Abrasion detection method and system for plugging column of ammeter connector

Abstract

The invention relates to the field of electric digital data processing, in particular to a plugging column abrasion detection method and system for an electric meter connector, and the method comprises the steps: obtaining the historical operation state data of a plugging column, setting a label, separating abrasion-related resistors, and carrying out the spectrum analysis of a voltage fluctuation signal to obtain a spectrum sequence. Dividing a spectrum sequence into a plurality of sub-segments by calculating curvature to judge segmentation points, calculating slopes and signal-to-noise ratios of the sub-segments, performing weighted fusion according to the signal-to-noise ratios to obtain a structural noise slope, training a neural network based on wear-related resistance and the structural noise slope, constructing a wear detection network, and inputting real-time operation state data. And accurate detection of the wear state of the plugging column is realized. Through real-time online monitoring and detailed wear state classification, the detection cost is remarkably reduced, meanwhile, the accuracy and reliability of equipment maintenance are improved, the service life of equipment is prolonged, and stable operation is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing. In particular, it relates to a method and system for detecting wear on the connector pins of an electricity meter. Background Technology

[0002] The meter's terminals are key components connecting the meter to external circuits. They mainly consist of conductive parts, insulating parts, and connecting terminals, used to transmit current, voltage, and signals, ensuring the meter's measurement accuracy and reliability. Their performance directly affects the meter's stable operation; however, long-term use can lead to wear, poor contact, corrosion, and other problems, requiring regular inspection and maintenance.

[0003] The meter terminals are a crucial component of the electricity meter system, and their performance directly affects the accuracy and reliability of the meter. Therefore, wear inspection of the meter terminals is essential, as it effectively detects potential faults, ensures the normal operation of the meter, and safeguards the stability and security of the power system.

[0004] However, numerous problems exist in the actual testing of electricity meter pin wear. Offline testing requires power outages and meter removal, which is not only costly but also prevents continuous online monitoring, severely impacting testing efficiency and timeliness. While infrared thermal imaging can detect anomalies to some extent, its sensitivity to early degradation phenomena (less than 10 mΩ) is insufficient due to the meter casing's obstruction and heat dissipation limitations, making it difficult to detect potential problems in a timely manner. Furthermore, ultrasonic or vibration methods require additional sensor installation, increasing installation complexity and making them susceptible to environmental noise interference, affecting the accuracy of the test results. Although current surge testing can be performed online, its inability to effectively distinguish between load fluctuations and actual wear conditions leads to a high false alarm rate, reducing the reliability of the test. Summary of the Invention

[0005] To address the problems of high cost and inability to monitor wear on actual electricity meter terminals during offline testing, insufficient sensitivity of infrared thermal imaging, complex installation and susceptibility to interference of ultrasonic or vibration methods, and high false alarm rate of current surge method, this invention provides solutions in the following aspects.

[0006] In a first aspect, a method for detecting wear on the plug terminals of an electricity meter includes: acquiring historical operating status data of the plug terminal nodes and setting tags; wherein the historical operating status data includes voltage fluctuation signals, contact resistance, and temperature data, and the tags include polishing, normal, and oxidation; separating wear-related resistance from the contact resistance of the plug terminal; performing spectral analysis on the voltage fluctuation signal to obtain a spectral sequence, analyzing the rate of change of the spectral sequence to calculate curvature, and determining whether a data point is a segmentation point based on the curvature; pre-setting the data range of the segmentation points to obtain multiple segmentation combinations, and dividing the spectral sequence into multiple sub-segments based on the segmentation combinations, traversing and calculating the residual sum of squares under each segmentation combination with a preset step size, and selecting the combination with the smallest residual sum of squares as the target segmentation combination; calculating the slope and signal-to-noise ratio of each sub-segment corresponding to the target segmentation combination, and weighting and fusing the sub-segment slopes according to the signal-to-noise ratio to obtain the structural noise slope; training a preset neural network based on the wear-related resistance and the structural noise slope to obtain a wear detection network, and inputting real-time operating status data into the wear detection network to obtain the wear status of the plug terminal.

