A cable stress intelligent sensing and early warning method and sensing and early warning device

Abstract

This invention relates to the field of early warning technology, and discloses a cable stress intelligent sensing and early warning method and device, comprising: a force acquisition module to acquire standardized multi-point force array data; a reference separation module to extract thermal effect baseline and contact deviation data and divide complete cycle periods; a deformation reconstruction module to generate potential deformation and bending deformation based on a swarm intelligence algorithm; an offset extraction module to acquire the offset of the eccentric hysteresis center; a stress assessment module to fuse the above parameters to amplify the average force value to obtain the equivalent stress index; a dynamic threshold module to generate a safety threshold based on historical data and calculate the over-limit ratio; and finally, an early warning module to output an alarm status. This invention can accurately locate hidden bending peaks induced by micro-rotation angles, achieving early warning.

Description

Technical Field

[0001] This invention relates to the field of early warning technology, and more specifically, to a method and device for intelligent sensing and early warning of cable stress. Background Technology

[0002] In underground tunnels and other laying environments, cables and their supports are subjected to the combined effects of uneven settlement, thermal expansion and contraction, changes in contact friction, and environmental loads over a long period. Existing stress monitoring methods mostly involve discretely arranging sensors at the central column of a flexible support to obtain the fulcrum reaction force and then comparing it with a fixed threshold to make a safety assessment.

[0003] However, when the subsidence is not just overall sinking, but is accompanied by a slight rotation of a column, deflection of the support arm, or a change in the angle of the mounting surface, the support relationship between the cable and the support arm will be reconstructed from a relatively uniform surface contact to an off-edge contact or a mixture of adhesion and slight slippage. In this case, the average support force read by discrete monitoring points may still be within the normal threshold range, but the bending stress near the local shoulder has risen sharply and formed a hidden peak.

[0004] Existing early warning methods generally rely on single-dimensional absolute force exceeding limits, lacking in-depth analysis of the evolution of single-column time series and its multi-dimensional parameter reconstruction. This conventional approach cannot accurately capture the eccentric hysteresis offset left on the force loop by micro-rotation angles and contact state changes, often only triggering alarms when destructive exceeding of absolute force occurs. This results in a significant lag in the current system's perception of hidden bending hotspots and early contact anomalies, easily leading to real missed detection risks and failing to meet the demand for highly timely early warning. Summary of the Invention

[0005] This invention provides a method and device for intelligent sensing and early warning of cable stress, which solves the technical problems mentioned in the background art.

[0006] In a first aspect, a cable stress intelligent sensing and early warning device includes: The force acquisition module is used to synchronously acquire and standardize the monitoring sequence of the support, and generate multi-point force array data; The reference separation module is used to extract thermal effect baseline and contact deviation data from the multi-point force array data, divide the complete cycle, and generate heating weight and cooling weight. The deformation reconstruction module is used to map the multi-point force array data through a swarm intelligence algorithm to generate potential deformation and bending deformation. The offset extraction module is used to extract the offset of the eccentric hysteresis center based on the thermal effect baseline, the potential deformation, the heating weight and the cooling weight within the complete cycle. The stress assessment module is used to fuse the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplify the average stress value of the multi-point force array data in the complete cycle to obtain the equivalent stress index. The dynamic threshold module is used to generate a dynamic safety threshold based on historical cyclic data and calculate the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold. The early warning module is used to determine the alarm column and early warning status based on the over-limit ratio, and to trigger a work order when the early warning status exceeds the limit.

[0007] Secondly, a cable stress intelligent sensing and early warning method, applied in any of the cable stress intelligent sensing and early warning devices described above, includes: The monitoring sequence of the stent is synchronously acquired and standardized to generate multi-point force array data; Thermal effect baseline and contact deviation data are extracted from the multi-point force array data, the complete cycle is divided, and heating weight and cooling weight are generated. The multi-point force array data is mapped using a swarm intelligence algorithm to generate potential deformation and bending deformation. Based on the thermal effect baseline, the potential deformation, the heating weight, and the cooling weight within the complete cycle, the eccentric hysteresis center offset is extracted. By fusing the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplifying the average force of the multi-point force array data within the complete cycle, an equivalent stress index is obtained. A dynamic safety threshold is generated based on historical cyclic data, and the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold is calculated. The alarm column and warning status are determined based on the over-limit ratio, and a work order is triggered when the warning status exceeds the limit.

[0008] The beneficial effects of this invention are as follows: by synchronously collecting multi-point stress array data, separating thermal effect baseline and contact deviation data, and using swarm intelligence algorithm to reconstruct potential deformation and bending deformation, it is possible to accurately extract the eccentric hysteresis center offset, and then integrate the local risk amplification factor to obtain the equivalent stress index. Finally, based on the dynamic safety threshold generated by historical data, the over-limit judgment is made. Without relying on manually setting fixed thresholds, it is possible to identify and accurately locate hidden dangers caused by local subsidence or cornering of the support, and effectively improve the timeliness of safety monitoring of underground cable operation. Attached Figure Description

[0009] Figure 1 This is a block diagram of a cable stress intelligent sensing and early warning device according to the present invention. Detailed Implementation

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

[0011] Example 1: As Figure 1 As shown, a cable stress intelligent sensing and early warning device includes: The force acquisition module is used to synchronously acquire and standardize the monitoring sequence of the support, and generate multi-point force array data; The reference separation module is used to extract thermal effect baseline and contact deviation data from the multi-point force array data, divide the complete cycle, and generate heating weight and cooling weight. The deformation reconstruction module is used to map the multi-point force array data through a swarm intelligence algorithm to generate potential deformation and bending deformation. The offset extraction module is used to extract the offset of the eccentric hysteresis center based on the thermal effect baseline, the potential deformation, the heating weight and the cooling weight within the complete cycle. The stress assessment module is used to fuse the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplify the average stress value of the multi-point force array data in the complete cycle to obtain the equivalent stress index. The dynamic threshold module is used to generate a dynamic safety threshold based on historical cyclic data and calculate the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold. The early warning module is used to determine the alarm column and early warning status based on the over-limit ratio, and to trigger a work order when the early warning status exceeds the limit.

[0012] Preferably, the force acquisition module is specifically used for: Obtain the position coordinates of the stent and map them to standardized position coordinates through range standardization. The original monitored forces within the data time window are acquired synchronously based on a unified time reference; when the data loss time does not exceed 10% of the median sampling interval of the thermal cycle, linear time interpolation is used to supplement the data. The original monitored force is mapped to a standardized instantaneous force through centralized standardization processing, and the difference between the original monitored forces at adjacent times is mapped to a standardized force increment. The standardized instantaneous forces of each support at the same moment are combined to generate the multi-point force array data; Wherein, the range standardization is the ratio of the difference between the input value and the minimum value to the sum of the range and the numerical stability term; the centering standardization is the ratio of the difference between the input value and the median of the window data to the sum of the interquartile range of the window data and the numerical stability term. The formula for calculating the standardized position coordinates is: ; The formula for calculating the standardized instantaneous force is as follows: ; The formula for calculating the standardized force increment is: ; The calculation formula for the multi-point force array data is as follows: ; In the formula, Indicates the first The standardized position coordinates of each bracket, Indicates the first The coordinates of the bracket position of each bracket. and These represent the maximum and minimum coordinates of all supports, respectively. Indicates the first The first support in The standardized instantaneous force at each moment, Indicates the first The first support in The original monitored force at each moment, and These represent the median and interquartile range of the original monitored force within the data time window, respectively. Indicates the first The first support in The standardized force increment at each moment, Indicates the first The difference in the original monitored force of each support at adjacent time points, and The interquartile range and the median of the window data, respectively, represent the difference. Indicates the first The multi-point force array data at each moment, Indicates the total number of stents. This represents the matrix transpose operation; This represents the numerical stability term.

