A method and apparatus for detecting degradation of a distribution box contact arc

Abstract

The application discloses a kind of distribution box contact arc degradation detection method and device, the steps of this method are: arc light voltage, arc current and three-way signal of acoustics impact are synchronously collected when contact opening and closing operation, five-dimensional sound arc feature vector including arc light peak voltage, arc light cumulative energy, arc light zero-crossing number, sound-light time delay difference and sound-light amplitude ratio are extracted;According to the dispersion degree between each dimension component operation, adaptively update detection threshold value;In sliding time window, fit each dimension drift rate and fuse to obtain comprehensive degradation score;According to the joint change mode of arc light feature and sound-light coupling feature, identify degradation type;According to the growth trend of comprehensive degradation score, extrapolate the number of remaining operations.The application also proposes a corresponding detection device, which can distinguish three degradation types of increasing contact resistance, insufficient spring pressure and contact material ablation, realize quantitative tracking of degradation trend and residual life prediction.

Description

Technical Field

[0001] This invention belongs to the field of power distribution equipment condition monitoring and fault early warning technology, specifically, it relates to a method and device for detecting arc degradation of contacts in a power distribution box. Background Technology

[0002] Distribution boxes are widely used switch control devices in low-voltage power distribution systems. Their internal contacts repeatedly perform opening and closing operations during long-term operation. Inevitably, arcing occurs during these operations. The high temperature and electromagnetic force of the arc gradually cause erosion of the contact surface material, accumulation of oxide film, and fatigue degradation of the mechanism springs. When this degradation accumulates to a certain extent, it leads to abnormally high contact resistance, decreased reliability of opening and closing, and in severe cases, may cause thermal failures or even fires. Therefore, effective monitoring and early warning of the degradation status of distribution box contacts are crucial for ensuring the safe operation of the power distribution system.

[0003] Existing methods for detecting contact condition mainly rely on periodic power outages for maintenance, judging the degradation state by manually measuring the contact resistance or observing the wear of the contact surface. Some online monitoring solutions collect voltage and current waveforms when the contacts operate, extracting single or a few characteristic parameters such as arc duration and arc voltage peak value, and determining whether the contacts need to be replaced based on fixed thresholds.

[0004] However, in practical applications, the aforementioned methods only focus on limited features in the arc voltage or current signals, neglecting the acoustic impact information accompanying the arc discharge process and the cross-modal coupling relationship between the arc signal and the acoustic signal. The acoustic impact signal contains rich information about the mechanical separation process of the contacts; when the contact gap characteristics or spring pressure change, the propagation delay and amplitude of the acoustic impact will change accordingly. Existing methods cannot capture this cross-modal information, resulting in insufficient ability to distinguish between different degradation mechanisms and difficulty in providing accurate early warnings in the early stages of degradation. Summary of the Invention

[0005] Existing contact degradation detection methods only utilize the limited characteristics of arc voltage or current, ignoring acoustic impact information and cross-modal coupling relationships. Furthermore, their fixed threshold judgment methods cannot adapt to individual differences and operating condition fluctuations, lacking the ability to track degradation trends and predict lifespan. Based on this, the following invention is proposed: A method and apparatus for detecting arc degradation of contacts in a distribution box, wherein the method comprises the following steps: Step S01: Simultaneously collect arc voltage signal, contact arc current signal and acoustic impact signal of the distribution box contacts during opening and closing operations; Step S02: Extract the five-dimensional acoustic arc feature vector from the acquired signal ( , , , , ),in This is the peak voltage of the arc. The accumulated arc energy is the time integral of the product of the arc voltage and the contact arc current. This refers to the number of zero-crossings of the arc light. The acoustic-optical time delay difference is defined as the time interval between the triggering moment of the arc voltage signal and the peak moment of the acoustic impact amplitude during the same operation. The acoustic-optic amplitude ratio is defined as the ratio of the peak amplitude of the acoustic impact to the peak voltage of the arc during the same operation. The ratio; of which and The cross-modal coupling characteristics characterize the physical relationship between the arc energy release process and the acoustic impact propagation process; Step S03: For each component of the five-dimensional acoustic arc feature vector, based on the recent continuous... In the next operation, the first Inter-operational dispersion of eigenvalues upper limit of dispersion during the initial commissioning and normal operation phase The ratio is used to calculate the adaptive coefficient. And update the detection thresholds for each dimension in an exponentially smooth manner. ;in For the recent The standard deviation of the sample size reflects the increased randomness of the arcing process caused by contact degradation. This is the threshold sensitivity coefficient, calibrated during the initial commissioning phase; Step S04: For each component of the five-dimensional acoustic arc feature vector, perform linear fitting on the changing trend of each component with the number of operations within the sliding time window, and take the absolute value of the fitting slope as the drift rate. The degradation contribution weights are obtained by normalizing the drift rates of each dimension. ; using the current values ​​of each component relative detection threshold Normalization bias Calculate the overall degradation score ; Step S05: Based on the five-dimensional acoustic arc feature vector ( , , The arc characteristic pattern of () and ( , The combined change pattern of acoustic-optical coupling characteristics is used to identify the contact degradation type and output a degradation type label; among which, the increase in contact resistance is manifested as... Continue to rise Increase, and No significant trend change Basically stable; insufficient spring pressure type manifests as increase, The trend is decreasing, and It shows an extending trend. Increased fluctuations in the operating area; ablation of the contact material is manifested as... Continuously increasing The trend is extending, and The coefficient of variation increases; Step S06: Based on the comprehensive degradation score As the number of operations increases, the slope is extrapolated to a preset contact maintenance threshold. Output the estimated number of remaining operations. And degradation type tags.

