Method for detecting breakage of film-covered wire

Abstract

The application relates to the technical field of fault prediction and health management, and discloses a film-covered wire breakage detection method, which comprises the following steps: a static impedance reference model containing viscoelastic characteristics is constructed, and a time constant representing the physical relaxation speed of insulating material is extracted; a phase correction operation is performed on an instruction motion posture coordinate by using an instant motion speed vector and the time constant, equivalent physical posture coordinates are obtained; theoretical electrical characteristic parameters are inquired in the static impedance reference model by using the equivalent physical posture coordinates, and residual characteristic sequences are differentially calculated with measured values to evaluate health factors; the application corrects kinematic coordinates by introducing a material time constant, eliminates measurement reference deviation caused by the viscoelastic hysteresis of high polymer material under high dynamic working conditions, and accurately captures and trends forecasts weak early damage signals.

Description

Technical Field

[0001] This invention relates to a method for detecting broken membrane-covered wires, belonging to the field of fault prediction and health management technology. Background Technology

[0002] In current automated systems of industrial robots, high-speed placement machines, and precision medical imaging equipment, flexible flat cables and flexible printed circuits, among other membrane-wrapped wires, serve as core components for dynamic electrical connections. These wires undergo high-frequency reciprocating bending with the motion actuators. Existing membrane-wrapped wire continuity detection technologies generally employ DC resistance monitoring or low-frequency impedance measurement methods, setting fixed resistance or impedance thresholds. When the monitored value exceeds the threshold, an open circuit is identified. Based on static threshold detection, these methods can identify complete conductor breakage faults when the equipment is stationary or operating at low speeds. However, in high-speed, high-acceleration dynamic operating scenarios, the geometry of the membrane-wrapped wire changes drastically, and its distributed parameters fluctuate in real time with bending and torsion. This leads to significant benign jumps in line impedance. In the early stages of fatigue microcrack initiation inside the conductor, the changes in impedance are weak and usually submerged in the background impedance fluctuations caused by geometric deformation. Existing monitoring methods struggle to distinguish between benign fluctuations caused by motion changes and malignant drifts caused by material damage in time-varying signals. Increasing the alarm threshold to avoid motion interference fails to capture early micro-damage signals and can only detect complete fracture failure. Lowering the threshold to increase sensitivity is prone to false alarms due to normal equipment operation. In addition, during long-term service, the aging of insulation materials, changes in environmental temperature and humidity, and the viscoelastic hysteresis effect of polymer materials exacerbate the drift and uncertainty of the reference value.

[0003] Besides the limitations of conventional detection methods, existing general health management strategies also suffer from insufficient specificity when dealing with such highly dynamic and strongly coupled specific physical field conditions. For example, Chinese invention patent CN114781667A discloses a multi-device full lifecycle PHM health management and predictive maintenance platform. The solution constructs a database of information on multiple types of equipment and a hierarchical maintenance system, focusing on achieving macro-standardization and process management of massive data through the Industrial Internet. However, when dealing with membrane envelope components with viscoelastic characteristics of polymer materials, the general algorithm model has obvious applicability boundaries. Such platforms lack physical-level decoupling mechanisms for high-speed motion accompanied by dynamic geometric deformation and material hysteresis effects. Relying on general sensor data or statistical probability models, they cannot fundamentally eliminate background impedance noise caused by changes in motion posture. When weak impedance changes caused by early microcracks are submerged in large-amplitude dynamic benign fluctuations, such management platforms, lacking dynamic benchmark correction capabilities, struggle to accurately capture micro-damage characteristics.

[0004] Therefore, the technical problem to be solved by this invention is to construct a membrane envelope health monitoring mechanism that can decouple motion attitude interference in real time, adaptively offset environmental and aging drift, and accurately correct dynamic hysteresis errors, so as to achieve early identification of microscopic damage and lifetime prediction in the context of strong noise. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for detecting membrane wrapping thread breakage, comprising the following steps: A static impedance reference model incorporating viscoelastic characteristics is constructed. With the membrane envelope in a defined initial healthy state, the drive device traverses the preset working path, synchronously collects the motion attitude coordinates and ambient temperature of the membrane envelope endpoints, and records the reference electrical characteristic parameters in the corresponding states. A static mapping relationship is established with motion attitude coordinates and ambient temperature as index keys and reference electrical characteristic parameters as values. The material hysteresis time constant is extracted. By comparing the impedance change response characteristics of the membrane-coated wire under the first and second motion speed conditions, the time constant characterizing the physical relaxation speed of the membrane-coated wire insulation material is calculated. The first motion speed produces a measurable viscoelastic hysteresis, and the second motion speed is lower than the first motion speed to simulate a quasi-static process. Real-time kinematics calculation and phase correction: During the operation of the membrane envelope, the current command motion attitude coordinates are collected in real time, the rate of change of the command motion attitude coordinates with time is calculated to obtain the instantaneous motion velocity vector, and the instantaneous motion velocity vector and time constant are used to perform phase correction operation on the command motion attitude coordinates to obtain the equivalent physical attitude coordinates that characterize the current actual geometric deformation of the membrane envelope. Dynamic baseline generation and differential assessment utilize equivalent physical attitude coordinates as corrected index keys to query the reference electrical characteristic parameters corresponding to the current moment in the static impedance reference model and define them as theoretical electrical characteristic parameters. The measured electrical characteristic parameters and theoretical electrical characteristic parameters are differentially calculated to obtain the residual characteristic sequence that eliminates the interference of material deformation hysteresis. Based on the monotonic evolution trend of the residual characteristic sequence, the health factor characterizing the degree of damage is calculated.

[0006] Preferably, in the real-time kinematics calculation and phase correction steps, the phase correction operation is strictly performed according to the following mathematical relationship, so as to use the time constant to offset the geometric deformation hysteresis caused by the viscoelasticity of the polymer material during high-dynamic motion: , where P eff (t) is defined as the equivalent physical attitude coordinates, P cmd v(t) is defined as the commanded motion attitude coordinates, v(t) is defined as the instantaneous motion velocity vector, τ is defined as the time constant, and k is defined as a weighting coefficient preset based on the damping characteristics of the membrane envelope material.