[0013] The i-th bracket is identified by its serial number, which is established according to the order in which the brackets are laid out along the line.

[0014] The support position coordinates are the numerical values ​​of the installation position of each support on the cable channel baseline, used to characterize the spatial relative relationship between supports. These coordinates can be obtained through total station measurement, laser distance measurement, 3D scanning modeling, or construction layout verification.

[0015] Standardized position coordinates are dimensionless position quantities obtained by standardizing the position coordinates of the support through range standardization, and are used to unify the spatial scale under different channel length conditions.

[0016] A unified time reference is a time reference used after all support monitoring sequences are aligned to the same clock, ensuring that cross-support data can be compared moment by moment. A time reference of 1 minute is preferred, as it covers the minute-level evolution of cable thermal and contact changes without significantly amplifying instantaneous noise.

[0017] A data time window is a continuous time segment used to calculate the median, interquartile range, and standardized results; it is the direct sampling range for various statistics. Preferably, it is 24 to 72 hours, a range that typically covers at least one complete thermal runoff cycle.

[0018] The k-th time point is the discrete sampling time number under a unified time base.

[0019] The original monitored force is the actual force value collected by the support monitoring unit before dimensionless processing. It can be obtained through a tension sensor, strain gauge force sensor, weighing load monitoring unit, or a reaction force sensor installed at the base of the support arm.

[0020] Data missing duration is the length of time during which there is no data between adjacent valid sampling points.

[0021] The median sampling interval of a thermal cycle is the median of the time difference between adjacent effective samples within a thermal cycle, used to characterize the typical sampling rhythm of that cycle.

[0022] Ten percent of the median sampling interval during the thermal cycling cycle is the upper limit for determining whether missing segments can be repaired using linear time interpolation. The preferred range is 5 to 15. This is because direct interpolation can maintain continuity when the missing segment is too small, while excessively large missing segments will distort the baseline.

[0023] Standardized instantaneous force is a dimensionless force value obtained after the original monitored force has been centrally standardized, and is used to eliminate the magnitude difference between different stents.

[0024] The difference in the original monitored force is the difference in the original monitored force at adjacent sampling times of the same support, and is used to characterize the rate of change of force.

[0025] The standardized force increment is a dimensionless change in the difference of the original monitored force after being centrally standardized, and is used to enhance the ability to detect sudden changes.

[0026] Multi-point force array data is a vectorized result obtained by combining the standardized instantaneous forces of each support at the same moment in a fixed order.

[0027] The maximum coordinate of all supports is the maximum value among all support position coordinates, which is used as the upper boundary for range standardization.

[0028] The minimum coordinate of all supports is the minimum value among all support position coordinates, used as the lower boundary for range standardization.

[0029] The median of the original monitored force window data is the median of the original monitored force sequence within the data time window, and is used to establish a robust center.

[0030] The interquartile range of the original monitored force window data is the difference between the upper and lower quartiles of the original monitored force sequence within the data time window, and is used to characterize the robustness of dispersion.

[0031] The median of the difference window data is the median of the original monitored force difference sequence within the data time window, and is used to establish a robust center for the force change.

[0032] The interquartile range of the window data for the difference is the interquartile range of the original monitored force difference sequence within the data time window, used to characterize the robustness and dispersion of the change.

[0033] The total number of supports is the number of supports included in the current monitoring channel, used to determine the dimension of the multi-point force array data.

[0034] The numerical stability term is a small positive number added to the standardized denominator to prevent numerical divergence caused by a denominator of 0 or too small. It is preferably between 0.0001 and 0.01, a range that provides stable calculation without significantly compressing the effective differences.

[0035] In practice, the center projection point of the first bracket at the entrance of the cable channel is taken as the zero point of the coordinate system, and the direction of cable extension is taken as the positive direction of the coordinate system. When there is a bend in the channel, the cumulative mileage is first calculated along the design center line, and then the center of each bracket is projected onto the unfolded line to obtain the position coordinates of the bracket.

[0036] In practice, a reaction force sensor or strain gauge force sensor is installed at the root of each support arm. The sensor range is preferably 1.5 to 2 times the design load capacity of the support, and the sampling frequency is preferably 1 to 4 times per minute. The acquisition link includes a sensor, a signal conditioning module, a clock synchronization module, and an edge acquisition terminal.

[0037] In practice, all acquisition terminals complete time synchronization through the same time source when they start up. The time source can be the main station clock, network time server, or edge gateway time synchronization module. After the data of each bracket is uploaded, it is first resampled according to a unified time reference, and then enters the same time combination process. The synchronization tolerance is first controlled within one sampling period, and further optimized to not exceed 30 seconds.

[0038] In practice, the data time window adopts a sliding window mechanism, with the window length preferably ranging from 24 to 72 hours, and the sliding step size preferably set to 1 or 5 sampling periods. At each new moment, the earliest sample in the window is deleted and the latest sample is added, and the median and interquartile range of the window data are recalculated.

[0039] In practice, the duration of missing data is determined by subtracting one normal sampling period from the difference between the first and last valid timestamps of the missing segment; the median sampling interval of the thermal cycling cycle is obtained by calculating the median of the adjacent valid sampling time differences within the most recent number of identified complete cycling cycles; linear time interpolation is only allowed when the duration of missing data is no more than 10% of the median sampling interval of the thermal cycling cycle.

[0040] In practice, centralized standardization processing is performed independently for each support within a data time window. The input value is subtracted from the median of the corresponding sequence for that window, and then divided by the sum of the interquartile range and the numerical stability term of the corresponding sequence for that window. For standardized instantaneous stress, the reference sequence is the original monitoring stress window sequence. For standardized stress increments, the reference sequence is the original monitoring stress difference window sequence.

[0041] In practice, the numerical stability term is taken as a fixed small positive number or a small positive number that scales proportionally with the window interquartile range. 0.001 or 0.1% of the window interquartile range is preferred, and the larger value is taken. The multi-point force array data is sorted by the support position coordinates from smallest to largest and arranged sequentially, maintaining the same order and dimension at all times.