[0006] Further, in step S01, the arc voltage signal is acquired by an arc voltage sensor installed near the contact gap, and the acoustic impact signal is acquired by a piezoelectric acoustic sensor installed on the distribution box housing, with a sampling rate of not less than 100kHz; the three signals are synchronously triggered by a unified clock source to eliminate time base deviation between channels.

[0007] Furthermore, in step S02, the arc light accumulates energy. The calculation uses arc voltage signal With contact arc current signal The product of arc duration The time integral within, i.e. ;in For the first time the arc voltage signal exceeds the preset trigger threshold At that moment, The arc voltage signal fell back to its last value. The following moments.

[0008] Furthermore, in step S02, the acousto-optic time delay difference The measurement was performed when the arc voltage signal first exceeded the preset trigger threshold. The moment of arrival is taken as the arc trigger reference time, and the moment when the acoustic impact signal envelope amplitude reaches its peak value in the same operation is taken as the acoustic arrival time. The difference between the two is... When the contacts are in good contact The typical value range is 0.3 to 0.8.

[0009] Furthermore, in step S03, the upper limit of dispersion during the initial commissioning phase is... Before the initial commissioning and normal operation phase In the next operation, the first Maximum value of the standard deviation sequence of eigenvalues; threshold sensitivity coefficient according to Calibration, among which For the first The initial detection threshold for dimensional features, This is the safety margin factor.

[0010] Further, in step S04, when the drift rate of a certain component... Below the preset minimum drift threshold At that time, the drift rate of that dimension is set to zero so that it does not participate in the degradation contribution weight allocation, thus avoiding the impact of noise fluctuations on the overall degradation score. Interference.

[0011] Furthermore, in step S05, the degradation type identification employs a multi-dimensional feature trend criterion for joint determination; for , , Calculate the sign of the trend slope and the coefficient of variation within the sliding time window, respectively. , Calculate the trend slope sign for each; combine the five-dimensional trend slope sign with the coefficient of variation to form a discriminant vector, and match it with the preset discriminant templates for the three degradation types, taking the one with the highest matching degree as the recognition result.

[0012] Furthermore, in step S06, the remaining number of operations... The calculation uses a comprehensive degradation score. In recent The linear fit slope in the next operation record ,according to Extrapolation yields; when If the number of operations is less than or equal to zero, it is determined that the contact is not currently showing a deterioration trend, and the remaining number of operations is not output.

[0013] Furthermore, this invention also proposes a device for detecting arc degradation of contacts in a distribution box, comprising: a multimodal acquisition module, including an arc voltage sensor, an arc current sensor, and an acoustic sensor, for simultaneously acquiring arc voltage signals, contact arc current signals, and acoustic impact signals during contact operation; a feature extraction module for calculating a five-dimensional acoustic arc feature vector from the multimodal signals; and an adaptive threshold module for continuously updating the detection thresholds for each dimension based on the recent inter-operational dispersion of each component. The degradation assessment module is used to calculate the degradation contribution weight. Overall Degradation Score and contact degradation type labels; prediction output module, used to extrapolate the overall degradation score. The growth trend reaches the contact repair threshold, outputting the estimated remaining number of operations. And degradation type tags.