[0007] Preferably, the dynamic baseline generation and differential evaluation steps further include an adaptive drift compensation process based on the anchor point state. This process includes: pre-setting at least one specific attitude in the device's operating trajectory as the anchor point attitude, and recording the initial anchor point impedance value corresponding to the anchor point attitude in the initial healthy state; when the equivalent physical attitude coordinates are detected to match the anchor point attitude, calculating the deviation between the current measured electrical characteristic parameters and the initial anchor point impedance value, and defining the deviation as a global drift factor; and using the global drift factor to globally weight and correct the theoretical electrical characteristic parameters output by the static impedance reference model, so that the static impedance reference model follows the non-damaging aging trend of the membrane envelope.

[0008] Preferably, the membrane envelope contains at least two parallel conductors. The dynamic baseline generation and differential evaluation steps specifically include a topological second-order differential process: selecting one conductor in the membrane envelope as the target conductor and selecting another conductor physically adjacent to the target conductor as the reference conductor; calculating the primary residuals between the measured electrical characteristic parameters of the target conductor and the reference conductor at the current moment and the theoretical electrical characteristic parameters obtained based on their equivalent physical attitude coordinates; calculating the algebraic difference between the primary residuals of the target conductor and the primary residuals of the reference conductor to obtain the second-order topological residual features; using the second-order topological residual features as a residual feature sequence to eliminate common-mode environmental interference and mechanical vibration noise acting on the entire membrane envelope.

[0009] Preferably, the measured electrical characteristic parameters include a first frequency impedance parameter and a second frequency impedance parameter, with the second frequency being higher than the first frequency; the dynamic baseline generation and differential evaluation steps include a dual-frequency sensitivity gating process, calculating the first residual between the measured and theoretical values ​​of the first frequency impedance parameter, and the second residual between the measured and theoretical values ​​of the second frequency impedance parameter; using the ratio of the second residual to the first residual to construct a damage discrimination logic, when the amplitude of the second residual increases while the amplitude of the first residual remains within a preset range, causing the ratio to exceed the damage judgment threshold, it is confirmed that the residual characteristic sequence is mainly caused by conductor micro-damage and the residual characteristic sequence is retained; when the first residual and the second residual increase simultaneously, causing the ratio to remain in the non-damage range, it is determined to be motion posture deviation interference and the residual characteristic sequence at that moment is suppressed.

[0010] Preferably, the step of extracting the material hysteresis time constant specifically includes: in the offline calibration stage, the drive device performs reciprocating motion at a first motion speed and records a first impedance change curve, the drive device performs the same reciprocating motion at a second motion speed and records a second impedance change curve; calculate the phase lag of the first impedance change curve relative to the second impedance change curve on the time axis; and determine the time constant using the ratio of the phase lag to the first motion speed.

[0011] Preferably, the health factor calculation process includes: performing sliding window integration on the residual feature sequence to obtain the cumulative damage energy value; monitoring the slope of the cumulative damage energy value over time; and determining that the membrane envelope has entered the late stage of fatigue failure and generating an early warning signal when the slope exceeds a preset micro-damage warning threshold and exhibits an irreversible monotonous growth characteristic.

[0012] Preferably, the reference electrical characteristic parameter is the AC impedance modulus at a specified high frequency. The selection of the specified high frequency is based on ensuring that the skin depth of the membrane conductor is less than one-third of the conductor thickness, thereby enhancing the sensitivity to the detection of microcracks on the conductor surface. The static impedance reference model is stored in the form of a multidimensional lookup table. The lookup process includes linear interpolation calculation based on the equivalent physical attitude coordinates and ambient temperature in the multidimensional lookup table.

[0013] Preferably, before the dynamic baseline generation and differential evaluation steps, a reference model initialization process is also included, in which the device is placed in a stationary state and all preset temperature points are traversed to obtain the pure temperature drift curve in the zero-velocity state; the ambient temperature dimension in the static impedance reference model is pre-calibrated using the pure temperature drift curve to establish the initial zero point of the static mapping relationship.

[0014] Preferably, after the step of generating the early warning signal, a closed-loop control intervention step is also included. In response to the health factor exceeding the preset safety limit, a deceleration command is sent to the equipment controller to forcibly reduce the operating acceleration and maximum speed of the equipment, reduce the dynamic mechanical stress acting on the membrane envelope, and delay the fracture process until maintenance and replacement are completed.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention constructs a dynamic impedance benchmark model that includes motion posture and environmental variables. The macroscopic impedance fluctuations caused by geometric bending and torsion during the reciprocating motion of the membrane envelope are defined as the inherent expected characteristics of the system. Real-time monitoring uses differential operations to extract the expected characteristics from the measured signal, obtaining the residual sequence caused only by microscopic damage to the conductor. Based on the decoupling mechanism of deterministic imaging, the influence of dynamic masking effect is eliminated in principle. This enables the measurement unit to lock the early microcrack signal at the micro-ohm level under the background of benign geometric impedance fluctuations at the level of hundreds of milliohms, overcoming the detection problem that benign deformation and malignant damage cannot be distinguished by static threshold in high-frequency dynamic scenarios.

[0016] 2. By utilizing the inherent anchor point attitude of the equipment's operating trajectory, an online adaptive drift compensation loop is established for the benchmark model. By capturing the long-term impedance deviation trend under specific low-stress anchor points of the cable, the loop distinguishes between global impedance drift caused by insulation material aging and connector contact surface oxidation and local impedance mutation caused by fatigue damage. Non-destructive aging factors are fed back to the benchmark model for dynamic correction, ensuring that the dynamic baseline automatically follows the evolution of the cable's physical characteristics throughout its entire life cycle. This avoids the risk of false alarms caused by the accumulation of long-term environmental factors and maintains the reliability of long-term operational fault prediction.

