Integrated constant torque intelligent T-type wire clamp online monitoring and early warning system

Abstract

The present application relates to state alarm technical field, specifically to integrated constant torque intelligent T type wire clamp online monitoring and early warning system, the system comprises: reference construction module uses steady state data to fit non-radiation heat response reference, radiation decoupling module calculates residual and smooths to remove solar radiation component, thermal resistance analysis module generates transient thermal resistance characteristic spectrum by using current mutation decomposition, loosening judgment module calculates contact layer heat response deviation according to minimum time constant term, fatigue prediction module accumulates cold state impedance increment and generates early warning information in combination with thermal deviation, in the present application, by constructing the regression reference of current square term and temperature rise, using residual smoothing, effectively separating solar radiation interference, adopting multiple exponential decomposition technology, accurately stripping the contact layer and thermal characteristics, quantitatively evaluating the clamp fastening state according to the decay rate deviation, and combining the cumulative damage of cold state impedance, the deterioration trend of the equipment is deeply analyzed and predicted, and early warning information is output.

Description

Technical Field

[0001] This invention relates to the field of status alarm technology, and in particular to an integrated constant torque intelligent T-type wire clamp online monitoring and early warning system. Background Technology

[0002] The field of status alarm technology involves a comprehensive technical system for real-time detection of system equipment operating parameters or environmental conditions, triggering warning signals when the detected indicators exceed safety thresholds. In overhead power distribution lines, T-type clamps or parallel groove clamps are key hardware for ensuring power transmission and the connection of lead wires. Their contact performance directly affects the reliability of power supply. In early applications, distribution lines mainly used traditional JB-type parallel groove clamps or piercing clamps, coupled with a maintenance mode of regular manual inspections. Traditional clamps were mostly asymmetrical in structure, and the installation quality depended entirely on the operators. The feel of the wire clamp is crucial; using a regular wrench to tighten it can easily lead to loose connections due to insufficient torque, or damage to the wires due to excessive torque. Maintenance personnel primarily rely on handheld infrared thermometers or observing temperature-indicating wax strips to diagnose faults. This method is not only limited by inspection cycles and has monitoring blind spots, but is also highly susceptible to environmental interference. For example, under strong solar radiation or windy conditions, the surface temperature of the clamp cannot accurately reflect the heating status of the internal contact points, leading to missed or incorrect fault diagnoses. To fundamentally solve the reliability problem of physical connections, current improved technology has developed and applied an integrated constant temperature... The torque-intelligent T-clamp features a significantly optimized mechanical structure. It employs a specially designed shear-type bolt structure, where the torque head on the nut automatically breaks off when the tightening torque reaches a preset standard value (e.g., a specific number of Newton-meters), ensuring consistent installation torque at the hardware level and eliminating human error. It also utilizes a symmetrical flow-guiding structure to accommodate bidirectional power flow and integrates a mechanical overheat ejection device. When the clamp temperature is too high, an internal temperature-changing element drives a red warning tube to pop out, providing a visual indication of overheating faults. Based on the stable mechanical connection characteristics and visual alarm function of this integrated constant torque intelligent T-clamp, an online monitoring and early warning system is proposed to further enhance the digital perception capabilities of the power distribution network. This system combines the standardized contact state of the constant torque clamp with a radiation decoupling algorithm to eliminate environmental interference, uses transient thermal resistance analysis to deeply analyze the microscopic contact changes inside the clamp, and predicts metal fatigue life based on historical thermal cycle data. This upgrades the system from a single mechanical post-event warning to all-weather, high-precision proactive intelligent prediction, accurately quantifying the clamp's health and remaining lifespan. Summary of the Invention

[0003] To address the technical problems existing in the prior art, this invention provides an integrated constant torque intelligent T-type wire clamp online monitoring and early warning system, which includes: The benchmark construction module filters the clamp temperature monitoring data in the range of no light and stable load current, calculates the real-time temperature rise by combining the ambient temperature, fits the square term of load current with the real-time temperature rise, and generates a non-radiative thermal response benchmark. The radiation decoupling module calculates the theoretical temperature rise under the non-radiative thermal response benchmark, calculates the deviation between the real-time temperature rise data and the theoretical temperature rise to obtain a residual sequence and smooths it, obtains the solar radiation temperature rise component, subtracts the component from the real-time temperature rise to obtain the radiation decoupling temperature rise data; The thermal resistance analysis module extracts the radiation decoupling temperature rise data segment when the current change rate data exceeds the standard, and generates a transient thermal resistance characteristic spectrum through multiple exponential fitting decomposition. The loosening judgment module marks the minimum component of the transient thermal resistance characteristic spectrum as the contact layer thermal response characteristic, calculates the deviation between the decay rate of the contact layer thermal response characteristic and the standard data, and generates thermal response deviation data. The fatigue prediction module calculates the ratio of radiation decoupling temperature rise data to current at the valley point, obtains cold-state equivalent contact impedance data, and accumulates the increment to obtain cumulative damage data. Combined with the thermal response deviation data, it calculates the clamp health index and outputs clamp maintenance early warning information.

[0004] As a further aspect of the present invention, the non-radiative thermal response benchmark includes the Joule thermal conversion coefficient, the convective heat dissipation coefficient, and the linear regression determination coefficient; the radiation decoupling temperature rise data includes the net Joule thermal temperature rise sequence, the synchronization timestamp sequence, and the residual correction factor; the transient thermal resistance characteristic spectrum includes the differential structure function curve, the integral structure function curve, and the thermal capacity and thermal resistance distribution map; the thermal response deviation data includes the absolute deviation value of the decay rate, the relative change rate of the amplitude weight, and the thermal response consistency coefficient; and the clamp maintenance early warning information includes the early warning level identifier, the estimated remaining lifespan, and the tightening torque adjustment parameters.