[0014] Compared with existing technologies, this invention has the following advantages: It introduces the synchronous acquisition of three signals—arc voltage, arc current, and acoustic impact—to construct a five-dimensional acoustic arc feature vector containing cross-modal coupling characteristics. This overcomes the limitations of traditional methods that rely solely on a single mode of the arc signal, effectively distinguishing between three different degradation mechanisms: increased contact resistance, insufficient spring pressure, and contact material ablation. Furthermore, it employs an adaptive threshold rolling update mechanism based on inter-operation dispersion, overcoming the shortcomings of fixed thresholds that cannot adapt to individual equipment differences and operating condition fluctuations. Finally, it achieves quantitative tracking of degradation trends through a comprehensive degradation scoring method that fuses multi-dimensional feature drift rates, and extrapolates the remaining number of operations, providing maintenance personnel with a basis for preventative maintenance decisions. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for detecting arc degradation of contacts in a distribution box according to the present invention; Figure 2 This is a flowchart of the five-dimensional acoustic arc feature extraction process of the present invention; Figure 3 This is a schematic diagram of the arc degradation detection device for the contact of the distribution box according to the present invention. Detailed Implementation

[0016] The present invention will be further described in detail below with reference to the accompanying drawings, so that those skilled in the art can more clearly understand the present invention. The present invention proposes a method and apparatus for detecting arc degradation of contacts in a distribution box, wherein the method is as follows... Figure 1 As shown, the steps are as follows: Step S01: Simultaneously acquire multi-mode signals during the opening and closing operation of the distribution box contacts. The purpose of this step is to simultaneously acquire three raw data streams—arc voltage signal, contact arc current signal, and acoustic impact signal—at the instant the contacts perform opening and closing operations, providing a complete data foundation for the subsequent extraction of the five-dimensional acoustic arc feature vector.

[0017] The arc voltage signal is acquired using an arc voltage sensor installed near the contact gap. This sensor directly measures the arc voltage generated across the gap during contact separation, and its measurement range should cover the arc voltage amplitude that may occur during the process from normal contact to complete separation. The arc voltage sensor should be installed as close to the contact gap as possible to minimize the attenuation effect of lead inductance on the high-frequency arc signal.

[0018] The arc current signal of the contact is acquired through an arc current sensor, typically a Hall effect current sensor or a Rogowski coil, which is installed on the conductor of the circuit containing the contact. The arc current signal records the change in current flowing through the contact circuit during arc discharge. This signal, together with the arc voltage signal, is used to calculate the accumulated energy of the arc. Arc cumulative energy is a key parameter for measuring the degree of thermal shock of the electric arc to the contact material during a single operation.

[0019] The acoustic shock signal is acquired by a piezoelectric acoustic sensor installed on the distribution box housing. At the moment of contact opening and closing, both mechanical separation and arc discharge generate broadband acoustic shock waves, which propagate along the housing and are received by the sensor. The frequency response range of the piezoelectric acoustic sensor must cover the main frequency bands of the contact mechanical shock and arc acoustic radiation.

[0020] The synchronous acquisition of the three signals adopts a unified clock source synchronous triggering mechanism, meaning that the three acquisition channels share the same hardware trigger signal and clock reference to eliminate time base deviation between channels. The sampling rate is no less than 100kHz to ensure the accuracy of subsequent acousto-optic delay. The measurement accuracy reaches the microsecond level. Each opening and closing operation triggers a complete multimodal data acquisition window, and the acquisition duration should cover a margin of time from the start of contact separation to the complete extinguishing of the arc.

[0021] This step outputs three synchronous raw signals: arc voltage signal. Contact arc current signal and acoustic impact signal This is used for subsequent feature extraction.

[0022] Step S02: Calculate the five-dimensional acoustic arc feature vector based on multimodal acquisition signals. This step extracts the five-dimensional acoustic arc feature vector from the three synchronous raw signals output in step S01. , , , , ),like Figure 2 As shown, this vector comprehensively characterizes the intensity and duration of the arc discharge process during a single opening and closing operation, as well as the cross-modal coupling relationship between the arc and the acoustic impact.

[0023] 1. Peak voltage of arc Extraction Arc peak voltage Defined as the arc voltage signal in a single operation The maximum absolute value within the arc duration. First, determine the start and end times of the arc duration: the moment when the arc voltage signal first exceeds the preset trigger threshold. The time is the start time The arc voltage signal fell back to its last value. The following times are the termination times. Trigger threshold Typically, the residual voltage peak value across the contacts under normal contact conditions is taken as 3 to 5 times to eliminate contact noise interference during contact closure. The maximum absolute value of the arc voltage signal within the interval is... As the contacts degrade, the increased contact resistance will lead to... It shows an upward trend.

[0024] 2. Arc light accumulated energy Calculation Arc light accumulated energy Using arc voltage signal With contact arc current signal The product of arc duration Time integral within:

[0025] This energy value reflects the total heat energy released by the electric arc during a single operation. In a digital acquisition system, the above formula uses the sampling interval... Discretize into summation form: As the contact material ablation worsens or the contact resistance increases, both the duration and intensity of the arc discharge tend to increase. It will continue to increase.