[0017] 3. A phase correction logic based on the viscoelastic hysteresis characteristics of materials is introduced to solve the spatiotemporal misalignment problem of cable mechanical deformation lagging behind motion commands under high dynamic conditions. By combining the real-time motion velocity vector and the material relaxation time constant, an equivalent physical attitude coordinate representing the current actual geometric shape of the cable is calculated and constructed. This coordinate replaces the original command coordinate as the reference model query index. The processing path eliminates the pseudo residual signal caused by the shape response delay of polymer materials from the physical level, ensuring that the robot maintains an extremely low background noise level consistent with the stationary state when performing high acceleration actions such as emergency stop and sharp turn, and realizing blind-spot-free monitoring of the dynamic process of the equipment in the full speed domain. Attached Figure Description

[0018] Figure 1 This is a flowchart of the membrane wrapping wire breakage detection method of the present invention; Figure 2 This is a comparison diagram of the impedance response characteristics before and after phase correction in this invention; Figure 3 This is a state transition diagram of the operating logic of the monitoring system of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention discloses a method for detecting membrane envelope wire breakage, including steps such as constructing a static impedance reference model incorporating viscoelastic characteristics, extracting the material hysteresis time constant, real-time kinematic calculation and phase correction, dynamic baseline generation, and differential evaluation. Addressing the technical problem that impedance fluctuations caused by geometric deformation of the membrane envelope under dynamic conditions can mask early, weak damage signals, this method performs the step of constructing a static impedance reference model incorporating viscoelastic characteristics. With the membrane envelope in a defined initial healthy state, the driving device is controlled to traverse a preset working path, which covers the global attitude space of the device. During this process, the motion attitude coordinates of the membrane envelope endpoints and the ambient temperature are simultaneously acquired. A high-precision impedance measurement unit is used to record the reference electrical characteristic parameters under the corresponding state. The reference electrical characteristic parameters are selected from a specified high-frequency range. The AC impedance modulus is set such that the skin depth of the membrane conductor is less than one-third of the conductor thickness. For example, a frequency band of 10MHz to 100MHz is selected. The high-frequency skin effect is used to enhance the detection sensitivity of microcracks on the conductor surface. The system establishes a static mapping relationship with motion attitude coordinates and ambient temperature as index keys and reference electrical characteristic parameters as values. This static mapping relationship is stored in the form of a multidimensional lookup table. The query process uses linear interpolation calculation based on the index key. In addition, a reference model initialization process is performed before building the model. The device is placed in a stationary state and all preset temperature points are traversed to obtain the pure temperature drift curve under zero velocity state. This curve is used to pre-calibrate the ambient temperature dimension in the static impedance reference model and establish the initial zero point of the static mapping relationship.

[0021] Given the viscoelastic characteristics of the membrane-coated wire insulation material, which causes the actual deformation of the cable during high-speed movement to lag behind the motion command, this method performs a step to extract the material's hysteresis time constant. During offline calibration, the drive device is controlled to perform reciprocating motion at a first speed and a first impedance change curve is recorded. This first speed is set to be sufficient to produce measurable viscoelastic hysteresis, for example, 80% of the device's rated maximum speed. The drive device is then controlled to perform the same reciprocating motion at a second speed and a second impedance change curve is recorded. This second speed is lower than the first speed to simulate a quasi-static process, for example, 5% of the device's rated maximum speed. The phase lag of the first impedance change curve relative to the second impedance change curve on the time axis is calculated, and the ratio of this phase lag to the first speed is calculated to determine the time constant τ characterizing the physical relaxation speed of the membrane-coated wire insulation material. During actual operation, real-time kinematic calculation and phase correction steps are performed. The system continuously acquires the current commanded motion attitude coordinates P. cmd (t), calculate the rate of change of this coordinate with time to obtain the instantaneous motion velocity vector v(t), and use the instantaneous motion velocity vector and time constant to perform phase correction operation on the commanded motion attitude coordinates to obtain the equivalent physical attitude coordinates P that characterize the current actual geometric deformation of the membrane envelope.eff (t), the phase correction operation is strictly performed according to the following mathematical relation: Where k is a weighting coefficient preset based on the damping characteristics of the membrane envelope material, the correction amount in the time dimension is used to offset the geometric deformation lag caused by the viscoelasticity of the polymer material during high dynamic motion, and to ensure the spatiotemporal alignment of the model index key with the actual physical state of the cable; the dynamic baseline generation and differential evaluation steps are performed, using the equivalent physical attitude coordinates P eff (t) is used as the corrected index key. The reference electrical characteristic parameter corresponding to the current time is obtained by querying the static impedance reference model and it is defined as the theoretical electrical characteristic parameter. The measured electrical characteristic parameter and the theoretical electrical characteristic parameter are differentially calculated to obtain the residual characteristic sequence that eliminates the interference of material deformation hysteresis.

[0022] To eliminate the impact of global impedance drift caused by material aging or environmental factors on monitoring accuracy, this step integrates an adaptive drift compensation process based on anchor point status. At least one specific posture in the equipment's operating trajectory is pre-set as the anchor point posture. The initial anchor point impedance value corresponding to the anchor point posture in the initial healthy state is recorded. When the equivalent physical posture coordinates match the anchor point posture, the deviation between the current measured electrical characteristic parameters and the initial anchor point impedance value is calculated. This deviation is defined as the global drift factor. The global drift factor is used to globally weight and correct the theoretical electrical characteristic parameters output by the static impedance reference model, ensuring that the model baseline follows the non-destructive aging trend of the membrane envelope. For common-mode interference in the environment, this step further employs a topological second-order difference process. For a membrane envelope containing at least two parallel conductors, one of them is selected as the target conductor, and the other conductor physically adjacent to the target conductor is selected as the reference conductor. The primary residuals between the measured electrical characteristic parameters of the target conductor and the reference conductor at the current moment and the theoretical electrical characteristic parameters obtained based on their equivalent physical attitude coordinates are calculated respectively. The algebraic difference between the primary residuals of the target conductor and the primary residuals of the reference conductor is calculated to obtain the second-order topological residual features. These second-order topological residual features serve as the final residual feature sequence, used to eliminate common-mode environmental interference and mechanical vibration noise acting on the entire membrane envelope.