[0042] Preferably, the reference separation module is specifically used for: Identify the local extreme values ​​of the standardized instantaneous force of each support, and construct the upper and lower force envelopes through shape-preserving interpolation; The mean values ​​of the upper force envelope and the lower force envelope are subjected to the centering and standardization process to generate the thermal effect baseline. The difference between the standardized instantaneous force and the thermal effect baseline is subjected to the centering and standardization process to extract the contact deviation data; The difference between the thermal effect baselines at adjacent time points is subjected to the centering and standardization process to obtain the change in the thermal effect baseline. When the change in the thermal effect baseline shows two consecutive positive and negative inflections and includes a complete rise and fall process, it is defined as a complete cycle. Based on the hyperbolic tangent transform result of the baseline change in thermal effect, the heating weight and the cooling weight are generated respectively through the range normalization process. The formula for calculating the thermal effect baseline is: ; The formula for calculating the contact deviation data is: ; The formula for calculating the baseline change of the thermal effect is: ; The formula for calculating the heating weight is: ; The formula for calculating the cooling weight is: ; In the formula, Indicates the first The first support in The baseline of the thermal effect at each moment; Indicates the first The first support in The upper limit envelope of the force at each moment; Indicates the first The first support in The lower limit envelope of the force at each moment; Indicates the first A data time window at a given moment; This refers to the centralized and standardized processing; Indicates the first The first support in The contact deviation data at each moment; Indicates the first The first support in The standardized instantaneous force at each moment; Indicates the first The first support in The baseline change of the thermal effect at each time point; Indicates the first The baseline of the thermal effect of the support at the previous moment; Indicates the first The first support in The heating weight at each moment; Represents the hyperbolic tangent function; This indicates the range standardization process; Indicates the first The first support in The cooling weight at each moment.

[0043] Local extrema are local peaks and valleys formed by comparing adjacent moments in a standardized instantaneous force sequence, and are used to extract the slowly varying envelope.

[0044] The upper limit envelope of the force is a sequence of upper boundaries formed by shape-preserving interpolation of local peak values, used to characterize the high-level trend of force changes.

[0045] The lower bound envelope of the stress is a sequence of lower boundaries formed by shape-preserving interpolation of local valleys, used to characterize the low-level trend of stress changes.

[0046] The thermal effect baseline is a slowly varying benchmark obtained by centering and standardizing the mean values ​​of the upper and lower force envelopes, and is used to describe the force background dominated by thermal expansion and contraction.

[0047] Contact deviation data are the results of the standardized instantaneous force deviation from the thermal effect baseline, used to characterize changes in contact state and local abnormal forces.

[0048] The change in thermal effect baseline is the result of centering and standardizing the difference in thermal effect baseline between adjacent time points, and is used to characterize the direction and rate of change in thermal state.

[0049] A complete cycle is a closed time interval formed by the baseline change in thermal effect undergoing one complete heating process and one complete cooling process.

[0050] The heating weight is a weight value obtained by processing the change in the thermal effect baseline through hyperbolic tangent transformation and range standardization, and is used to highlight the sample during the heating phase.

[0051] The cooling weight is a weight value obtained by processing the change in the thermal effect baseline through hyperbolic tangent transformation and range standardization, and is used to highlight the sample during the cooling stage.

[0052] In practice, the standardized instantaneous stress sequence of each stent is first smoothed by the median of three points, and then the comparison method of adjacent three points is used to identify local peaks and local valleys. When a point is greater than the two points before and after it and the difference from the average of the last five points is not less than 0.2, it is judged as a local peak. When a point is less than the two points before and after it and the difference from the average of the last five points is not less than 0.2, it is judged as a local valley.

[0053] In practice, shape-preserving segmented cubic interpolation is used to generate the upper limit envelope of force for the local peak sequence, and the same method is used to generate the lower limit envelope of force for the local valley sequence. If there are fewer than two local peaks in a certain time window, the slope of the end of the upper limit envelope of the previous window is used as the boundary extension. If there are fewer than two local valleys, the same processing is performed on the lower limit envelope of force.

[0054] In practice, the thermal effect baseline retains only the low-frequency trend caused by the combined effects of ambient temperature, current-carrying temperature rise, and slow-changing support deformation, while the contact deviation data retains only the high-frequency deviation relative to the thermal effect baseline. In practice, the upper and lower envelope mean values ​​are first calculated, and then the mean values ​​are centered and standardized to obtain the thermal effect baseline. Subsequently, the thermal effect baseline is subtracted from the standardized instantaneous force and centered and standardized again to obtain the contact deviation data.

[0055] In practice, when the change in the thermal effect baseline changes from a continuous positive value to a continuous negative value, and there are at least 3 sampling points before and after the change, it is considered that a positive-to-negative transition from rising to falling has occurred; when the change in the thermal effect baseline changes from a continuous negative value to a continuous positive value, and there are at least 3 sampling points before and after the change, it is considered that a negative-to-positive transition from falling to rising has occurred; in order to suppress noise, the thermal effect baseline change with an absolute value less than 0.05 is uniformly regarded as 0.

[0056] In practice, the first consecutive positive sampling point after a negative-to-positive transition is taken as the starting point of the cycle, and the cycle ends with the consecutive negative segment after the subsequent positive-to-negative transition, and then the last consecutive negative sampling point before the next negative-to-positive transition is taken as the ending point of the cycle, thus forming a complete rise and fall process; only when the duration of both the heating segment and the cooling segment is not less than one-tenth of the window length is it considered a complete cycle.

[0057] In detail, the generation range, saturation processing, and zero-value processing of the heating and cooling weights include: first, the change in the thermal effect baseline is fed into a hyperbolic tangent transform, so that large positive values ​​approach 1 and large negative values ​​approach -1, and then mapped to the interval between 0 and 1 respectively before performing range standardization processing; the heating weight is mainly close to 1 in the heating segment and close to 0 in the cooling segment; the cooling weight is the opposite; if the change in all thermal effect baselines is the same in a certain period, then the heating and cooling weights are both set to 0.5.

[0058] Preferably, the deformation reconstruction module is specifically used for: Spline interpolation is performed based on the standardized position coordinates of each support and the corresponding standardized instantaneous force to generate a continuous spatial function; The first and second derivatives of the continuous spatial function are calculated and then the centering and normalization processes are performed to generate the spatial slope and the bending deformation, respectively. Initialize the projection weight vector with non-negative elements and a sum of one; The projected weight vector is used to weight the standardized instantaneous force, the contact deviation data, the spatial slope, and the bending deformation, and then the centering and standardization process is performed to generate the potential deformation. The potential deformation and the hysteresis area, center drift, and high-frequency fragmentation of the thermal effect baseline during the complete cycle are calculated to construct a fitness function. The fitness function is used to optimize the projection weight vector, and the potential deformation corresponding to the optimal projection weight vector is output. The formula for calculating the continuous space function is as follows: ; The formula for calculating the spatial slope is: ; The formula for calculating the bending deformation is: ; The formula for calculating the potential deformation is: ; The fitness function is calculated using the following formula: ; in, Indicates the first The first support in The area of ​​the loop within each complete cycle is calculated using the following formula: ; Indicates the first The first support in The center drift amount within each complete cycle is calculated using the following formula: ; Indicates the first The first support in The high-frequency fragmentation degree within each complete cycle is calculated using the following formula: ; In the formula, Indicates the first The continuous space function at time n, This indicates the spline interpolation operation. Indicates the first The standardized position coordinates of each bracket, Indicates the first The first support in The standardized instantaneous force at each moment; Indicates the first The first support in The spatial slope at each time point, This refers to the centralized and standardized processing; Indicates the first The first support in The bending deformation at each moment; Indicates the first The first support in The potential deformation at each moment. Represents the projection weight vector, , , and These represent the normalized weight elements corresponding to the projected weight vector. Indicates the first The first support in The contact deviation data at each moment; Represents the fitness function; This represents the pure mathematical absolute value operation. Indicates the baseline of the thermal effect. Indicates the first The complete cycle contains a set of times. This represents the number of samples in the set. and They represent the mathematical operations in the first place. The arithmetic mean extracted within each complete cycle.