[0005] As a further aspect of the present invention, the benchmark construction module includes: The data filtering and acquisition submodule monitors the surface temperature of the clamp, ambient temperature and load current in real time and records the timestamp information corresponding to the sampling time. Based on the timestamp information, it extracts the monitoring data within the time interval without light, calculates the fluctuation variance of the load current and filters the electrical stable operation data segment with fluctuation variance less than the preset stable threshold, and generates a pure steady-state monitoring sample. The temperature difference calculation and processing submodule, based on the pure steady-state monitoring sample, indexes and analyzes the surface temperature of the clamp and the ambient temperature at the same sampling time, calculates the temperature difference between the two, constructs a relative temperature rise time series according to the time order, and generates a measured temperature rise data series. The benchmark fitting generation submodule calls the pure steady-state monitoring sample, extracts the monitored load current, performs a power operation on the load current to obtain the current square term value, constructs the current square term value into a regression input vector in time order, calls the measured temperature rise data sequence as the regression output vector, performs binary linear regression fitting calculation, calculates the slope coefficient and intercept parameter of the regression equation, and generates a non-radiative thermal response benchmark.

[0006] As a further aspect of the present invention, the process of obtaining the preset stability threshold is specifically as follows: The historical load current monitoring sequence of the clamp is retrieved, and the historical load current monitoring sequence is sliced ​​using a sliding time window to construct multiple current sampling segments containing continuous sampling points. The numerical variance of each current sampling segment is calculated to construct a variance dataset characterizing the current fluctuation state distribution. A Gaussian mixture model is used to estimate the probability density of the variance dataset and fit it to generate a bimodal distribution curve containing steady-state characteristic peaks and transient characteristic peaks. The steady-state characteristic peak located in the low value range of the bimodal distribution curve is identified, and the mathematical expectation and standard deviation of the steady-state characteristic peak are analyzed. The sum of the mathematical expectation and three times the standard deviation is calculated, and the sum is set as the preset stability threshold.

[0007] As a further aspect of the present invention, the radiation decoupling module includes: The theoretical temperature rise calculation submodule obtains the clamp load current during the daytime monitoring period, and, in conjunction with the non-radiative thermal response benchmark, calculates the expected temperature rise value of the target clamp load current intensity under an ideal environment without light, generating theoretical Joule thermal temperature rise data. The radiation component extraction submodule calls the real-time temperature rise data and the theoretical Joule heat temperature rise data to perform difference calculation, constructs a residual time series, and performs weighted moving average smoothing filtering on the residual time series to generate solar radiation temperature rise components. The decoupling data generation submodule calls the real-time temperature rise data and performs time-domain data alignment with the solar radiation temperature rise component, and subtracts the solar radiation temperature rise component from the real-time temperature rise data point by point to generate radiation decoupling temperature rise data.

[0008] As a further aspect of the present invention, the thermal resistance analysis module includes: The rate of change calculation and judgment submodule collects real-time load current and calculates the difference quotient between current values ​​of adjacent sampling points, constructs a load current rate of change sequence, calls preset step trigger reference data, compares the load current rate of change sequence with the judgment threshold, filters out abrupt change time points that exceed the judgment threshold, and generates a step trigger time index vector. The transient data extraction submodule receives radiation-decoupled temperature rise data and aligns the radiation-decoupled temperature rise data with the step trigger time index vector in the time domain, locks the starting position of the step change, extracts the temperature rise response value sequence within a preset time window after the starting position, and generates a transient temperature rise evolution data segment. The feature spectrum fitting and generation submodule constructs a function expression with multiple exponential superpositions for the transient temperature rise evolution data segment, calculates the fitting residual between the function expression and the transient temperature rise evolution data segment, minimizes the fitting residual by performing iterative approximation operations on the exponential decay coefficient and amplitude coefficient in the function expression, analyzes the values ​​of multiple time constant terms and the corresponding amplitude weight terms, and generates a transient thermal resistance feature spectrum.

[0009] As a further aspect of the present invention, the process of obtaining the step trigger reference data is specifically as follows: Retrieve the load current monitoring sequence of the clamp during its historical normal operation cycle, perform a first-order difference operation on the load current monitoring sequence, extract the current change increment of adjacent sampling points, perform absolute value processing on the current change increment, construct a statistical sample of change amplitude, calculate the mathematical expectation and standard deviation of the statistical sample of change amplitude, and based on the statistical characteristics of normal distribution, calculate the sum of the mathematical expectation and 3 times the standard deviation, and set the calculated sum as the step trigger reference data.

[0010] As a further aspect of the present invention, the loosening detection module includes: The feature sorting and filtering submodule calls the transient thermal resistance feature spectrum, extracts the value of each time constant item in the spectrum, sorts the value of each time constant item in ascending order, identifies and locks the time constant item with the smallest value, and generates the index of the smallest time constant component. The contact feature extraction submodule extracts the corresponding independent exponential function component from the transient thermal resistance feature spectrum based on the minimum time constant component index, marks it as the thermal characteristic component of the contact layer, and separates the corresponding amplitude weight term value and decay rate value to generate the thermal response feature vector of the contact layer. The deviation quantization calculation submodule calls the preset fastening state reference data and thermal conductivity standard data, obtains the amplitude weight term value and decay rate value from the thermal response feature vector of the contact layer, compares the amplitude weight term value with the fastening state reference data, calculates the difference between the decay rate value and the thermal conductivity standard data, and generates thermal response deviation data.

[0011] As a further aspect of the present invention, the process of obtaining the preset fastening state reference data and thermal conductivity standard data is specifically as follows: A standard wire clamp prototype is selected, and the contact surface fastening bolts of the wire clamp prototype are tightened to the rated torque using a torque calibration device. A step current excitation is applied to the wire clamp prototype in a constant temperature environment, and the standard temperature rise response sequence generated by the wire clamp prototype is collected. Multiple exponential function superposition fitting decomposition is performed on the standard temperature rise response sequence to obtain a reference feature spectrum containing multiple independent time constant terms. The time constant term with the smallest value in the reference feature spectrum is selected, and the amplitude weight term value corresponding to the time constant term with the smallest value is extracted and set as the fastening state reference data. The decay rate value corresponding to the time constant term with the smallest value is extracted and set as the thermal conductivity standard data.