[0023] To distinguish between benign motion interference and malignant damage signals, this step also employs a dual-frequency sensitivity gating process. The measured electrical characteristic parameters include a first-frequency impedance parameter and a second-frequency impedance parameter, with the second frequency being higher than the first frequency. The first residual between the measured and theoretical values ​​of the first-frequency impedance parameter, and the second residual between the measured and theoretical values ​​of the second-frequency impedance parameter, are calculated. The ratio of the second residual to the first residual is used to construct a damage discrimination logic: when the amplitude of the second residual increases while the amplitude of the first residual remains within a preset range, causing the ratio to exceed the damage judgment threshold, it is confirmed that the residual characteristic sequence is mainly caused by conductor micro-damage, and this residual characteristic sequence is retained; when the first and second residuals increase simultaneously, causing the ratio to remain within a preset range... In the non-damage zone, the problem is identified as motion posture deviation interference, and the residual feature sequence at that moment is suppressed. Finally, the health factor is calculated based on the residual feature sequence, and closed-loop control is executed. The residual feature sequence is subjected to sliding window integration to obtain the cumulative damage energy value. The slope of the cumulative damage energy value over time is monitored. When the slope exceeds the preset micro-damage warning threshold and shows an irreversible monotonous growth characteristic, the membrane envelope is determined to have entered the late stage of fatigue failure, and an early warning signal is generated. In response to the early warning signal or the health factor exceeding the preset safety limit, the system sends a deceleration command to the equipment controller to forcibly reduce the equipment's operating acceleration and maximum speed, reduce the dynamic mechanical stress acting on the membrane envelope, and delay the fracture process.

[0024] Example 1: In a typical automotive electronics surface mount production line application scenario, the end effector of a six-axis industrial robot performs high-frequency reciprocating pick-and-place operations, with a peak joint angular velocity of 180.0 degrees / second and a running acceleration of 2.0g. The flexible flat cable connecting each joint undergoes nonlinear bending and torsion with the robot's high-speed movement. The cable's insulation and sheath are made of polyimide, which exhibits viscoelastic properties. This results in a millisecond-level time lag between the cable's mechanical deformation response under high dynamic stress and the commanded motion coordinates fed back by the motor encoder. This spatiotemporal misalignment caused by the material's physical properties, under uncompensated monitoring mode, leads to a discrepancy between the theoretical electrical characteristic parameters output by the static impedance reference model and the measured values ​​at the current moment. A pseudo-residual on the order of 50.0 milliohms is generated. This pseudo-residual amplitude is higher than the impedance change of 2.0 to 5.0 milliohms caused by early microcracks in the conductor, causing the weak damage signal to be masked by dynamic geometric noise. During the offline calibration phase, the system executes the material hysteresis time constant extraction program, drives the robot to run the standard operation trajectory at 100% and 5% of the rated speed respectively, and simultaneously records the impedance response curves under the two sets of working conditions. The system determines the time constant τ, which characterizes the physical relaxation speed of the insulation material of this specific batch of flexible flat cables, to be 12.5 milliseconds by calculating the phase offset of the high dynamic curve relative to the quasi-static curve on the time axis. During the online monitoring phase, the system reads the current command motion posture coordinates P from the robot controller at a sampling frequency of 1.0 kHz. cmd The system calculates the instantaneous velocity vector v(t) using real-time differential calculation. The system utilizes the preset weighting coefficient k and the time constant τ obtained from the above calibration, and applies the phase correction formula... Real-time calculations are performed, where the weighting coefficient k is set to 1.0 in this embodiment. When a robot joint is in an acceleration phase, the command motion posture coordinate P is... cmd When the indicated value (t) is 45.00 degrees and the instantaneous motion velocity vector v(t) is 100.00 degrees / second, the system calculates a correction term of 1.25 degrees, thus obtaining the equivalent physical attitude coordinates P that characterize the current actual physical bending state of the cable. eff (t) is 43.75 degrees. This calculation result shows that the actual geometry of the cable remains at a position equivalent to 43.75 degrees due to the viscoelastic hysteresis of the material.

[0025] The system uses the equivalent physical attitude coordinates of 43.75 degrees and the current ambient temperature of 40.0℃ as index keys to look up the theoretical electrical characteristic parameter of 1.250 ohms in the static impedance reference model. Simultaneously, the high-frequency impedance measurement unit acquires the measured electrical characteristic parameter at 10.0MHz, which is 1.254 ohms. The system calculates the primary residual of 4.0 milliohms by subtracting the measured value from the theoretical value. Compared to the theoretical value of 1.300 ohms obtained by directly looking up the table using the uncorrected command motion attitude coordinates of 45.00 degrees, and the resulting pseudo-residual of 46.0 milliohms, this phase correction mechanism suppresses dynamic geometric noise to a background level below 1.0 milliohms, revealing the true residual of 4.0 milliohms. To further confirm the source of the residual, the system performs the same processing procedure on the physically adjacent reference conductor, calculating the primary residual of the reference conductor to be 0.5 milliohms. The system then executes topology two... The second-order topological residual characteristic of 3.5 milliohms is obtained by subtracting the 0.5 milliohm residual of the reference conductor from the 4.0 milliohm residual of the target conductor. This is used to eliminate the common-mode temperature drift of the primary side measurement circuit and the contact resistance fluctuation caused by the overall mechanical vibration of the cable. The system retrieves the impedance data at a low frequency of 10.0 kHz for dual-frequency sensitivity comparison and confirms that the low-frequency residual component is 0.1 milliohms. The asymmetry of the amplitude of the high-frequency and low-frequency residuals is consistent with the skin effect characteristics of microcracks on the conductor surface. Based on the above judgment, the system confirms that microcracks have induced inside the cable conductor and includes the 3.5 milliohm residual value in the cumulative damage energy model. When the slope of the cumulative damage energy value changes irreversibly and monotonically, and exceeds the micro-damage warning threshold, the system determines that the membrane-coated wire has entered the late stage of fatigue failure, sends a warning signal to the main control computer of the production line and triggers an active deceleration protection strategy, limiting the maximum running acceleration of the robot to 50%.