[0026] 3. Number of arc zero crossings Statistics Arc zero crossing times Defined as arc voltage signal During the arc duration The number of times the arc crosses the zero-voltage line. In digital signal processing, the number of zero-crossings is counted by detecting the sign change of adjacent sampling points point by point. The number of arc zero-crossings is closely related to the stability and arc-burning characteristics of the arc: under normal contact conditions, the arc experiences fewer zero-crossing reignition processes; when insufficient spring pressure causes unstable contact separation speed, the arc will frequently extinguish and reignite. Significantly increased.

[0027] 4. Audio-visual time delay difference Measurement Acoustic-optical time delay difference Defined as the time interval between the trigger moment of the arc voltage signal and the peak moment of the acoustic impact amplitude during the same operation. Specifically, it is defined as the time interval from when the arc voltage signal first exceeds the preset trigger threshold. The moment As the reference time for arc triggering; for acoustic impact signals Perform a Hilbert transform to extract its envelope, and take the moment when the envelope amplitude reaches its peak. As the acoustic arrival time; the difference between the two is... The acoustic-optic time delay difference reflects the propagation delay from the generation of the arc light to the arrival of the acoustic impact peak at the sensor. Its value is influenced by the contact gap size, spring preload, and the acoustic propagation path of the housing. When the contact gap increases due to ablation or spring relaxation, leading to a greater separation distance, This trend will continue.

[0028] 5. Sound-light amplitude ratio Calculation Acousto-optic amplitude ratio Defined as the peak amplitude of the acoustic impact and the peak voltage of the arc during the same operation. The ratio, that is ,in This represents the peak amplitude of the acoustic impact signal envelope. The acousto-optic amplitude ratio characterizes the relative magnitude of the acoustic impact response per unit arc voltage intensity. Under good contact conditions, The typical value range is 0.3 to 0.8. When insufficient spring pressure weakens the mechanical impact of the contacts, the peak amplitude of the acoustic impact decreases, while the peak arc voltage may remain or even increase. The trend will be reduced.

[0029] This step outputs a five-dimensional acoustic arc feature vector. , , , , This is used for subsequent adaptive threshold updates and degradation assessments.

[0030] Step S03: Adaptive threshold rolling update based on inter-operation dispersion This step targets each component of the five-dimensional acoustic arc feature vector and utilizes its statistical fluctuation characteristics in recent continuous operations to adaptively adjust the detection threshold of each dimension. This allows the threshold to dynamically change with the contact degradation process, avoiding false alarms under normal operating conditions and responding promptly when degradation occurs.

[0031] 1. Dispersion between operations Calculation The 5th feature vector of the five-dimensional acoustic arc Dimensional components ( , respectively corresponding , , , , Take the most recent consecutive The set of eigenvalues ​​of this dimension in this operation Calculate its standard deviation as the interoperational dispersion: in For this In the next operation, the first The mean of the eigenvalues. Number of recent consecutive operations. The value ranges from 20 to 50 times, with the specific value determined based on the operating frequency of the distribution box: a larger value is used when the operating frequency is high to enhance statistical stability, and a smaller value is used when the operating frequency is low to improve response sensitivity. (Inter-operation dispersion) This reflects the degree of random fluctuation in the arcing process during recent contact operation; contact degradation will lead to increased uncertainty in the arcing process. It then rises.

[0032] 2. Upper limit of dispersion during the initial commissioning phase calibration Upper limit of dispersion during the initial commissioning phase The calibration phase is determined after the distribution box is first put into operation. The calibration phase is the earlier one. The operation ( Typically, 100 to 200 times are taken, and it is divided into several lengths. The sliding sub-windows calculate the first sub-window within each sub-window. The standard deviation of the eigenvalues ​​is taken as the maximum standard deviation of all sub-windows. This value represents the upper bound of the dispersion of the contacts under normal operating conditions, and is used to normalize the dispersion changes in subsequent operating phases.

[0033] 3. Adaptive coefficient With detection threshold Update Adaptive coefficients Based on the current operational dispersion Upper limit of dispersion in the initial commissioning phase The ratio is determined as follows:

[0034] when consistently below hour, Approaching zero indicates that the contact is within the normal fluctuation range; when contact degradation leads to... Exceed hour, A saturation value of 1 indicates that the fluctuations caused by degradation have exceeded the normal range.