[0007] By fusing voltage fluctuation signals, contact resistance, and temperature data, wear-related resistances are accurately separated. Spectral analysis of the voltage fluctuation signals is performed to calculate curvature and identify segmentation points. The spectral sequence is then divided, and the slope and signal-to-noise ratio of each segment are calculated. Finally, the structural noise slope is obtained through weighted fusion. Based on this, a neural network is trained to construct a wear detection model, enabling real-time and accurate monitoring of the wear state of connector pins. This achieves online monitoring, reduces detection costs, improves detection accuracy and reliability, provides a scientific basis for equipment maintenance, extends equipment life, and ensures stable operation.

[0008] Preferably, the calculation method for the wear-related resistance includes: Obtain the difference between the current temperature and the initial temperature of the plug, and use the difference as the temperature change. Add 1 to the sum of the product of the temperature change and the temperature coefficient of resistance, and then multiply the sum by the initial resistance value to obtain the temperature-related resistance. Use the difference between the total contact resistance of the plug measured in real time and the temperature-related resistance as the real-time wear-related resistance.

[0009] By calculating the temperature-dependent resistance of the connector and separating the wear-dependent resistance from the total contact resistance, the resistance change information caused by wear can be accurately extracted, thus more accurately reflecting the actual wear degree of the connector. This calculation method takes into account the influence of temperature changes on resistance, avoids misjudgments caused by temperature fluctuations, and improves the accuracy and reliability of wear detection.

[0010] Preferably, the step of obtaining the spectral sequence includes: Select a preset window function to divide the time-domain signal into multiple time segments. Apply a short-time Fourier transform to the signal in each time segment to convert the time-domain signal into a frequency-domain signal, and obtain the frequency components and amplitudes within the time segment. The power spectral density of each time segment is obtained by taking the square of the modulus of the transformation result within each time segment. The power spectral densities of all time segments are combined to construct a spectrum sequence, where each row corresponds to a time segment and each column corresponds to a frequency point.

[0011] It can effectively capture the frequency characteristics of voltage fluctuation signals in different time segments, providing rich frequency domain information for subsequent wear state analysis, which helps to more accurately identify the wear characteristics of the plug, thereby improving the sensitivity and accuracy of wear detection.

[0012] Preferably, the method for obtaining the segmentation point includes: The frequency axis in the spectral sequence is converted to a logarithmic scale, the logarithmic frequency of each frequency point is calculated, and the natural logarithm of the power spectral density at each frequency point in the spectral sequence is taken to obtain the logarithmic power spectral density. The first and second derivatives of the logarithmic power spectrum sequence are calculated, and the absolute value of the second derivative is normalized to obtain the curvature. Data points with curvature greater than a preset threshold are used as segmentation points.

[0013] Curvature is obtained by using the second derivative to accurately identify abrupt changes or inflection points in the spectral sequence. These points typically correspond to significant changes in signal characteristics. Data points with curvature greater than a preset threshold are identified as segmentation points, which can accurately divide the spectral sequence into different characteristic segments. This provides a more reliable basis for subsequent signal analysis and wear condition assessment, improves the sensitivity to detecting changes in signal characteristics, enhances the ability to identify the wear condition of connector pins, and improves the accuracy and reliability of the detection.

[0014] Preferably, the step of obtaining the partition combination includes: Using any dividing point as a reference, a preset number of data points are obtained on the left and right sides of the dividing point to form a target partition set. Based on each data point in the target partition set, a corresponding partition combination is generated.

[0015] Preferably, the signal-to-noise ratio is calculated using the following methods: For each segment, the signal energy and residual energy are calculated separately, and the ratio of signal energy to residual energy is used as the signal-to-noise ratio of each segment.