[0026] Example 2: This example establishes a membrane envelope dynamic impedance testing platform in a controlled laboratory environment. The aim is to objectively verify the signal extraction capability, weak damage identification accuracy, and sensitivity to key parameter boundaries of the method of this invention under complex physical field coupling conditions by introducing a multi-dimensional comparative experimental system. The experimental platform consists of a six-degree-of-freedom motion simulation stage, a programmable temperature and humidity environmental chamber, and a high-precision broadband impedance analyzer. The motion simulation stage accurately reproduces the motion trajectory of industrial robot joints, with an angular velocity control accuracy of 0.1 degrees / second and a maximum acceleration set to 5.0g. The environmental chamber simulates temperature cycling conditions from -20.0℃ to 80.0℃, with a temperature control accuracy of 0.5℃. The impedance analyzer's measurement frequency range covers 10.0Hz to 100.0MHz, with a basic measurement accuracy of 0.05%. The membrane-coated cable sample used in the experiment was a 1.5-meter-long polyimide-insulated flexible flat cable containing 12 parallel conductors. To simulate the electromagnetic environment of a real industrial site, Gaussian white noise with a signal-to-noise ratio of 20dB and a power frequency interference signal with a frequency of 50.0Hz were actively coupled into the signal acquisition circuit to construct a noise background with a high degree of engineering realism. The experiment focused on setting the core sampling parameter of the system, namely the impedance sampling frequency. The determination of this parameter followed the following engineering decision-making logic: it was identified that the impedance fluctuation frequency caused by the motion of the membrane-coated cable was mainly concentrated in the fundamental frequency and its low harmonic range of the motion command, and the physical relaxation process of the material had a specific time scale; secondly, the technical trade-off was that increasing the sampling rate could more accurately capture transient lag, but it would increase the computational load of the real-time data processor; according to the Nyquist sampling theorem and control theory, the sampling frequency should be at least 10 times the effective bandwidth of the measured signal to avoid aliasing and ensure phase reconstruction accuracy. Under the experimental conditions, the main motion frequency was approximately 2.0Hz, considering the material relaxation time constant. The time interval is 12.5 milliseconds (corresponding to a characteristic frequency of approximately 80.0 Hz). To ensure a complete analysis of this hysteresis characteristic, the sampling frequency is set to 1.0 kHz.

[0027] To comprehensively evaluate the technical effects of this invention and the synergistic effects of its key features, the experimental design included a multi-dimensional control system comprising a control group and a gradient verification group. The specific groupings are as follows: Control Group A: Using the traditional fixed threshold detection method, only the impedance magnitude at a single frequency (10.0 kHz) was monitored, without any phase correction or differential mechanism, to simulate existing technical benchmarks. Partially missing control group B: Introduced a time constant-based... The phase correction algorithm is described, but it only performs single-frequency monitoring on a single conductor, lacking topological second-order difference and dual-frequency gating mechanisms. This is used to verify the phase correction's effect on suppressing dynamic geometric noise. Sample group C of this invention fully executes the phase correction, topological second-order difference, and dual-frequency sensitivity gating processes to verify the synergistic effect of the complete technical solution. Gradient verification groups D1-D3 all adopt the complete technical solution of this invention, setting the depth of the pre-fabricated microcracks to 2.0% (early initiation) and 10.0% (mid-stage) of the conductor thickness, respectively. The experiment used two parameters to verify the gradient response law of the technical effect as the severity of the problem changes: 40.0% (late-stage expansion) and 40.0% (late-stage failure). During the experiment, the motion simulation table was driven to perform sinusoidal reciprocating motion at an angular velocity of 100.0 degrees / second, while the ambient temperature rose from 25.0℃ to 60.0℃ at a rate of 2.0℃ / minute. At the 10,000th cycle, microcracks of the corresponding depth were pre-introduced on the surface of the target conductor using a precision scratcher. Experimental data showed that in the healthy stage before crack introduction, the affected... The combined effects of motion geometry deformation and ambient temperature drift resulted in a large periodic fluctuation in the measured impedance value of control group A, ranging from 45.0 milliohms to 55.0 milliohms, accompanied by a significant temperature baseline drift and an extremely low signal-to-noise ratio. Control group B, by introducing phase correction, successfully suppressed the periodic fluctuations caused by motion hysteresis to within 5.0 milliohms, demonstrating the effectiveness of phase correction in suppressing dynamic geometric noise. Due to the lack of a common-mode suppression mechanism, the impedance curve still showed an overall drift of approximately 8.0 milliohms with increasing temperature. After phase correction and topological second-order difference processing, the residual characteristic sequence of sample group C of this invention remained stable at a background noise level of less than 0.5 milliohms throughout the entire temperature-changing motion process, effectively eliminating most of the geometric and environmental interferences and improving the signal-to-noise ratio to over 20.0 dB, demonstrating a strong synergistic effect among the various technical features. When a microcrack was introduced into the system, the responses of each group of schemes showed differences, as shown in Table 1, which records the changes in key characteristic parameters and judgment results of each group before and after the introduction of the crack.

[0028] Table 1: Comparison of characteristic parameter responses before and after crack introduction For sample group C of this invention, when the crack depth is 5.0%, the residual at the second frequency (10.0MHz) jumps to 3.8 milliohms, while the residual at the first frequency (10.0kHz) remains at around 0.4 milliohms. The ratio of the two reaches 9.5, far exceeding the preset damage judgment threshold (set to 3.0). This asymmetry in high and low frequency response directly verifies the skin effect sensitivity mechanism of high-frequency signals to microscopic defects on the conductor surface, thereby eliminating misjudgments that may be caused by random fluctuations in contact resistance (usually synchronous changes across the entire frequency band). Furthermore, the gradient verification of the gradient law of the data analysis technology effect of groups D1-D3 shows that as the crack depth increases from 2.0% to 40.0%, the cumulative damage energy value calculated by sample group C of this invention shows a clear exponential growth trend, with a strong positive correlation with the crack depth. Specifically, the residual peak value corresponding to a crack depth of 2.0% is 1.2 milliohms, and that corresponding to a depth of 10.0% is 7.5 milliohms. At a depth of 40.0%, the impedance corresponds to 28.4 milliohms. When the crack depth exceeds 60.0%, the rate of increase in impedance change reaches a saturation inflection point, which is consistent with the physical law that the current density distribution of the remaining cross-section of the conductor tends to a steady state. This nonlinear characteristic is used to determine the late-stage failure warning threshold of the health factor, proving that the present invention has the ability to quantitatively perceive and respond to damage at different stages. The experimental results show that by introducing a phase correction and topological difference mechanism based on material viscoelasticity, the method of the present invention can improve the measurement signal-to-noise ratio under dynamic conditions from less than 0.1 dB in the control group A to more than 20.0 dB in the sample group C. The dual-frequency gating logic effectively separates the micro-damage signal from the macro-motion noise, realizing the deterministic capture of weak early damage in a strong interference environment. At the same time, the gradient verification results confirm the intrinsic correlation between the technical effect and the degree of damage, verifying the effectiveness and creativity of the static impedance benchmark model combined with the dynamic correction strategy in practical engineering applications.