[0035] Detection thresholds for each dimension Updated using exponential smoothing:

[0036] in The threshold sensitivity coefficient is calibrated as follows: during the initial commissioning phase, based on the first... Initial detection threshold for dimensional features With safety margin factor according to Calculated. Initial detection threshold. The safety margin factor is determined based on the contact qualification criteria provided by the equipment manufacturer or industry standards. Typically, a value of 2 to 4 is used; a larger value results in a more conservative threshold. This calibration method ensures that under normal operating conditions... Maintaining stability, but when degradation occurs. Able to follow The value is increased and adjusted upwards in a timely manner to balance detection sensitivity and false alarm suppression.

[0037] This step outputs the five-dimensional detection threshold. ( (This information is used for subsequent comprehensive degradation scoring.)

[0038] Step S04: Comprehensive degradation score based on five-dimensional acoustic arc feature drift rate This step utilizes the trend drift information of the five-dimensional acoustic arc feature vector within a sliding time window, integrates the degradation contribution of each dimension component, and calculates the comprehensive degradation score. This provides a unified indicator for the quantitative assessment of the degree of degradation.

[0039] 1. Drift rate Fitting The 5th feature vector of the five-dimensional acoustic arc Dimensional components, take the nearest one Feature value sequence in the next operation record Using the operation number as the independent variable and the eigenvalues ​​as the dependent variable, a linear fit is performed using the least squares method. Let the fitted line be... Then the fitting slope This reflects the average rate of change of the eigenvalue of this dimension with the number of operations. Drift rate Take the absolute value of the fitted slope, i.e. The number of operations covered by the sliding time window. The value range is from 30 to 100 times. When the value is too small, the fitting result is greatly affected by short-term fluctuations. When the value is too large, the response delay to the degradation trend increases.

[0040] To avoid noise fluctuations interfering with the degradation score, when the drift rate of a certain dimension component... Below the preset minimum drift threshold When this occurs, the drift rate in that dimension is set to zero. Minimum drift threshold. The sum of the mean and standard deviation of the absolute values ​​of the linear fitting slopes of each component during the initial commissioning phase can be taken as the value. The upper limit of the normal fluctuation range is used as the noise filtering benchmark.

[0041] 2. Degradation Contribution Weight Calculation Normalizing the drift rates in each dimension yields the degradation contribution weights: When the drift rate in all dimensions is zero (i.e., the contact does not exhibit any dimensional degradation trend), let all That is, each dimension is equally weighted. Degradation contribution weight. This allows dimensions with higher drift rates to dominate the overall score, thus automatically focusing on the feature direction where degradation is most significant.

[0042] 3. Normalization bias Compared with the overall degradation score Calculation Current values ​​of each component relative detection threshold The normalization bias is:

[0043] This indicates that the component in this dimension has exceeded the detection threshold. This indicates that the score is still within the threshold. (Comprehensive Degradation Score) The sum of weighted biases for each dimension:

[0044] The larger the value, the more severe the contact degradation. A consistently negative value indicates that all characteristics of the contact are within the normal range.

[0045] This step outputs the overall degradation score. and the contribution weight of degradation in each dimension This information is intended for use in subsequent degradation type identification and remaining lifetime prediction.

[0046] Step S05: Degradation type identification based on the joint change pattern of arc light and acousto-optic coupling features This step is based on the arc light characteristics in the five-dimensional acoustic arc feature vector ( , , The variation pattern and acoustic-optical coupling characteristics () , The combined combination of change patterns identifies the contact degradation type and outputs a degradation type label.

[0047] 1. Extraction of multidimensional feature trend criteria right , , The three arc light characteristic components are calculated separately within the sliding time window. Trend slope symbol in the next operation record (Positive sign indicates an upward trend, negative sign indicates a downward trend, and zero indicates no significant trend) and coefficient of variation .right , Calculate the sign of the trend slope of each of the two acousto-optic coupling characteristic components. The five-dimensional trend slope signs mentioned above are combined with the three-dimensional coefficient of variation to form a discriminant vector.

[0048] 2. Physical mechanisms and characteristic patterns of the three types of degradation (1) Increased contact resistance. The physical cause is the accumulation of oxide film on the contact surface and hardening of the contact surface. During long-term operation, an oxide layer forms on the contact surface under the high temperature of the electric arc, and at the same time, the arc thermal effect causes local hardening of the contact surface. Both of these factors together lead to a reduction in the effective contact area and an increase in contact resistance. This manifests at the characteristic level as follows: It continues to rise due to increased contact resistance. The increase is due to the increase in arc discharge intensity. No significant changes No significant trend change (the mechanical structure of the contact has not deteriorated). Basically stable (acoustic impact characteristics unchanged).