[0012] As a further aspect of the present invention, the fatigue prediction module includes: The impedance characteristic calculation submodule monitors the real-time load current of the clamp and identifies the sampling time point when the load current is at its lowest value, marking it as the cold reference point. It calculates the ratio of the radiation decoupling temperature rise data to the square term of the load current at the cold reference point, generating cold equivalent contact impedance data. The damage accumulation analysis submodule calculates the numerical difference between the cold equivalent contact impedance data in the current monitoring cycle and the adjacent previous monitoring cycle, generates single thermal cycle plastic residual data, and performs a time-based cumulative summation operation on the single thermal cycle plastic residual data to generate cumulative damage data. The integrated early warning decision submodule calls the accumulated damage data and the thermal response deviation data to construct a multi-dimensional state mapping matrix. The accumulated damage data and thermal response deviation data are mapped as state variables to the multi-dimensional state mapping matrix to perform weighted fusion calculation, generate the clamp health index, determine the clamp maintenance needs, and generate clamp maintenance early warning information.

[0013] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: A non-radiative thermal response benchmark was constructed by linear regression fitting of the load current square term and temperature rise data. By calculating the residual sequence between the real-time temperature rise and the theoretical benchmark and performing smoothing, the temperature rise interference caused by solar radiation was effectively separated and eliminated. The transient thermal resistance characteristics were analyzed using the multiple exponential decomposition technique under the step response. The thermal response characteristics of the contact layer were accurately extracted based on the time constant difference. The quantitative assessment of the fastening state was achieved by comparing the contact layer decay rate with the standard data. Combined with the analysis of the equivalent contact impedance increment at the cold trough moment, the cumulative thermal damage was analyzed, and the accurate prediction and proactive maintenance early warning of the degradation trend of the clamp contact performance were realized. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the integrated constant torque intelligent T-type wire clamp online monitoring and early warning system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the benchmark construction module in this invention; Figure 4 This is a flowchart of the radiation decoupling module in this invention; Figure 5 This is a flowchart of the thermal resistance analysis module in this invention; Figure 6 This is a flowchart of the loosening detection module in this invention; Figure 7 This is a flowchart of the fatigue prediction module in this invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0018] This invention provides an integrated constant torque intelligent T-type wire clamp online monitoring and early warning system, such as... Figure 1-2 The diagram shown illustrates an integrated constant torque intelligent T-type wire clamp online monitoring and early warning system. This system includes: The benchmark construction module filters the clamp temperature monitoring data in the range of no light and stable load current, calculates the real-time temperature rise by combining the ambient temperature, fits the square term of load current with the real-time temperature rise, and generates a non-radiative thermal response benchmark. The radiation decoupling module calculates the theoretical temperature rise under the non-radiative thermal response benchmark, calculates the deviation between the real-time temperature rise data and the theoretical temperature rise to obtain the residual sequence and smooths it, obtains the solar radiation temperature rise component, subtracts the component from the real-time temperature rise to obtain the radiation decoupling temperature rise data; The thermal resistance analysis module extracts radiation decoupling temperature rise data segments when the current change rate data exceeds the standard, and generates transient thermal resistance characteristic spectra through multiple exponential fitting decomposition. The loosening judgment module marks the minimum component of the transient thermal resistance characteristic spectrum as the thermal response characteristic of the contact layer, calculates the deviation between the decay rate of the thermal response characteristic of the contact layer and the standard data, and generates thermal response deviation data. The fatigue prediction module calculates the ratio of radiation decoupling temperature rise data to current at the valley point, obtains cold-state equivalent contact impedance data and accumulates the increment to obtain cumulative damage data. Combined with thermal response deviation data, it calculates the clamp health index and outputs clamp maintenance early warning information.

[0019] The non-radiative thermal response benchmark includes the Joule thermal conversion coefficient, convective heat dissipation coefficient, and linear regression determination coefficient. The radiation decoupling temperature rise data includes the net Joule thermal temperature rise sequence, the synchronization timestamp sequence, and the residual correction factor. The transient thermal resistance characteristic spectrum includes the differential structure function curve, the integral structure function curve, and the thermal capacity and thermal resistance distribution map. The thermal response deviation data includes the absolute deviation value of the decay rate, the relative change rate of the amplitude weight, and the thermal response consistency coefficient. The clamp maintenance early warning information includes the early warning level indicator, the estimated remaining lifespan, and the tightening torque adjustment parameters.

[0020] Specifically, such as Figure 2 , 3 As shown, the benchmark building block includes: The data filtering and acquisition submodule monitors the surface temperature of the clamp, ambient temperature and load current in real time and records the timestamp information corresponding to the sampling time. Based on the timestamp information, it extracts the monitoring data within the time interval without light, calculates the fluctuation variance of the load current and filters the electrical stable operation data segment with fluctuation variance less than the preset stable threshold, and generates a pure steady-state monitoring sample. The module retrieves the full-time load current monitoring history of the intelligent T-clamp over the past 30 consecutive monitoring days. Using a sliding time window with a time span of 15 minutes and a step size of 1 minute, the historical data is overlaid and sliced ​​to construct an analysis sequence consisting of 1440 sample sets containing current values ​​within local time periods. For each sample set, the statistical variance of all current values ​​is calculated, and the calculated variance is mapped to a fluctuation index characterizing the severity of current fluctuations within that time period. Frequency distribution statistics are performed on all fluctuation indices in the analysis sequence to construct a fluctuation probability density histogram. This module further employs a Gaussian mixture model to estimate the probability density of the fluctuation variance dataset. This model consists of two Gaussian distribution components, corresponding to the background noise distribution and the load jump distribution, respectively. Parameter iteration is performed using the expectation-maximization algorithm. In the E-step, the posterior probability of each variance data point belonging to the steady-state noise component is calculated. In the M-step, the mean, covariance, and mixture weights of the Gaussian components are updated based on the posterior probability until the increment of the log-likelihood function is less than 0.001, at which point iteration stops. The module generates a bimodal probability density curve by fitting the curve. It analyzes the background noise peak in the low-value range of the bimodal probability density curve and extracts the expected value and standard deviation of the background noise peak. For example, the expected value is 0.5 square amperes and the standard deviation is 0.1 square amperes. Based on the statistical characteristics of the normal distribution, the sum of the expected value and 3 times the standard deviation is calculated, that is, the sum of 0.5 and 3 multiplied by 0.1 is calculated to obtain 0.8 square amperes. This 0.8 square amperes is established as the preset stable threshold. During real-time operation, this module monitors the surface temperature of the clamp, ambient temperature, and load current, and records the timestamp information corresponding to the sampling time. Based on the timestamp information, it extracts the monitoring data within the period of no sunlight from 7 pm to 5 am the next day, calculates the fluctuation variance of the real-time load current, and if the current fluctuation variance within a certain 15-minute time window is 0.4 square amperes, which is less than the preset stability threshold of 0.8 square amperes, then the data segment is determined to be an electrically stable operation data segment, which is retained and a pure steady-state monitoring sample is generated; otherwise, it is discarded.