[0059] The continuous spatial function is a continuous force distribution function constructed based on the standardized position coordinates and standardized instantaneous forces of each support, and is used to reconstruct the spatial force pattern between discrete support points.

[0060] The independent variable of a continuous spatial function is a spatial position variable that takes values ​​along a standardized position coordinate axis. It is used to describe the continuous stress state at any position between supports.

[0061] The spatial slope is the result of centering and standardizing the first derivative of the continuous spatial function at the support location, and is used to characterize the trend of force variation along the spatial direction.

[0062] The bending deformation is the result of centering and standardizing the second derivative of the continuous spatial function at the support location. It is used to characterize the force curvature and potential bending concentration near the support.

[0063] The projection weight vector is a set of non-negative weights that are weighted and fused from standardized instantaneous force, contact deviation data, spatial slope, and bending deformation. It is used to control the proportion of potential deformation. Preferably, all four elements are initialized to 0.25. Because uniform initial values ​​do not pre-set feature preferences, they facilitate stable convergence in the optimization process.

[0064] The normalized weight elements are the individual weights in the projected weight vector corresponding to the standardized instantaneous force, contact deviation data, spatial slope, and bending deformation, respectively, and their sum is kept to 1.

[0065] Potential deformation is a comprehensive representation of multiple features fused under the action of the projection weight vector and then processed by centering and standardization. It is used to characterize structural deformation that has not yet been directly manifested but has left traces in the stress.

[0066] The fitness function is an objective function that uses the area of ​​the loop, the center drift, and the high-frequency fragmentation to evaluate the merits of the projected weight vector and guides the optimization process.

[0067] The hysteresis loop area is the area enclosed by the hysteresis loop formed by the potential deformation and thermal effect baseline within a complete cycle. It is used to characterize the cumulative offset and energy-sampling effect present during the cycle.

[0068] The center drift is a quantitative result of the change in the geometric center of adjacent complete cycles, used to characterize the long-term offset trend of the lap center.

[0069] High-frequency fragmentation is an average measure of the fluctuation intensity of potential deformation at adjacent sampling points over a complete cycle.

[0070] The m-th complete cycle is the m-th complete rise and fall cycle numbered sequentially by time.

[0071] The set of times contained in the m-th complete cycle is the set of all discrete sampling times within that cycle.

[0072] The number of samples in the set is the number of valid sampling points contained in the m-th complete cycle.

[0073] The mean value of the potential deformation is the arithmetic mean of the potential deformation within the m-th complete cycle, which is used to represent the location of the potential deformation center in that cycle.

[0074] The mean value of the thermal effect baseline is the arithmetic mean of the thermal effect baseline within the m-th complete cycle, which is used to represent the location of the thermal effect center in that cycle.

[0075] In practice, continuous spatial functions are constructed using natural cubic spline interpolation first. When the number of supports is less than 4, shape-preserving piecewise cubic interpolation is used instead. Natural boundary conditions with a second derivative of 0 are used at both ends of the boundary. If the standardized instantaneous force of a support deviates from the mean of the three neighboring supports by more than 3 times the interquartile range of the window data, then the point is first subjected to outlier clamping before participating in spline interpolation.

[0076] In practice, at each sampling time, the derivative of the continuous space function is analyzed, and the first and second derivatives are taken at the standardized coordinates of each support. Then, the derivative sequence is centered and standardized. When piecewise splines are used, the derivative of each segment is calculated according to the analytical expression, and the left and right limit values ​​are consistent at the connection of segments.

[0077] In practice, the initial value of the projection weight vector is 0.25 for all four elements, and all elements are required to be non-negative and their sum is always 1. After each optimization update, the negative values ​​are truncated to 0, and then the sum of each element is normalized again. If an element is consistently less than 0.02 after normalization and does not improve for 20 consecutive generations, it is allowed to remain at its minimum value, but the feature must not be deleted directly.

[0078] In practice, the particle swarm optimization algorithm or differential evolution algorithm is used to optimize the projected weight vector. When using the particle swarm optimization algorithm, the population size is preferably 20 to 40, the maximum number of iterations is preferably 100 to 300, and both the global learning factor and the individual learning factor are preferably 1.5 to 2.0. The termination condition is that the fitness function improvement is less than 0.001 for 20 consecutive generations, or the maximum number of iterations is reached.

[0079] In practice, the centralized and standardized processing of potential deformation adopts the same data time window as the original monitored force and is maintained independently for each support. Whenever the window is updated, the median and interquartile range of the potential deformation sequence are recalculated. If the effective samples in a window are less than 80% of the designed sample number of the window, the potential deformation will not be output for that window.

[0080] In practice, the loop area is summed by the trapezoid formed by adjacent sampling points within a complete cycle and the absolute value is taken, without the need for additional points to close the loop at the beginning and end of the cycle; the center drift is taken as the average value of the potential deformation in the previous complete cycle and the average value of the thermal effect baseline; the high-frequency fragmentation is taken as the sum of the absolute values ​​of the differences between adjacent potential deformations within the same complete cycle divided by the number of samples minus 1.

[0081] In practice, the fitness function takes the loop area multiplied by 1, adds the center drift, divides by 1, and adds the high-frequency fragmentation, so that large loops and obvious drifts are rewarded, while high-frequency jitter is penalized. After optimization, the projected weight vector with the largest fitness function is output, and all potential deformations are recalculated using this vector. For the first complete cycle, since there is no previous cycle, the center drift is temporarily set to 0, and the historical comparison term is not involved.

[0082] Preferably, the offset extraction module is specifically used for: The mean values ​​of the thermal effect baseline and the potential deformation within the complete cycle are calculated and then the centering and normalization process is performed to form the current cycle geometric center. Weights are constructed by combining the lap area within the complete historical cycle, and the historical reference geometric center is calculated by weighted geometric median; Calculate the Euclidean distance between the current cyclic geometric center and the historical reference geometric center, and perform absolute value standardization to obtain the unidirectional drift distance; The potential deformation is weighted and averaged using the heating weight and the cooling weight respectively, and then the centering and standardization process is performed to obtain the heating branch center and the cooling branch center respectively. Calculate the absolute difference between the heating branch center and the cooling branch center and perform absolute value standardization to obtain the branch asymmetry. The unidirectional drift distance and the branch asymmetry are multiplied by one and then fused, and then the centering and standardization process is performed to extract the eccentric hysteresis center offset. The formula for calculating the current cyclic geometric center is: ; ; ; The formula for calculating the historical reference geometric center is: ; The formula for calculating the unidirectional drift distance is: ; The calculation formulas for the heating branch center and the cooling branch center are as follows: ; ; The formula for calculating the branch asymmetry is: ; The formula for calculating the eccentric hysteresis center offset is: ; In the formula, This represents the mean value of the baseline thermal effect described in the mathematical operations. This indicates the centralized standardization process. Indicates the first The complete cycle contains a set of times. This represents the number of samples in the set. Indicates the baseline of the thermal effect; This represents the mean value processing result of the potential deformation quantity described in the mathematical operation. This represents the potential deformation amount; Indicates the first The current cycle geometry center of the complete cycle; Indicates the historical reference geometric center, This represents the operation of the weighted geometric median. Indicates the first The geometric center corresponding to the complete cycle of the history. Indicates the first The lap area corresponding to the complete cycle of the history; This represents the unidirectional drift distance. This indicates the absolute value standardization process. Indicates Euclidean distance; This indicates the center of the heating branch. This indicates the heating weight; This indicates the cooling branch center. This represents the cooling weight; Indicates the degree of asymmetry of the branch; This represents the offset of the eccentric hysteresis center.