[0049] (2) Insufficient Spring Pressure Type. Its physical root cause is spring fatigue and loose connections in the operating mechanism. After numerous operating cycles, the spring experiences metal fatigue, resulting in decreased preload and unstable contact separation speed and insufficient closing contact force. Characteristically, this manifests as: The number of cases increases due to frequent arc extinguishing and reignition caused by unstable contact separation speed. The trend decreases due to reduced mechanical impact. The lengthening trend is due to the prolonged separation process. The unstable separation rate leads to increased fluctuations between operations (increased coefficient of variation).

[0050] (3) Contact material ablation type. Its physical cause is arc thermal melting and erosion, leading to deeper pits on the contact surface and material loss, resulting in increased contact gap and prolonged breaking time. Characteristically, it manifests as: The arc duration increases due to the increased gap size, and continues to grow. The acoustic propagation path changes due to the increased gap, resulting in a trend of lengthening. The coefficient of variation increases because the irregular ablation of the surface causes the arc path to be different each time.

[0051] 3. Degradation type determination The extracted discriminant vector is matched against the preset discriminant templates for the three degradation types mentioned above. The discriminant templates are defined by the expected range of the trend slope sign and coefficient of variation for each dimension. The matching degree is measured by the proportion of the dimension meeting the criteria to the total criterion dimensions. The result with the highest matching degree is selected. If the highest matching degree is lower than a preset confidence threshold, an "uncertain" label is output. If the matching degrees for multiple degradation types are similar, a "compound degradation" label is output, and each candidate type is listed.

[0052] This step outputs a degradation type label, which is then used in conjunction with subsequent remaining lifetime prediction steps.

[0053] Step S06: Predicting remaining lifespan based on the growth trend of the comprehensive degradation score This step is based on the overall degradation score. As the number of operations increases, the remaining number of operations before the extrapolation contact reaches the maintenance threshold provides maintenance personnel with a time window for preventative maintenance.

[0054] 1. Overall degradation score growth slope Calculation Take the overall degradation score In recent The sequence in the next operation record Using the operation number as the independent variable and the comprehensive degradation score as the dependent variable, a linear fit was performed using the least squares method. The slope obtained from the fit is the growth slope. . This indicates that the degradation trend continues to worsen. The larger the value, the faster the degradation rate; This indicates that the contact is not currently showing a deterioration trend or that the deterioration trend has slowed down.

[0055] 2. Remaining number of operations extrapolation when At that time, calculate the remaining number of operations using linear extrapolation:

[0056] in The preset contact maintenance threshold represents the critical level at which the contact degradation has reached the point where maintenance or replacement is necessary. The determination of the contact life is based on the equipment manufacturer's contact life specifications or historical operation and maintenance experience, and is usually set when the equipment is first put into operation. If the contact does not currently show a degradation trend, the remaining number of operations is not output, and only the degradation type label is retained as reference information.

[0057] 3. Output Results This step ultimately outputs the estimated number of remaining operations. And the degradation type label obtained in step S05. It provides maintenance personnel with quantified maintenance time windows, and the degradation type label indicates the main cause of degradation. The combination of the two can guide the development of targeted repair plans: for the increased contact resistance, focus on checking the oxidation of the contact surface and polishing or replacing the contact plate; for the insufficient spring pressure, focus on checking the operating mechanism spring and connecting parts and adjusting or replacing them; for the contact material ablation, focus on assessing the thickness of the remaining contact material and replacing the contact.

[0058] This invention also proposes a device for detecting arc degradation of contacts in a distribution box. This device is used to perform the aforementioned method for detecting arc degradation of contacts in a distribution box, such as... Figure 3 As shown, it includes a multimodal acquisition module, a feature extraction module, an adaptive threshold module, a degradation assessment module, and a prediction output module.

[0059] The multimodal acquisition module includes three types of sensor units: an arc voltage sensor, an arc current sensor, and a piezoelectric acoustic sensor, as well as a unified clock synchronization trigger control unit. The arc voltage sensor is installed near the contact gap to collect the arc voltage signal generated in the contact gap during opening and closing operations. Its measurement range covers the arc voltage amplitude during the process from normal contact to complete separation of the contacts. The arc current sensor uses a Hall effect current sensor or a Rogowski coil, installed on the conductor of the circuit where the contacts are located, to collect the arc current signal flowing through the contact circuit during arc discharge. The piezoelectric acoustic sensor is installed on the outer surface of the distribution box housing to collect the acoustic impact signal generated at the moment of contact opening and closing. Its frequency response range covers the main frequency bands of contact mechanical impact and arc acoustic radiation. A unified clock synchronization trigger control unit provides the same hardware trigger signal and clock reference for the three acquisition channels, ensuring a sampling rate of no less than 100kHz and eliminating time base deviation between channels. Each opening and closing operation is uniformly triggered by this control unit to complete a full multimodal data acquisition window.