[0016] Preferably, the model structure of the preset neural network includes an input layer, a hidden layer, and an output layer. The input layer contains two neurons, corresponding to wear-related resistance and structural noise slope, respectively. The hidden layer is designed with several neurons as needed. The output layer contains three neurons, corresponding to polishing-dominated, normal, and oxidation-dominated states, respectively. The wear-related resistance and structural noise slope of each data point in the spectrum sequence corresponding to the historical operating state data are used as training set data to train the neural network. The weights and biases of the network are adjusted to ensure that the network can accurately classify the wear state, which includes polishing-dominated state, normal state, and oxidation-dominated state.

[0017] Secondly, a wear detection system for the plug pins of an electricity meter connector includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned wear detection method for the plug pins of the electricity meter connector is implemented.

[0018] The present invention has the following effects: 1. This invention acquires historical operational status data of the plug-in nodes and processes and analyzes it using a neural network model, enabling real-time monitoring of the wear status of the meter plug-in nodes during actual operation. Compared to traditional offline testing, which requires frequent disassembly and interruption of operation, this invention avoids the problems of equipment downtime and reduced work efficiency caused by offline testing, significantly reducing testing costs.

[0019] 2. This invention can not only determine whether the connector pins are worn, but also further subdivide the wear state into polishing-dominated states, normal states, and oxidation-dominated states. This detailed classification helps to more accurately understand the actual condition of the connector pins, facilitating targeted maintenance measures. Maintenance personnel can formulate reasonable maintenance plans based on the test results, prevent potential failures in advance, extend equipment lifespan, and improve the stability and reliability of equipment operation. Attached Figure Description

[0020] Figure 1 This is a flowchart of steps S1-S4 in a method for detecting wear on the plug pins of an electric meter connector according to an embodiment of the present invention.

[0021] Figure 2 This is a structural block diagram of a wear detection system for the plug pins of an electric meter connector according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0023] Reference Figure 1 A method for detecting wear on the pins of an electricity meter connector includes steps S1-S4, as detailed below: S1: Obtain historical operating status data of the plug-in node and set tags; the historical operating status data includes: voltage fluctuation signal, contact resistance and temperature data, and the tags include polishing, normal and oxidation.

[0024] It should be noted that the meter terminals are key components connecting the meter to external circuits, primarily functioning to transmit current and voltage signals, and are typically made of metal. When meter terminals wear down, their contact resistance and interface roughness change, leading to fluctuations in the terminal's node voltage. Changes in contact resistance can reflect the degree of wear, while structural noise within the node voltage fluctuations can indicate the type of wear. It's important to note that contact resistance is not only related to wear but also affected by temperature; therefore, collecting temperature data can provide a basis for subsequently eliminating temperature-related resistance variations.

[0025] Voltage fluctuation signal: A high-precision analog front-end circuit is set inside the meter to collect the voltage signal of the contact area of ​​the metal contact of the plug at a sampling frequency of 300Hz. The built-in 0.2Hz-120Hz bandpass filter retains the structural noise, ensuring that the structural noise components in the 1Hz-100Hz frequency band are completely captured, and suppressing low-frequency drift and high-frequency aliasing noise.

[0026] Contact resistance: A four-wire micro-ohmmeter is set up near the meter's plug-in terminal. A sinusoidal micro-current with an amplitude not exceeding 10mA and a frequency not less than 1kHz is injected into the plug-in terminal. The potential difference between the two ends of the plug-in terminal is collected synchronously through two pairs of independent leads. The Kelvin connection method is used to eliminate the influence of lead resistance. The contact resistance of the plug-in terminal is obtained after digital processing by an ADC (Analog-to-Digital Converter).

[0027] Temperature data: A high-precision, small-sized PT1000 platinum resistance temperature sensor was used for temperature acquisition. The sampling frequency was set to 100Hz, and the temperature of the plug was obtained after ADC digital processing.

[0028] Based on the actual wear condition of the connector pins, the collected data is labeled with three states: polished, normal, and oxidized. The labels are set by those skilled in the art based on the condition of the connector pins.

[0029] Wear on the meter terminals affects their contact resistance and interface roughness, leading to node voltage fluctuations. Changes in contact resistance can reflect the degree of wear, while structural noise in node voltage fluctuations can reflect the type of wear. Since contact resistance is also affected by temperature, temperature data is collected to subsequently filter out temperature-dependent resistance variations.