[0083] The current cyclic geometric center is a two-dimensional center point composed of the mean value of the thermal effect baseline and the mean value of the potential deformation within the current complete cycle, used to represent the overall hysteresis position of the current cycle.

[0084] The historical reference geometric center is a robust reference center calculated from the geometric center of a complete historical cycle using the weighted geometric median, and is used as a benchmark for comparison in the current cycle.

[0085] The one-way drift distance is the result of the absolute value normalization of the Euclidean distance between the current loop geometry center and the historical reference geometry center, and is used to characterize the offset magnitude of the loop center.

[0086] The heating branch center is the heating stage center value obtained by performing a weighted average on the potential deformation using heating weights and then performing a centering and standardization process.

[0087] The cooling branch center is the cooling stage center value obtained by performing a weighted average on the potential deformation using cooling weights and then standardizing it by centering.

[0088] Branch asymmetry is the result of the absolute difference between the heating branch center and the cooling branch center after absolute value standardization, and is used to characterize the degree of hysteresis branch skewness.

[0089] The eccentric hysteresis center offset is a comprehensive offset index obtained by fusing unidirectional drift distance and branch asymmetry and then centering and standardizing it. It is used to reveal the eccentric hysteresis phenomenon caused by micro-rotation or contact reconstruction.

[0090] The geometric center corresponding to the complete cycle of the h-th history is the two-dimensional geometric center result of the complete cycle with history number h.

[0091] In practice, the horizontal coordinate of the current cyclic geometric center is taken from the mean value of the thermal effect baseline, and the vertical coordinate is taken from the mean value of the potential deformation. Both are at a dimensionless unified scale before entering the geometric center, so they can be directly formed into two-dimensional points. To ensure comparability across cycles, the horizontal and vertical coordinates adopt the historical reference scale of the same line and the same support.

[0092] In practice, the historical reference geometric center is solved by iterative weighted geometric median. The initial point is the arithmetic mean of all historical geometric centers. Then, the points are updated by multiplying the inverse distance from each historical geometric center to the current estimated point by its weight. The process stops when the coordinate change in two consecutive iterations is less than 0.001. To suppress the influence of abnormal cycles, if the distance between a single historical geometric center and the median center exceeds three times the median of all historical distances, its weight is multiplied by 0.5.

[0093] In practice, the weights for constructing the historical loop area are obtained by dividing 1 by 1 and adding the historical loop area. Then, the weights of all historical cycles are normalized so that their sum is 1. In this way, the smaller the loop area, the higher the weight of the stable cycle. When the system has only one historical cycle, the geometric center of the historical reference geometric center is directly taken as the cold start result.

[0094] In practice, the unidirectional drift distance is first calculated by Euclidean distance between the current cyclic geometric center and the historical reference geometric center, and then absolute value standardization is performed using the historical geometric center distance sequence of the same support as the reference set. If there are fewer than 5 historical valid distance samples, the median and interquartile range are temporarily calculated using all existing samples.

[0095] In practice, the warm branch center only uses samples with a heating weight greater than 0, and the cooling branch center only uses samples with a cooling weight greater than 0. When the sum of the weights of a certain branch is less than 0.05, it indicates that the heating segment or cooling segment of that cycle is not identified enough. At this time, the overall mean of the potential deformation of that cycle is used to replace the branch center, and that cycle is recorded as a low confidence cycle.

[0096] In practice, the unidirectional drift distance and branch asymmetry are first incremented by 1, then multiplied and subtracted by 1 to obtain the fusion value. Finally, the eccentric hysteresis center offset is formed by centering and standardizing the set of historical complete cycles. When this value is continuously positive and increases cycle by cycle, it indicates that the overall offset of the return line center and the branch asymmetry are enhanced at the same time, which usually corresponds to the micro-rotation of the support arm, unilateral wear or local jamming.

[0097] Preferably, the stress assessment module is specifically used for: The bending deformation, the contact deviation data, and the standardized instantaneous force are respectively processed by absolute value standardization to obtain the mean value within the complete cycle. Based on the historical cycle data, the centering standardization process is performed to generate the bending deformation concentration, contact anomaly, and cycle force mean value, respectively. The eccentric hysteresis center offset, the bending deformation concentration and the contact anomaly are all added by one and then multiplied. The product is then subtracted by one and the centralization and standardization processing is performed based on the historical cyclic data to generate a local risk amplification factor. The local risk amplification factor is incremented by one and multiplied by the mean cyclic force. Based on the historical cyclic data, the centralization and standardization processing is performed to output the equivalent stress index. The formula for calculating the bending deformation concentration is: ; The formula for calculating the contact anomaly degree is: ; The formula for calculating the average value of the cyclic force is as follows: ; The formula for calculating the local risk amplification factor is as follows: ; The formula for calculating the equivalent stress index is as follows: ; In the formula, Indicates the first The first support in The concentration of bending deformation within each complete cycle. This indicates the centralized standardization process. Indicates time, Indicates the first The complete cycle contains a set of times. This represents the number of samples in the set. This indicates the absolute value standardization process. This indicates the amount of bending deformation. This refers to the historical cyclic data; Indicates the first The first support in The degree of contact anomaly within each complete cycle. This represents the contact deviation data; Indicates the first The first support in The average cyclic force within each complete cycle. This indicates the standardized instantaneous force; This represents the local risk amplification factor. This indicates the offset of the eccentric hysteresis center; The equivalent stress index, which is dimensionless, represents the relative risk intensity used for early warning determination.

[0098] Historical cycle data is a collection of historical monitoring and calculation results accumulated by the system according to a complete cycle, which is used to construct a robust reference scale and dynamic comparison baseline.

[0099] Bending deformation concentration is the result of bending deformation amount after absolute value standardization, periodic mean normalization, and data standardization based on historical cycle data. It is used to characterize the degree of local bending anomaly concentration.

[0100] Contact anomaly is the result of contact deviation data after absolute value standardization, periodic mean normalization, and historical cyclic data standardization, and is used to characterize the intensity of contact state anomalies.

[0101] The average cyclic force is the result of standardizing the instantaneous force by absolute value standardization, averaging within the period, and data standardization based on historical cyclic data. It is used to characterize the overall force level of the period.