[0021] The temperature difference calculation and processing submodule, based on pure steady-state monitoring samples, indexes and parses the surface temperature of the clamp and the ambient temperature at the same sampling time, calculates the temperature difference between the two and constructs a relative temperature rise time series according to the time order, and generates a measured temperature rise data series. Based on the generated pure steady-state monitoring samples, the module iterates through the index keys in the sample database to accurately locate and parse the data pairs of clamp surface temperature and ambient temperature at the same sampling time. It then performs a subtraction operation, subtracting the ambient temperature value from the clamp surface temperature value to eliminate the influence of ambient base temperature fluctuations on the temperature rise analysis. The module arranges all calculated temperature differences according to the timestamp order, constructing a relative temperature rise time series. This is then combined with the load current data at the corresponding time to form a mapping relationship, generating a measured temperature rise data series. For example, at sampling time T1, the monitored clamp surface temperature is 45.5 degrees Celsius, and the ambient temperature is 25.2 degrees Celsius. This module performs the operation of subtracting 25.2 from 45.5 to obtain a temperature difference of 20.3 degrees Celsius, and writes this 20.3 degrees Celsius into the corresponding position of the measured temperature rise data sequence. This module performs linear interpolation to fill in any missing values ​​in the sequence that may be caused by packet loss during transmission. That is, it uses the values ​​of two known data points before and after the missing time point, combined with the time interval weight, to calculate the intermediate value. For example, if the temperature rise at the previous time point is 20.3 degrees Celsius and the temperature rise at the next time point is 20.5 degrees Celsius, the value of the missing point in the middle is calculated to be 20.4 degrees Celsius, so as to ensure the continuity and integrity of the data sequence and provide a standardized input stream without discontinuities for subsequent regression fitting.

[0022] The benchmark fitting generation submodule calls the pure steady-state monitoring sample, extracts the monitored load current and performs a power operation on the load current to obtain the current square term value, constructs the current square term value as a regression input vector in time order, calls the measured temperature rise data sequence as the regression output vector, performs binary linear regression fitting calculation, calculates the slope coefficient and intercept parameter of the regression equation, and generates a non-radiative thermal response benchmark. This module calls upon the load current data recorded in the pure steady-state monitoring samples, performs a power operation on each current sampling point (i.e., calculates the square of the current value) to reflect the physical nature of the thermal effect of the current, and arranges the obtained current square terms in chronological order to construct an N-row, 1-column regression input vector, where N is the total number of sample points. Simultaneously, this module calls upon the measured temperature rise data sequence, setting it as an N-row, 1-column regression output vector to construct a binary linear regression model. This model assumes that the temperature rise is proportional to the square of the current, i.e., the temperature rise equals the slope coefficient multiplied by the square of the current plus the intercept parameter. This module uses the least squares algorithm to perform fitting calculations, solving the coefficients of the regression equation through matrix operations. The process involves constructing the product of the transpose of the input matrix and the input matrix, inverting this product matrix, and then multiplying the inverse matrix by the transpose of the input matrix and the output vector in sequence. This allows for the analytical determination of the optimal slope coefficient and intercept parameter of the regression equation. For example, after processing a sample containing 1000 sets of data, the slope coefficient is calculated to be 0.0005 degrees Celsius per square ampere, and the intercept parameter is 0.2 degrees Celsius. This means that under ideal baseline conditions of no light and no wind, for every 1 ampere increase in current, for every 1 unit increase in the square term, the temperature rise will increase by 0.0005 degrees Celsius. This linear equation is then generated as a non-radiative thermal response baseline, which is used in subsequent steps to predict the theoretical Joule thermal temperature rise under any current.

[0023] Specifically, such as Figure 2 , 4 As shown, the radiation decoupling module includes: The theoretical temperature rise calculation submodule obtains the clamp load current during the daytime monitoring period, and, in conjunction with the non-radiative thermal response benchmark, calculates the expected temperature rise value of the target clamp load current intensity under ideal conditions without light, generating theoretical Joule thermal temperature rise data. The module acquires real-time load current data of the line clamps during the daytime monitoring period (e.g., 08:00 to 16:00 daily), and performs forward inference calculations using the non-radiative thermal response benchmark model generated in the previous stage. First, it performs a squaring operation on each load current value collected during the daytime to obtain a current square sequence. Then, it substitutes each value in this current square sequence into the linear equation of the non-radiative thermal response benchmark, multiplying the current square value by the slope coefficient 0.0005, and adding the product to the intercept parameter 0.2, thereby calculating the predicted load current at that moment under conditions of no solar radiation interference. The module calculates the theoretical Joule temperature rise for a given time period. For example, if the load current is 200 amperes at a certain moment during the day, the module calculates 200 squared to get 40,000 square amperes, multiplies it by 0.0005 to get 20 degrees Celsius, adds 0.2 degrees Celsius, and finally generates the theoretical Joule temperature rise data for that moment as 20.2 degrees Celsius. The module repeats the above process for the current data throughout the day to generate a complete theoretical Joule temperature rise data curve. This curve represents the ideal temperature change trajectory when the line clamp is only affected by the heating of the current, and serves as a reference for subsequent stripping away the influence of solar radiation.