[0030] The two wear modes, interface oxidation and micro-pit polishing, have opposite effects on contact resistance and structural noise. A single parameter is insufficient to accurately determine the wear state. It is necessary to comprehensively consider the changes in contact resistance and structural noise, and combine them with other factors such as temperature for analysis.

[0031] S2: Separate the wear-related resistance from the contact resistance of the plug-in pin; perform spectral analysis on the voltage fluctuation signal to obtain the spectral sequence, analyze the rate of change of the spectral sequence to calculate the curvature, and determine whether the data point is a segmentation point based on the curvature.

[0032] The calculation methods for wear-related resistance include: Obtain the difference between the current temperature and the initial temperature of the plug, and use the difference as the temperature change. Add 1 to the sum of the product of the temperature change and the temperature coefficient of resistance, and then multiply the sum by the initial resistance value to obtain the temperature-related resistance. Use the difference between the total contact resistance of the plug measured in real time and the temperature-related resistance as the real-time wear-related resistance.

[0033] Specifically, the temperature-dependent resistance satisfies the following relationship: ; In the formula, Indicates the first Temperature-dependent resistance at any given time This indicates the initial resistance value of the connector pin. The temperature coefficient of resistance of the connector pin is determined by the actual material of the connector pin. Indicates the first Temperature of the connector pin at any given time. This indicates the measured temperature of the plug during initial use.

[0034] Specifically, the wear-related resistance satisfies the following relationship: ; In the formula, Indicates the first Wear-related resistance over time Indicates the first Total contact resistance at any given time Indicates the first Temperature-dependent resistance at any given time.

[0035] It should be noted that the measured temperature and contact resistance when the meter was first put into use are used as a reference. The theoretical resistance value at the current temperature is calculated by combining the temperature coefficient of resistance of the plug material. The measured contact resistance is then subtracted from the theoretical value to obtain the wear-related resistance that purely reflects the wear state.

[0036] Select a preset window function to divide the time-domain signal into multiple time segments. Apply a short-time Fourier transform to the signal in each time segment to convert the time-domain signal into a frequency-domain signal, and obtain the frequency components and amplitudes within the time segment. The power spectral density of each time segment is obtained by taking the square of the modulus of the transformation result within each time segment. The power spectral densities of all time segments are combined to construct a spectrum sequence, where each row corresponds to a time segment and each column corresponds to a frequency point.

[0037] For example, a preset window function (such as a Hanning window or a rectangular window) with a length of 2 seconds can be selected and adjusted as needed to analyze the frequency components of the signal and their changes over time in the frequency domain.

[0038] It should be noted that the Fourier transform is used to obtain the spectral sequence in this embodiment. It can also be obtained by wavelet transform and autocorrelation coefficient. The above methods are well known to those in the art and will not be described in detail.

[0039] The power spectral density was obtained through short-time Fourier transform, and the structural noise slope was accurately extracted using a piecewise weighted fitting method to avoid bias caused by single-band fitting. Simultaneously, the least squares algorithm was used to suppress outlier interference, ensuring the accuracy of the slope extraction. Since different wear modes (interface oxidation and micro-pit polishing) have opposite effects on the structural noise slope, the change in the structural noise slope becomes a key indicator for distinguishing wear types, and its magnitude also reflects the rate of wear change.

[0040] The frequency axis in the spectral sequence is converted to a logarithmic scale, the logarithmic frequency of each frequency point is calculated, and the natural logarithm of the power spectral density at each frequency point in the spectral sequence is taken to obtain the logarithmic power spectral density. The first and second derivatives of the logarithmic power spectrum sequence are calculated, and the absolute value of the second derivative is normalized to obtain the curvature. Data points with curvature greater than a preset threshold are used as segmentation points.

[0041] For example, the preset threshold is 0.25, which can be adjusted according to specific circumstances.