[0029] Example 3: This example combines Figures 1 to 3 A method for detecting broken membrane-covered wires is described, such as... Figure 1As shown, in the multi-dimensional data real-time acquisition step, the command motion attitude coordinates, ambient temperature, and dual-frequency impedance parameters are acquired. Then, the data enters the real-time kinematics calculation and phase correction stage. The material hysteresis time constant is used to eliminate viscoelastic hysteresis. Next, based on the equivalent physical attitude coordinates, theoretical electrical characteristic parameters are queried to perform dynamic baseline generation and differential evaluation. This process combines the multi-dimensional lookup table mapping relationship provided by the static impedance reference model. The adaptive drift compensation module based on the anchor point state performs global weighted correction to correct non-damaging aging trends. The topological second-order differential process is executed to calculate the algebraic difference between the residuals of the target and reference conductors to eliminate common-mode interference. The processed data enters the dual-frequency sensitivity gating process. The high-low frequency residual ratio logic is used to identify micro-damage. Then, the health factor is calculated to obtain the cumulative damage energy value of the residual characteristic sequence. Finally, in the closed-loop control intervention stage, an early warning signal is generated and the equipment operating speed is reduced.

[0030] like Figure 2 As shown, this chart uses time (in seconds) as the horizontal axis and impedance (in Ω) as the vertical axis, displaying impedance change curves of three different types within a time window of 0 to 10 seconds. The theoretical impedance curve for commanded attitude, represented by a dashed line, exhibits high amplitude fluctuations; the measured impedance curve, represented by a dotted line, is located in the middle region; and the corrected theoretical impedance curve, represented by a solid line, shows an evolution pattern that more closely matches the physical true value by introducing a time constant to compensate for phase lag in the command coordinates. Figure 3 As shown, the system starts in the offline calibration and initialization state. During this stage, the material time constant is extracted and the static impedance reference is established. After calibration, the system enters the real-time dynamic monitoring state, inputs the command coordinates and dual-frequency impedance, and processes phase correction and second-order difference. When the system matches the anchor point attitude, it jumps to the adaptive drift compensation state to calculate the global drift factor and output the corrected reference model. After the model is corrected, it returns to the monitoring state. If the residual amplitude is abnormal, it enters the micro-damage logic discrimination state to execute dual-frequency sensitivity gating to judge benign interference and real damage. If it is determined to be motion or common-mode interference, it returns to the monitoring state. If cumulative damage is confirmed and the health factor slope exceeds the limit, it enters the early warning and closed-loop control intervention state, outputs a micro-damage early warning signal and executes a forced reduction in operating speed until it finally enters the maintenance shutdown state.

[0031] Example 4: Weighting coefficient k calibration is performed during the equipment installation and commissioning phase: drive the robot joints to perform a single-axis step motion at 50% of the rated speed, and simultaneously collect the command position sequence P at a sampling frequency of 1.0kHz. cmd (t) and the measured impedance response sequence Z meas (t), initially set k to 1.0, substitute into P cmd (t) to calculate the equivalent physical attitude coordinates P using the phase correction formula. eff (t), retrieve the corresponding theoretical impedance sequence Z in the static impedance reference model.theo (t), calculate Z theo (t) and Z meas (t) The cross-correlation coefficient of the time axis is iteratively adjusted in the interval [0.5, 1.5] with a step size of 0.01. The k value that maximizes the cross-correlation coefficient is selected as the joint solidification operation parameter. This calibration linearizes the nonlinear viscoelastic damping characteristics of a specific batch of insulating materials into a single scalar coefficient, so that the phase of the corrected theoretical impedance curve coincides with the measured curve to the greatest extent. The damage judgment threshold is determined based on the background noise statistical characteristics under the healthy state of the equipment. During the period when the equipment is confirmed to be in the initial healthy state and in normal production operation, no less than 1000 motion cycles of dual-frequency residual data are continuously collected. The statistical distribution of the ratio of high frequency to low frequency residuals in the dataset is calculated. The mean μ and standard deviation σ of the ratio distribution are extracted. According to the statistical 3σ criterion, the damage judgment threshold is set to μ+3σ. This setting filters out 99.73% probability of random contact resistance fluctuations and environmental electromagnetic interference, and retains statistical outliers. During the online monitoring stage, when the real-time calculated residual ratio exceeds the threshold for 5 consecutive sampling cycles, the damage accumulation counting logic is triggered. No manual intervention is required to adapt to different electromagnetic environment background noise levels.

[0032] Example 5: This example specifically illustrates the mathematical mechanism for constructing cumulative damage energy values ​​based on residual feature sequences, and the standardized engineering procedure for determining micro-damage warning thresholds through offline accelerated aging tests. Addressing the lack of specific operational procedures in the health factor calculation process, this example constructs a three-segment calculation model of residual-energy-trend. The system performs sliding window integration on the real-time acquired residual feature sequences, defining the width of the sliding window as N sampling points (e.g., N=1000, corresponding to a time length of 1.0 second). For the current time t, the system extracts a segment of the residual sequence from t-N+1 to time t, and calculates the residual energy within that segment. The integral or sum of the absolute values ​​of the differences is defined as the instantaneous damage energy E(t). Through the accumulation over time, the random fluctuations of single-point residuals are smoothed, enhancing the ability to capture persistent weak damage signals. The system calculates the slope S(t) of the instantaneous damage energy E(t) over time, i.e., S(t) = dE(t) / dt. This slope directly reflects the rate of damage development and is the core indicator for judging the trend of damage evolution. Finally, the system compares the calculated S(t) with the preset micro-damage warning threshold. Only when S(t) continuously exceeds the threshold and shows a monotonically non-decreasing trend is the accumulated damage considered effective and the health factor updated.