[0024] The radiation component extraction submodule calls the real-time temperature rise data and the theoretical Joule thermal temperature rise data to perform difference calculation, constructs the residual time series, performs weighted moving average smoothing filtering on the residual time series, and generates the solar radiation temperature rise component. The module calls real-time daytime temperature rise data (i.e., measured temperature rise including both current heating and solar radiation effects) and the newly generated theoretical Joule thermal temperature rise data, and performs a difference operation at the same timestamp. That is, it subtracts the theoretical Joule thermal temperature rise data from the real-time temperature rise data to construct a residual time series. This residual series physically includes the thermal effect caused by solar radiation and random measurement noise. In order to extract the pure solar radiation temperature rise component, this module performs weighted moving average smoothing filtering on the residual time series. The moving window length is set to 30 sampling points, and the data within the window are assigned weight values ​​that conform to a Gaussian distribution. That is, the closer the data is to the center point, the higher the weight, and the farther away the data is, the lower the weight. The weighted average value within the window is calculated by convolution operation to filter out high-frequency random noise. For example, the original residual value at a certain moment is 5.8 degrees Celsius. After weighted smoothing and combining the data trend of the preceding and following moments, the corrected value is 5.6 degrees Celsius. This smoothed series is generated as the solar radiation temperature rise component, which accurately reflects the solar thermal shock intensity that changes over time.

[0025] The decoupled data generation submodule calls the real-time temperature rise data and the solar radiation temperature rise component to perform time-domain data alignment, and subtracts the solar radiation temperature rise component from the real-time temperature rise data point by point to generate radiation decoupled temperature rise data. The module retrieves the raw real-time temperature rise data and the smoothed solar radiation temperature rise component, performing a strict time-domain data alignment operation to ensure precise matching of the timestamps for each data point. It then performs point-by-point subtraction, subtracting the corresponding solar radiation temperature rise component from each value in the real-time temperature rise data, thus numerically removing the superimposed effect of external solar radiation on the clamp temperature. For example, at a certain time T2, the sensor measures a clamp temperature rise of 35.5 degrees Celsius, and the calculated solar radiation temperature rise component at that time is 5.6 degrees Celsius. The module subtracts 5.6 from 35.5, yielding 29.9 degrees Celsius. This calculated result of 29.9 degrees Celsius is the radiation-decoupled temperature rise data for that time, representing the clamp's true thermal response after removing the influence of ambient light, generated only by the load current and its own contact resistance. By performing this operation on data from all time periods, the module generates a clean radiation-decoupled temperature rise data curve. This curve eliminates environmental meteorological interference, providing a high signal-to-noise ratio data foundation for subsequent accurate analysis of the clamp's contact status.

[0026] Specifically, such as Figure 2 , 5 As shown, the thermal resistance analysis module includes: The rate of change calculation and judgment submodule collects real-time load current and calculates the difference quotient between current values ​​of adjacent sampling points, constructs a load current rate of change sequence, calls preset step trigger reference data, compares the load current rate of change sequence with the judgment threshold, filters out abrupt change time points that exceed the judgment threshold, and generates a step trigger time index vector. The module collects load current monitoring sequences of intelligent T-clamps during historical normal operation cycles (e.g., the past year). It then performs a first-order difference operation on this sequence, calculating the difference between the current value at the current moment and the current value at the previous moment. It extracts the current change increments from adjacent sampling points and takes the absolute value of all increments to construct a statistical sample of the change amplitude. This module calculates the expected value and standard deviation of this sample. For example, if the expected value of the current change is calculated to be 2 amps and the standard deviation to be 1.5 amps, based on the statistical properties of a normal distribution, it calculates the sum of the expected value and three times the standard deviation, i.e., 2 plus 3 multiplied by 1.5, resulting in 6. The 6.5 amps threshold is set as the step trigger reference data. In real-time monitoring, the module collects the real-time load current and calculates the difference quotient (i.e., the derivative of current with respect to time) between the current values ​​of adjacent sampling points, constructs a load current change rate sequence, and compares each value in the sequence with the 6.5 amps step trigger reference data. If the current change rate at a certain moment is 15 amps per second, which is significantly greater than the 6.5 amps threshold, the module determines that a current change has occurred at that moment, selects that time point, and records it in the step trigger time index vector as the trigger signal for subsequent transient thermal resistance analysis.

[0027] The transient data extraction submodule receives radiation-decoupled temperature rise data and aligns the radiation-decoupled temperature rise data with the step trigger moment index vector in the time domain. It locks the starting position of the step change, extracts the temperature rise response value sequence within a preset time window after the starting position, and generates a transient temperature rise evolution data segment. The module receives radiation-decoupled temperature rise data and aligns the data stream with the step trigger moment index vector in the time domain, locking each step trigger moment determined to be a current surge. Using the locked step surge start position as zero, the module extracts a sequence of temperature rise response values ​​within a preset time window (e.g., 60 minutes). This extraction aims to capture the complete dynamic process of the clamp's temperature transitioning from an old steady state to a new steady state after a current surge. For example, if a current step is detected at 10:00:00, the module extracts radiation-decoupled temperature rise data from 10:00:00 to 11:00:00, containing 3600 sampling points (assuming a 1Hz sampling rate), generating a transient temperature rise evolution data segment. This data segment removes redundant information from the steady-state period, retaining only the dynamic response portion containing the clamp's thermal impedance characteristics. The data segment is then formatted as a time-temperature two-dimensional array and transmitted to the next-level module for feature decomposition.

[0028] The feature spectrum fitting and generation submodule constructs a function expression with multiple exponential superpositions for the transient temperature rise evolution data segment, calculates the fitting residual between the function expression and the transient temperature rise evolution data segment, minimizes the fitting residual by performing iterative approximation operations on the exponential decay coefficient and amplitude coefficient in the function expression, analyzes the values ​​of multiple time constant terms and the corresponding amplitude weight terms, and generates the transient thermal resistance feature spectrum. For the extracted transient temperature rise evolution data segment, a multi-exponential superposition function expression is constructed. This expression is in the form that the temperature change over time equals the sum of multiple exponential terms, each containing an amplitude coefficient and a decay term (time divided by the time constant) with the natural constant e as its base. This module calculates the fitting residual between the initial function expression and the actual transient temperature rise evolution data segment, i.e., it calculates the sum of squares of the differences between the model's predicted value and the measured value at each time point. The Levenberg-Marquardt iterative algorithm is used to adjust the exponential decay coefficients (corresponding to time constants) in the function expression. The module continuously approximates the measured curve using the reciprocal of the constant and the amplitude coefficient until the descent gradient of the fitting residual is less than 1 multiplied by 10 to the power of -6. The module ultimately resolves three independent time constant terms and their corresponding amplitude weight terms. For example, the resolution results are: the first term is a time constant of 50 seconds with an amplitude of 0.5, the second term is a time constant of 300 seconds with an amplitude of 0.3, and the third term is a time constant of 1200 seconds with an amplitude of 0.2. These parameters are combined to generate the transient thermal resistance characteristic spectrum, where different time constants correspond to the thermal response speeds of different physical structural layers (such as the contact layer, the body layer, and the convection layer) inside the clamp.