[0042] By separating wear-related resistances, the influence of temperature on contact resistance is eliminated, making subsequent wear condition assessments more accurate. Spectral analysis reveals the frequency components in voltage fluctuation signals and their changes over time, providing a basis for determining wear types. The calculation of curvature and the determination of segmentation points help divide the spectral sequence into different sub-segments for more detailed analysis of signal characteristics across different frequency ranges.

[0043] S3: Set the data range of the preset segmentation point to obtain multiple segmentation combinations, and divide the spectrum sequence into multiple sub-segments based on the segmentation combinations. Use a preset step size to traverse and calculate the residual sum of squares under each segmentation combination, and select the combination with the smallest residual sum of squares as the target segmentation combination. Calculate the slope and signal-to-noise ratio of each sub-segment corresponding to the target segmentation combination, and weight and fuse the sub-segment slopes according to the signal-to-noise ratio to obtain the structural noise slope.

[0044] Using any dividing point as a reference, a preset number of data points are obtained on the left and right sides of the dividing point to form a target partition set. Based on any data point in each target partition set as the target point, the target points in each target partition set are obtained in turn to generate the corresponding partition combination.

[0045] Calculate the slope and signal-to-noise ratio of each segment corresponding to the target division combination, and then weight and fuse the slopes of the segments according to the signal-to-noise ratio to obtain the structural noise slope.

[0046] For each segment, the signal energy and residual energy are calculated separately, and the ratio of signal energy to residual energy is used as the signal-to-noise ratio of each segment.

[0047] By analyzing the power spectrum characteristics of voltage fluctuation signals, a piecewise weighted fitting method is used to accurately extract the structural noise slope and calculate its change relative to the initial state. This change effectively reflects the change in the micro-roughness of the contact interface and is a key indicator for determining the wear type (interface oxidation or micro-pit polishing). Furthermore, this invention uses the signal-to-noise ratio of each frequency band as a weight for comprehensive analysis, significantly improving the reliability of slope extraction and providing an accurate data foundation for the subsequent construction of the wear index, thereby achieving precise monitoring and evaluation of the wear state.

[0048] By traversing the data range of preset segmentation points and preset step sizes, multiple segmentation combinations are generated. This fully considers the impact of different segmentation methods on spectral sequence analysis, thereby finding the optimal segmentation combination. The calculation of the residual sum of squares is used to evaluate the fitting effect of different segmentation combinations. Selecting the combination with the smallest residual sum of squares as the target segmentation combination ensures a more reasonable division of the spectral sequence. The slope and signal-to-noise ratio of each segment are calculated, and the slopes of the segments are weighted and fused according to the signal-to-noise ratio. This comprehensively considers the signal quality and noise level of different segments, obtaining a more accurate structural noise slope, and providing a more reliable data foundation for subsequent wear status judgment.

[0049] S4: Train a preset neural network based on wear-related resistance and structural noise slope to obtain a wear detection network. Input real-time operating status data into the wear detection network to obtain the wear status of the plug.

[0050] The pre-defined neural network model structure includes an input layer, a hidden layer, and an output layer. The input layer contains two neurons, corresponding to the wear-related resistance and the structural noise slope, respectively. The hidden layer is designed with several neurons as needed. The output layer contains three neurons, corresponding to polishing-dominated, normal, and oxidation-dominated states, respectively.

[0051] The wear-related resistance and structural noise slope of each data point in the spectrum sequence corresponding to the historical operating status data are used as training data to train the neural network. The weights and biases of the network are adjusted to ensure that the network can accurately classify the wear state, which includes: polishing-dominated polishing state, normal state, and oxidation-dominated oxidation state.

[0052] Neural networks can learn the features and patterns in historical operating status data. The wear detection network obtained through training can accurately classify real-time operating status data, thereby achieving high-precision detection of the wear status of the plug pins.

[0053] By using wear-related resistance and structural noise slope as inputs to a neural network, the ability of these two key parameters to characterize the wear state can be fully utilized, thereby improving the accuracy and reliability of wear detection.

[0054] This invention also provides a system for detecting wear on the connector pins of an electricity meter. For example... Figure 2 As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a method for detecting wear on the plug pins of an electricity meter connector according to the first aspect of the present invention. The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, the settings and functions of which are known in the art and will not be described further here.