[0060] The feature extraction module receives three synchronous raw signals output by the multimodal acquisition module and calculates a five-dimensional acoustic arc feature vector from them. , , , , The module performs the following operations sequentially: First, it determines the start and end times of the arc duration based on the arc voltage signal. and Extract the peak voltage of the arc within this time interval. Then, the arc voltage signal and the contact arc current signal are multiplied point by point and... The accumulated energy of the arc light is obtained by summing within the interval. ; Statistical arc voltage signal in Number of zero crossings within the interval The envelope of the acoustic impact signal is extracted, and the difference between the peak time of the envelope and the arc trigger time is used to obtain the acousto-optic time delay difference. Finally, the acousto-optic amplitude ratio was calculated using the ratio of the peak amplitude of the acoustic impact envelope to the peak voltage of the arc. The feature extraction module passes the five-dimensional acoustic arc feature vector to the adaptive thresholding module and the degradation assessment module.

[0061] The adaptive thresholding module receives the five-dimensional acoustic arc feature vector output by the feature extraction module and maintains a storage array of recent continuous features. A sliding window cache for the eigenvalues ​​of each operation. This module calculates the inter-operation dispersion for each component dimension. Compare it with the pre-defined upper limit of dispersion during the initial commissioning phase. Comparison to obtain adaptive coefficients And update the detection thresholds for each dimension in an exponentially smooth manner. Calibration parameters during the initial commissioning phase ( Threshold sensitivity coefficient Initial detection threshold

[0062] This module automatically calibrates and stores the data during the initial commissioning of the equipment. The adaptive threshold module updates the five-dimensional detection threshold. Transferred to the degradation assessment module.

[0063] The degradation assessment module receives the five-dimensional acoustic arc feature vector output by the feature extraction module and the five-dimensional detection threshold output by the adaptive threshold module, and performs two functions. The first function is a comprehensive degradation score: linear fitting of each component within a sliding time window to obtain the drift rate. Normalized to the degradation contribution weight Calculate the normalized deviation of each dimension. The weighted sum is used to obtain the overall degradation score. The second function is degradation type identification: It calculates the trend slope sign and coefficient of variation for each component to form a discrimination vector, which is then matched with preset discrimination templates for three degradation types (increased contact resistance, insufficient spring pressure, and contact material ablation) to output a degradation type label. The degradation assessment module will then integrate the degradation score. The degradation type label is passed to the prediction output module.

[0064] The prediction output module receives the comprehensive degradation score output by the degradation assessment module. and degradation type tags, for In recent The linear fitting slope in the next operation record Perform calculations. When At that time, according to Extrapolation yields the estimated number of remaining operations. And output it along with the degradation type label; when If the current contact does not show a degradation trend, only a degradation type label is output for reference. The output of the prediction output module can be transmitted to the upper-level monitoring system or the local display terminal via the communication interface for maintenance personnel to make maintenance decisions.

Claims

1. A method for detecting arc degradation of contacts in a distribution box, characterized in that, Includes the following steps: Step S01: Simultaneously collect arc voltage signal, contact arc current signal and acoustic impact signal of the distribution box contacts during opening and closing operations; Step S02: Extract the five-dimensional acoustic arc feature vector from the acquired signal ( , , , , ),in This is the peak voltage of the arc. The accumulated arc energy is the time integral of the product of the arc voltage and the contact arc current. This refers to the number of zero-crossings of the arc light. The acoustic-optical time delay difference is defined as the time interval between the triggering moment of the arc voltage signal and the peak moment of the acoustic impact amplitude during the same operation. The acoustic-optic amplitude ratio is defined as the ratio of the peak amplitude of the acoustic impact to the peak voltage of the arc during the same operation. The ratio; of which and The cross-modal coupling characteristics characterize the physical relationship between the arc energy release process and the acoustic impact propagation process; Step S03: For each component of the five-dimensional acoustic arc feature vector, based on the recent continuous... In the next operation, the first Inter-operational dispersion of eigenvalues upper limit of dispersion during the initial commissioning and normal operation phase The ratio is used to calculate the adaptive coefficient. And update the detection thresholds for each dimension in an exponentially smooth manner. ;in For the recent The standard deviation of the sample size reflects the increased randomness of the arcing process caused by contact degradation. This is the threshold sensitivity coefficient, calibrated during the initial commissioning phase; Step S04: For each component of the five-dimensional acoustic arc feature vector, perform linear fitting on the changing trend of each component with the number of operations within the sliding time window, and take the absolute value of the fitting slope as the drift rate. The degradation contribution weights are obtained by normalizing the drift rates of each dimension. ; Current values ​​of each component relative detection threshold Normalization bias Calculate the overall degradation score ; Step S05: Based on the five-dimensional acoustic arc feature vector ( , , The arc characteristic pattern of () and ( , The acoustic-optic coupling characteristics of the joint change pattern are used to identify the contact degradation type and output a degradation type label. The increase in contact resistance manifests as follows: Continue to rise Increase, and No significant trend change Basically stable; insufficient spring pressure type manifests as increase, The trend is decreasing, and It shows an extending trend. Increased fluctuations in the operating area; ablation of the contact material is manifested as... Continuously increasing The trend is extending, and The coefficient of variation increases; Step S06: Based on the comprehensive degradation score As the number of operations increases, the slope is extrapolated to a preset contact maintenance threshold. Output the estimated number of remaining operations. And degradation type tags.