[0029] Specifically, such as Figure 2 , 6 As shown, the looseness detection module includes: The feature sorting and filtering submodule calls the transient thermal resistance feature spectrum, extracts the value of each time constant item in the spectrum, sorts the value of each time constant item in ascending order, identifies and locks the time constant item with the smallest value, and generates the index of the smallest time constant component. For example, for the three time constants of 50 seconds, 300 seconds, and 1200 seconds obtained from the above analysis, the module sorts them as [50, 300, 1200]. Based on the principle of heat transfer, the module identifies and locks the time constant with the smallest value, namely the 50-second component. Since the heat capacity of the contact resistance is extremely small, its thermal response speed is the fastest. Therefore, the smallest time constant directly corresponds to the thermal characteristics of the wire clamp contact interface. The module generates an index pointing to the smallest value (50 seconds) and its corresponding position of the smallest time constant component. Through this index, this specific physical process is separated from the mixed thermal response and clearly marked as the key feature reflecting the contact state, eliminating the interference of the long tail effect of large heat capacity body and environmental convection heat dissipation.

[0030] The contact feature extraction submodule extracts the corresponding independent exponential function component from the transient thermal resistance feature spectrum based on the minimum time constant component index, marks it as the thermal characteristic component of the contact layer, and separates the corresponding amplitude weight term value and decay rate value to generate the thermal response feature vector of the contact layer. Based on the minimum time constant component index, the corresponding independent exponential function component is accurately extracted from the transient thermal resistance characteristic spectrum and marked as the contact layer thermal characteristic component. This module separates the amplitude weight term value and the exponential decay rate value (i.e., taking the reciprocal of the time constant) corresponding to this component. For example, the amplitude weight corresponding to a 50-second time constant is extracted as 0.5 and the decay rate is 0.02 (1 divided by 50). These two values ​​constitute the contact layer thermal response characteristic vector [0.5, 0.02]. Among them, the amplitude weight term reflects the proportion contribution of the contact thermal resistance to the total thermal resistance, and the decay rate reflects the heat conduction speed of the contact layer. This vector compresses the complex multidimensional thermal resistance spectrum into two core indicators characterizing the health status of the contact interface, providing a direct mathematical object for subsequent deviation quantification.

[0031] The deviation quantification calculation submodule calls the preset fastening state reference data and thermal conductivity standard data, obtains the amplitude weight term value and decay rate value from the thermal response feature vector of the contact layer, compares the amplitude weight term value with the fastening state reference data, calculates the difference between the decay rate value and the thermal conductivity standard data value, and generates thermal response deviation data. Select a calibration-grade intelligent T-clamp of the same model as the monitored object. Using a torque calibration device, tighten the fastening bolts on the contact surface of the standard clamp prototype to the rated torque (e.g., 40 Nm). Apply a step current excitation from 0 Amperes to 600 Amperes to the standard clamp prototype in a constant temperature environment of 25 degrees Celsius. Collect the standard temperature rise response sequence and perform multiple exponential decomposition. Analyze to obtain the reference characteristic spectrum. Select the time constant term with the smallest value (e.g., 45 seconds), extract its corresponding amplitude weight term value (e.g., 0.45), and set it as the reference data for the tightened state. Extract its attenuation rate value (e.g., 0.022). The data is set as the thermal conductivity standard data. In real-time analysis, the module calls the above-mentioned preset data and obtains the measured amplitude weight (0.5) and measured attenuation rate (0.02) from the real-time generated contact layer thermal response feature vector. The measured amplitude weight 0.5 is compared with the fastening state reference data 0.45, and the deviation is calculated to be 0.05. The measured attenuation rate 0.02 is compared with the thermal conductivity standard data 0.022, and the numerical difference is calculated to obtain a value difference of -0.002. The module combines these two differences to generate thermal response deviation data, which quantifies the degree of deterioration of the current wire clamp contact state relative to the ideal fastening state.

[0032] Specifically, such as Figure 2 , 7 As shown, the fatigue prediction module includes: The impedance characteristic calculation submodule monitors the real-time load current of the clamp and identifies the sampling time point when the load current is at its lowest value, marking it as the cold reference point. It calculates the ratio of the radiation decoupling temperature rise data to the square term of the load current at the cold reference point, generating cold equivalent contact impedance data. The real-time load current of the monitoring clamp is used to identify the sampling time point when the load current is at its lowest value (usually between 3 am and 4 am) through the peak-valley search algorithm, and this point is marked as the cold reference point. The module extracts the radiation decoupling temperature rise data and load current data at the cold reference point, for example, the temperature rise is 5 degrees Celsius and the current is 100 amperes. The module performs a division operation to calculate the ratio of the radiation decoupling temperature rise data (5) to the square term of the load current (10000), that is, 5 divided by 10000, to obtain 0.0005 (this value is the equivalent thermal impedance, the physical dimension is degrees Celsius per square ampere, and here it is used as a generalized impedance feature). The calculation result is generated as cold equivalent contact impedance data. The cold point is selected for calculation to eliminate the error caused by the nonlinear change of material resistivity at high temperature, and to ensure that the impedance feature only reflects the physical structure change of the contact surface (such as microscopic spot degradation), and to provide a pure reference impedance index.

