Non-uniform wear morphology identifies the wear distribution area of ​​local ablation and the corresponding spalling depth based on whether the wear depth level is higher or lower than the arc time level.

9. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 1, characterized in that, The method involves fusing the actual wear morphology of the contacts with the predicted abnormal drift of the integral baseline zero point, and using a neural network algorithm to generate a primary and secondary state collaborative assessment, thereby obtaining the state collaborative analysis results of the pole-mounted circuit breaker, including: The wear distribution area and spalling depth in the actual wear morphology of the contact are converted into wear morphology feature vectors, and the prediction result of the zero-position abnormal drift of the integral baseline is converted into a drift state feature vector. The first and second fusion feature vectors are spliced ​​and fused together. A state collaborative assessment network is constructed using a neural network algorithm, which outputs three classification nodes: the remaining life status of the corresponding contacts, the risk level of zero drift of the integral baseline, and maintenance and disposal suggestions. These are then integrated to form the state collaborative analysis results of the pole-mounted circuit breaker.

10. The method for coordinated state analysis of primary and secondary integrated pole-mounted circuit breakers according to claim 9, characterized in that, The integrated results of the pole-mounted circuit breaker status collaborative analysis include: Extract the wear distribution area and spalling depth from the actual wear morphology of the contact, and the drift tendency level from the zero-position abnormal drift prediction result of the integral baseline, and analyze the remaining life range of the contact's opening and closing operations. If the drift state is abnormal and the lifetime range is short, the interference level is determined to be high interference; if the drift state is not abnormal and the lifetime range is long, the interference level is determined to be low interference; the other combinations are determined to be medium interference. Long cycles correspond to good conditions, medium cycles to average conditions, and short cycles to poor conditions. High interference corresponds to high risk, medium interference corresponds to medium risk, and low interference corresponds to low risk. Based on the combination of remaining life status and drift risk level, maintenance and disposal recommendations are matched and integrated to obtain the results of collaborative analysis of pole-mounted circuit breaker status.

Images