[0102] The local risk amplification factor is a comprehensive risk amplification obtained by multiplying the offset of the eccentric hysteresis center, the concentration of bending deformation, and the degree of contact anomaly by 1, subtracting 1, and then standardizing it again. It is used to amplify the impact of local anomalies on the overall risk.

[0103] The equivalent stress index is a comprehensive stress index obtained by multiplying the local risk amplification factor by 1 and the mean value of cyclic stress, and then centering and standardizing it. It is used as the core quantity for threshold determination and early warning output.

[0104] In practice, the absolute value standardization process first takes the absolute value of the input sequence, then subtracts the median of the absolute value sequence of the reference set from the input value, and divides it by the sum of the interquartile range and the numerical stability term of the absolute value sequence of the reference set; the bending deformation, contact deviation data and standardized instantaneous force are all based on the corresponding sequences within the same complete historical cycle of the support as the reference set.

[0105] In practice, historical cyclic data will be retained in priority for the most recent 30 to 90 complete cycles, and updated according to the first-in-first-out principle after each new cycle is formed. When a cycle has more than 20 missing samples or when known abnormal events such as manual maintenance or external collisions occur, only the original records of that cycle will be kept and will not be included in the statistical reference.

[0106] In practice, all three indicators are aggregated using a single stent and a single complete cycle as the unit. The average of the standardized absolute values ​​at each moment within the cycle is calculated first, and then compared with historical cycle data to obtain a centralized standardized output. Multiple stents should not be mixed and averaged, nor should multiple cycles be directly merged.

[0107] In practice, the offset of the eccentric hysteresis center, the concentration of bending deformation, and the degree of contact anomaly are uniformly limited to a reasonable range formed by three times the interquartile range above and below the historical median before entering the product fusion. Those exceeding the range are treated as boundary values ​​to prevent a single outlier from causing abnormally high risk results. After fusion, the data is again processed by centralization and standardization according to the historical cyclic data to keep the results of different stents comparable.

[0108] In practice, when there are fewer than 5 valid historical complete cycles, the local risk amplification factor and the equivalent effect index are calculated using only the currently obtained historical samples to calculate the temporary median and the temporary interquartile range; when there are fewer than 3 historical samples, the formal limit violation judgment is not output for the time being, and only the observation status is output.

[0109] In practice, the equivalent stress index is used to represent relative risk rather than directly representing the material strength value. The equivalent stress index is preferentially divided into three levels: attention, warning and emergency. The attention zone corresponds to the dynamic safety threshold but is higher than the median level of historical health records. The warning zone corresponds to the dynamic safety threshold being close to or exceeding the dynamic safety threshold. The emergency zone corresponds to the dynamic safety threshold being significantly exceeded for multiple consecutive periods.

[0110] Preferably, the dynamic threshold module is specifically used for: The eccentric hysteresis center offset in the historical cycle data is sorted in ascending order according to the value. The historical cycle data with a ranking not greater than the ranking threshold are extracted to construct a historical health record. The ranking threshold is the floor value of the current complete cycle number minus one and divided by two. Extract the corresponding equivalent efficacy index from the historical health records, and calculate the median and interquartile range of the equivalent efficacy of the records; The dynamic safety threshold is generated by adding the median of the equivalent stress of the archives to the interquartile range of the equivalent stress of the archives. Divide the equivalent stress index of the current complete cycle by the sum of the dynamic safety threshold and the numerical stability term, and output the over-limit ratio. The formula for calculating the historical health records is as follows: ; The formula for calculating the dynamic security threshold is: ; The formula for calculating the over-limit ratio is: ; In the formula, Indicates the first The first support in The historical health records for each complete cycle; This represents the time sequence number in the historical cyclic data; This indicates the sequence number of the current complete cycle. The ranking function represents ascending order sorting; This indicates the offset of the eccentric hysteresis center under the corresponding sequence number; This represents the floor function; Indicates the first The first support in The dynamic security threshold for each complete cycle; A mathematical operator for calculating the median; This indicates the equivalent stress index under the corresponding serial number; Represents the mathematical operator for calculating interquartile range; Indicates the first The first support in The over-limit ratio for each of the complete cycle periods; Indicates the first The equivalent stress index of the stent in the current complete cycle; This represents the numerical stability term.

[0111] Historical health records are a set of relative health cycles selected from historical cyclic data after sorting by the offset of the eccentric hysteresis center.

[0112] The ranking threshold is the upper limit of the selection calculated based on the current complete cycle number.

[0113] The median equivalent stress in historical health records is the median value of the equivalent stress index within historical health records, used to characterize the typical stress level of a health sample.

[0114] The interquartile range of archival equivalence is the interquartile range of the equivalence index within historical health records, used to characterize the natural range of fluctuations in health samples.

[0115] The dynamic safety threshold is an adaptive safety margin obtained by adding the median of the equivalent stress of the archives to the interquartile range of the equivalent stress of the archives.

[0116] The limit exceedance ratio is the ratio of the equivalent stress index of the current complete cycle to the dynamic safety threshold. It is used to directly determine whether the limit has been exceeded and the degree of exceeding the limit.

[0117] In practice, historical health records are established independently for each stent. First, the eccentric hysteresis center offset of the complete historical cycle of the stent is sorted from smallest to largest. If the same value is found, the cycle with the earlier time and higher data completeness is retained first. To ensure the stability of the threshold, historical health records retain at least 3 complete cycles. If there are less than 3 after filtering according to the ranking threshold, the records are supplemented to 3 starting from the smallest offset.

[0118] In practice, the ranking threshold is obtained by subtracting 1 from the current complete cycle number, dividing by 2 and rounding down. However, if the result is less than 3, it is uniformly treated as 3. In the first 5 complete cycles of the system, the dynamic safety threshold adopts a temporary threshold mode, which is directly constructed by adding the interquartile range to the median of all currently obtained healthy samples.

[0119] In practice, when the number of historical health records is no less than 4, the median of the equivalent efficacy of the records is taken directly from the center position, and the interquartile range of the equivalent efficacy of the records is calculated by subtracting the lower quartile from the upper quartile; when the number of records is 3, the median of the equivalent efficacy of the records is taken as the middle value, and the interquartile range of the equivalent efficacy of the records is taken as half of the maximum value minus the minimum value.

[0120] In practice, when the sum of the median of the equivalent statistic of the records and the interquartile range of the equivalent statistic of the records is less than 0.05, the dynamic safety threshold is forcibly set to 0.05; if all values ​​in the health records are the same, the dynamic safety threshold is the larger of the same value and 0.05.

[0121] In practice, the over-limit ratio is updated using a complete cycle as the update unit, and is updated once for each new cycle. A single over-limit ratio greater than 1 indicates that the current cycle is over-limit. If the ratio is greater than 1 for two consecutive complete cycles, it is determined to be a stable over-limit. If the ratio is greater than 1 for three consecutive cycles and shows an upward trend, it is upgraded to a high-confidence over-limit.

[0122] In practice, historical health records must be established independently for each stent, and the eccentric hysteresis center offset and equivalent stress index of different stents must not be mixed and sorted. The reason is that the installation conditions, stress background and local contact state of different stents are different, and mixing them will mask individual differences.