[0033] The damage accumulation analysis submodule calculates the numerical difference between the cold equivalent contact impedance data in the current monitoring cycle and the adjacent previous monitoring cycle, generates single thermal cycle plastic residual data, and performs a time-based cumulative summation operation on the single thermal cycle plastic residual data to generate cumulative damage data. The module reads the cold-state equivalent contact impedance data (0.00052) calculated for the current monitoring cycle (e.g., day N) and calls the cold-state equivalent contact impedance data (0.00050) from the adjacent previous monitoring cycle (day N-1). It then performs a subtraction operation to calculate the numerical difference, i.e., 0.00052 minus 0.00050, resulting in an increment of 0.00002. This increment is generated as single-cycle plastic residual data, which characterizes the irreversible increase in impedance at the clamp contact surface due to fretting wear or plastic deformation after one diurnal thermal expansion and contraction cycle. The module performs a time-based cumulative summation operation on the single-cycle plastic residual data, accumulating the residuals from all historical cycles. For example, after 100 days of operation, the accumulated residual sum reaches 0.002. This accumulation result generates cumulative damage data, which records the cumulative fatigue damage degree of the clamp throughout its entire lifespan from installation to the present, reflecting the long-term evolution trend of contact performance.

[0034] The integrated early warning decision submodule calls the cumulative damage data and thermal response deviation data to construct a multi-dimensional state mapping matrix. It then maps the cumulative damage data and thermal response deviation data as state variables to the multi-dimensional state mapping matrix to perform weighted fusion calculation, generate the clamp health index, determine the clamp maintenance needs, and generate clamp maintenance early warning information. Using cumulative damage data (e.g., 0.002) and thermal response deviation data (e.g., amplitude deviation 0.05), a multidimensional state mapping matrix is ​​constructed. This matrix maps the cumulative damage data to a long-term aging factor and the thermal response deviation data to a short-term loosening factor, performing a weighted fusion calculation. For example, setting the weight of the long-term aging factor to 0.4 and the weight of the short-term loosening factor to 0.6, the calculation formula is: Health Index = 100 - (Normalized value of long-term aging factor × 40 + Normalized value of short-term loosening factor × 60). Assuming the normalized long-term factor is 20 and the short-term factor is 30, then 100 - (20 × 0.4 + 30 × 0.6) = 100 - (8 + 18) = 74; The generated cable clamp health index is 74; This module compares the index with the preset maintenance level range. If the index is between 70 and 80, the cable clamp maintenance requirement is determined to be "recommended inspection". If the index is below 60, it is determined to be "emergency maintenance". Finally, a cable clamp maintenance early warning message containing specific health scores and maintenance suggestions is generated and sent to the monitoring terminal through the communication interface. Please refer to Table 1, which lists the calculation example data of the cable clamp health index at different operating stages.

[0035] Table 1. Example Table of Cable Clamp Health Status Assessment As shown in Table 1, with the increase of accumulated damage data and thermal resistance deviation, the calculated wire clamp health index decreases significantly, accurately reflecting the evolution of the equipment from good to deterioration.

[0036] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An integrated constant torque intelligent T-type wire clamp online monitoring and early warning system, characterized in that, The system includes: The benchmark construction module filters the clamp temperature monitoring data in the range of no light and stable load current, calculates the real-time temperature rise by combining the ambient temperature, fits the square term of load current with the real-time temperature rise, and generates a non-radiative thermal response benchmark. The benchmark construction module includes: The data filtering and acquisition submodule monitors the surface temperature of the clamp, ambient temperature and load current in real time and records the timestamp information corresponding to the sampling time. Based on the timestamp information, it extracts the monitoring data within the time interval without light, calculates the fluctuation variance of the load current and filters the electrical stable operation data segment with fluctuation variance less than the preset stable threshold, and generates a pure steady-state monitoring sample. The temperature difference calculation and processing submodule, based on the pure steady-state monitoring sample, indexes and analyzes the surface temperature of the clamp and the ambient temperature at the same sampling time, calculates the temperature difference between the two, constructs a relative temperature rise time series according to the time order, and generates a measured temperature rise data series. The benchmark fitting generation submodule calls the pure steady-state monitoring sample, extracts the monitored load current and performs a power operation on the load current to obtain the current square term value, constructs the current square term value into a regression input vector according to the time sequence, calls the measured temperature rise data sequence as the regression output vector, performs a binary linear regression fitting calculation, calculates the slope coefficient and intercept parameter of the regression equation, and generates a non-radiative thermal response benchmark. The radiation decoupling module calculates the theoretical temperature rise under the non-radiative thermal response benchmark, calculates the deviation between the real-time temperature rise data and the theoretical temperature rise to obtain a residual sequence and smooths it, obtains the solar radiation temperature rise component, subtracts the component from the real-time temperature rise to obtain the radiation decoupling temperature rise data; The thermal resistance analysis module extracts the radiation decoupling temperature rise data segment when the current change rate data exceeds the standard, and generates a transient thermal resistance characteristic spectrum through multiple exponential fitting decomposition. The loosening judgment module marks the minimum component of the transient thermal resistance characteristic spectrum as the contact layer thermal response characteristic, calculates the deviation between the decay rate of the contact layer thermal response characteristic and the standard data, and generates thermal response deviation data. The fatigue prediction module calculates the ratio of radiation decoupling temperature rise data to current at the valley point, obtains cold-state equivalent contact impedance data, and accumulates the increment to obtain cumulative damage data. Combined with the thermal response deviation data, it calculates the clamp health index and outputs clamp maintenance early warning information.

2. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 1, characterized in that, The non-radiative thermal response benchmark includes the Joule thermal conversion coefficient, the convective heat dissipation coefficient, and the linear regression determination coefficient. The radiation decoupling temperature rise data includes the net Joule thermal temperature rise sequence, the synchronization timestamp sequence, and the residual correction factor. The transient thermal resistance characteristic spectrum includes the differential structure function curve, the integral structure function curve, and the thermal capacity and thermal resistance distribution map. The thermal response deviation data includes the absolute deviation value of the decay rate, the relative change rate of the amplitude weight, and the thermal response consistency coefficient. The clamp maintenance early warning information includes the early warning level identifier, the estimated remaining lifespan, and the tightening torque adjustment parameters.

3. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 1, characterized in that, The process of obtaining the preset stability threshold is as follows: The historical load current monitoring sequence of the clamp is retrieved, and the historical load current monitoring sequence is sliced ​​using a sliding time window to construct multiple current sampling segments containing continuous sampling points. The numerical variance of each current sampling segment is calculated to construct a variance dataset characterizing the current fluctuation state distribution. A Gaussian mixture model is used to estimate the probability density of the variance dataset and fit it to generate a bimodal distribution curve containing steady-state characteristic peaks and transient characteristic peaks. The steady-state characteristic peak located in the low value range of the bimodal distribution curve is identified, and the mathematical expectation and standard deviation of the steady-state characteristic peak are analyzed. The sum of the mathematical expectation and three times the standard deviation is calculated, and the sum is set as the preset stability threshold.

4. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 1, characterized in that, The radiation decoupling module includes: The theoretical temperature rise calculation submodule obtains the clamp load current during the daytime monitoring period, and, in conjunction with the non-radiative thermal response benchmark, calculates the expected temperature rise value of the target clamp load current intensity under an ideal environment without light, generating theoretical Joule thermal temperature rise data. The radiation component extraction submodule calls the real-time temperature rise data and the theoretical Joule heat temperature rise data to perform difference calculation, constructs a residual time series, and performs weighted moving average smoothing filtering on the residual time series to generate solar radiation temperature rise components. The decoupling data generation submodule calls the real-time temperature rise data and performs time-domain data alignment with the solar radiation temperature rise component, and subtracts the solar radiation temperature rise component from the real-time temperature rise data point by point to generate radiation decoupling temperature rise data.

5. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 4, characterized in that, The thermal resistance analysis module includes: The rate of change calculation and judgment submodule collects real-time load current and calculates the difference quotient between current values ​​of adjacent sampling points, constructs a load current rate of change sequence, calls preset step trigger reference data, compares the load current rate of change sequence with the judgment threshold, filters out abrupt change time points that exceed the judgment threshold, and generates a step trigger time index vector. The transient data extraction submodule receives radiation-decoupled temperature rise data and aligns the radiation-decoupled temperature rise data with the step trigger time index vector in the time domain, locks the starting position of the step change, extracts the temperature rise response value sequence within a preset time window after the starting position, and generates a transient temperature rise evolution data segment. The feature spectrum fitting and generation submodule constructs a function expression with multiple exponential superpositions for the transient temperature rise evolution data segment, calculates the fitting residual between the function expression and the transient temperature rise evolution data segment, minimizes the fitting residual by performing iterative approximation operations on the exponential decay coefficient and amplitude coefficient in the function expression, analyzes the values ​​of multiple time constant terms and the corresponding amplitude weight terms, and generates a transient thermal resistance feature spectrum.

6. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 5, characterized in that, The process of obtaining the step trigger reference data is as follows: Retrieve the load current monitoring sequence of the clamp during its historical normal operation cycle, perform a first-order difference operation on the load current monitoring sequence, extract the current change increment of adjacent sampling points, perform absolute value processing on the current change increment, construct a statistical sample of change amplitude, calculate the mathematical expectation and standard deviation of the statistical sample of change amplitude, and based on the statistical characteristics of normal distribution, calculate the sum of the mathematical expectation and 3 times the standard deviation, and set the calculated sum as the step trigger reference data.

7. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 5, characterized in that, The looseness detection module includes: The feature sorting and filtering submodule calls the transient thermal resistance feature spectrum, extracts the value of each time constant item in the spectrum, sorts the value of each time constant item in ascending order, identifies and locks the time constant item with the smallest value, and generates the index of the smallest time constant component. The contact feature extraction submodule extracts the corresponding independent exponential function component from the transient thermal resistance feature spectrum based on the minimum time constant component index, marks it as the thermal characteristic component of the contact layer, and separates the corresponding amplitude weight term value and decay rate value to generate the thermal response feature vector of the contact layer. The deviation quantization calculation submodule calls the preset fastening state reference data and thermal conductivity standard data, obtains the amplitude weight term value and decay rate value from the thermal response feature vector of the contact layer, compares the amplitude weight term value with the fastening state reference data, calculates the difference between the decay rate value and the thermal conductivity standard data, and generates thermal response deviation data.

8. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 7, characterized in that, The process of obtaining the preset fastening state reference data and thermal conductivity standard data is as follows: A standard wire clamp prototype is selected, and the contact surface fastening bolts of the wire clamp prototype are tightened to the rated torque using a torque calibration device. A step current excitation is applied to the wire clamp prototype in a constant temperature environment, and the standard temperature rise response sequence generated by the wire clamp prototype is collected. Multiple exponential function superposition fitting decomposition is performed on the standard temperature rise response sequence to obtain a reference feature spectrum containing multiple independent time constant terms. The time constant term with the smallest value in the reference feature spectrum is selected, and the amplitude weight term value corresponding to the time constant term with the smallest value is extracted and set as the fastening state reference data. The decay rate value corresponding to the time constant term with the smallest value is extracted and set as the thermal conductivity standard data.

9. The integrated constant torque intelligent T-type wire clamp online monitoring and early warning system according to claim 7, characterized in that, The fatigue prediction module includes: The impedance characteristic calculation submodule monitors the real-time load current of the clamp and identifies the sampling time point when the load current is at its lowest value, marking it as the cold reference point. It calculates the ratio of the radiation decoupling temperature rise data to the square term of the load current at the cold reference point, generating cold equivalent contact impedance data. The damage accumulation analysis submodule calculates the numerical difference between the cold equivalent contact impedance data in the current monitoring cycle and the adjacent previous monitoring cycle, generates single thermal cycle plastic residual data, and performs a time-based cumulative summation operation on the single thermal cycle plastic residual data to generate cumulative damage data. The integrated early warning decision submodule calls the accumulated damage data and the thermal response deviation data to construct a multi-dimensional state mapping matrix. The accumulated damage data and thermal response deviation data are mapped as state variables to the multi-dimensional state mapping matrix to perform weighted fusion calculation, generate the clamp health index, determine the clamp maintenance needs, and generate clamp maintenance early warning information.

Images