2. The method for detecting arc degradation of contacts in a distribution box according to claim 1, characterized in that: In step S01, the arc voltage signal is acquired by an arc voltage sensor installed near the contact gap, and the acoustic impact signal is acquired by a piezoelectric acoustic sensor installed on the distribution box housing, with a sampling rate of not less than 100kHz; the three signals are synchronously triggered by a unified clock source to eliminate time base deviation between channels.

3. The method for detecting arc degradation of contacts in a distribution box according to claim 2, characterized in that: In step S02, the arc light accumulates energy. The calculation uses arc voltage signal With contact arc current signal The product of arc duration The time integral within, i.e. ;in For the first time the arc voltage signal exceeds the preset trigger threshold At that moment, The arc voltage signal fell back to its last value. The following moments.

4. The method for detecting arc degradation of contacts in a distribution box according to claim 3, characterized in that: In step S02, the acousto-optic time delay difference The measurement was performed when the arc voltage signal first exceeded the preset trigger threshold. The moment of arrival is taken as the arc trigger reference time, and the moment when the acoustic impact signal envelope amplitude reaches its peak value in the same operation is taken as the acoustic arrival time. The difference between the two is... When the contacts are in good contact The typical value range is 0.3 to 0.

8.

5. The method for detecting arc degradation of contacts in a distribution box according to claim 1, characterized in that: In step S03, the upper limit of dispersion during the initial commissioning phase. Before the initial commissioning and normal operation phase In the next operation, the first Maximum value of the standard deviation sequence of eigenvalues; threshold sensitivity coefficient according to Calibration, among which For the first The initial detection threshold for dimensional features, This is the safety margin factor.

6. The method for detecting arc degradation of contacts in a distribution box according to claim 1, characterized in that: In step S04, when the drift rate of a certain component... Below the preset minimum drift threshold At that time, the drift rate of that dimension is set to zero so that it does not participate in the degradation contribution weight allocation, thus avoiding the impact of noise fluctuations on the overall degradation score. Interference.

7. The method for detecting arc degradation of contacts in a distribution box according to claim 1, characterized in that: In step S05, the degradation type identification employs a multi-dimensional feature trend criterion for joint determination; for , , Calculate the sign of the trend slope and the coefficient of variation within the sliding time window, respectively. , Calculate the trend slope sign for each; combine the five-dimensional trend slope sign with the coefficient of variation to form a discriminant vector, and match it with the preset discriminant templates for the three degradation types, taking the one with the highest matching degree as the recognition result.

8. The method for detecting arc degradation of contacts in a distribution box according to claim 1, characterized in that: In step S06, the remaining number of operations The calculation uses a comprehensive degradation score. In recent The linear fitting slope in the next operation record ,according to Extrapolation yields; when If the number of operations is less than or equal to zero, it is determined that the contact is not currently showing a deterioration trend, and the remaining number of operations is not output.

9. A device for detecting arc degradation of contacts in a distribution box, characterized in that, include: The multimodal acquisition module includes an arc voltage sensor, an arc current sensor, and an acoustic sensor, which are used to simultaneously acquire the arc voltage signal, the contact arc current signal, and the acoustic impact signal when the contact moves. The feature extraction module is used to calculate the five-dimensional acoustic arc feature vector from multimodal signals; The adaptive threshold module is used to continuously update the detection thresholds for each dimension based on the recent inter-operational dispersion of each component. ; The degradation assessment module is used to calculate the degradation contribution weight. Overall Degradation Score and contact degradation type labels; The prediction output module is used to extrapolate the overall degradation score. The growth trend reaches the contact repair threshold, outputting the estimated remaining number of operations. And degradation type tags.

Images