[0123] Preferably, the early warning module is specifically used for: The over-limit ratio of each stent is exponentially calculated and divided by the sum of the exponential calculation results of all stents to generate the positioning contribution weight. The positioning contribution weight of each bracket is multiplied by the corresponding standardized position coordinate and summed to generate a normalized alarm position; Extract the bracket corresponding to the maximum value of the over-limit ratio as the alarm column; perform the centralized standardization process on the maximum value of the over-limit ratio based on the historical over-limit ratio to generate the alarm intensity; The sign operation result of the difference between the maximum value of the over-limit ratio and one is extracted as the warning status; When the warning status is determined to be an over-limit, a work order is triggered that includes the alarm column, the normalized alarm location, the alarm intensity, the equivalent stress index, the dynamic safety threshold, and the over-limit ratio. The formula for calculating the location contribution weight is as follows: ; The formula for calculating the normalized alarm location is: ; The formula for extracting the alarm column is: ; The formula for calculating the alarm intensity is: ; The formula for calculating the warning status is: ; In the formula, Indicates the first The first support in The positioning contribution weight for each complete cycle. Represents an exponential function. Indicates the first The first support in The over-limit ratio for each complete cycle. This represents the sum of the exponential results of the over-limit ratios of all stents; Indicates the first The normalized alarm position for each complete cycle. Indicates the first The standardized position coordinates of each bracket; Indicates the first The alarm column for each complete cycle of the aforementioned alarm. This indicates the operation of extracting the index corresponding to the maximum value; Indicates the first The alarm intensity for each complete cycle, This indicates the centralized standardization process. This represents the maximum value of the over-limit ratio. This represents the historical limit-crossing ratio; Indicates the first The warning status for each complete cycle. Represents a sign function, when The time indicates that the warning status is determined to be out of limit.

[0124] The positioning contribution weight is a relative weight formed by exponentially calculating and normalizing the over-limit ratio of each support, and is used to reflect the contribution of each support to the overall alarm position.

[0125] The normalized alarm position is the result of a weighted sum of the positioning contribution weights of each bracket and the corresponding standardized position coordinates, used to output the relative alarm position on a line.

[0126] The alarm column is the support with the highest over-limit ratio, used to indicate the specific support location that should be checked first.

[0127] Alarm intensity is the result of centralized standardization of the maximum out-of-limit ratio relative to the historical out-of-limit ratio, and is used to characterize the relative severity of the current out-of-limit event.

[0128] The historical out-of-limit ratio is a set of historical out-of-limit ratios stored in the system, used to evaluate the degree of abnormality of the current maximum out-of-limit ratio relative to historical events.

[0129] The warning status is a result obtained by comparing the maximum value of the over-limit ratio with 1, and is used to trigger or suppress work order output.

[0130] A work order is a structured record of actions taken when an alarm status is determined to be out of bounds. It is used to send the alarm status, location, intensity, and determination criteria to the operation and maintenance process.

[0131] In practice, the over-limit ratio of each stent is first calculated exponentially, and then the results of all exponents are summed and normalized so that the stent with the higher over-limit ratio receives a greater positioning contribution weight. When the over-limit ratios of multiple stents are close, the positioning contribution weight will be naturally distributed among these stents instead of being hard-cut at a single point.

[0132] In practice, the normalized alarm position is first output as a relative position within the range of 0 to 1, and then the actual support position coordinates of all supports are used to calculate the on-site mileage position. When it is necessary to generate a work order text, the normalized alarm position can be converted into a segment description between the two nearest supports.

[0133] In practice, when the over-limit ratios of two or more supports are tied for the highest, the support with the smaller position coordinates and closer to the normalized alarm position is selected as the alarm support; if one of the parallel supports has maintained a high over-limit ratio for the most recent three complete cycles, that support is selected first.

[0134] In practice, the historical over-limit ratio is stored for each stent for the most recent 30 to 100 complete cycles. The current alarm intensity uses the sequence of the largest over-limit ratio to construct a centralized standardized reference. After each new cycle, the current largest over-limit ratio is added to the historical over-limit ratio set and updated according to the first-in-first-out principle.

[0135] In practice, when the maximum value of the over-limit ratio is greater than 1, the warning status is set to 1 and indicates that the limit has been exceeded; when the maximum value of the over-limit ratio is equal to 1, the warning status is set to 0 and indicates that the critical observation has been completed; when the maximum value of the over-limit ratio is less than 1, the warning status is set to -1 and indicates that the limit has not been exceeded. To avoid fluctuations, once the warning status enters the over-limit phase, it must be maintained for at least one complete cycle. Only when the subsequent two consecutive complete cycles fall back to below 1 is a reset allowed.

[0136] In practice, a work order must include at least the alarm column, normalized alarm location, alarm intensity, equivalent stress index, dynamic safety threshold, over-limit ratio, formation time, number of consecutive over-limit cycles, and recommended handling level. Work orders are first generated by the edge terminal and then sent to the monitoring master station and the operation and maintenance platform. To prevent the same anomaly from being repeatedly dispatched in a short period of time, if the alarm columns of two adjacent complete cycles are the same and the difference in normalized alarm location does not exceed 0.05, they are merged into the same event, and only the severity is updated.

[0137] Example 2: A cable stress intelligent sensing and early warning method, applied to any of the cable stress intelligent sensing and early warning devices described in this embodiment, comprising: The monitoring sequence of the stent is synchronously acquired and standardized to generate multi-point force array data; Thermal effect baseline and contact deviation data are extracted from the multi-point force array data, the complete cycle is divided, and heating weight and cooling weight are generated. The multi-point force array data is mapped using a swarm intelligence algorithm to generate potential deformation and bending deformation. Based on the thermal effect baseline, the potential deformation, the heating weight, and the cooling weight within the complete cycle, the eccentric hysteresis center offset is extracted. By fusing the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplifying the average force of the multi-point force array data within the complete cycle, an equivalent stress index is obtained. A dynamic safety threshold is generated based on historical cyclic data, and the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold is calculated. The alarm column and warning status are determined based on the over-limit ratio, and a work order is triggered when the warning status exceeds the limit.

[0138] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A cable stress intelligent sensing and early warning device, characterized in that, include: The force acquisition module is used to synchronously acquire and standardize the monitoring sequence of the support, and generate multi-point force array data; The reference separation module is used to extract thermal effect baseline and contact deviation data from the multi-point force array data, divide the complete cycle, and generate heating weight and cooling weight. The deformation reconstruction module is used to map the multi-point force array data through a swarm intelligence algorithm to generate potential deformation and bending deformation. The offset extraction module is used to extract the offset of the eccentric hysteresis center based on the thermal effect baseline, the potential deformation, the heating weight and the cooling weight within the complete cycle. The stress assessment module is used to fuse the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplify the average stress value of the multi-point force array data in the complete cycle to obtain the equivalent stress index. The dynamic threshold module is used to generate a dynamic safety threshold based on historical cyclic data and calculate the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold. The early warning module is used to determine the alarm column and early warning status based on the over-limit ratio, and to trigger a work order when the early warning status exceeds the limit.