[0033] To address the lack of systematic basis for setting micro-damage warning thresholds, this embodiment introduces a parameter calibration procedure based on accelerated aging tests throughout the entire life cycle. The aim is to determine the critical threshold distinguishing between benign fluctuations and malignant damage through offline experiments. The calibration process is as follows: A membrane envelope sample consistent with the actual application specifications is selected and placed in a composite stress accelerated aging test chamber containing temperature and humidity cycling and mechanical bending. The test chamber operates according to a preset accelerated aging profile until the sample fractures and fails. Throughout the aging process, the system continuously collects and calculates the slope of the instantaneous damage energy change S(t) of the sample. By analyzing the evolution curve of S(t) throughout the entire life cycle, the inflection point when the sample transitions from the stable phase to the rapid damage propagation phase is identified. The slope value corresponding to this inflection point is defined as the critical damage slope S. crit To ensure robustness in engineering applications, a safety factor α (e.g., 0.8) is introduced, and the minor damage warning threshold is set to... This procedure transforms empirical threshold setting into a quantitative determination process based on objective experimental data, ensuring the scientific validity and applicability of the thresholds. Furthermore, to verify the effectiveness of the aforementioned model and thresholds, this embodiment designs a gradient verification experiment, constructing a membrane envelope test set containing different degrees of damage, corresponding to 10%, 50%, and 90% of the entire lifespan. The aforementioned calculation model is used to process each test set, and the output health factor values ​​are recorded. Experimental data shows that as the degree of damage increases, the calculated health factor exhibits a clear non-linear growth trend: at the 10% lifespan stage, the health factor remains low and fluctuates; at the 50% lifespan stage, the health factor increases; and at the 90% lifespan stage, the health factor rises exponentially and stably triggers an early warning. This result confirms that the constructed calculation model can accurately reflect the true health status of the membrane envelope, and that the calibrated thresholds can effectively distinguish the damage characteristics at different stages.

[0034] Example 6: In large-scale industrial deployment scenarios involving heterogeneous batches of membrane-coated cables, to ensure that the monitoring system can adapt to the differences in electromagnetic characteristics caused by manufacturing tolerances between different individual cables, a standardized system initialization and baseline calibration procedure must be executed after physical installation. This procedure is automatically triggered by the equipment controller, driving the actuator to run a preset quasi-static scanning trajectory throughout the entire workspace, covering all typical joint angle combinations and displacement paths. The running speed is strictly limited to less than 5.0% of the rated speed to eliminate the effect of viscoelastic hysteresis. During this process, the system synchronously collects measured impedance data at each discrete attitude coordinate (θ, x) and the current ambient temperature T, and uses multiple regression or spline interpolation algorithms to construct an initial static impedance benchmark model Z for this specific cable instance. Ref=F(θ,x,T), thereby internalizing the cable's manufacturing tolerances and initial installation stress state into the model's inherent parameters. Simultaneously, the system automatically identifies and locks the specific posture with the minimum mechanical stress gradient as the anchor point state, continuously collecting the average impedance value within 5.0 seconds in this state as the initial anchor point impedance value, establishing the zero-point reference for full life-cycle drift compensation. The equipment executes a set of high-dynamic excitation actions with varying acceleration. By analyzing the phase hysteresis spectrum of the impedance response relative to the motion command, the insulation material hysteresis time constant τ of the individual cable is measured and calibrated. This pre-calibration process establishes a deterministic mapping relationship between the measurement system and the physical object, ensuring that the residual feature sequence extracted in subsequent online monitoring can accurately reflect the microscopic damage evolution inside the conductor, rather than being a product of individual cable differences or installation errors.

[0035] To address the dispersion in material relaxation characteristics across different batches of membrane-coated wire, this specification further integrates adaptive generation and verification logic for the parameter matrix. The system employs a stratified random sampling method to extract at least 30 representative samples from membrane-coated wire inventories of different production batches and storage periods, constructing an initial calibration sample set. Using a high-precision impedance analyzer and a temperature- and humidity-controlled chamber, static impedance spectroscopy scanning is performed on each sample across the entire temperature range of -20.0℃ to 80.0℃ to obtain the impedance distribution characteristics of the sample set at different temperature points. Dynamic excitation testing is then performed on each sample, recording the impedance response curve under step stress input. The physical relaxation time constant τ of the insulating material for each sample is extracted using nonlinear least squares fitting, and the mean constant of the sample set is calculated. and standard deviation σ τ Based on the aforementioned measured data, the system automatically generates an initial parameter matrix containing the temperature drift coefficient, material relaxation time constant, and their confidence intervals. Then, a validation sample set is introduced, containing test samples with known minor damage (such as pre-fabricated microcracks). The initial parameter matrix is ​​used to perform blind testing on the validation sample set, calculating the precision and recall of the broken wire prediction model. If the precision is below 98.0% or the recall is below 95.0%, parameter optimization iteration logic is automatically triggered. This involves adjusting the weight coefficients in the parameter matrix or introducing new compensation terms until the model performance meets the preset acceptance criteria. Finally, the validated parameter matrix is ​​solidified as the equipment's factory default configuration, and an engineering report containing detailed calibration data and validation results is generated as a benchmark file for the equipment's full lifecycle health management.