[0055] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for detecting wear on the pins of an electricity meter connector, characterized in that, include: Acquire historical operating status data of the plug-in terminal and set tags; the historical operating status data includes: voltage fluctuation signal, contact resistance and temperature data, and the tags include polishing, normal and oxidation; Wear-related resistance is obtained by separating the contact resistance of the plug-in pins; the voltage fluctuation signal is subjected to spectral analysis to obtain the spectral sequence, the rate of change of the spectral sequence is analyzed to calculate the curvature, and the data point is determined as a segmentation point based on the curvature. The data range of the preset segmentation points is used to obtain multiple segmentation combinations. Based on the segmentation combinations, the spectrum sequence is divided into multiple sub-segments. The residual sum of squares under each segmentation combination is calculated by traversing with a preset step size. The combination with the smallest residual sum of squares is selected as the target segmentation combination. The slope and signal-to-noise ratio of each sub-segment corresponding to the target segmentation combination are calculated. The slope of the sub-segment is weighted and fused according to the signal-to-noise ratio to obtain the structural noise slope. A pre-set neural network is trained based on wear-related resistance and structural noise slope to obtain a wear detection network. Real-time operating status data is input into the wear detection network to obtain the wear status of the plug.

2. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The calculation method for the wear-related resistance includes: Obtain the difference between the current temperature and the initial temperature of the plug, and use the difference as the temperature change. Add 1 to the sum of the product of the temperature change and the temperature coefficient of resistance, and then multiply the sum by the initial resistance value to obtain the temperature-related resistance. Use the difference between the total contact resistance of the plug measured in real time and the temperature-related resistance as the real-time wear-related resistance.

3. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The steps for obtaining the spectral sequence include: Select a preset window function to divide the time-domain signal into multiple time segments. Apply a short-time Fourier transform to the signal in each time segment to convert the time-domain signal into a frequency-domain signal, and obtain the frequency components and amplitudes within the time segment. The power spectral density of each time segment is obtained by taking the square of the modulus of the transformation result within each time segment. The power spectral densities of all time segments are combined to construct a spectrum sequence, where each row corresponds to a time segment and each column corresponds to a frequency point.

4. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The methods for obtaining the segmentation points include: The frequency axis in the spectral sequence is converted to a logarithmic scale, the logarithmic frequency of each frequency point is calculated, and the natural logarithm of the power spectral density at each frequency point in the spectral sequence is taken to obtain the logarithmic power spectral density. The first and second derivatives of the logarithmic power spectrum sequence are calculated, and the absolute value of the second derivative is normalized to obtain the curvature. Data points with curvature greater than a preset threshold are used as segmentation points.

5. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The steps for obtaining the partition combination include: Using any dividing point as a reference, a preset number of data points are obtained on the left and right sides of the dividing point to form a target partition set. Based on each data point in the target partition set, a corresponding partition combination is generated.

6. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The signal-to-noise ratio is calculated using the following methods: For each segment, the signal energy and residual energy are calculated separately, and the ratio of signal energy to residual energy is used as the signal-to-noise ratio of each segment.

7. The method for detecting wear on the plug pins of an electricity meter connector according to claim 1, characterized in that, The preset neural network model structure includes an input layer, a hidden layer, and an output layer. The input layer contains two neurons, corresponding to wear-related resistance and structural noise slope, respectively. The hidden layer is designed with several neurons as needed. The output layer contains three neurons, corresponding to polishing-dominated, normal, and oxidation-dominated states, respectively. The wear-related resistance and structural noise slope of each data point in the spectrum sequence corresponding to the historical operating state data are used as training data to train the neural network. The weights and biases of the network are adjusted to ensure that the network can accurately classify wear states, including polishing-dominated state, normal state, and oxidation-dominated state.

8. A system for detecting wear on the plug pins of an electricity meter connector, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the method for detecting pin wear of an electricity meter connector according to any one of claims 1-7.

Images