2. The intelligent sensing and early warning device for cable stress according to claim 1, characterized in that, The force acquisition module is specifically used for: Obtain the position coordinates of the stent and map them to standardized position coordinates through range standardization. The original monitored forces within the data time window are acquired synchronously based on a unified time reference; when the data loss time does not exceed 10% of the median sampling interval of the thermal cycle, linear time interpolation is used to supplement the data. The original monitored force is mapped to a standardized instantaneous force through centralized standardization processing, and the difference between the original monitored forces at adjacent times is mapped to a standardized force increment. The standardized instantaneous forces of each support at the same moment are combined to generate the multi-point force array data; Wherein, the range standardization is the ratio of the difference between the input value and the minimum value to the sum of the range and the numerical stability term; the centering standardization is the ratio of the difference between the input value and the median of the window data to the sum of the interquartile range of the window data and the numerical stability term.

3. The intelligent sensing and early warning device for cable stress according to claim 2, characterized in that, The reference separation module is specifically used for: Identify the local extreme values ​​of the standardized instantaneous force of each support, and construct the upper and lower force envelopes through shape-preserving interpolation; The mean values ​​of the upper force envelope and the lower force envelope are subjected to the centering and standardization process to generate the thermal effect baseline. The difference between the standardized instantaneous force and the thermal effect baseline is subjected to the centering and standardization process to extract the contact deviation data; The difference between the thermal effect baselines at adjacent time points is subjected to the centering and standardization process to obtain the change in the thermal effect baseline. When the change in the thermal effect baseline shows two consecutive positive and negative inflections and includes a complete rise and fall process, it is defined as a complete cycle. Based on the hyperbolic tangent transform result of the baseline change in the thermal effect, the heating weight and the cooling weight are generated respectively through the range normalization process.

4. The intelligent sensing and early warning device for cable stress according to claim 3, characterized in that, The deformation reconstruction module is specifically used for: Spline interpolation is performed based on the standardized position coordinates of each support and the corresponding standardized instantaneous force to generate a continuous spatial function; The first and second derivatives of the continuous spatial function are calculated and then the centering and normalization processes are performed to generate the spatial slope and the bending deformation, respectively. Initialize the projection weight vector with non-negative elements and a sum of one; The projected weight vector is used to weight the standardized instantaneous force, the contact deviation data, the spatial slope, and the bending deformation, and then the centering and standardization process is performed to generate the potential deformation. The potential deformation and the hysteresis area, center drift, and high-frequency fragmentation of the thermal effect baseline during the complete cycle are calculated to construct a fitness function. The fitness function is used to optimize the projection weight vector, and the potential deformation corresponding to the optimal projection weight vector is output.

5. The intelligent sensing and early warning device for cable stress according to claim 4, characterized in that, The offset extraction module is specifically used for: The mean values ​​of the thermal effect baseline and the potential deformation within the complete cycle are calculated and then the centering and normalization process is performed to form the current cycle geometric center. Weights are constructed by combining the lap area within the complete historical cycle, and the historical reference geometric center is calculated by weighted geometric median; Calculate the Euclidean distance between the current cyclic geometric center and the historical reference geometric center, and perform absolute value standardization to obtain the unidirectional drift distance; The potential deformation is weighted and averaged using the heating weight and the cooling weight respectively, and then the centering and standardization process is performed to obtain the heating branch center and the cooling branch center respectively. Calculate the absolute difference between the heating branch center and the cooling branch center and perform absolute value standardization to obtain the branch asymmetry. The unidirectional drift distance and the branch asymmetry are multiplied by one and then fused, and then the centering and standardization process is performed to extract the eccentric hysteresis center offset.

6. The intelligent sensing and early warning device for cable stress according to claim 5, characterized in that, The stress assessment module is specifically used for: The bending deformation, the contact deviation data, and the standardized instantaneous force are respectively processed by absolute value standardization to obtain the mean value within the complete cycle. Based on the historical cycle data, the centering standardization process is performed to generate the bending deformation concentration, contact anomaly, and cycle force mean value, respectively. The eccentric hysteresis center offset, the bending deformation concentration and the contact anomaly are all added by one and then multiplied. The product is then subtracted by one and the centralization and standardization processing is performed based on the historical cyclic data to generate a local risk amplification factor. The local risk amplification factor is incremented by one and multiplied by the mean cyclic force. Based on the historical cyclic data, the centralization and standardization process is performed to output the equivalent stress index.

7. The intelligent sensing and early warning device for cable stress according to claim 6, characterized in that, The dynamic threshold module is specifically used for: The eccentric hysteresis center offset in the historical cycle data is sorted in ascending order according to the value. The historical cycle data with a ranking not greater than the ranking threshold are extracted to construct a historical health record. The ranking threshold is the floor value of the current complete cycle number minus one and divided by two. Extract the corresponding equivalent efficacy index from the historical health records, and calculate the median and interquartile range of the equivalent efficacy of the records; The dynamic safety threshold is generated by adding the median of the equivalent stress of the archives to the interquartile range of the equivalent stress of the archives. The equivalent stress index of the current complete cycle is divided by the sum of the dynamic safety threshold and the numerical stability term to output the over-limit ratio.

8. The intelligent sensing and early warning device for cable stress according to claim 7, characterized in that, The early warning module is specifically used for: The over-limit ratio of each stent is exponentially calculated and divided by the sum of the exponential calculation results of all stents to generate the positioning contribution weight. The positioning contribution weight of each bracket is multiplied by the corresponding standardized position coordinate and summed to generate a normalized alarm position; The bracket corresponding to the maximum value of the over-limit ratio is extracted as the alarm column; The maximum value of the over-limit ratio is subjected to the centralized standardization process based on the historical over-limit ratio values ​​to generate the alarm strength; The sign operation result of the difference between the maximum value of the over-limit ratio and one is extracted as the warning status; When the warning status is determined to be an out-of-limit condition, a work order is triggered that includes the alarm column, the normalized alarm location, the alarm intensity, the equivalent stress index, the dynamic safety threshold, and the out-of-limit ratio.

9. A cable stress intelligent sensing and early warning method, applied in the cable stress intelligent sensing and early warning device according to any one of claims 1-8, characterized in that, include: The monitoring sequence of the stent is synchronously acquired and standardized to generate multi-point force array data; Thermal effect baseline and contact deviation data are extracted from the multi-point force array data, the complete cycle is divided, and heating weight and cooling weight are generated. The multi-point force array data is mapped using a swarm intelligence algorithm to generate potential deformation and bending deformation. Based on the thermal effect baseline, the potential deformation, the heating weight, and the cooling weight within the complete cycle, the eccentric hysteresis center offset is extracted. By fusing the eccentric hysteresis center offset, the bending deformation, and the contact deviation data, and amplifying the average force of the multi-point force array data within the complete cycle, an equivalent stress index is obtained. A dynamic safety threshold is generated based on historical cyclic data, and the over-limit ratio of the equivalent stress index relative to the dynamic safety threshold is calculated. The alarm column and warning status are determined based on the over-limit ratio, and a work order is triggered when the warning status exceeds the limit.

Images