[0036] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0037] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting broken wires in a membrane covering, characterized in that, Includes the following steps: A static impedance reference model incorporating viscoelastic characteristics is constructed. With the membrane envelope in a defined initial healthy state, the drive device traverses the preset working path, synchronously collects the motion attitude coordinates and ambient temperature of the membrane envelope endpoints, and records the reference electrical characteristic parameters in the corresponding states. A static mapping relationship is established with motion attitude coordinates and ambient temperature as index keys and reference electrical characteristic parameters as values. The material hysteresis time constant is extracted. By comparing the impedance change response characteristics of the membrane-coated wire under the first and second motion speed conditions, the time constant characterizing the physical relaxation speed of the membrane-coated wire insulation material is calculated. The first motion speed produces a measurable viscoelastic hysteresis, and the second motion speed is lower than the first motion speed to simulate a quasi-static process. Real-time kinematics calculation and phase correction: During the operation of the membrane envelope, the current command motion attitude coordinates are collected in real time, the rate of change of the command motion attitude coordinates with time is calculated to obtain the instantaneous motion velocity vector, and the instantaneous motion velocity vector and time constant are used to perform phase correction operation on the command motion attitude coordinates to obtain the equivalent physical attitude coordinates that characterize the current actual geometric deformation of the membrane envelope. Dynamic baseline generation and differential assessment utilize equivalent physical attitude coordinates as corrected index keys to query the reference electrical characteristic parameters corresponding to the current moment in the static impedance reference model and define them as theoretical electrical characteristic parameters. The measured electrical characteristic parameters and theoretical electrical characteristic parameters are differentially calculated to obtain the residual characteristic sequence that eliminates the interference of material deformation hysteresis. Based on the monotonic evolution trend of the residual characteristic sequence, the health factor characterizing the degree of damage is calculated.

2. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, In the real-time kinematics calculation and phase correction steps, the phase correction operation is strictly performed according to the following mathematical relationship, so as to use the time constant to offset the geometric deformation hysteresis caused by the viscoelasticity of the polymer material during high-dynamic motion: , where P eff (t) is defined as the equivalent physical attitude coordinates, P cmd v(t) is defined as the commanded motion attitude coordinates, v(t) is defined as the instantaneous motion velocity vector, τ is defined as the time constant, and k is defined as a weighting coefficient preset based on the damping characteristics of the membrane envelope material.

3. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, The dynamic baseline generation and differential evaluation steps also include an adaptive drift compensation process based on the anchor point state. This process includes: pre-setting at least one specific attitude in the equipment's operating trajectory as the anchor point attitude, and recording the initial anchor point impedance value corresponding to the anchor point attitude in the initial healthy state; when the equivalent physical attitude coordinates are detected to match the anchor point attitude, calculating the deviation between the current measured electrical characteristic parameters and the initial anchor point impedance value, and defining the deviation as a global drift factor; using the global drift factor to globally weight and correct the theoretical electrical characteristic parameters output by the static impedance reference model, so that the static impedance reference model follows the non-damaging aging trend of the membrane envelope.

4. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, The membrane envelope contains at least two parallel conductors. The dynamic baseline generation and differential evaluation steps specifically include a topological second-order differential process: one conductor in the membrane envelope is selected as the target conductor, and another conductor physically adjacent to the target conductor is selected as the reference conductor; the primary residuals between the measured electrical characteristic parameters of the target conductor and the reference conductor at the current moment and the theoretical electrical characteristic parameters obtained based on their equivalent physical attitude coordinates are calculated respectively; the algebraic difference between the primary residuals of the target conductor and the primary residuals of the reference conductor is calculated to obtain the second-order topological residual features; the second-order topological residual features are used as a residual feature sequence to eliminate common-mode environmental interference and mechanical vibration noise acting on the entire membrane envelope.

5. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, The measured electrical characteristic parameters include a first frequency impedance parameter and a second frequency impedance parameter, with the second frequency being higher than the first frequency. The dynamic baseline generation and differential evaluation steps include a dual-frequency sensitivity gating process, which calculates the first residual between the measured and theoretical values ​​of the first frequency impedance parameter, and the second residual between the measured and theoretical values ​​of the second frequency impedance parameter. Damage discrimination logic is constructed using the ratio of the second residual to the first residual. When the amplitude of the second residual increases while the amplitude of the first residual remains within a preset range, causing the ratio to exceed the damage judgment threshold, it is confirmed that the residual characteristic sequence is mainly caused by microscopic damage to the conductor, and the residual characteristic sequence is retained. When the first and second residuals increase synchronously, causing the ratio to remain in the non-damage range, it is determined to be motion posture deviation interference, and the residual characteristic sequence at that moment is suppressed.

6. The method for detecting broken wires in a membrane covering according to claim 1, characterized in that, The steps for extracting the material hysteresis time constant specifically include: in the offline calibration stage, the drive device performs reciprocating motion at a first motion speed and records the first impedance change curve; the drive device performs the same reciprocating motion at a second motion speed and records the second impedance change curve; the phase lag of the first impedance change curve relative to the second impedance change curve on the time axis is calculated; and the time constant is determined using the ratio of the phase lag to the first motion speed.

7. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, The health factor calculation process includes: performing sliding window integration on the residual feature sequence to obtain the cumulative damage energy value; monitoring the slope of the cumulative damage energy value over time; and determining that the membrane envelope has entered the late stage of fatigue failure and generating an early warning signal when the slope exceeds the preset micro-damage warning threshold and exhibits an irreversible monotonous growth characteristic.

8. The method for detecting broken wires in a membrane covering according to claim 1, characterized in that, The reference electrical characteristic parameter is the AC impedance modulus at a specified high frequency. The selection of the specified high frequency is based on ensuring that the skin depth of the membrane conductor is less than one-third of the conductor thickness, thereby enhancing the sensitivity to the detection of microcracks on the conductor surface. The static impedance reference model is stored in the form of a multidimensional lookup table. The lookup process includes linear interpolation calculations based on equivalent physical attitude coordinates and ambient temperature in the multidimensional lookup table.

9. The method for detecting broken wires in a membrane covering as described in claim 1, characterized in that, Before the dynamic baseline generation and differential evaluation steps, a reference model initialization process is also included, in which the device is placed in a stationary state and all preset temperature points are traversed to obtain the pure temperature drift curve under zero velocity state; the pure temperature drift curve is used to pre-calibrate the ambient temperature dimension in the static impedance reference model to establish the initial zero point of the static mapping relationship.

10. The method for detecting broken wires in a membrane covering according to claim 7, characterized in that, After generating the early warning signal, a closed-loop control intervention step is also included. In response to the health factor exceeding the preset safety limit, a deceleration command is sent to the equipment controller to forcibly reduce the equipment's operating acceleration and maximum speed, reduce the dynamic mechanical stress acting on the membrane envelope, and delay the fracture process until maintenance and replacement are completed.

Images